SNAI1 is upregulated during muscle regeneration and represses FGF21 and ATF3 expression by directly binding their promoters

During skeletal myogenesis, the zinc‐finger transcription factors SNAI1 and SNAI2, are expressed in proliferating myoblasts and regulate the transition to terminally differentiated myotubes while repressing pro‐differentiation genes. Here, we demonstrate that SNAI1 is upregulated in vivo during the early phase of muscle regeneration induced by bupivacaine injury. Using shRNA‐mediated gene silencing in C2C12 myoblasts and whole‐transcriptome microarray analysis, we identified a collection of genes belonging to the endoplasmic reticulum (ER) stress pathway whose expression, induced by myogenic differentiation, was upregulated in absence of SNAI1. Among these, key ER stress genes, such as Atf3, Ddit3/Chop, Hspa5/Bip, and Fgf21, a myokine involved in muscle differentiation, were strongly upregulated. Furthermore, by promoter mutant analysis and Chromatin immune precipitation assay, we demonstrated that SNAI1 represses Fgf21 and Atf3 in proliferating myoblasts by directly binding to multiple E boxes in their respective promoter regions. Together, these data describe a new regulatory mechanism of myogenic differentiation involving the direct repressive action of SNAI1 on ER stress and Fgf21 expression, ultimately contributing to maintaining the proliferative and undifferentiated state of myoblasts.


| INTRODUCTION
Skeletal muscle development is a complex process that involves myoblast proliferation as well as myofibers differentiation. A molecular switch involving various actors, including myogenic differentiation antigen (MyoD), SNAI1, and SNAI2, regulates the transition from proliferating myoblasts to terminally differentiated myotubes. Members of the SNAI family are zincfinger transcription factors that recognize and bind to E boxes (5′-CANNTG-3′) with a C/G-rich central dinucleotide and act as transcriptional repressors by recruiting HDAC1/2. 1-3 SNAI1 and 2 are expressed in proliferating myoblasts and rapidly turned off as muscle differentiation proceeds. It has been demonstrated that in proliferating myoblasts, the SNAI1-HDAC1/2 repressive complex is in competition with MyoD, a key regulator of myogenesis, for binding to a subset of G/Crich differentiation-specific E boxes, thereby preventing the MyoD occupancy on these regulatory elements and consequently blocking terminal differentiation. 4 Fibroblast Growth Factor 21 (FGF21) is an endocrine hormone expressed in numerous tissues including liver, brown adipose tissue, white adipose tissue, and pancreas. 5 Although the liver is generally considered the main site of FGF21 production, FGF21 may be considered a myokine given that many studies indicate that skeletal muscle can serve as a relevant source of its production, especially in response to insulin stimulation. 6 Moreover, it has been demonstrated that MyoD binds directly to the promoter region of Fgf21, leading to the activation of Fgf21 expression in differentiating mouse C2C12 myoblasts. 7,8 In this cellular context, FGF21 induces the expression of early myogenic genes such as MyoD, MYF4, and MYH1, as demonstrated by overexpression and silencing of FGF21. Moreover, overexpression of FGF21 promotes myoblast cell cycle exit and enhances myogenic differentiation as shown by increased fusion index and myotube length, whereas its negative modulation attenuates these processes. 8,9 In line with these observations, transgenic mice with muscle-specific FGF21 overexpression show decreased muscle mass while both mRNA and protein expression levels of FGF21 are dramatically increased in dystrophic skeletal muscle, underlining its role in muscle development and regeneration in vivo. 8,10 Independent of these transcriptional events, the endoplasmic reticulum (ER) stress, and ER stress-induced UPR occur transiently during myoblast terminal differentiation and myofiber formation, where they play a critical role in myogenesis. [11][12][13] ER is a specialized organelle required for its crucial role in the synthesis, assembly, folding, routing, and degradation of a large number of proteins and in Ca 2+ storage. In response to ER redox imbalance or disruption of ER Ca 2+ homeostasis, misfolded or unfolded proteins accumulate in the ER, 14 thereby triggering an ER stress response, termed the unfolded protein response (UPR), that serves as a compensatory process that aims to restore ER homeostasis in order to preserve cellular functions and survival.
Activation of UPR pathways is induced by three ER transmembrane sensors: PERK, inositol-requiring protein 1 (IRE1), and activating transcription factor 6 (ATF6). [15][16][17] In the context of muscle differentiation, IRE1α stimulates skeletal muscle regeneration via a RIDD (regulated IRE1dependent decay)-mediated decrease in myostatin concentration and impairment of myoblast differentiation. 13 In the early stage of myoblast differentiation, the PERK-ATF4 signaling pathway transactivates muscle differentiationassociated miRNAs, such as miR-128 which promotes myoblast differentiation and inhibits proliferation. 18,19 However, the PERK-ATF4 axis, through CHOP/DDIT3 induction, can also repress MyoD transcription, thereby delaying myoblast differentiation. 20 Lastly, ATF6, but not other ER stress sensors, is specifically activated during apoptosis in myoblasts and promotes myoblast differentiation by enhancing apoptosis of fusion-incompetent cells. 11,12 To better characterize the role of SNAI1 in myogenesis, we studied the expression of SNAI1 in mouse tibialis anterior muscle following bupivacaine-induced muscle degeneration and regeneration and we identified putative SNAI1 target genes by microarray expression analysis of SNAI1short hairpin (sh) RNA-silenced C2C12 myoblasts. These studies led to our discovery that SNAI1 represses the ER stress/UPR pathway as well as the expression of the prodifferentiation factor FGF21. Furthermore, by promoter mutation analysis and ChIP, we identified the binding sites for SNAI1 in the promoters of both FGF21 and the ER stress the direct repressive action of SNAI1 on ER stress and Fgf21 expression, ultimately contributing to maintaining the proliferative and undifferentiated state of myoblasts.

