Runx2 downregulates Lpl expression through super-silencer formation to alter lipid metabolism in Zhu Schwann cells after nerve injury
Zhaowei Zhu, Rui Kuang, Shouwen Su, Yujing Zhang, Guanggeng Wu, Yi Zhang, Vincent Pang, Xiang Zhou, Yan Yang, Ge Li, Bo He, Yangbin Xu
TL;DR
This study reveals how Runx2 changes lipid metabolism in Schwann cells after nerve injury by forming a super-silencer that suppresses Lpl gene expression.
Contribution
The first super-silencer landscape in Schwann cells with Runx2 overexpression is identified, linking it to lipid metabolism regulation.
Findings
Runx2 forms a super-silencer in the Lpl gene region, downregulating its expression in Schwann cells.
Lipid metabolism is disrupted in Runx2-overexpressing Schwann cells following nerve injury.
Modified molecular markers for repair-type Schwann cells are described.
Abstract
Phenotypic transformation of Schwann cells (SCs) plays a crucial role in nerve regeneration. Previous studies have demonstrated that Runx2 significantly influences the biological behavior of SCs. Nonetheless, the regulatory mechanisms that govern its epigenetic regulation are not yet fully elucidated. To facilitate this investigation, an adenovirus for the overexpression of Runx2 was constructed. Healthy adult Sprague–Dawley rats, weighing between 100 and 150 g and irrespective of sex, were randomly selected for the study. After establishing a model of sciatic nerve crush injury, tissue samples were harvested for histological analysis at both 4 and 7 days post-injury. In vitro, an Runx2-overexpressing SC line was established. Thorough analysis of transcriptome data, coupled with CUT&Tag sequencing of histones and transcription factors in SCs following Runx2 overexpression, was…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6- —the National Natural Science Foundation of China
- —the Natural Science Foundation of Guangdong Province
- —Guangdong Basic and Applied Basic Research Foundation
- —Special project of Dengfeng Program of Guangdong Provincial People’s Hospital
- —the Natural Science Foundation of Guangdong Province, China
- —the Guangzhou Basic Research Program Joint Funding Project for Municipal and University/Institute Collaboration
- —The Science and Technology Development Fund of Macau
- —the Guangzhou Science and Technology Plan Project
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsNerve injury and regeneration · Neurogenesis and neuroplasticity mechanisms · Mesenchymal stem cell research
Background
The regeneration of tissues after injury requires the repair or replacement of damaged or lost cells. Some animals, such as salamanders and newts, can regenerate limbs and other organs after injury. This process involves reprogramming cells at the injury site into pluripotent stem cells or progenitor/immature cells with positional information, repairing lost cells or tissue structures by reproducing developmental processes [1]. In adult mammals, peripheral nerves are among the few tissues capable of extensive regeneration.
The regeneration of nerve axons after peripheral nerve injury repair is a continuous and complex process. SCs play a crucial role in nerve axon regeneration [2]. Following nerve injury, SCs undergo dedifferentiation and reprogramming, reverting to an active state similar to “immature SCs” in early developmental stages. Subsequently, these cells proliferate and reform myelin sheaths, ultimately completing the critical repair process of nerve regeneration. This plasticity mechanism enables SCs to effectively promote axonal regeneration and functional recovery after peripheral nervous system injury [3].
In our preliminary studies, we performed whole-transcriptome sequencing of sciatic nerves from Sprague Dawley (SD) rats at 4 and 7 days after crush injury, and found that the expression level of Runx2 in nerves correlates with injury duration (database from Zhang (2021) [4]). Although some studies have attempted to explore the relationship between Runx2 and the state of SCs, there are currently only five publications on the role of Runx2 in peripheral nerves [3, 5–8].
Ding et al. discovered that Runx2 expression gradually increased after sciatic nerve crush, peaking at 1 week, suggesting that Runx2 may be involved in nerve axon growth and SCs differentiation and migration [5]. Wang et al. found that curcumin enhances SC proliferation and myelination through Runx2, promoting sciatic nerve injury repair [6]. Hu et al. also found that Runx2 promotes SC migration and myelin regeneration. These results all suggest that Runx2 plays a role in the process of SC dedifferentiation and reprogramming after peripheral nerve injury [7].
Through integrated multiomics analysis of Runx2-overexpressing (Runx2-OE) SCs combined with single-cell RNA sequencing (scRNA-seq) data from injured rat sciatic nerves, we discovered that Runx2 can promote phenotypic transformation of SCs by activating Sox2 [3]. Runx2 is highly expressed only in early repair-type Zhu1 SCs, while showing low to median expression in repair-type He SCs and late repair-type Zhu2 SCs. Sox2 is best known for its core role in maintaining embryonic stem cell self-renewal and pluripotency [9]. Research has also shown that it is one of the transcription factors that can help reprogram somatic cells into induced pluripotent stem cells [10, 11].
After SCs regain stem cell-like properties and become immature-like SCs, their biological functions inevitably change. To further investigate the epigenetic mechanisms by which Runx2 influences SCs, we established Runx2-OE SCs and conducted multiomics analysis. The results suggest that effects of Runx2 on cellular phenotype may not depend on histone modifications but rather directly target closed chromatin, affecting chromatin accessibility. Simultaneously, this influences the recruitment of other transcription factors, affecting the construction of key transcriptional elements, particularly influencing the status of negative regulatory factors, thereby influencing cellular functions [3].
To further explore how Runx2 affects cell fate through negative factors, this study aims to use histone modification (H3K27me3) chromatin accessibility using targeted endonuclease and tagmentation (CUT&Tag) and transcription factor (Runx2 and CTCF) CUT&Tag sequencing, combined with mRNA sequencing results from Runx2-OE SCs, to gain a deeper understanding of the epigenetic regulatory mechanisms of how Runx2 affecting SCs under injury conditions. The raw sequencing data were shared with those in He et al. [3].
