Circular RNAs in Myogenesis

Abstract

Skeletal muscles have an immense ability to regenerate from the muscle stem cells called satellite cells. The process of skeletal muscle regeneration is called myogenesis, which starts with activation of quiescent satellite cells immediately after muscle injury followed by proliferation and fusion of myoblasts into myotubes. Myogenesis is orchestrated through by the expression of a specific set of genes at each step regulated by complex gene regulatory networks. Besides the well-established roles of transcription factors, increasing evidence demonstrated that circular (circ)RNAs modulate gene expressions during myogenesis and involve in muscle-related diseases. Here we review the recent findings of circRNAs involved in myogenesis.

1. Introduction

Skeletal muscle is an indispensable tissue that plays a crucial role in locomotion, metabolism, and homeostasis, representing ~40 % of the body weight in adults. Skeletal muscle is composed of multinucleated contractile myofibers. Myofibers in skeletal muscle are generated by fusion of multiple mononucleated myoblasts through a process known as myogenesis [1, 2]. Myogenesis occurs in two phases of life, first during the embryonic development and second, upon muscle damage in adults. During embryogenesis, the skeletal muscle originates from myogenic progenitor cells (MPCs) that express the transcription factors paired box 7 (Pax7) and paired box 3 (Pax3), are committed to generate myoblasts [2–6]. The committed myoblasts undergo proliferation, growth, elongation, and finally fuse to form the multinucleated myotubes. Myogenesis is transcriptionally regulated by the well-established myogenic regulatory factors (MRFs), including MyoD, Myf5, myogenin, and MRF4 [2, 4, 6]. MPCs also produce a subpopulation of postnatal muscle stem cells called satellite cells, reside between the sarcolemma and basal lamina of muscle fiber and contribute to post-injury muscle regeneration in adult muscle [7–9]. Satellite cells express Pax7 and remain in the quiescent state. They are immediately activated after muscle injury and divide into proliferating myoblasts marked by co-expression of Pax7 and MyoD [6, 9]. A small portion of the activated satellite cells suppress MyoD expression and return to the quiescent state to maintain the satellite cell population in the muscle. Proliferating myoblasts induce the expression of MRFs and suppress Pax7 expression that induces differentiation and fuse to the existing muscle fibers or forms new multinucleated myotubes, leading to repair of the damaged muscles [10]. Failure of skeletal muscle development during embryonic stage leads to embryonic lethality, while the incapability to maintain or restore the injured skeletal muscle in adults leads to muscle diseases or death [11]. Dysregulation of gene expression during myogenesis can be harmful, causing atrophy and other muscle pathologies [12, 13]. Thus, elucidation of the molecular factors governing myogenesis is of extreme importance. Besides the general regulation of myogenesis by the MRFs, recent studies established the role of posttranscriptional regulation of gene expression during myogenesis [14]. Myogenesis is regulated posttranscriptionally via RNA-binding proteins (RBPs) and noncoding (nc)RNAs [15–19]. Recent findings established that ncRNAs, including rRNAs, tRNAs, micro (mi)RNAs, long noncoding (lnc)RNAs, and circular (circ)RNAs play a critical role in the regulation of gene expression.

CircRNAs represent a large family of covalently closed continuous loop RNAs ubiquitously expressed in eukaryotes [20–22]. Although they were discovered decades ago, they were considered as byproducts of splicing errors without any significant function. Only recently, thousands of circRNAs were discovered with the advent of high throughput RNA sequencing (RNA-Seq) and novel bioinformatic tools, moved circRNA to the forefront of RNA research. Increasing pieces of evidence suggest that circRNAs regulate gene expression by acting as a sponge for micro (mi)RNAs, RNA Binding Proteins (RBPs), and can be translated into proteins [22–27]. Furthermore, recent studies showed that circRNAs are implicated in several diseases and can potentially be used for diagnosis and therapeutics [28–30]. Moreover, circRNAs have emerged as key players in various cellular events, including cell proliferation, survival, and differentiation. In this review, we summarize circRNA biogenesis, classification, regulatory functions, and highlight the emerging roles of circRNAs in myogenesis.