K E Y W O R D S
E box, endoplasmic reticulum stress, gene silencing, microarray analysis, muscle regeneration, myoblast differentiation, promoter, skeletal muscle, transcription marker Activating Transcription Factor 3 (ATF3) genes. Together, our data describe a new regulatory mechanism of myogenic differentiation involving the direct repressive action of SNAI1 on ER stress and FGF21 expression, ultimately contributing to maintain the proliferative and undifferentiated states that characterize myoblasts.

| Animals and in vivo assays
Animal experiments were performed according to the guidelines for animal experimentation and welfare provided by Padua University and approved by the Padua University ethics committee. For the in vivo analysis, three-month old mice received a pre-emptive dose of the analgesic Carprofen and then were anesthetized with Isoflurane. Skin adjacent to the tibial anterior muscle was shaved, wiped clear of debris with sterile water, and sterilized with alternating scrubs of Iodine/Betadine and alcohol three times. Briefly, 25 μl of bupivacaine 0.5% w/v were injected into the tibialis anterior muscles to induce acute skeletal muscle regeneration through a single intramuscular injection. 21 Following bupivacaine injection, mice were sacrificed at different time points and the tibialis anterior muscles were dissected, frozen, and processed for further analysis.

| Muscle embedding and cryosectioning
After dissection, skeletal tibialis anterior muscles were embedded with a minimum amount of Tissue-Tek O.C.T (Sakura Finetek, USA). The embedded muscles were frozen by placing them into the cooled 2-methylbutane for 5 min and then the muscle samples were transferred to a −80°C freezer for storage. Ten microliters-thick cryosections were processed for Hematoxylin and Eosin staining.

| RNA extraction and RT-qPCR
Total RNA was extracted from C2C12 cells using RNeasy® Plus Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer's instruction. Total RNAs were quantified by QIAxpert system spectrophotometer by measuring the UV/VIS absorption spectrum and used to evaluate the gene expression by means of Brilliant III Ultra-Fast SYBR Green RT-qPCR Master Mix (Agilent Technologies, Santa Clara, CA, USA) was used. Primers for RT-qPCR of mouse genes are shown in Table 1.

| Western blot analysis
Total cell extracts were prepared as previously described. 24 The protein concentration in each sample was measured using Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA). Primary and secondary antibodies used for immunoblotting are reported in Table 2.

| Microarray analysis
RNA was exctracted and purified as previously described. 25 Microarray analysis was performed on the Illumina Platform (Illumina, San Diego, CA, USA) and the results were analyzed using the BeadStudio Gene ExpressionModule (GX). Data were background adjusted and quartile normalized using default parameters in the Genome BeadStudio Software. Probes with p-value < .05 in all the three samples of at least one condition were selected for downstream analysis. Differentially expressed (DE) probes were calculated on the Galaxy website (usega laxy.org) using the Limma package (bioconductor-limma v.3.48.0) with the limma-voom method and Benjamin and Hochberg adjusted p-value (adjP) calculation. Differentially expressed probes were defined as adjP < .05 and |logFC| > 0.5. Volcano Plot and Heatmaps were generated in Python v3.4.5.