Methods
To investigate peripheral nerve regeneration mechanisms, adult Sprague–Dawley (SD) rats (100–150 g, mixed-sex cohort) were subjected to standardized sciatic nerve crush injury. Primary Schwann cells (SCs) were isolated from 3- to 5-day-old neonatal SD rats via enzymatic dissociation of sciatic nerves, followed by purification and expansion in vitro. All experimental protocols complied with the ethical guidelines of Sun Yat-Sen University’s Experimental Animal Administration Committee, incorporating perioperative analgesia, daily postoperative monitoring, and predefined humane endpoints. Complete details of all reagents (including manufacturers and dosages), instruments, and software are provided in Supplementary Tables S2–S3.
Spatial–temporal profiling of Runx2 expression in crushed sciatic nerves via multiplex immunofluorescence
Adult Sprague–Dawley rats (100–150 g, both sexes, n = 3/group) were allocated into: normal controls (no surgery), 4-day post-crush (PI4d), and 7-day post-crush (PI7d) cohorts. The sciatic nerve crush model was established according to previous reports from literature [3, 4]. Focal sciatic nerve injury was induced by triple 30-s vascular clamp compressions at the piriformis muscle inferior border. Injured nerve segments were harvested at designated timepoints. Animals in the normal group underwent the same operation but without sciatic nerve clamp operation. The samples were processed and stained according to previous literature reports [3, 12]. Fixed 4% paraformaldehyde (PFA) samples underwent sequential immunofluorescence staining with anti-Runx2, NF200, CD31, and S100b. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Confocal imaging captured z-stacks from two randomized sections per sample.
Using ImageJ software, we measured the total fluorescence intensity of Runx2 in each field of view. The fluorescence intensity values were normalized by dividing the total Runx2 fluorescence intensity of each field by the average Runx2 fluorescence intensity of the control group, yielding relative fluorescence intensity values. GraphPad software was then used to analyze whether there were statistically significant differences in relative Runx2 intensity between groups and to generate statistical graphs. Since the data did not follow a normal distribution, the Kruskal–Wallis test was employed for analysis.
Morphology changes of Runx2-overexpression SCs (Zhu1 SCs) in vitro
A recombinant adenoviral vector (Ad-Runx2-EGFP) expressing Runx2 under cytomegalovirus (CMV) promoter was generated. Control vector expressing enhanced green fluorescent protein (EGFP) only (Ad-EGFP) was constructed in parallel.
In accordance with previous methods [3, 12], SCs from 3- to 5-day-old neonatal SD rats were isolated and purified. P0 SCs were expanded in DMEM/F12 + 10% fetal bovine serum (FBS) + 2 μM forskolin + 10 ng/mL neuregulin 1 (NRG1) at 37 °C/5% CO_2_. Cells at 90% confluence were dissociated for transduction. SCs transduced with Ad-Runx2-EGFP or Ad-EGFP (multiplicity of infection, MOI = 200) were subjected to imaging for 72 h using laser confocal microscopy. Morphometric parameters (cell area, elongation factor, and process complexity) were captured.
mRNA profiling of Runx2-OE SCs
Runx2-OE and control (EGFP) SCs (n = 2 biologically independent replicates) underwent paired-end RNA sequencing by Shanghai Biotechnology Corporation [3]. Library preparation utilized the SMART-Seq^®^ HT kit, with quality control via Qubit 4.0 and Agilent 2100 Bioanalyzer. HISAT2 (version 2.0.4) aligned reads to Rnor_6.0 genome, followed by fragments per kilobase of transcript per million mapped reads (FPKM) normalization [13]. Differential expression was determined by edgeR (p < 0.05, |log_2_fold change (FC)|> 1). Gene set enrichment analysis identified enriched KEGG pathways.
High-resolution epigenomic mapping of Runx2 regulatory landscapes
The CUT&Tag assay was performed as described previously with modifications [3, 14] and was performed by Guangzhou Huayin Health Medical Group Co., Ltd. (Guangdong, China). For H3K27me3, Runx2, and CTCF profiling, 1 × 10⁶ SCs underwent CUT&Tag following the manufacturer’s protocol with critical modifications. The raw sequencing data were shared with those in He et al. [3].
Establishment and functional analysis of the super-silencer library
The rank ordering of super-enhancers (ROSE) algorithm was adapted to identify SSs by using H3K27me3 CUT&Tag data. The methodology has been described previously [15], and named ReSE. Specifically, the H3K27me3 peak files identified by MACS2 and the BAM files were used as inputs for the algorithm. Intergenic and intronic H3K27me3 peaks within 4 kb were stitched together to define a single silencer entity spanning a genomic region. Isolated peaks (without neighboring peaks within 12.5 kb) were considered individual silencers. Super-silencers (SSs) were defined as silencers with H3K27me3 intensity above a cutoff where the slope of the distribution plot of H3K27me3 intensity equals 1. The remaining silencers were classified as typical silencers (TSs). SSs and TSs were annotated using the annotate Peaks.pl tool from HOMER software (version 4.11). Motif enrichment analysis for SS and TS was performed using the findMotifsGenome.pl tool from HOMER software (version 4.11). A protein–protein interaction (PPI) network was constructed according to the intersection among SS in both groups and downregulated sequences from mRNA-seq.
Dual-luciferase assay to verify the binding sites of the Runx2 peak region and Lpl SS
To elucidate the specific mechanism by which Runx2 acts to inhibit Lpl, we employed the FIMO component of the MEME software suite to scan and identify Runx2 binding sites within the Lpl SS, with the detailed protocols reported in a previous study [3]. On the basis of the identified binding site sequences, we designed the following luciferase reporter constructs. The sequences of these reporters are listed in Supplementary Table S1.
