The human circ-ZNF609 is derived from ZNF609. Circ-ZNF609 showed higher expression in myotubes than in myoblasts and knockdown it by siRNA reduced myoblast proliferation ^[11][14]. In addition, the mouse orthologue circ-zfp-609 interacted with miR-194-5p. It is known that miR-194-5p represses BCLAF1 expression and promotes myoblasts differentiation ^[12][51]. Interestingly, circ-ZNF609 contains a 753-nucleotide open reading frame, and it is the first protein-coding circRNA identified in skeletal muscle, but so far, the function of the protein is totally unknown. Together, circ-zfp-609 inhibits myoblasts differentiation by sponging miR-194-5p and upregulation of BCLAF1.

During chicken muscle development, RBFOX2 generated 11 isoforms of circRNAs, in which circRBFOX2.2-3 and circRBFOX2.2-4 were derived from exon2-3 and exon 2-4 respectively. Both of the circRNAs were expressed differentially during chicken muscle development. It was determined that circRBFOX2 contained mir-206 binding sites ^[6][48]. Previous research has proved that mir-206 involved in the cell cycle by repressing CCND2 (cyclin D2), which is an indispensable factor in cell cycle progression ^[13][14][52,53]. In addition, circRBFOX2 negatively regulated miR-206 expression by an unknown mechanism ^[6][48]. In summary, circRBFOX2 can sponge miR-206 and negatively regulate miR-206 expression, thus increasing CCND2 expression and promoting myoblasts proliferation.

CircSVIL is implicated as a positive regulator of myogenesis. It is generated from exon 6 to 14 of supervillin (SVIL). The abundance of circSVIL in skeletal muscles sharply increased from E10 to E15 during chicken embryonic development and maintained at high abundance in the later stage. Four binding sites for miR-203 in circSVIL were predicted using miRanda and RNAhybrid. Further, luciferase reporter assay and Ago2 RNA immunoprecipitation showed circSVIL and miR-203 interacted with each other. It is well known that miR-203 targets c-JUN, which is an essential factor for cell proliferation ^[15][54]. miR-203 can also inhibit the expression of MEF2C, which is an important regulator of muscle development ^[16][55]. Together, miR-203 has been implicated as a negative regulator of myoblast proliferation and differentiation. CircSVIL acts as a decoy of miR-203, thus playing a positive role in myogenesis.

CircLMO7, derived from LMO7, was highly expressed in skeletal muscle tissue. Overexpression of circLMO7 significantly decreased the expression of MyoD and myogenin (MyoG), suggesting that circLMO7 inhibited myoblast differentiation. Further analysis revealed that circLMO7 overexpression increased myoblasts proliferation and protected them from apoptosis. By performing luciferase reporter assay, it was found that circLMO7 interacted with miR-378a-3p. It is well known that miR-378a-3p inhibits HDAC4 expression ^[17][56]. HDAC4 can decrease the transcription of MEF2A and act as a repressor of myoblast differentiation ^[18][57]. In summary, circLMO7 can serve as a decoy for miR-378a-3p, results in higher expression of HDAC4 and decreases expression of MEF2A, thus promoting myoblasts differentiation.

CircFUT10, generated by FU10, predominantly expressed in bovine skeletal muscle tissue. It showed higher expression levels in embryonic skeletal muscles than adult skeletal muscles. Overexpression of circFUT10 inhibited cell proliferation, induced myoblasts apoptosis and enhanced myoblast differentiation. RNAhybrid and TargetScan analysis revealed that circFUT10 contained three miR-133a binding sites. Further luciferase assay confirmed that circFUT10 interacted with miR-133a ^[19][58]. MiR-133a has been shown to be important in myogenesis and targets serum response factor (SRF), which is an inhibitor of myoblast proliferation ^[20][59]. In summary, circFUT10 sponges miR-133a, leading to the enhancement of SRF expression. Thus, inhibiting myoblast proliferation and promoting differentiation.