| Plasmids and site-directed mutagenesis
Based on the mouse Fgf21 promoter sequence, specific primers were designed (Table 3) to amplify a full-length promoter, spanning from −1764 bp with respect to the transcription start site to +166 bp, and shorter promoter fragments progressively excluding the putative binding sites for SNAI1, previously identified by bioinformatic analysis. In total, four promoter fragments were amplified by PCR from Mus musculus genome using the Q5® High- Fidelity Table 5 for primer sequences). After the amplification, the PCR products were cloned into pCR®-Blunt II-TOPO®, using the Zero Blunt® TOPO® PCR Cloning kit (Invitrogen-Thermo Fisher Scientific, Waltham, MA, USA handbook. After 24 h, the GM was replaced with DM, and the day after, the cells were harvested in Passive Lysis Buffer (Promega Corp., Madison, WI) and the luciferase activity was measured by the Nano Dual-luciferase report™ assay system (Promega Corp., Madison, WI), following the manufacturer's instructions. Luciferase activity was normalized to the activity of the Renilla luciferase as an internal control. C2C12 transfection by electroporation was used to overexpress SNAI1 in order to analyze Fgf21 and Atf3 expression. C2C12 cells were trypsinized and 2.0 × 10 6 cells·point −1 were centrifuged three times for 5 min, at 190× g with DMEM/F12 supplemented with 2.5% (v/v) FBS and 0.25% (w/v) bovine serum albumin (BSA, Sigma-Aldrich, Milan, Italy). Then, the cells were resuspended in 200 μl of the same medium, and 10 μg of DNA (pCMV6-entry vector or pcMV6-SNAI1) were added. Each mix was put into an electroporation cuvette (Gene Pulser electroporation cuvette, Gap Width 0.2 cm, Bio-Rad; Hercules, CA, USA). Electroporation was performed using the following parameters: 290 V, 1.000 μF, 200 Ω. After the electroporation, the cuvette content was resuspended with pre-warmed complete C2C12 medium and plated in 6-well dishes. After 24 h the GM was replaced with fresh medium and again after 1 day with DM for the other 24 h before luciferase assay.

| ChIP-qPCR
For the ChIP-qPCR analysis, C2C12 cells transfected with pCMV6 or pCMV6-SNAI1 by electroporation as above and the CUT&RUN Assay Kit (Cell Signaling, Danvers, MA, USA) was used following the manufacturer instructions. For the immunoprecipitation, we used the anti-FLAG M2 antibody. Following targeted chromatin digestion by the enzyme pAG-MNase, the reaction was stopped with the Stop Buffer containing digitonin, RNAse A, and Spike-in DNA. The sample normalization spike-in DNA is fragmented genomic DNA from the yeast S. cerevisiae that facilitates normalization between samples and between experiments during qPCR analysis.
In the end, input and enriched chromatin samples were collected with the phenol/chloroform extraction followed by ethanol precipitation. The purified, enriched DNA was quantified by qPCR with QuantiTec SYBR Green PCR Master Mix (Qiagen, Hilden, Germany), following the manufacturer's instructions. Oligonucleotides users for qPCR are listed in Table 6. qPCR amplification reaction of spike-in DNA was performed for sample normalization and was analyzed using the Percent Input Method. 26 Signals obtained from each immunoprecipitation were expressed as a percent of the total input chromatin.

| Snai1 silencing
For the RNA interference-mediated knockdown of SNAI1, we used the lentiviral transfer plasmid pLKO.1 from the TRC shRNA library (Open Biosystems, Huntsville, AL) expressing specific shRNA for mouse SNAI1 (Table 7)

| Statistical analysis
The data analysis was performed using Prism 9 statistical software (GraphPad Software Inc., San Diego, CA). Significance was estimated by one-way or two-way ANOVA. The significant differences were estimated using Tukey's multiple comparison test (*p < .05; **p < .01; ***p < .001; ****p < .0001, n.s., not significant). Student's t test was used to confirm significant differences between treatments. Two-tailed probabilities of less than .05 were considered significant.