Lpl SS region verification:
Wild-type (WT) vector: pGL3-promoter/WT;
Empty vector control: pGL3-promoter;
Mutant 1 vector: pGL3-promoter/Mut1 (1772-1786)
Mutant 2 vector: pGL3-promoter/Mut2 (1107-1121).
These reporter constructs were prepared as viral vectors and cotransfected into 293 T cells along with either the Runx2 overexpression plasmid (Runx2-OE) or a control plasmid (pcDNA3.1) [3]. The luciferase reporter system was constructed by Dongzebio Co., Ltd. All results are presented as the mean ± standard error of the mean (SEM), with n = 4 plates of cells per condition.
Spatial–temporal dynamics of Lpl in Schwann cell states
Leveraging a previously published sciatic nerve scRNA-seq atlas (GSE216665) [3, 16], we implemented a focused interrogation strategy for Lpl regulation: First, Sox10 ^+^ Plp1 ^+^ Schwann cells (n = 8421 cells) were isolated from integrated datasets (naïve, PI3d, PI12d, and PI60d) using Seurat’s subset function. For each cell type (Table 1), marker genes were defined as differentially expressed mRNAs (DEmRNAs) with an adjusted p value < 0.05 and a log_2_-fold change > 0.25 compared with the remaining cell types. For each group, DEmRNAs of each cell type were defined as genes with an adjusted p value < 0.05 and a log_2_-fold change > 0.25 between the two groups.Table 1. Specific gene markers of different SCs (He’s classification) [3]NominationZhuHeRemakMyelinating/myelinHZMarkerP75NTRP75NTREgr2Egr2BdnfMki67GfapMbpGdnfTop2aCdh19MpzErbb3Cdk1Scn7aPrxRunx2MagCldn19SCs are annotated on the basis of the expression of the marker genes Sox10, Plp1, and S100bMyelin SCs are defined as SCs that formed mature myelin in normal nerves or injured nerves, while myelinating SCs refer to SCs that are in the process of forming myelin sheaths but have not yet completed myelinationHZ SCs indicate SC markers but do not belong to any other classification**Zhu SCs include Zhu1 SCs and Zhu2 SCs, with Runx2 serving as a marker that distinguishes Zhu1 SCs from Zhu2 SCs
Analysis of in-bulk tissue transcriptome data of nerve crush injury in SD rats
The transcriptomic sequencing data of SD rats’ sciatic nerve crush injury models were downloaded from the GEO database (accession no. GSE162548) [4]. The dataset includes samples from normal sciatic nerves (NC) and crushed nerves at 4 days (PI4d) and 7 days (PI7d) post-injury (n = 2). We conducted reanalysis of transcriptome data following previously reported methods [4, 17]. To compare gene expression levels among different genes and different samples, reads were converted into FPKM for normalization of gene expression levels. The Stringtie (version 1.3.0), (trimmed mean of M values (TMM) method and Perl scripts were used to calculate the FPKM value of each gene. The average expression values of Lpl at different injury times were obtained, and line graphs were drawn for intergroup comparison.
Images, data processing, and statistical analysis
The data were processed by using SPSS and GraphPad Prism. One-way analysis of variance (ANOVA) with the Tukey multiple-comparison method was used to check the specific differences among groups. The chi-squared test was used to analyze percentage data, setting the test standard α as 0.05.
The Shapiro‒Wilk test was used to analyze the luciferase reporter system, setting p > 0.05. The Brown–Forsythe test was used to judge whether homogeneity of variances was met or not (p > 0.05). Finally, corrected one-way ANOVA, followed by Dunnett’s test for pairwise comparisons, was performed. The vector or WT group as control was used and compared with the mutant groups.
Results
Expression of Runx2 after nerve injury and analysis of differentially expressed genes in Runx2-OE SCs
Immunofluorescence staining for CD31/NF200/S100 and Runx2 was performed at the injury site at different time points after peripheral nerve crush injury. The results revealed a significant increase in Runx2 expression in SCs during the early phase of injury, with noticeable coexpression in the S100-positive areas, which were predominantly localized in SCs (Fig. 1a–c). Quantitative analysis of Runx2 immunofluorescence staining revealed that the ratios for PI4d and PI7d groups were significantly elevated compared with the normal group (p < 0.05), but no statistically significant difference was observed between PI4d and PI7d groups (Fig. 1d).Fig. 1. Histological observation of Runx2 expression in SD rats following sciatic nerve crush injury at 4 and 7 days, and transcriptome analysis of SCs after Runx2-overexpression (Runx2-OE). A–C Immunofluorescence images of peripheral nerves in normal nerves (scale bars represent 100 μm, A) and at 4 days (B) and 7 days (C) post-injury (n = 3). In all panels, green represents NF200 ^+ ^axons (green) and Runx2 ^+^ (red) in 1, CD31 ^+^ endothelial cells (green) and Runx2 ^+^ (red) in 2, and S100 ^+^ SCs (green) and Runx2 ^+^ (red) in 3. Blue denotes nuclei (DAPI ^+^). D Bar chart depicting quantitative analysis of Runx2 immunofluorescence staining in peripheral nerves at different time points post-injury. E Morphological observation of SCs after Runx2-OE. Red arrows in EGFP group point to control SCs, while yellow arrows in Runx2-OE group point to SCs after Runx2 overexpression. Scale bars represent 50 μm. F Heatmap showing the distribution of differentially expressed mRNAs (DEmRNAs) between the two groups. Red indicates upregulation, and green indicates downregulation, n = 2. G Ridge plot showing GSEA analysis of DEmRNAs, with red boxes highlighting significantly enriched pathways, the color gradient from red to blue indicates increasing p values. G1 KEGG enrichment analysis, G2 GO enrichment analysis. *p < 0.05, **p < 0.01, ns indicates no significance
The morphologies of these in vitro-cultured cells were examined using light microscopy. It could be found that SCs in the EGFP group (designated as He SCs) exhibited characteristic bipolar elongated morphology, whereas SCs in the Runx2-OE group (designated as Zhu SCs) displayed a rounded cell shape with short protrusions (Fig. 1e).