2. Circular RNAs

CircRNAs were first discovered in plant viroids by electron microscopy then in hepatitis delta virus (HDV) and eukaryotic cells [31–33]. Price and colleagues were among the first to hypothesize the biogenesis of circRNAs by 5’ to 3’ backsplicing [34]. Later in 1992, Puttaraju and Been reported the generation of circRNA by ligating the 5’ splice site to the upstream 3’ splice site of a self-splicing group I intron [35]. Endogenous circRNAs were observed in humans and termed as scrambled exons [36, 37]. Further studies identified circular RNAs in the nonpoly(A) RNA fractions both in human and rodents. In 1993 the circular Sry transcript was found to be explicitly expressed in testes and localized in the cytoplasm [38]. Another study by Danan et al. discovered that the housekeeping non-coding RNAs, snoRNAs, and RNase P RNA generate circular RNAs in archaea [39]. Although several circRNAs were identified between the 1970s and 1990s, their abundance and function were not appreciated until the advent of high-throughput RNA-sequencing (RNA-Seq). Novel computational tools were employed to detect backsplice reads in RNA-seq for global analysis of circRNA expression [40]. Since backsplice reads can be generated from trans-splicing events and template switching of RT during cDNA synthesis, circRNA enrichment methods were used to enrich circRNA by RNase R digestion or depletion of poly(A) RNA and rRNA before RNA-seq [41–44]. CircRNA enrichment followed by high throughput RNA-seq revealed that circRNA are widely expressed in various organisms, developmental stages, and tissues [20, 21, 42].

CircRNA classification and biogenesis

CircRNAs are categorized into three types: exonic, intronic, and exon-intron circRNAs depending on the portion of the parental transcript generating the circRNAs. Several studies identified numerous Exonic (E)circRNAs based on the presence of backsplice sequences [21]. A subset of intronic lariats with specific sequences near the 5′ splice site and branchpoint resistant to the debranching process were found to generate stable circular intronic (ci)RNAs [45]. Recently, a large pool of circRNAs generated from intronic sequences was identified by RNA-seq and termed as Intronic (I)circRNAs [43]. A few circRNAs containing both exonic and intronic sequences were identified to date and termed as Exon-Intron (EI)circRNAs [46].

Different types of circRNAs are generated from the primary transcript by various mechanisms. The canonical splicing machinery generates exonic circRNAs by joining a downstream 5’splice site to an upstream 3’ splice site by a process known as backsplicing [47]. The rate of backsplicing is very slow compared to the linear splicing and positively correlates with the rate of transcription by the polymerase II [48]. Backsplicing is promoted by the complementary sequences like ALU elements in the flanking introns of the circularized exon [21]. RBPs, including Quaking 1 (QKI), muscleblind (MBL), FUS, and nuclear factor 90 (NF90) are reported to promote the circularization of the exonic sequence through binding to and pairing the flanking intron[21, 49–53]. A recent study by Zhang et al. showed that alternate 5’ donor splice site and 3’ acceptor splice sites could generate multiple circRNAs from a single gene locus [54]. Interestingly, backsplicing coupled with alternate linear splicing can generate several circRNA isoforms with the same backsplice site [54, 55]. CiRNAs generated from the intronic lariats contains a sequence motif which makes them resistant to debranching [45].

CircRNA functions

CircRNAs control gene expression at different levels, including transcription, pre-mRNA splicing, mRNA translation, and protein function (Figure 1). Two EIcircRNAs, circEIF3J and circPAIP2 were reported to be localized in the nucleus and interact with U1 snRNP and RNA Pol II in the promoter region to enhance transcription of the target genes [46]. Some ciRNAs such as c-sirt7 and ci-ankrd52 generated from intronic lariats interacts with the Pol II complex and positively regulate the transcription of their parental genes with unknown mechanism [45]. Other circRNAs were reported to regulate mRNA splicing [50]. Since backsplicing utilizes the same splicing machinery as linear splicing, circRNA splicing competes with pre-mRNA splicing, however with lower efficiency [50]. Backsplicing generating exonic circRNAs from the premRNA also leads to exon skipping and generation of splice variants of mature mRNA. For instance, DMD circular RNA is generated from the DMD gene during the pre-mRNA splicing leading to multiple exon skipping and generation of short DMD mRNA [56].

Figure 1: Possible role of circRNAs in gene regulation.