| SNAI1 is upregulated during early stages of skeletal muscle regeneration and in proliferating C2C12 myoblasts
While SNAI1 has been reported to be expressed in proliferating myoblasts in vitro, 4 we sought to evaluate the expression profile of SNAI1 during myogenesis in vivo. As such, a mouse model of myotoxin-induced muscle degeneration was used, i.e., bupivacaine-induced muscle injury, that serves as a useful in vivo experimental model for studying factors involved in muscle differentiation and muscle plastic adaptation to functional demand. 21 In this model, regeneration requires the activation of undifferentiated satellite cells, which proliferate, differentiate into myoblasts expressing muscle-specific markers, fuse into myotubes, and finally mature into myofibers. 21 Briefly, bupivacaine was injected in the mouse tibialis anterior muscles, with tissues collected at different stages of muscle regeneration for mRNA and protein content analyses. After bupivacaine injection, proliferation of mononuclear cells, mainly inflammatory, and myogenic cells, was most active within 1-4 days of injection. 27 Myogenic cell differentiation and new myotube formation were observed ∼5-6 days post injection. Ten days after the injection, the overall architecture of the muscle was restored, although most regenerated myofibers were smaller and displayed central myonuclei ( Figure 1A). SNAI1 mRNA and protein were induced early in regenerating muscle (2-4 days after the injury), corresponding with the maximum activation of satellite cells, as defined by MyoD expression (Figure 1B,C). SNAI1 mRNA expression rapidly decreased at day 6, coinciding with the beginning of the myotubes differentiation program, as defined by the increased expression of embryonic skeletal muscle myosin heavy chain (MYH3) and concurrent decline of MyoD expression, supporting a potential role for SNAI1 during the early stages of muscle regeneration. Similarly, SNAI2 was maximally induced at day 3 after the injury, but its expression was maintained up to day 10 ( Figure 1B). In primary skeletal myoblasts, SNAI1 is known to repress genes upregulated during muscle terminal differentiation with SNAI1 knockdown leading to the precocious cell cycle withdrawal and terminal differentiation. 4 To define the role of SNAI1 during earlier stages of myogenic differentiation, we used the C2C12 cell line as a widely used and wellestablished in vitro model for studying mammalian muscle differentiation. Additionally, in the presence of growth factors, these cells are considerably less prone to spontaneous differentiation than primary myoblasts, a feature that optimizes their use as a model to study the transcriptome of proliferating myoblasts. This is confirmed by the lower expression levels of p21 (CIP1/WAF1), a key inducer of cellcycle arrest, and myogenin (MYF4) under growth conditions ( Figure 1D,E). SNAI1 was expressed in proliferating C2C12 cells maintained in a growth medium and was rapidly downregulated after switching to a low-serum differentiation medium ( Figure 1D), showing an expression profile that resembles that obtained with primary myoblasts ( Figure 1E). 4 SNAI1 mRNA expression profile did not correlate with the protein levels ( Figure 1F), however, suggesting a posttranslational regulatory mechanism underlying its downregulation during myoblasts-to-myotube differentiation. Indeed SNAI1 is a short-lived protein since it is rapidly ubiquitinated and degraded by the 26S proteasome system. Its degradation is triggered by phosphorylation, such as that mediated by protein kinase 1D (PKD1), a kinase rapidly upregulated during myoblast differentiation or by GSK-3β. 28,29 Moreover, with regard to discrepancies between SNAI1 mRNA and protein expression, SNAI1 can also act as a repressor of its own promoter via a negative feedback loop. 30 Similarly, SNAI2 was expressed in proliferating myoblasts and rapidly downregulated after switching to a differentiation medium ( Figure 1E). Conversely, in C2C12 cells, the SNAI2 protein shows a different expression profile with an increased expression on day 1 after switching to a differentiation medium and a more persistent expression that could explain why this cell line is less prone to differentiation ( Figure 1D).

stress-and cell proliferation-related gene expression in C2C12 myoblasts
To identify genes regulated by SNAI1 in proliferating myoblasts, C2C12 cells were transduced with lentiviral particles for the expression of three different small hairpin RNAs (shRNA) in order to downregulate SNAI1 expression. A substantial reduction of SNAI1 protein level was observed in all three shRNAs (Figure 2A). When cultured in a growth medium, the SNAI1-silenced C2C12 cells did not show alterations in the proliferation rate or spontaneous differentiation to myotubes ( Figure S1). Microarray gene expression profile analysis was performed with mRNA extracted from C2C12 cells (Table S1) and 180 significantly downregulated and 218 significantly upregulated genes were found in all three SNAI1 knockdown cell lines as compared with control transduced cells (adjP < .05 and |logFC| > 0.5) (Table S2 and Figure 2B,C). To begin defining the biological categories affected by SNAI1 silencing, the differentially expressed genes (DEGs) were subjected to gene ontology (GO) enrichment analysis performed using the online Database for Annotation, Visualization, and Integrated Discovery (DAVID) analysis platform (Table S3). 31 To identify only the context pertinent and non-redundant enriched GO terms in biological processes, a selected list of significantly enriched terms (pvalue < .05) was plotted using REVIGO. 32 This analysis identified two main biological processes significantly enriched in this gene set both as p-value and number of DEG; endoplasmic reticulum (ER) stress (GO:0034976 and GO:0030968; ER stress and UPR are closely related processes) and positive regulation of cell proliferation (GO:0008284) ( Figure 2D-F). Changes in gene expression were validated by real-time quantitative PCR (RT-qPCR) for a selected group of genes, including Fgf21, Atf3, DNA damage-inducible transcript 3 protein (Ddit3), heat shock protein family A (Hsp70) member 5 (Hspa5/Bip), hepatocyte growth factor (Hgf), ectonucleotide pyrophosphatase/phosphodiesterase 2 (Enpp2), and plateletderived growth factor receptor alpha (Pdgfra) ( Figure 2G). As expected, DAVID analysis did not identify pathways involved in myoblast terminal differentiation, confirming that the ability of C2C12 to maintain an undifferentiated phenotype in high serum conditions is not dependent on SNAI1 expression even though the expression of a large group of genes involved in cell proliferation is perturbed.

