Single-cell analysis reveals the lncRNA-MEG3/miRNA-133a-3p/PRRT2 axis regulates skeletal muscle regeneration and myogenesis

Skeletal muscle is the largest motor and metabolic organ of the body, which has a robust capacity for regeneration following injury or disease. Delayed regeneration after skeletal muscle injury reduces muscle contractility and leads to dysfunction of innervation. Therefore, identifying the regulation components in skeletal muscle regeneration and determining their molecular mechanisms are important to discover novel therapeutic markers for muscular diseases. Long non-coding RNA (LncRNA) has been implicated in skeletal muscle regeneration. Recent developed singlecell RNA sequencing (scRNA-seq) provides a higher resolution of cellular differences than bulk RNA-seq. Here, we reanalyzed single-cell transcriptomes data of skeletal muscle regeneration and identified lncRNA maternally expressed gene 3 (lncRNA-MEG3) was highly expressed in muscle satellite cells (MuSCs). Further study showed that lncRNAMEG3 regulates skeletal muscle regeneration via sponging miR-133a-3p to regulate proline-rich transmembrane protein 2 (PRRT2) expression level. These results suggested that lncRNA-MEG3 might be a potential target for skeletal muscle diseases. To identify critical lncRNAs associated with muscle regeneration at single-cell level, we re-analyzed scRNA-seq dataset generated by De Micheli (Fig. S1A). By combining the scRNA-seq atlas (Fig. S1B) and histology of regenerating muscle (Fig. S1C), we discovered that day 5 post-injury was a critical time point for muscle regeneration. Subsequently, we investigated the expression of lncRNAs on day 5 following injury. 764 lncRNAs were expressed in at least one cell (Fig. S1D) and six lncRNAs highly expressed in MuSCs, myofibroblasts and fibro/adipogenic progenitors (FAPs) with lncRNA-MEG3 exhibited the greatest abundance in