(a) EIcircRNA interacts with the U1 snRNP and the target gene promoter to induce transcription by RNA Pol II. (b) CiRNAs bind to elongating RNA Pol II to promote transcription. (c) CircRNAs act as miRNA sponge to enhance translation of miRNA target genes. (d) CircRNAs act as decoys for RBPs. (e) CircRNAs act as scaffolds for protein complexes to control protein functions. (f) CircRNAs with IRES can translate into proteins.

CircRNAs are exceptionally stable due to lack of free 5’ or 3’ ends, which makes them resistant to exonucleases [57]. The presence of miRNA response elements (MREs) and RBP binding sites in several circRNAs make them potent competitive endogenous RNAs (ceRNAs) that regulate the functions of miRNAs and RBPs [26]. For instance, ciRS-7 contains more than 70 binding sites for miR-7 and thereby inhibits miR-7 activity leading to the upregulation of miR-7 target gene expression [58]. Besides ciRS-7, more than 100 circRNAs have been reported to inhibit miRNA binding to target mRNAs and hence regulate various cellular processes like cell growth, proliferation, differentiation, and senescence. In addition, recent studies have focused on investigating the network of circRNA-miRNA-mRNA in disease diagnosis and therapies [28, 30].

Besides miRNA sponging by circRNAs, high-throughput RNA CLIP-seq data analysis revealed the association of RBPs with circRNAs, which may influence the function of interacting RBPs [59, 60]. Although thousands of interactions between circRNAs and RBPs are identified in the CLIP-seq studies, only a handful of functional circRNA-RBP interactions are experimentally validated. For example, circMbl was found to regulate alternative splicing of mbl pre-mRNA by acting as a decoy for the splicing factor MBL protein. [50]. Recently, circPABPN1 (circular Poly(A) Binding Protein Nuclear 1) was reported to inhibit the interaction between HuR and its target PABPN1 mRNA leading to decrease in PABPN1 translation [23]. Circ-Foxo3 regulates cell cycle progression by interacting with the cell cycle regulators cyclin-dependent kinase 2 (CDK2) and p21 (CDKN1A) [61].

CircRNAs are believed to be noncoding in nature due to lack of 5’ cap and poly(A) tail required for translation. As most of the circRNAs are originated form exons and localized in the cytoplasm, there is a possibility of circRNA translation. For example, some circRNAs are predicted to be translated due to the presence of an internal ribosomal entry site (IRES) [59, 62]. Further, artificial circRNA with an infinite open reading frame (ORF) was found to be translated efficiently in the in vitro translation system [63]. CircZNF609 was one of the first circRNA reported to harbor the IRES and translate into a protein that regulates myogenesis in muscle cells [64]. A recent study reported the translation of circRNAs with N6-methyladenosine (m6A) modifications that can act as IRES for translation initiation [25]. However, systematic investigation is required to identify the proteins translated from circRNAs and their functions in physiology and pathology.

3. CircRNAs regulating myogenesis

Accumulating evidence established that circRNAs are critical regulators of gene expression in various pathophysiological processes. Skeletal muscle myogenesis is tightly regulated by the timely expression of various MRFs, which are known to be regulated posttranscriptionally by RBPs, miRNAs, and lncRNAs. However, the role of circRNAs in myogenesis is only emerging in the last few years. A recent report demonstrated differentially expressed circRNAs during skeletal muscle aging in monkeys, but their roles remain to be explored [65]. Another study identified differential expression of circRNAs in differentiating myoblasts and suggested the translation potential of circRNAs [66]. CircRNAs are also shown to be differentially expressed in prenatal and postnatal skeletal muscle [67]. Here, we summarize the current understanding of the role of circRNAs in myogenesis (Figure 2).

Figure 2:

Schematic representation of the role of circRNAs in skeletal muscle differentiation.

DMD circular RNA:

As the name indicates, DMD circular RNA is generated from the exon-45–55 of dystrophin (DMD) gene. A total of 8 multiple exon skipping (MES) events reported for the DMD gene [56]. Biogenesis of circRNA often leads to skipping of circRNA exon from the mature mRNAs leading to the generation of splice variants. Indeed, the MES involving exon 45–55 generates exon 44–56-connected endogenous mRNA and DMD circRNA. Interestingly, the upstream posttranscriptional introns were found to be critical for intra-intronic interactions, which promote MES and backsplicing leading to DMD circRNA generation [56]. About 63% of DMD patients with a deletion mutation (exon 45–55) in DMD gene results in the production of truncated dystrophin protein [68]. However, the DMD circRNA biogenesis could lead to a shorter functional dystrophin protein which could rescue the DMD phenotype into an asymptomatic or mild phenotype [68].