| SNAI1 represses FGF21 and ATF3 expression in proliferating myoblasts by directly interacting with their gene promoters
As during myogenesis, SNAI1 acts as a transcriptional repressor, 4  genes in order to identify new target genes. Among two of the most upregulated transcripts were Fgf21 and Atf3. During muscle cell differentiation, Fgf21 is positively regulated by the transcription factor MyoD, thereby activating the expression of early myogenic genes, promoting cell cycle exit, and enhancing myogenic differentiation. 9 In turn, ATF3 is a well-known marker of ER stress activation, 33 a response required for early myoblast differentiation. Given SNAI1 expression during myogenesis and the effect of FGF21 and ER stress response on myogenic differentiation, we first examined their expression pattern during muscle regeneration in vivo. In bupivacainetreated muscle, both FGF21 and ATF3 mRNA and protein were strongly induced ( Figure 3A-D). In this experimental model, SNAI1 expression was partially overlapping with FGF21 and ATF3 expression. Nevertheless, these data may remain consistent with the repressive role of SNAI1 on these genes as myoblasts are only partially synchronized in their differentiation process. Indeed, when the expression of FGF21 and ATF3 mRNA were monitored in synchronized C2C12 cultures, following the switch to a low serum medium their expression was strongly induced, as opposed to the fall in SNAI1 protein expression as well as in mRNA expression of two of the most downregulated genes following SNAI1 silencing, Hgf, and Pdgfra ( Figure 3E-G). Moreover, ectopic overexpression of SNAI1 in differentiating myoblasts induced a consistent decrease of both FGF21 and ATF3 mRNA levels, thereby confirming the role of SNAI1 in their negative regulation ( Figure 4A-C).
In an effort to characterize the promoter regions that might be involved in this regulation, we performed a bioinformatic analysis that identified the potential E box binding sites for SNAI1 (CACCTG, CACGTG, and CAGCTG) in the mouse Fgf21 promoter region sequence. Three clusters of putative SNAI1 binding sites were centered on positions −1659 (cluster A), −1000 (cluster B), and +21 (cluster C) with respect to the transcriptional start site (TSS; +1). To determine the functional activity of these E boxes relative to SNAI1 binding, we generated luciferase reporter constructs carrying Fgf21 promoter regions of different lengths, spanning from position +166 to −1764, −1681, −1091, and −926 ( Figure 5A), and the reporter vectors were co-transfected together with pCMV6-SNAI1 expression vector or a pCMV6-empty control vector in C2C12 myoblasts. In this system, both the full length and each of the 3 other promoters with progressive 5′ deletions decreased their activity when SNAI1 was overexpressed, but the full length −1764 promoter proved to be the most responsive to the repression by SNAI1 (Figure 5B,D). To investigate which E boxes contained in cluster C are directly bound by SNAI1, we introduced single and multiple point mutations altering the nucleotide sequence of one, two, or all the three E boxes in the reporter vector harboring the −926 promoter. We observed that only the triple point-mutant was no longer responsive to SNAI1 repressive activity ( Figure 5C,D), suggesting that all three binding sites play a role in SNAI1 recruitment to the Fgf21 promoter. To further verify and characterize the direct binding of SNAI1 to the Fgf21 promoter, ChIP followed by qPCR was performed in C2C12 cells transiently transfected for the expression of Flag-tagged SNAI1. As shown, SNAI1 effectively binds to the Fgf21 promoter and the E box clusters interested are the cluster A and C, but not B. No significant signal increase was observed with primers annealing in exon 2 sequence that was used as a negative control ( Figure 5E). Similar analyses performed for the Atf3 promoter demonstrate the binding of SNAI1 to E boxes located at −1873 and between −1320 and −1221 with respect to the TSS ( Figure 6).
As FGF21 expression can be induced both by ER stress 34,35 and the transactivation action of MyoD, SNAI1 could serve to repress FGF21 expression mediated by ER stress stimuli other than those associated with myoblast differentiation programs per se. To test this hypothesis, C2C12 cells were treated with thapsigargin, a well-characterized ER stress-inducing agent, and FGF21 mRNA expression was measured at different time points. The results showed an induction of the level of FGF21 mRNA in cycling C2C12 cells as a consequence of the ER stress induction ( Figure 7A). Finally, C2C12 cells were transiently co-transfected with the −1764 Fgf21 promoter reporter plasmid and pCMV6-empty vector or pCMV6-SNAI1 to overexpress SNAI1. After 24 h from the transfection, cells were treated with thapsigargin. Measurement of luciferase activity showed that Fgf21 promoter activity was upregulated in response to the thapsigargin treatment and SNAI1 was able to repress this induction ( Figure 7B). Moreover, using specific inhibitors of the three ER stress pathways, we find that in differentiating myoblasts, FGF21 transcription is downstream of the IRE1 pathway, while ATF3 and DIDIT3/CHOP expression depend on the PERK-ATF4 axis ( Figure 7C-E). In addition, we analyzed the myoblast differentiation by means of western blot analysis of MYH3 expression in C2C12 cells treated with the ER stress inhibitors. As shown in Figure 7F, ATF6 and IRE1 pathways, but not the PERK-ATF4 axis, are involved in the promotion of myoblast differentiation.