Smart-seq mRNA sequencing analysis identified 1317 upregulated and 1345 downregulated genes in Runx2-OE SCs, with notable upregulation of Runx2, Notch, Mki67, and Pten and downregulation of Il1rl1, Lpl, and Egfr, among others (Fig. 1f). GSEA analysis revealed negative correlations between Runx2-OE SCs and KEGG pathways related to lipid and atherosclerosis (normalized enrichment score, NES = −2.10, p < 0.001) and hypoxia-inducible factor (HIF)−1 signaling pathway (NES = −2.16, p < 0.001), while GO terms were associated with positive regulation of response to external stimulus (NES = −2.27, p < 0.001) and inflammatory response (NES = −2.34, p < 0.001) (Fig. 1G1, G2), suggesting that Runx2 might influence the biological properties of SCs by altering cellular lipid metabolism and immune responses.
Analysis of key regulatory transcription factors Runx2 and CTCF in Runx2-OE SCs
Further CUT&Tag-seq analysis was performed to examine Runx2 and CTCF binding capacity in Runx2-OE SCs. In Runx2 CUT&Tag-seq results, peaks in the Runx2-OE group were primarily concentrated in the Transcription Start Site (TSS) region (Fig. 2a), with 11.32% located in the promoter region (Fig. 2b); the corresponding Upset plot is shown in Fig. 2c. Runx2 signals were significantly higher in the OE group compared with the control group within Runx2-upregulated peak regions, while showing an opposite pattern in Runx2-downregulated peak regions (Fig. 2d). The upregulated peaks in the Runx2-OE group were mainly enriched in pathways including synaptic vesicle cycle, neurotrophin signaling pathway, WNT signaling pathway, cAMP signaling pathway, signaling pathway regulating pluripotency of stem cells, and PI3K–AKT signaling pathway (Fig. 2e).Fig. 2. Runx2 and CTCF Cut&Tag sequencing in Runx2-OE transfected SCs. A–E Distribution and number of Runx2 peaks in Runx2-OE SCs. A Heatmap showing the distance between Runx2 peaks and transcription start sites (TSS) after Runx2-OE. B Pie chart illustrating the feature distribution of Runx2 peaks in the Runx2-OE and vector groups; C Upset plot visualizing feature distribution of Runx2 peaks. D Comparative analysis of the distribution of Runx2 signals in the Runx2-OE group versus the EGFP control group, centered on the upregulated/downregulated Runx2 CUT&Tag peaks (±3 kb). The purple symbol denotes the control group (EGFP SCs), whereas the green symbol represents the Runx2-OE group. E List of KEGG enrichment pathways showing genes with Runx2 peaks in Runx2-OE SCs. F–J Distribution and number of CTCF peaks in Runx2-OE SCs. F Heatmap showing the distance between CTCF peaks and TSS after Runx2-OE. G Pie chart illustrating the feature distribution of CTCF peaks in the Runx2-OE and vector groups; H Upset plot visualizing feature distribution of CTCF peaks. I Comparative analysis of the distribution of CTCF signals in the Runx2-OE group versus the EGFP control group, centered on the upregulated/downregulated Runx2 CUT&Tag peaks (±3 kb). The purple symbol denotes the control group (EGFP SCs), whereas the green symbol represents the Runx2-OE group. J List of KEGG enrichment pathways showing genes with CTCF peaks in Runx2-OE SCs. In E and J, the circle size represents the number of genes, and the color gradient from red to blue indicates increasing p values. The red boxes indicate key pathways
In CTCF CUT&Tag-seq analysis, no significant difference was observed in the number of peaks at TSS regions between OE and control groups (Fig. 2f), with 6.15% located in the promoter region in the OE group (Fig. 2g); the corresponding Upset plot is shown in Fig. 2h). CTCF signals were notably higher in the OE group within Runx2-upregulated peak regions, while showing an opposite pattern in Runx2-downregulated peak regions, suggesting potential significant interactions between Runx2 and CTCF in the former peak regions (Fig. 2i). The upregulated peaks in the Runx2-OE group were primarily enriched in pathways including aldosterone-regulated sodium reabsorption, ubiquitin mediated proteolysis, and signaling pathway regulating pluripotency of stem cells (Fig. 2j).
