The main function of the heart is to regulate blood output in response to the changing hemodynamics of the body, which enables the heart to meet the needs of the body under normal and stress conditions. To successfully accomplish this task in the presence of a prolonged increase in preload or afterload, cardiac muscle cells usually undergo a certain volume increase, a condition known as hypertrophy (Nakamura and Sadoshima, 2018). It is well known that although cardiac hypertrophy begins as an adaptive response, chronic cardiac hypertrophy can eventually develop into heart failure and eventual death. Since cardiac hypertrophy is a key risk factor for cardiac failure, there is a pressing need to study its molecular biological mechanism in the progression of cardiac hypertrophy and discover its potential therapeutic targets.

In recent years, mounting evidence shows that the non-coding RNAs play a crucial role in cardiovascular diseases, especially in the process of cardiac hypertrophy (Poller et al., 2018). Non-coding RNA is a type of RNA that lacks protein-coding functions, which can modulate biological processes by regulating the expression of coding RNAs in various ways (Matsui and Corey, 2017). Non-coding RNAs are mainly divided into micro RNA (miRNA), circular RNA (circRNA), long non-coding RNA (lncRNA), and extracellular RNA (exRNA) (Ling et al., 2013; Sato-Kuwabara et al., 2015; St. Laurent et al., 2015; Ebbesen et al., 2016). The interactions between these RNAs eventually form a competing endogenous RNA (ceRNA) regulatory network, a concept proposed by Salmena (Salmena et al., 2011). Different from other non-coding RNAs, circRNAs are not vulnerable to the degradation of RNA enzymes and are more stable than linear RNAs (Jeck et al., 2013). CircRNA has a variety of molecular functions in sponge miRNA, regulatory transcription, regulatory RNA binding proteins, and even coding proteins (Chen et al., 2021). What gets the most attention is that some circRNAs contain miRNA response elements (MREs), making them miRNA sponges that reduce miRNA binding to mRNA (Hansen et al., 2013). To date, circRNAs have been involved in the occurrence and progression of various diseases, including cancer, neurological diseases, and cardiovascular diseases (Lei et al., 2020; Lim et al., 2020; Mehta et al., 2020). In recent years, researchers have begun to pay attention to the role of circRNA in regulating the physiological and pathological processes of cardiac hypertrophy (Li et al., 2020). CircRNA HRCR acts as an endogenous miR-223 sponge, isolating and inhibiting miR-223 activity, resulting in increased ARC expression (Wang et al., 2016). CircRNA Slc8a1 can act as an endogenous sponge for miR-133a in cardiomyocytes (Lim et al., 2019). However, few studies have delved into the mechanism underlying the regulation of cardiac hypertrophy by the circRNA-related ceRNA network.

Therefore, in this study, we downloaded mRNA and miRNA expression data related to cardiac hypertrophy from a public database (Aggarwal et al., 2014). This database included three controls and three samples of myocardial hypertrophy. The researchers of this database performed microarray analysis of mRNA and high-throughput sequencing of miRNA. We used Limma p and DESeq2 packages to screen out differentially expressed mRNAs and miRNAs. Then, we used TarBase and miRTarBase to forecast the targeting miRNAs of the differentially expressed genes (DEGs). Subsequently, the predicted miRNAs and differentially expressed microRNAs (DEMs) were intersected, and then circBank was used to predict the related circRNAs of the miRNAs obtained from the intersection. Finally, a circRNA-related ceRNA network was established. To further obtain the key ceRNA network, we used the Degree module in CytoHubba for analysis and obtained 4 mRNAs, 13 miRNAs, and 5 circRNAs. We then used PCR to detect the expression levels of these mRNAs and miRNAs in animal models of cardiac hypertrophy. Finally, a hub ceRNA network, including three mRNAs, four miRNAs, and four circRNAs, was obtained. In this study, a circRNA-mediated ceRNA regulatory network related to cardiac hypertrophy was established for the first time. This ceRNA regulatory network will provide new data for further elucidation of the mechanism of myocardial hypertrophy and open new avenues of research into therapeutic targets and drug development of cardiac hypertrophy.

In our study, we screened out DEGs and DEMs based on data from the public datasets GSE60291 and GSE60292, respectively. We set up a PPI network for the obtained DEGs. It is worth noting that several known proteins associated with myocardial hypertrophy, such as Myc, Fos, and Egr1, were present at the center of the network (Brand et al., 1993; Ahuja et al., 2010; Palomer et al., 2020). In addition, other proteins in the network have also been shown to be implicated in cardiovascular diseases. Scott D Shapiro et al. found that exogenous injection of CCNA2 virus can induce myocardial regeneration after infarction through cytoplasmic division of adult cardiomyocytes (Shapiro et al., 2014). Jin et al. found that DUSP1 reduces cardiac ischemia/reperfusion injury by inhibiting Mff-mediated mitochondrial fission and BNIP3-related mitochondrial autophagy through the JNK pathway (Jin et al., 2018). To a certain extent, these research results indicated that the DEGs screened by us can well reflect the differentially expressed proteins in cardiac hypertrophy. This point was further confirmed in the results of GO analysis and KEGG analysis. These results showed that DEGs were concentrated in several classical pathways, including the MAPK pathway, the estrogen receptor pathway, and the p53 signaling pathway. In addition, the erbB signaling pathway also appeared in the results. ErbB signaling has been associated with cancer in previous studies (Yarden and Sliwkowski, 2001). In recent years, more and more studies have shown that the erbB signaling pathway plays an important role in cardiac hypertrophy and heart failure. Blockage of erbB-1 signaling in the heart leads to blockage of erbB-2 signaling, and blockage of both leads to cardiac dysfunction (Rajagopalan et al., 2008). In neonatal, adolescent, and adult cardiomyocytes, cardiac hypertrophy is caused by induction of constitutively active ERBB2 (D'Uva et al., 2015). Therefore, based on our results, we hypothesize that the erbB-related signaling pathway will become a potential research direction of cardiac hypertrophy in the future.