| DISCUSSION
Skeletal muscles possess a unique ability to regenerate completely the following damage. After acute injury or trauma, degeneration is induced, followed by regeneration processes that involve different cell populations, including a resident population of muscle cell progenitors, i.e., satellite cells. In the absence of environmental stimuli, satellite cells that are located in a specialized niche between the sarcolemma and the basal lamina are maintained in a quiescent state. Following injury or in various disease states, satellite cells are activated, proliferate, and migrate to the damaged areas, where they eventually differentiate, giving rise to newly formed myofibers. 36 Our data show that in vivo, during the muscle degenerationregeneration process induced by bupivacaine, the transcriptional repressor, SNAI1, is upregulated, both at the protein and mRNA levels, at 3-4 days post-injury when quiescent satellite cells are activated and assume a proliferative state, confirming its possible involvement as a negative regulator in the muscle differentiation process. In this process, the MRFs such as MyoD, MYF4, MYF5, and MRF4 are later engaged where they play a crucial role in committing satellite cells to terminal differentiation. 37 To prevent premature differentiation, SNAI1 and SNAI2 compete with MyoD for the same binding sequences and act as terminal differentiation repressors by recruiting HDAC1/2 to target genes. 3,4 To gain further insight into the specific role that SNAI1 plays during myogenesis and to identify new SNAI1 target genes, we silenced SNAI1 in the C2C12 myoblast cell line. As opposed to primary myoblasts, 4 C2C12 cells do not terminally differentiate in the presence of growth factors, and as a consequence of SNAI1 knockdown, maintain a proliferative rate similar to that of control cells. As such, these cells serve as an optimal model to study SNAI1-dependent effects on the transcriptome of a pure population of proliferating myoblasts. Using three different shRNAs to avoid silencing artifacts, we identified almost 400 DEGs. The relatively low number of DEGs identified is explained by the selectivity of the analysis performed and also by the partially overlapping repressor activities of SNAI1 and SNAI2, which allowed us to identify the subset of genes regulated exclusively by SNAI1. In proliferating C2C12 cells, SNAI2 is likely sufficient to maintain the undifferentiated phenotype, a conclusion consistent with the more stable expression of SNAI2, compared with primary myoblasts ( Figure 1D,E), and the absence of enriched GO terms related to myoblast terminal differentiation after SNAI1 silencing.
ER stress, the UPR, and regulation of cell proliferation were the main biological processes identified as significantly enriched by DAVID analysis following SNAI1 targeting. Among cell proliferation-related genes, two of the most downregulated genes were Hgf and Pdgfra, whose expression during myoblasts differentiation was inversely related to SNAI1 expression. Interestingly, it has been described that HGF released during muscle repair promotes the quiescent satellite cells to enter the cell cycle while inhibiting muscle cell differentiation by positively regulating the expression of the myogenic-inhibitory transcription factor, Twist1, which negatively regulates the cell-cycle inhibitor, p27. 38,39 Furthermore, in vitro studies have described an autocrine loop involving HGF and its receptor, c-met, wherein signaling results in the silencing of MyoD and MYF4 gene expression, thereby inhibiting myotube formation. 40 Similarly, PDGFRα signaling can impair terminal myogenic differentiation in C2C12 cells 41 while other genes downregulated following SNAI1 silencing, such as Notch1, Androgen receptor (Ar), Cdk2, Peg3, c-Jun, and Rarb, are also known to promote myoblast proliferation and inhibit myotube differentiation. [42][43][44][45][46][47] These observations suggest that SNA1 inhibits myoblast differentiation by indirectly promoting the expression of genes involved in myoblast proliferation. Interestingly, another gene belonging to the GO term "regulation of cell proliferation," is Fgf21, which in a contrary fashion, is upregulated in SNAI1 knockdown cells. However, FGF21 is an atypical member of the FGF family and, in addition to its function as a regulator of energy metabolism, is significantly increased under certain conditions, such as muscular dystrophy 48 and exercise. 