Changes in chromatin H3K27me3 modification status in Runx2-OE SCs
To better understand the histone methylation modifications in SCs following Runx2-OE, we performed H3K27me3 CUT&Tag-seq analysis (Fig. 3a). The results indicated that Runx2-OE cells exhibited more peaks compared with the control group (vector) (66,918 versus 60,009; Fig. 3B1), while the vector group showed greater peak accumulation in the transcription start site (TSS) region than the Runx2-OE group (Fig. 3a). In Runx2-OE cells, peaks were predominantly concentrated in distal intergenic regions (75.41%), with 4.6% located in promoter sequences (Fig. 3B2); the corresponding Upset plot is shown in Fig. 3B3.Fig. 3H3K27me3 Cut&Tag sequencing in Runx2-OE transfected SCs. A Heatmap showing the distance between H3K27me3 peaks and transcription start sites (TSS) after Runx2 overexpression. B Distribution and number of H3K27me3 peaks in Runx2-OE SCs. B1 Bar graph showing differences in H3K27me3 peak numbers between different groups; B2 Pie chart illustrating the feature distribution of H3K27me3 peaks in the Runx2-OE and vector groups; B3 Upset plot visualizing feature distribution of H3K27me3 peaks. C List of functional enrichment pathways showing genes with H3K27me3 peaks in Runx2-OE SCs. C1 KEGG pathway analysis. The circle size represents the number of genes, and the color gradient from red to blue indicates increasing p values. C2 GO analysis. D HOMER-predicted motifs responsive to Runx2-OE H3K27me3 data
Enrichment analysis of target genes corresponding to signals in Runx2-OE revealed that KEGG pathways were mainly concentrated in sulfur metabolism, axon guidance, serotonergic synapse, calcium signaling pathway, and neuroactive ligand-receptor interaction (Fig. 3C1). GO analysis suggested that these signals were primarily enriched in processes regulating cell connections and synaptic assembly, as well as lipid metabolism processes such as triglyceride lipase and sterol esterase activity (Fig. 3C2). HOMER motif scanning of Runx2-OE SCs revealed potential binding of Runx_AML (8.52%, Fig. 3d).
Establishment of Runx2-related super-silencer landscape in Runx2-OE SCs
Following Cai’s methodology [15], we performed ReSE analysis on the H3K27me3 CUT&Tag results from both groups of cells. We identified 775 methylation-enriched sites in the control group (EGFP SCs) and 756 of these sites in the Runx2-OE group (Fig. 4A1). In accordance with Cai et al.’s definition, we considered these sites as SSs. In the control group, 17.17% of these SSs were located around promoters, whereas in Runx2-OE SCs, this proportion was 12.43% (Fig. 4A2).Fig. 4. Super-silencer (SS) landscape construction in* Runx2*-OE transfected SCs. A Distribution of SS region in Runx2-OE SCs. A1 Hotkey plot of SS sorted by the ROSE algorithm in Runx2-OE and EGFP groups; A2 Pie chart illustrating the feature distribution of SS in the Runx2-OE and EGFP groups. B List of GO enrichment pathways showing genes with SS in Runx2-OE SCs. The red boxes indicate key pathways. C List of KEGG enrichment pathways showing genes with SS in Runx2-OE SCs. The red boxes indicate key pathways. The circle size represents the number of genes, and the color gradient from red to blue indicates increasing p values. D HOMER-predicted motifs responsive to Runx2-OE SS. E Visualization of the Tmem132e gene showing SS, with the grey box indicating key elements (E) within the SS, and the red box indicating CTCF peaks area
GO and KEGG functional enrichment analyses of these genes revealed that, in Runx2-OE cells, genes associated with SS were involved in biological functions such as the regulation of neuronal differentiation, synaptic assembly, motor organ morphology, and cell fate determination. These genes were localized in presynaptic/postsynaptic membranes and axon terminals, influencing the activity of transcriptional regulatory complexes and DNA-binding transcriptional repressor/activator molecules (Fig. 4b). KEGG analysis suggested that these genes could affect SCs pluripotency pathways, the WNT signalling pathway, the cAMP signalling pathway, and neuroactive ligand‒receptor interactions (Fig. 4c).
HOMER scanning of motifs binding in SS regions revealed that the known motifs mainly included Snail1 (17.55%) and ATHB34 (16.09%), while the predicted binding motifs were RNF4F (42.15%), ZNF682 (19.81%), Sol1 (15.69%), and bHLH10 (12.37%). Other motifs are shown in Fig. 4d. Visualization of the regulatory regions of Tmem132e (Fig. 4e) genes using Integrative Genomics Viewer (IGV) software revealed SS regions around the coding sequences. These regions contained increased Runx2 signals, along with enriched H3K27me3 peaks.
Runx2 downregulates lipoprotein lipase (Lpl) expression through SS formation, interfering with Schwann cell lipid metabolism
Intersection analysis of SS solely existing in Runx2-OE group and downregulated sequences from mRNA-seq revealed six genes that presented downregulated expression and SS formation following Runx2-OE. These genes were Lpl, Tlr2, Tmem132e, Mylk3, Xkr6, and Thbd (Fig. 5a). The construction of a PPI network with these genes showed that Lpl could form an interaction network with Thbd and Tlr2. Moreover, Lpl was associated with cellular cholesterol metabolism and glycolipid metabolism processes, suggesting that changes in Lpl expression would affect SC function through altering lipid metabolism (Fig. 5b). Meanwhile, Thbd and Tlr2 are two molecules with different functions, participating in the coagulation–anticoagulation system and innate immune response processes, respectively.Fig. 5. Super-silencer landscape construction in* Runx2-*OE transfected SCs. A Intersection diagram of SS target sequences uniquely present in Runx2-OE SCs and downregulated in transcriptome data, with arrows indicating genes influenced