CircZNF609:

The human circZNF609 originates from the exon 2 of ZNF609 gene. Interestingly, an ORF is present in circZNF609 which begins with the start codon as the linear mRNAs and ends with an in-frame stop codon created upon circularization [64]. CircZNF609 is one of the first endogenous circRNAs reported to associate with polyribosomes for translation. Moreover, CircZNF609 has been shown to promote myoblast proliferation [64]. Additionally, the mouse ortholog circZfp609 inhibits myoblast differentiation by sponging miR-194–5p through its four potential binding sites. miR-194–5p is known to drive myoblast differentiation by targeting a transcription factor BCLAF1 (Bcl associated factor 1) [69]. Expression of the myogenic transcription factors like Myf5 and MyoG positively correlate with miR-194–5p overexpression and negatively correlate with circZfp609 in C2C12 myoblast cells. In sum, circZfp609 acts as a decoy for miR-194–5p to promote expression of BCLAF1 and thereby suppresses myoblast differentiation [69].

CircLMO7:

CircLMO7 is expressed predominantly in bovine muscle tissue and produced from the host gene LMO7. CircLMO7 inhibits myoblast differentiation by sponging microRNA miR-378a-3p, which targets the 3’ UTR of HDAC4 mRNA to inhibit its expression [70]. miR-378a-3p promotes myoblast differentiation and inhibits proliferation by suppressing HDAC4 expression [71]. Indeed, HDAC4 silencing promotes expression of myogenic transcription factors, including MyoD and MEF2C promoting myogenesis. CircLMO7 sequesters miR-378a-3p leading to inhibition of apoptosis marked by upregulation of Bcl2. Together, circLMO7 binds to and inhibits miR-378a-3p activity in muscle cells enhancing myoblast proliferation and repressing myogenesis and apoptosis [70].

CircFGFR4:

CircFGFR4 is highly abundant in bovine muscle tissue and originates from its host gene FGFR4. CircFGFR4 contains eight putative miR-107 binding sites that can act as a ceRNA and inhibit miR-107 activity [72]. Overexpression of miR-107 inhibits myoblasts differentiation and apoptosis. It targets the 3’UTR and suppresses the expression of Wnt3a, which is known to promote muscle fiber formation during prenatal and postnatal myogenesis. Simultaneously, miR-107 inhibited apoptosis marked by increased Bcl-2 and decreased Caspase-9 expression in myoblasts [72]. Together, circFGFR4 promotes myoblast differentiation through the upregulation of Wnt3a [72].

CircFUT10:

CircFUT10 originates from the host gene FUT10 and highly expressed in bovine embryonic skeletal muscle tissue compared to adult skeletal muscle. Mechanistically, circFUT10 contains three miR-133a binding sites and hence may act as ceRNA in muscle tissue [73, 74]. Overexpression of circFUT10 inhibits myoblast proliferation by arresting cells at G1/G0 phase. It promotes myoblast differentiation by upregulating the expression of serum response factor (SRF) [74]. CircFUT10 also promotes apoptosis in bovine myoblast cells by sponging miR-133a which regulates Bcl-2 [74].

CircSVIL:

Exonic circSVIL originates from exon 6–14 region of SVIL (Supervillin) located on chromosome 2 in chicken. CircSVIL is highly expressed in leg muscle during the late developmental stages of the chicken embryo [75]. CircSVIL contains four putative binding sites for miR-203, which is highly expressed in C2C12 myoblast as well as in chicken skeletal muscle. miR-203 targets and suppresses the expression of c-JUN and MEF2C. c-JUN is well known to be involved in the MAPK pathway or JNK pathway leading to the expression of Cyclin D1 and drive cell proliferation [75]. Similarly, MEF2C is a well-known transcription factor involved in muscle differentiation. Overexpression of circSVIL increase the number of cells in the S and G2/M cell cycle phases compared to G0/G1 phases indicating its role in myoblast proliferation. Interestingly, circSVIL is highly expressed in differentiated myotubes compared to growing C2C12 myoblasts. Together, circSVIL promotes both myoblast proliferation and differentiation by sponging miR-203 and increasing expression of its targets c-JUN and MEF2C [75].