49 Furthermore, in the context of myoblast differentiation, Fgf21 is a direct transcriptional target of MyoD, which activates the expression of the early myogenic genes, promotes cell cycle exit, and enhances myogenic differentiation. 8,9,50 Here, we described for the first time that FGF21 is upregulated, both as mRNA and protein, in vivo during muscle regeneration, supporting its potential role in the regulation of myogenic differentiation. Interestingly, FGF21 inhibits macrophage-mediated inflammation by activating NRF2 and suppressing the NF-κB signaling pathway 51 while promoting anti-inflammatory M2 macrophage polarization. 52 F I G U R E 4 FGF21 and ATF3 expression was repressed by overexpression of SNAI1 in differentiating myoblasts. (A) C2C12 myoblasts were transiently transfected with the pCMV6-emty plasmid or with the pCMV6-SNAI1 plasmid and exogenous SNAI1 protein expression was detected by Western blot analysis by means of anti-FLAG antibody. Myoblasts differentiation was induced 48 h post-transfection by the medium switch, which corresponds to day 0 of differentiation. Histone H3 expression was used to confirm equal loading. Fgf21 (B) and Atf3 (C) expression was analyzed by RT-qPCR at 24, 48, and 72 h after transfection corresponding to differentiation day −1, day 0, and day 1, respectively. The values were normalized to the Gapdh mRNA levels. Asterisks indicate significant differences (p-values) between expression in control (pCMV6) and SNAI1 overexpressing (pcMV6-SNAI1) C2C12 cells at the indicated days after the switch to differentiation medium. Data were evaluated from three independent experiments using the one-way analysis of variance ANOVA. F I G U R E 5 SNAI1 regulates Fgf21 promoter activity by binding directly to E box sequences. (A) Schematic representation of the Fgf21 gene and the four promoter fragments, which were cloned upstream of the luciferase reporter gene. The blue and yellow boxes represent the E Box CACCTG and CACGTG/CAGCTG, respectively. Positions of the identified E boxes with respect to the TSS are indicated and referred to the first bp of the CANNTG consensus. The E boxes are grouped into three clusters. Point mutations of cluster C E boxes are referred as to m1, m3, and m3. The amplicon obtained after ChIP-qPCR is represented as red lines. Pink boxes represent the exon sequences. (B, C) C2C12 myoblasts were transiently co-transfected with the pCMV6-emty or with the pCMV6-SNAI1 plasmid and the reporter plasmids carrying the indicated Fgf21 promoter fragments (B) or E box point mutants of the −926 promoter fragment (C). Luciferase activity was assayed 24 h after the switch to DM and normalized to the Renilla luciferase internal control and expressed as fold change relative to cells co-transfected with the plasmid carrying the full length (−1764) promoter and the pCMV6-emty vector. (D) SNAI1 transcription factor repressive activity is here expressed as relative residue activity of each promoter mutant obtained by normalizing luciferase activity of cells co-transfected with pCMV6-SNAI1 to that of cells co-transfected with empty control vector. p-Values represent statistical differences between the relative residue activity of each promoter and the relative residue activity of the promoter −926 carrying point mutations of all three E boxes of cluster C (−926 m1,2,3). (E) SNAI1 binds both E box clusters A and C in the Fgf21 promoter region. pCMV6-SNAI1 or empty control vectors were transiently co-transfected in C2C12 and ChIP assay with anti-FLAG antibody followed by qPCR was performed. Upon normalization to spike-in DNA, results were expressed as a signal relative to input. Data were evaluated from three independent experiments using the twoway (B, C, and E) and one-way (D) ANOVA. Macrophages massively infiltrate muscle injury sites and participate in the skeletal muscle regeneration process. [53][54][55][56][57] Pro-inflammatory monocytes recruited after skeletal muscle injury stimulate myogenic cell proliferation, but later transition into an anti-inflammatory state that supports myogenesis by promoting myoblast terminal differentiation. 58,59 Taken together, these findings allow us to speculate that myoblast-derived FGF21 promotes myogenic differentiation both by acting directly on myoblast cells and indirectly promoting macrophage switching from a pro-inflammatory M1 to anti-inflammatory M2 phenotype.