by Runx2-related SS. B Protein–protein interaction (PPI) network of target genes of Runx2-related SS. C Single-cell RNA sequencing (scRNA-seq) and bulk tissue transcriptome analysis of SD rats at different time periods after nerve crush injury. The scRNA-seq data includes 3 days (PI3d), 12 days (PI12d), and 60 days (PI60d) post-injury compared with those of normal nerves (NC). Data source: GSE216665. The bulk tissue transcriptome data includes normal nerves (NC), 4 days post-injury (PI4d), and 7 days (PI7d) post-injury data. Data source: GSE162548. C1 UMAP distribution plot of different cell types in scRNA-seq data; C2 Violin plot and bubble plot illustrating the expression changes of Lpl in SCs during different injury time. Yellow, green, and purple indicate gradually decreasing gene expression compared with the mean. The bubble size represents the proportion of cells expressing the gene within each cell type, with larger areas indicating higher expression proportions. C3 Line graph illustrating the expression changes of Lpl in the injury segment at different time points after peripheral nerve injury. D Visualization of Lpl gene. D1 IGV graph of the Lpl gene showing SS, with the grey box indicating key elements (E) within the SS and the red box indicating CTCF peaks area. D2 HOMER-predicted motifs responsive to CTCF CUT&Tag-seq in Runx2-OE SCs; E Luciferase reporter assay to investigate the mechanism by which Runx2 downregulates Lpl expression through the formation of a super-silencer (SS) (n = 4). E1 Schematic of the luciferase reporter plasmid construction and the Lpl-SS gene sequence (reused from He et al. [3]); E2 Bar graph showing the luminescence response in 293 T cells for the Lpl-SS wild type (WT) and vector control group; E3 Bar graph comparing the luminescence response in 293 T cells for mutant versions of the Lpl-SS region (Mut1 and Mut2) and the wild type (WT) before and after overexpression of Runx2; *p < 0.05, **p < 0.01, ***p < 0.005
We reanalyzed previously published bulk tissue transcriptome data (GSE162548) from SD rat sciatic nerve crush injury at 4 days (PI4d) and 7 days (PI7d) [4] and found that Lpl expression decreased in the early post-injury period (PI4d) and remained lower than normal nerves (NC) at PI7d, indicating that Lpl expression is suppressed in the early stages of peripheral nerve injury (Fig. 5C3). To observe the distribution and function of Lpl in peripheral nerve injury, we reanalyzed the GSE216665 dataset. After annotation, the cell clusters in GSE216665 could be divided into ten cell populations, as shown in Fig. 5C1. Analysis of Lpl expression in SC clusters at different time points after injury revealed that Lpl expression began to increase in SCs primarily at 12 days post-injury (PI12d) and continued through 60 days post-injury (PI60d) (Fig. 5C2). Further observation of Lpl changes in different SCs subtypes showed that the Lpl gene was mainly expressed in He and HZ SCs at PI60d (Fig. 5C1).
Visualization of the regulatory regions of Lpl (Fig. 5D2) using IGV software revealed SS regions around the coding sequences. This region contained enriched H3K27me3 and Runx2 peaks in the promoter area. Meanwhile, some CTCF signals could be found around the SS region in these genes whether Runx2-OE or not. Further HOMER scanning of CTCF CUT&Tag for motif analysis revealed that Runt family-related transcription factors, including Runx1, Runx2, Runx, and Runnx-AML, showed high occurrence frequencies (9.70–13.38%, as shown in Fig. 5D2).
To verify the existence of Lpl SS after Runx2-OE, we employed a dual-luciferase reporter assay. Different luciferase reporter plasmids were cotransfected with either the Runx2-OE plasmid or the control plasmid (pcDNA3.1) into 293 T cells (Fig. 5E1). Statistical analysis of the luciferase reporter system data via the Shapiro‒Wilk test indicated that the data satisfied the assumption of normality (p > 0.05). However, the Brown–Forsythe test suggested that the assumption of homogeneity of variances was not met (p > 0.05). The results revealed that the fluorescence intensity in the wild-type (WT) group was significantly lower than that in the empty vector group, suggesting that Runx2 can bind to the selected Lpl SS fragment and inhibit Lpl gene expression (Fig. 5E2).
Corrected one-way ANOVA of the WT, Mut1, and Mut2 groups revealed an F statistic of 185.4 with a corrected P < 0.0001. Pairwise comparisons revealed statistically significant differences between the WT and Mut1 groups and between the WT and Mut2 groups (P < 0.0001). Upon Runx2 overexpression, the WT group presented a marked decrease in fluorescence intensity compared with that in the control group, indicating a strong inhibitory interaction between Runx2 and the Lpl SS region in the WT construct. Interestingly, when the predicted binding sites within the SS were mutated, both the Mut1 and Mut2 groups presented reduced fluorescence intensity compared with their respective controls following Runx2 overexpression. However, this reduction was less pronounced than that in the WT group, suggesting that Mut1 and Mut2 retain some capacity to interact with Runx2, albeit with diminished efficacy (Fig. 5E3).
These findings further corroborate that Runx2 suppresses Lpl expression by promoting the formation of an SS in the Lpl transcriptional regulatory region and binding to this SS. Moreover, through IGV observation of Tmem132e and Lpl (Figs. 4e and 5D1), high CTCF signals were detected near the SS regions of target genes, but these signal peaks showed little change before and after Runx2-OE. We hypothesize that, after Runx2-OE, the three-dimensional structure of the SS regulatory region is altered, resulting in shortened distances between CTCF and regulatory regions, further suppressing the expression of target genes.