It was noted that the expression level of circSNX29 was much higher in embryonic skeletal muscle than adult skeletal muscle. In addition, nucleoplasmic separation assay showed that it is enriched in the cytoplasm. Overexpression of circSNX29 inhibited myoblasts proliferation and facilitated differentiation. RNA hybrid showed that circSNX29 contained nine potential miR-744 binding sites. Then luciferase screening assay proved that circSNX29 directly interacted with miR-744 ^[21][60]. Next, it was found that Wnt5a and CaMKIId were the targets of miR-744. MiR-744 dramatically inhibited their expression levels and led to the activation of the non-canonical Wnt pathways, which are essential for myoblast self-renewal and muscle fibers growth ^[22][61]. Together, circSNX29 acts as a miR-744 sponge and increases Wnt5a and CaMKIId expression results in the activation of non-canonical Wnt pathways and myoblasts differentiation.

CircFGFR4 was highly expressed in bovine skeletal muscle. Overexpression of circFGFR4 induced cell apoptosis and promoted myoblasts differentiation. RNAhybrid and TargetScan revealed that circFGFR4 contained 18 putative miR-107 binding sites. Luciferase assay and RNA pull-down assays confirmed the interaction of miR-107 and circFGFR4. Next, wnt3a was identified as the target of miR-107 ^[23][62]. Inhibition of Wnt3a repressed myotube forming and protected myoblast from apoptosis. However, whether this is a common function of Wnt3a remains to be determined, as it has been reported that the expression of Wnt3a switched muscle stem cells from a myogenic to a fibrogenic lineage and increased connective tissue deposition ^[24][63]. In summary, circFGFR4 acts as a miR-107 sponge and increases Wnt3a expression, leading to bovine primary myoblasts differentiation.

CircFGFR2, generated by exon 3-6 of FGFR2, was found differentially expressed during chicken embryo skeletal muscle development ^[25][64]. Flow cytometry analysis of the cell cycle and EdU assays demonstrated that circFGFR2 accelerated myoblast proliferation. Meanwhile, circFGFR2 positively regulated myoblasts differentiation. The results of luciferase reporter assay and biotin-coupled miRNA pull-down assay suggested that circFGFR2 interacted with miR-133a-5p and miR-29b-1-5p. Further investigation discovered that miR-133a-5p and miR-29b-1-5p inhibited chicken myoblast proliferation and differentiation. Despite miR-133a-5p, miR-29 is another important myogenesis regulator, which can reduce proliferation and facilitate differentiation of myoblasts by targeting AKT ^[26][27][65,66]. Altogether, circFGFR2 acts as miR-133a-5p and miR-29b-1-5p sponge to promote skeletal muscle proliferation and differentiation.

During chicken skeletal muscle development, HIPK3 generated 11 isoforms of circRNAs ^[28][67]. CircHIPK3 was produced by the third exon of HIPK3 and differentially expressed among chicken myogenesis. CircHIPK3 promoted the proliferation and differentiation of myoblasts. In circHIPK3, three potential binding sites for miR-30a-3p were identified through miRDB and RNAhybrid analyses. Luciferase assay suggested that circHIPK3 could act as a sponge of miR-30a-3p. Further investigation discovered that MiR-30a-3p was differentially expressed during chicken skeletal muscle development and suppressed myoblasts differentiation by targeting MEF2C ^[16][28][29][55,67,68]. In summary, circHIPK3 sponges miR-30a-3p, thus increasing MEF2C expression and skeletal muscle differentiation. However, the underlying mechanism of circHIPK3 promoting myoblast proliferation needs further study.

The dystrophy gene is among the first genes identified to be able to generate circRNAs in skeletal muscle ^[30][69]. It is the largest human gene, consisting of 79 exons. Frame-shifting deletions or nonsense mutations of dystrophy lead to Duchenne muscular dystrophy (DMD), which is a severe muscular disease characterized by progressive muscle degeneration and weakness. In contrast to DMD, Becker muscular dystrophy (BMD) patients have milder symptoms, since they can express truncated, but partially functional protein. Mutation ranging from exon 45 to 55 of the gene represents nearly 60% of DMD/BMD cases ^[31][70]. Transforming a DMD phenotype into a BMD phenotype by 45–55 exon skipping has been proposed a new treatment strategy. Recently, Hitoshi et al. reported that eight distinct patterns of circRNAs derived from 45–55 exons and their biogenesis was related to exons skipping ^[32][71]. These results suggested that artificial and specific increase the expression levels of these circRNAs by exon skipping might have the possibility to improve or cure DMD patients. Thus, further study for the mechanism of circRNA biogenesis will be a benefit for the treatment of these muscular diseases.