Skeletal muscle is the largest motor and metabolic organ of the body, which has a robust capacity for regeneration following injury or disease. Delayed regeneration after skeletal muscle injury reduces muscle contractility and leads to dysfunction of innervation. Therefore, identifying the regulation components in skeletal muscle regeneration and determining their molecular mechanisms are important to discover novel therapeutic markers for muscular diseases. Long non-coding RNA (LncRNA) has been implicated in skeletal muscle regeneration. Recent developed singlecell RNA sequencing (scRNA-seq) provides a higher resolution of cellular differences than bulk RNA-seq. Here, we reanalyzed single-cell transcriptomes data of skeletal muscle regeneration and identified lncRNA maternally expressed gene 3 (lncRNA-MEG3) was highly expressed in muscle satellite cells (MuSCs). Further study showed that lncRNA-MEG3 regulates skeletal muscle regeneration via sponging miR-133a-3p to regulate proline-rich transmembrane protein 2 (PRRT2) expression level. These results suggested that lncRNA-MEG3 might be a potential target for skeletal muscle diseases.
To identify critical lncRNAs associated with muscle regeneration at single-cell level, we re-analyzed scRNA-seq dataset generated by De Micheli (Fig. S1A). 1 By combining the scRNA-seq atlas (Fig. S1B) and histology of regenerating muscle (Fig. S1C), we discovered that day 5 post-injury was a critical time point for muscle regeneration. Subsequently, we investigated the expression of lncRNAs on day 5 following injury. 764 lncRNAs were expressed in at least one cell (Fig. S1D) and six lncRNAs highly expressed in MuSCs, myofibroblasts and fibro/adipogenic progenitors (FAPs) with lncRNA-MEG3 exhibited the greatest abundance in MuSCs (Fig. 1A). For confirmation, we re-analyzed other scRNA-seq datasets for skeletal muscle 2,3 and obtained consistent results (Fig. S1E). During muscle regeneration, lncRNA-MEG3 expression was induced on day 3 and peaked on day 5 post injury (Fig. 1B). Interestingly, based on reanalysis of scRNA-seq data published by He et al, 4 we found that lncRNA-MEG3 was upregulated in MuSCs transitioned from quiescent to differentiated during embryonic development (Fig. S1F, G). Real time quantitative PCR (RT-qPCR) results showed that lncRNA-MEG3 was abundantly expressed in skeletal muscle at postnatal day 0 and downregulated from postnatal day 0 to day 65 (Fig. S1H, I). Together, these findings suggested that lncRNA-MEG3 might be a potential regulator of skeletal muscle regeneration.
Next, a regeneration model by tibialis anterior (TA) muscle injection was used to determine the role of lncRNA-MEG3 during skeletal muscle regeneration in vivo (Fig.  S2AeC). In the groups of lncRNA-MEG3 knockdown, the peak of newly formed myofibers appeared at 7 days post injury, which was 2 days delayed compared to the control group ( Fig. 1C; Fig. S2DeH). Single-cell transcriptome analysis showed that immune cells were increased in the lncRNA-MEG3 knock-down group at 5 days post injury ( Fig.  S2IeK) further indicating that lncRNA-MEG3 knockdown delayed the kinetics of skeletal muscle regeneration.
Given skeletal muscle regeneration relies on the skeletal muscle satellite cell myogenesis, we used functional gain and loss to study the effects of lncRNA-MEG3 on primary myoblasts (Fig. S3). The cell-counting-kit-8 (CCK-8) assay, 5-ethynyl-2 0 -deoxyuridine (EdU)-staining, immunofluorescence (IF), RT-qPCR and Western blotting results showed that lncRNA-MEG3 knockdown significantly improved the C2C12 proliferation and inhibited differentiation, while cells treated with lncRNA-MEG3 overexpression vector showed the opposite phenomenon Peer review under responsibility of Chongqing Medical University.
The localization of lncRNA in the cell is assumed to be a marker for determining the regulatory mechanisms of lncRNA. 5 We found lncRNA-MEG3 is mainly expressed in cytoplasm of C2C12 myotube (Fig. 1G). So, we speculated that the lncRNA-MEG3 might act as ceRNA in myogenesis. Microarray profiling of lncRNAs and miRNAs expression in TA muscle from day 0 to day 65 were performed ( Fig. S6AeD; Table S1, S2) to identify miRNAs interact with lncRNA-MEG3. Accordingly, miR-133a-3p was selected from four myomiRs using RNA immunoprecipitation (RIP)-qPCR (Fig. 1H). This was further confirmed by RNA antisense purification (RAP)-qPCR (Fig. 1I). Subsequently, the luciferase activity of pmirGLO-lncRNA-MEG3-WT was declined after miR-133a-3p over-expression, which was rescued by increasing the concentration of lncRNA-MEG3 (Fig. S6EeG). Furthermore, the expression level changes of miR-133a-3p was opposite with lncRNA-MEG3 during skeletal muscle development, C2C12 myoblasts differentiation, and regeneration (Fig. S6HeJ). To further determine the regulatory relationship, we subsequently conducted the rescue experiments. The results confirmed that lncRNA-MEG3 repressed the effect of miR-133a-3p on C2C12 proliferation and differentiation (Fig. S7, S8).
To identify genes that were involved in the ceRNA network, we performed microarray profiling of the mRNA transcriptome in lncRNA-MEG3 knockdown C2C12 and skeletal muscle at postnatal day 0 and day 65 ( Fig. S9AeC and Table S3e5). Interestingly, only PRRT2 was identified after bioinformatic analysis (Fig. 1J). In addition, the expression levels of PRRT2 and lncRNA-MEG3 were positively correlated during skeletal muscle development and C2C12 differentiation, and the PRRT2 mRNA expression level was upregulated after lncRNA-MEG3 overexpression (Fig. S9DeH). Further experiments showed that PRRT2 is a target gene of miR-133a-3p and it could inhibit C2C12 proliferation and promote differentiation (Fig. S9IeL, S10, 11).

Ethics declaration
All animal procedures were performed according to the guidelines of Institutional Animal Care and Use Committee of Agricultural Genomics Institute at Shenzhen, Chinese Academy of Agricultural Sciences.

Author contributions
Zhonglin Tang designed the study and revised the manuscript. Zishuai Wang and Yilong Yao wrote the manuscript. Yilong Yao, Zishuai Wang and Yun Chen performed molecular and cellular experiments; Liyuan Wang and Yun Chen performed the single cell assay; Lei Liu and Zishuai Wang performed single cell RNA-seq analysis; Yun Chen helped with cell culture, cell transfection, qPCR, vector construction, EdU-staining and western blotting. Guoqiang Yi, Yalan Yang, and Dazhi Wang helped revise the manuscript.

Conflict of interests
The author(s) declare that they have no conflict of interest. The mechanism graph of the lncRNA-MEG3 regulatory network. Data are expressed as mean values AE SEM, and a paired two-tailed Student's t-test was used to analyze the statistical significance between two groups. **P < 0.01, and *P < 0.05.

Consent for publication
All authors have agreed to publish this manuscript.