CircRBFOX2:

A total of 11 circRNAs were found to be generated from the RBFOX2 gene and termed as circRBFOX2 in chicken muscle. Interestingly circRBFOX2s are differentially expressed during chicken skeletal muscle development [76]. Four of the circRBFOX2 isoforms harbor binding sites for miR-1a-3p and miR-206. Knockdown of miR-206 or overexpression of circRBFOX2 promotes proliferation of chicken skeletal muscle cells. Interestingly, cyclin D2 is a target of miR-206 and is upregulated with overexpression of circRBFOX2. These findings suggest that circRBFOX2 promotes proliferation of chicken myoblasts by enhancing the expression of cyclin D2 through sponging miR-206 [76].

CircFGFR2:

The fibroblast growth factor receptor 2 (FGFR2) gene generates circFGFR2 containing the exon 3–6. CircFGFR2 is differentially expressed in the skeletal muscle during chicken embryogenesis [77]. Overexpression of circFGFR2 promotes proliferation and differentiation of myoblasts, while circFGFR2 silencing inhibits both proliferation and differentiation. Furthermore, circFGFR2 targets both miR-133a-5p and miR-29b-1–5p. Overexpression of circFGFR2 suppresses the activity as well as the levels of both miRNAs. Interestingly, both the miRNAs inhibit myoblast proliferation and differentiation, and circFGFR2 diminishes the inhibitory effects of these miRNAs [77]. In sum, circFGFR2 promotes proliferation and differentiation of chicken skeletal muscle by acting as a decoy for miR-133a-5p and miR-29b-1–5p.

CONCLUDING REMARKS AND PERSPECTIVES

CircRNAs, a large class of ncRNAs ubiquitously expressed in eukaryotic cells. However, the molecular regulators of backsplicing to generate circRNAs is not fully understood. A large number of circRNAs expressed in muscle cells with unknown functions. However, circRNAs are emerging as a crucial regulator of gene expression. CircRNAs appear to regulate gene expression by a variety of mechanisms, including acting as transcriptional regulator, miRNA sponges, RBP sponges, and templates for protein translation. Besides their role in gene regulation, circRNAs are also abundantly found in body fluids which makes them a potential diagnostic and prognostic biomarker for diseases. Although the current knowledge of circRNA biology is at a very early stage, few circRNAs have been implicated in various pathophysiological process including cancer, diabetes, aging, and muscular dystrophy. Here, we have provided an overview of the molecular mechanisms of circRNA functions and involvement in skeletal muscle cell development and differentiation (Figure 2; Table 1). For instance, circZfp609 and CircFGFR4 regulate myogenesis by sponging miR-194–5p and miR-107 respectively. Future studies will uncover more functional complexes between circRNAs and these trans-acting factors (miRNAs and RBPs) in myogenesis. Together with these factors, circRNAs may function in an interplay to regulate gene expression in myogenesis. In addition, circRNA turnover and the roles in specific subcellular locations remain to be studied. Urgent investigations are warranted to uncover the implications of circRNAs in myogenesis. Therefore, a better understanding of the molecular mechanisms by which circRNAs regulate myogenesis will eventually help the design of novel therapeutic strategies for muscle diseases.

Table 1:

List of circRNAs regulating myogenesis

Circular RNA	Species	Parental Gene	Target miRNA	miRNA Target	Reference
DMD circRNA	Homo sapience	DMD			[56]
CircZNF609	Homo sapience	ZNF609			[64]
CircZfp609	Mus musculus	Zfp609	miR-194-5p	BCLAF1	[69]
CircLMO7	Bos taurus	LMO7	miR-378a-3p	HDAC4	[70]
CircFGFR4	Bos taurus	FGFR4	miR-107	WNT3A	[72]
CircFUT10	Bos taurus	FUT10	miR-133a	SRF	[73, 74]
CircSVIL	Gallus gallus	SVIL	miR-203	C-JUN, MEF2C	[75]
CircRBFOX2	Gallus gallus	RBFOX2	miR-206	CCND2	[76]
CircFGFR2	Gallus gallus	FGFR2	miR-133a-5p miR-29b-1-5p		[77]

Highlights.

• Circular RNAs regulate gene expression in skeletal muscle cells by sponging miRNAs.

• Circular RNAs are differentially expressed during myogenesis.

• Proliferation and differentiation of muscle cells are regulated by circular RNAs acting as miRNA sponges.