In considering additional genes upregulated in response to SNAI1 silencing, we also found Atf3, Ddit3/ Chop, and Hspa5/Bip, transcripts coding for key molecules linked to ER stress and ER UPR-related processes. [60][61][62] Similarly to what we observed for FGF21, we find that ATF3 is upregulated during muscle regeneration in vivo and myoblast differentiation in vitro when SNAI1 expression decreases. Furthermore, we found that in differentiating myoblasts, SNAI1 overexpression downregulates Fgf21 and Atf3 transcripts and negatively regulates their promoter regions wherein SNAI1 binds multiple E boxes organized in clusters. However, due to methodological limitations, we are not able to discriminate among the E boxes recognized by SNAI1within the same cluster as they are confined to a DNA region less than 200 bp in size, and the resolution limit of the chromatin complexes cleaved by the pAG-Mnase. In any case, the fact that the basal activity of the full-length Fgf21 promoter decreased drastically in the −926 deletion mutant is in agreement with the observation presented by Liu et al. 9 where they have described the first E box of the cluster B in position −1067 as a binding site for the transactivator, MyoD, and an essential element for the full promoter activity. In our study, SNAI1 binding sites did overlap with the MyoD-targeted F I G U R E 6 SNAI1 regulates Atf3 promoter activity by binding directly to E box sequences. (A) Schematic representation of the Atf3 gene and the two promoter fragments, which were cloned upstream of the luciferase reporter gene. (B) C2C12 myoblasts were transiently co-transfected with the pCMV6-emty or with the pCMV6-SNAI1 plasmid and the reporter plasmids carrying the indicated Atf3 promoter fragments. Luciferase activity was assayed 24 h after the switch to DM and normalized to the Renilla luciferase internal control and expressed as fold change relative to cells co-transfected with the plasmid carrying the full length (−1943) promoter and the pCMV6-emty vector. (C) SNAI1 binds both E box at −1873 and the E box cluster centered at −1270. pCMV6-SNAI1 or empty control vectors were transiently co-transfected in C2C12 and ChIP assay with anti-FLAG antibody followed by qPCR was performed. Upon normalization to spike-in DNA, results were expressed as a signal relative to input. Data were evaluated from three independent experiments using two-way ANOVA. E box in the Fgf21 promoter, suggesting that their opposite effect on the Fgf21 transcription is not based on a simple competition for the binding sites. Similarly, we also found that SNAI1 directly binds the Atf3 promoter. Together, these results underline the complex effects of SNAI1 on gene expression and function during the myoblast proliferation and differentiation programs engaged in responses to muscle injury.
In conclusion, our findings and previous studies support a model wherein SNAI1 can contribute to the undifferentiated phenotype of myoblasts by indirectly promoting the expression of Hgf and Pdgfra as well as several other pro-proliferative genes while directly repressing the expression of the MyoD target genes involved in myoblast terminal differentiation and the pro-differentiation myokine, FGF21 ( Figure 8). As FGF21 is directly involved in myoblast differentiation, we outline a new mechanism by which SNAI1 inhibits myoblast terminal differentiation. We also described that SNAI1 negative regulates the expression of genes such as Atf3, Ddit3/Chop, and Hspa5, which are downstream of the PERK-dependent arm of ER stress. Because DIDT3/CHOP directly repressed the transcription of the MyoD gene, 20 in our model, we speculate that SNAI1 represses DIDT3/CHOP to contribute to the maintenance of myoblast identity and the delicate equilibrium between proliferation and differentiation ( Figure 8). Unlike DIDT3/ CHOP, less is known about the role of HSPA5 and ATF3 in the terminal differentiation of myoblasts and further studies are necessary to address this issue. F I G U R E 8 Scheme of the regulatory network connecting SNAI1, FGF21, and ER stress in the myoblasts. SNAI1 directly represses Fgf21 expression that in turn promotes myoblast cell cycle exit and terminal differentiation. SNAI1 also represses the expression of ER stress target genes downstream of the IRE1 sensor and the terminal differentiation program driven by MyoD. HGF expression, consistently with its role in promoting myoblast proliferation, is indirectly activated by SNAI1.