Discussion
The Runt-related transcription factor (Runx) family includes Runx1, Runx2, and Runx3, all sharing a highly homologous Runt-associated DNA-binding domain [3]. Among them, Runx2 (also termed Osf2/Cbfa1, Pebp2αA, or AML-3) acts as a transcription factor critically involved in tumorigenesis, metastasis, and invasion [18], as well as in osteogenic and chondrogenic differentiation/maturation [19]. It also modulates adipogenic and epithelial differentiation processes in adipose-derived stem cells [20]. Owens et al. demonstrated that Runx2 overexpression induces epithelial–mesenchymal transition (EMT) and suppresses differentiation in mammary cells, whereas Runx2 knockout inhibits breast cancer cell invasion/migration and improves survival rates [21], underscoring its central role in cellular proliferation and differentiation. Notably, Fang et al. [22] identified Runx2 as a master regulator of fibrotic gene expression in pulmonary fibrosis. Conditional Runx2 knockout reduces pathological fibroblast activation, suppresses extracellular matrix (ECM) deposition, and mitigates fibrotic progression, highlighting its pivotal role in alveolar fibroblast phenotypic switching [22]. In the central nervous system (CNS), Runx2 is predominantly expressed in gliomas and astrocytes [23, 24]. Tiwari et al. further revealed that Runx2 regulates astrocyte differentiation and promotes cellular maturation [24]. Kalinski et al. performed scRNA-seq on mouse sciatic nerve injury models and found that, 3 days post-injury, SCs mainly divided into three types: SC1, SC2, and SC3. Among these, SC3 highly expresses transcription regulators including Sox4, Runx2, Hmga1, Jun, and the POU family member Pou3f1 [25], but they did not conduct further research on Runx2.
Runx2 influences nerve injury repair by regulating SC phenotype and metabolic processes
Ding et al. found that Runx2 could stimulate SC differentiation and enhance SC proliferation and migration ability through activating the Akt/GSk3β pathway, ultimately promoting nerve regeneration [5]. Hu et al. demonstrated using a SC-specific Runx2 knockout mouse model with sciatic nerve injury that Runx2 deletion impaired SC migration and caused myelination deficits [7]. While this study established that Runx2 is essential for SCs remyelination, it did not comprehensively elucidate its regulatory mechanisms, particularly within epigenetic contexts. Hung et al. performed chromatin immunoprecipitation sequencing (ChIP-seq) analysis on rat peripheral nerve injury models and found that activated c-Jun binds to Runx2 enhancer sequences, inducing Runx2 upregulation and participating in regulation of SC activation and myelin breakdown and reorganization [8].
We classified reprogrammed SCs post-injury into three subtypes: He SCs (proliferative SCs), HZ SCs (transition SCs), and Zhu SCs (promyelinating-like SCs). We found that Runx2 expression was first detected in Zhu SCs, corresponding to the earliest appearance of Zhu1 SCs in the pseudotime trajectory. Subsequently, Runx2 expression in SCs gradually decreased as He SCs emerged. By day 60, Runx2 expression in Zhu SCs had declined to levels comparable to other SC subtypes. Eventually, all SCs differentiated into either myelin SCs or Remak SCs [3]. This manifestation indicates that Runx2 is an important molecule in maintaining the phenotype of Zhu1 SCs, or in other words, Runx2 is a marker that distinguishes Zhu1 SCs (Runx2 positive) from Zhu2 SCs (Runx2 negative).
In this experiment, we found extensive overlap between SCs marker S100 and Runx2 staining regions. Additionally, 72 h after infecting rat primary SCs with Runx2 overexpression adenovirus (Runx2-OE) and corresponding control virus (EGFP), Runx2-OE SCs showed notably flatter cell bodies compared with controls, with retracted processes at both ends and disappearance of intercellular network structures. Smart-seq mRNA sequencing analysis of differentially expressed genes revealed that genes differentially expressed in Runx2-OE SCs were mainly related to lipid metabolism and inflammatory response, while key transcription factor (Runx2 and CTCF) target genes were enriched in signaling pathways regulating pluripotency of stem cells. Furthermore, inhibitory histone modification signals (H3K27me3) after OE were enriched in regulatory regions of lipid metabolism-related target genes, further suggesting that Runx2 might influence SC fate by regulating cellular lipid metabolism processes (Table 1).
Runx2 can downregulate lipoprotein lipase (Lpl) in SCs through the formation of a super-silencer
There are usually two different regions in gene sequences, namely activation regions and suppression regions, which are jointly involved in regulating the expression of cell identity-related genes [15]. Activated regions associated with cell self-renewal are often regulated by super-enhancers (SEs). SEs are regions enriched in H3K27ac or transcription factors, which can greatly increase the expression of target genes [26]. Conversely, cell lineage-specific genes are often associated with repressive transcription elements, particularly in the CpG regions of chromosomal structure [27]. CpG regions are often enriched in H3K27me3 and the catalytic group of the PRC2 complex [28]. Huang et al. reported that regions enriched in H3K27me3 act as silencers, which could function through the connection between the H3K27me3-DNaseI high-response region and gene expression [29]. Cai et al. defined regions enriched in H3K27me3 as an SS via methods that search for SEs. They also reported that, unlike conventional silencers, SSs interact with chromatin through loop formation, inhibiting the function of long segments of chromatin [15]. CUT&Tag sequencing analysis of H3K27me3 in Runx2-OE cells indicated that Runx2 affects the proportion of genes undergoing histone modification.
In accordance with Cai’s method [15], we investigated the results of CUT&Tag sequencing analysis of H3K27me3 in two groups of cells via ReSE analysis and found 756 SSs in SCs after Runx2 upregulation. The proportion of SSs around the promoter was 17.17% in EGFP cells and 12.43% in Runx2-OE SCs. GO enrichment and KEGG pathway analyses revealed that the genes with SSs after Runx2 overexpression participate in synapse assembly, motor organ morphology, cell fate, and other biological functions. These genes are expressed in the presynaptic/postsynaptic membrane and affect the function of SCs through the stem cell pluripotency pathway, the WNT signaling pathway, the cAMP signaling pathway, and neuroactive ligand‒receptor interactions. Considering that SS primarily functions by affecting gene expression, we performed intersection analysis between SS that only appeared in Runx2-OE cells and downregulated genes after OE, ultimately establishing a Runx2-dependent super-silencer landscape in SCs. This landscape primarily includes six genes: Lpl, Tlr2, Tmem132e, Mylk3, Xkr6, and Thbd. Through expanded PPI network analysis, we found that Lpl, Thbd, and Tlr2 could form an interaction network. Moreover, Lpl is connected to both cellular cholesterol metabolism and glycolipid metabolism processes, suggesting that changes in Lpl expression would alter SC functional states by affecting lipid metabolism.