The ceRNA network has been described as a complex post-transcriptional endogenous regulatory network in which circRNAs, lncRNAs, and other RNAs act as sponges for miRNAs to regulate mRNA expression. LncRNA ZFAS1 promotes the transformation of lung fibroblasts into myofibroblasts and iron function through the miR-150-5p/SLC38A1 axis (Yang et al., 2020). LncRNA Hoxaas3 regulates Runx1 by targeting miR-450b-5p and promotes lung fibroblast activation and fibrosis (Lin et al., 2020). In the ceRNA network, circRNA, as the only circular non-coding RNA, can compete with the targeting mRNA of related miRNA through its own miRNAs elements, ultimately affecting the mRNA expression level (Arnaiz et al., 2019). We used the public website prediction tool to predict the targets of DEGs and DEMs, and through the aforementioned screening method, we ultimately obtained a positive regulatory ceRNA network and a negative regulatory ceRNA network. Then, the Cytohubba plug-in of Cytoscape was used to further screen the key ceRNA network. The expression levels of mRNAs and miRNAs in the network were verified by animal experiments. Finally, three differentially expressed mRNAs (Col12a1, Thbs1, and Tgfbr3) and four differentially expressed miRNAs (miR-20a-5p, miR-27b-3p, miR-342-3p, and miR-378a-3p) were obtained.

In the heart, Col12a1 usually appears as a marker of myocardial fibrosis (Ning et al., 2017). Our results suggested that Col12a1 may play a key part in the process of cardiac hypertrophy through a ceRNA regulatory network. In addition, Thbs1 is a physiological regulator of TGFβ activation and is involved in the regulation of the myocardial fibrosis–related ceRNA network. One study shows that LncRNA RNF7 can promote cardiac fibrosis via the miR-543/Thbs1 axis and activation of TGFβ1 (Ouyang et al., 2020). In some AngII-induced cardiovascular disease models, the expression level of THBS1 was significantly increased (Bao et al., 2020; Jana et al., 2020). This trend is consistent with our results and further supports that Thbs1 may be a key regulatory site in the cardiac hypertrophy–related ceRNA network. In contrast to Thbs1, Tgfbr3 is a potential negative regulator of the TGFβ signaling pathway. The expression level of Tgfbr3 is negatively correlated with the degree of myocardial fibrosis (Chu et al., 2011). Overexpression of Tgfbr3 reduces collagen production in fibroblasts by inhibiting miR-21 expression (Liang et al., 2012). In addition, Tgfbr3 is an important drug target. A drug study has shown that Simvastatin could reduce cardiac fibrosis induced by infarction via upregulating the expression of Tgfbr3 (Sun et al., 2015).

Relevant miRNAs were also acquired. A study also indicated that miR-27b-3p was significantly increased in auricular tissue and angII-stimulated atrial fibroblasts in patients with atrial fibrillation. In this study, it is worth noting that miR-27b-3p promotes atrial fibrosis by targeting Tgfbr3 to activate Smad3 signaling in atrial fibroblasts (Yang et al., 2019). The targeting interaction relationship between miR-27b-3p and Tgfbr3 appeared in our prediction results, which further indicated that the miR-27b-3p/Tgfbr3 axis may play a significant role in myocardial hypertrophy. The expression level of plasma miR-20a-5p was proportional to the degree of left ventricular remodeling and dilation after myocardial infarction (Gao et al., 2019). The level of miR-342-3p in circulating miRNAs in heart failure model mice was significantly reduced (Kaneko et al., 2017). MiR-378a-3p could prevent myocardial apoptosis induced by ischemia-reperfusion injury through TRIM55/DUSP1/JNK signaling (Tan et al., 2020). These proven miRNAs might provide therapeutic targets for myocardial hypertrophy.

In recent years, circRNA has attracted increased scholarly attention in the study of myocardial hypertrophy. Garikipati et al. showed that the circRNA circFNDC3b regulates cardiac repair after myocardial infarction via the FUS/VEGF-A axis (Garikipati et al., 2019). Gan et al. confirmed that the circRNA_101,237 mediates anoxia/reoxygenation injury by targeting let-7a-5p/IGF2BP3 in cardiomyocytes (Gan et al., 2020). The samples in GSE60291 and GSE60292 belong to Homo sapiens. The species specificity of circRNAs may account for the lack of suitable experimental methods to prove the expression level of circRNAs selected in animal experiments. But the expression levels of these circRNA-related mRNAs and miRNAs in our ceRNA hub network have been proved by PCR. Therefore, the circRNAs (circ_0002702, circ_0110,609, circ_0013751, and circ_0047959) predicted on this basis may play a key role in the process of myocardial hypertrophy, providing a new inspiration for the clinical diagnosis and treatment of myocardial hypertrophy.

In conclusion, our study established and analyzed the circRNA-related ceRNA network involved in cardiac hypertrophy disorders for the first time and successively verified the differentially expressed miRNAs and mRNAs in animal experiments. The analysis results provided a new therapeutic target and a new idea for the regulatory mechanism of the cardiac hypertrophy–related ceRNA network. Nonetheless, to further clarify the expression levels of these RNAs in cardiac hypertrophy, it is still necessary to verify the ceRNAs with clinical samples and reveal the specific regulatory mechanism of the ceRNA network related to cardiac hypertrophy.