In the early stage of peripheral nerve injury, Runx2 reduces Lpl expression in Zhu1 SCs, affecting lipid metabolism
Visualization of the Lpl gene revealed the presence of SS regions in its transcription zone, and certain regions (1772–1786 and 1107–1121) were found to bind with Runx2 to induce a response. After the overall SS region and these two segments were knocked out, the inhibitory effect of Runx2 on Lpl expression was significantly weakened, further confirming that Runx2 can interfere with Lpl expression by forming a SS.
Studies have shown that myelin lipid metabolism can provide energy for starved axons, highlighting the critical role of lipid metabolism in the nervous system [30, 31]. Lipoprotein lipase (LPL), a key hydrolase, plays a central role in lipid metabolism and energy homeostasis [32]. Rothe et al. demonstrated that LDL receptor-mediated pathways internalize intraneural lipoproteins into SCs and peripheral neurons, repurposing cholesterol/cholesterol esters (CEs) from myelin degradation for membrane biosynthesis during remyelination and axonal regrowth. Following peripheral nerve injury, the breakdown of axons and myelin sheaths releases large amounts of cholesterol, CEs, and sphingomyelin. Therefore, during the early phase of injury, SCs need to focus on processing these lipids, rather than exogenous lipids [33]. For this reason, The expression of Lpl, which processes exogenous lipids, decreases in Zhu1 SCs during the early stage of injury.
In the later phase of peripheral nerve injury, when degraded myelin-derived lipids are depleted, exogenous lipid supplementation becomes necessary. Huey et al. proposed that LPL-mediated hydrolysis of exogenous triglycerides (TGs) serves as a major source of free fatty acids (FFAs) for SCs, likely playing a pivotal role in peripheral nervous system myelination. As a TG-specific hydrolase, LPL catabolizes TGs within lipoproteins (e.g., chylomicrons, very low-density lipoprotein(VLDL)) to generate FFAs and monoacylglycerols, thereby regulating local fatty acid availability to support late-stage myelin biosynthesis and repair [34, 35]. Therefore, Lpl expression increases in SCs during the late stage of injury.
In summary, after peripheral nerve injury, Zhu1 SCs, with high expression of Runx2, preferentially utilize cholesterol/CEs derived from myelin breakdown during the early phase by downregulating Lpl expression to minimize exogenous lipid utilization. By inhibiting the expression of Lpl, Runx2 can weaken the ability of SCs to metabolize lipids, promoting Zhu1 SCs to better adapt to the characteristics of lipid metabolism after engulfing disintegrated myelin sheaths in the early stage of injury.
Runx2 binding to SS may induce three-dimensional structural changes in target gene regulatory regions, downregulating target genes through CTCF-assisted regulation
From the visualization of Tmem132e and Lpl (Figs. 4E and 5D1), we observed high CTCF signals near the SS regions of target genes. These signals showed little change in peak values before and after Runx2-OE. We hypothesized that, after Runx2 binds to SS, it may alter the three-dimensional structure of the regulatory region, leading to shortening distances between CTCF and regulatory regions, thereby assisting in the suppression of target gene expression.
CCCTC-binding factor (CTCF) is a nuclear DNA-binding protein first discovered by Lobanenkov et al. in 1990. CTCF negatively regulates chicken c-myc expression through interactions with three regularly spaced repeats of the CCCTC DNA motif [36]. CTCF is a multivalent DNA-binding protein containing 11 zinc fingers and is universally expressed in most vertebrate tissues [37]. CTCF can bind to tens of thousands of conserved DNA sites, with over 30% located in intronic regions and over 50% in intergenic regions [38]. Conventional CTCF target sites are highly conserved and mainly located in intergenic regions, while cell type-specific sites are primarily found in introns [39]. CTCF has also been identified as an RNA-binding protein, with CTCF–RNA interactions involving CTCF dimerization, long-range genomic binding sites, chromatin looping, and gene regulation [40]. CTCF can also regulate various post-translational modification processes by affecting transcription factors, chromatin remodeling factors, methylation regulators, histone modification factors, and splicing factors [41]. Owing to its unique role in chromatin architecture, CTCF is known as the “genome architect,” capable of regulating the expression of various molecules and thereby affecting their epigenetic functions [42]. CTCF can act as a transcription factor directly activating or suppressing target gene expression, or as a chromatin insulator indirectly interfering with enhancer/silencer–promoter contacts [43]. Furthermore, it can regulate gene expression by mediating chromatin looping and altering the spatial distance between genomic sites [44–46].
However, these hypotheses regarding whether CTCF plays a role and whether there is a synergistic effect between its function and SS still require more experimental validation, such as CTCF knockout experiments and chromosome conformation capture (3C) or high-throughput chromosome conformation capture (HiC) experiments in Runx2-OE SCs to examine chromatin structure. If hypotheses are confirmed, they would provide deeper insights into the epigenetic regulatory mechanisms in SC metabolism and enable precise regulation of SC functions in response to age and environmental changes in the future, promoting rapid SC response after nerve injury and timely adjustment to facilitate rapid peripheral nerve regeneration.
Conclusions
In the early stage after peripheral nerve injury, Runx2 is upregulated in SCs, causing the phenotypic conversion of SCs to Zhu1 SCs, and establishes a super-silencer (SS) within the transcriptional regulatory region of Lpl, leading to the downregulation of Lpl expression.
Supplementary Information
Additional file 1