ETM Experimental and Therapeutic Medicine 1792-0981 1792-1015 D.A. Spandidos ETM-0-0-8748 10.3892/etm.2020.8748 Articles Identification of a lncRNA-miRNA-mRNA network based on competitive endogenous RNA theory reveals functional lncRNAs in hypertrophic cardiomyopathy TaoLichan 1 * ShiJia 2 * HuangXiaoli 3 * HuaFei 3 YangLing 1 Department of Cardiology, The Third Affiliated Hospital of Soochow University, Changzhou, Jiangsu 213003, P.R. China Department of Neurosurgery, The Third Affiliated Hospital of Soochow University, Changzhou, Jiangsu 213003, P.R. China Department of Endocrinology, The Third Affiliated Hospital of Soochow University, Changzhou, Jiangsu 213003, P.R. China Correspondence to: Dr Fei Hua, Department of Endocrinology, The Third Affiliated Hospital of Soochow University, 185 Juqian Road, Changzhou, Jiangsu 213003, P.R. China czhuafei@vip.sina.com

Dr Ling Yang, Department of Cardiology, The Third Affiliated Hospital of Soochow University, 185 Juqian Road, Changzhou, Jiangsu 213003, P.R. China 690643672@qq.com

*Contributed equally

 08 2020 13 05 2020 20 2 1176 1190 03 06 2019 04 12 2019 Copyright © 2020, Spandidos Publications 2020

Hypertrophic cardiomyopathy (HCM) is an autosomal dominant disease that affects 1 in every 200 people in the general population, leading to cardiac ischemia, heart failure and increased risk of sudden death. Recently, accumulating evidence has suggested that long noncoding RNAs (lncRNAs) may serve specific roles in various biological processes and participate in the pathology of various diseases, including HCM. Although a large number of lncRNAs have been detected, the functions of lncRNAs in HCM are still unknown. In the present study, a global triple network based on competitive endogenous RNA (ceRNA) theory was constructed using data from the National Center for Biotechnology Information Gene Expression Omnibus. Furthermore, Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway analyses of mRNAs in the lncRNA-microRNA (miRNA)-mRNA network were performed using the Cytoscape plugins, BiNGO and Database. The lncRNA-miRNA-mRNA network was composed of 30 lncRNA nodes, 94 mRNA nodes and 8 miRNA nodes. Subsequently, hub nodes and the number of relationship pairs were analyzed and showed that 5 lncRNAs (ENST00000597346.1, ENST00000458178.1, ENST00000544461.1, ENST00000567093.1 and ENST00000571219.1) were closely related to HCM. Cluster module analysis and Random Walk with Restart of the ceRNA network further confirmed the potential role of two lncRNAs (ENST00000458178.1 and ENST00000567093.1) in HCM. The present study provides a new strategy for identifying potential pathways associated with HCM or other diseases. Furthermore, lncRNA-miRNA pairs may be regarded as candidate diagnostic biomarkers or potential therapeutic targets for HCM.

 long non-coding RNA microRNA mRNA competing endogenous RNA hypertrophic cardiomyopathy Introduction

Hypertrophic cardiomyopathy (HCM) is one of the most common genetic heart disorders with a prevalence of 1 in 200 worldwide (0.6% of the general population) (1). HCM is characterized by increased thickness of the ventricular wall. Over 1,500 mutations in at least 11 genes, encoding components of the cardiac sarcomere-associated proteins, have been identified as potential cause for HCM (2-4). MYH7 and MYBPC3, encoding β-myosin heavy chain and myosin-binding protein C, respectively, are the 2 most common genes involved, accounting for 50-60% of the HCM families. However, in about 40% of HCM patients, the causal genes remain to be identified, highlighting the need for precision medicine as genetic diversity of HCM-associated individuals (3,4). The clinical outcomes of HCM are diverse, ranging from no symptoms to cardiac ischemia, cardiac arrhythmia, congestive heart failure and other organ system dysfunctions (5). HCM is also associated with an increased risk of sudden death, heart failure and thromboembolic events, which results in an annual mortality rate of 1% worldwide (6).

A challenging characteristic of diagnosing HCM is the lack of association between genotype and phenotype, as family members carrying the same mutation develop different symptoms (7). Therefore, HCM is regarded as a highly complex disease due to its diverse clinical presentations, heterogeneous phenotypes, large number of mutations and broad spectrum of complications (4,7). The development of HCM likely results from a combination of endogenous genetic mutations with an exogenous decline in protein-protective mechanisms, and environmental factors such as lifestyle, blood pressure and physical exercise (8). In this regard, it is crucial to identify transcriptional profiles, epigenetic modifications and post-translational modifications which may initiate the onset of HCM.

In the last decade, accumulating evidence has indicated that rather than being transcriptional noise, diverse noncoding RNAs (ncRNAs) act as important regulators of HCM initiation and progression at the post-transcriptional level (9-15). To date, studies have primarily focused on small, noncoding endogenous RNAs (9,10). MicroRNAs (miRNAs) are 22-25 nucleotides in length and regulate gene expression by directly targeting mRNAs for degradation or translational repression (10). Several miRNAs, such as miR-21, miR-1, miR-29a, miR-133a and miR-130b have been described as important regulators of HCM in murine and human hearts (9,11-14). Furthermore, these miRNAs can be detected in the bloodstream, highlighting the possibility of their use as circulating biomarkers of HCM (11,14,15). However, the role of miRNAs in the progression of HCM remains to be elucidated.

In addition to miRNAs, long noncoding RNAs (lncRNAs) are known as important regulators of cardiac pathology (16,17). lncRNAs are transcripts >200 nucleotides in length with no protein-coding capacity. Based on their diverse biochemical roles, lncRNAs can perform their functions via RNA-DNA, RNA-RNA or RNA-protein interactions (18). Notably, lncRNAs have been reported to competitively interact with miRNAs and thus inhibit target mRNA degradation by a competitive endogenous RNA (ceRNA) regulatory mechanism (19,20). However, little is known regarding the function of lncRNAs in HCM, particularly in human hearts. Current studies of lncRNAs in HCM primarily describe expression profiling by RNA sequencing or microarray technology (21,22). However, the underlying mechanisms involving miRNAs or mRNAs are still unclear. Therefore, considering the large number lncRNAs expressed in the heart and limited knowledge of their function physiologically and pathophysiologically, it is hypothesized that functionally-related lncRNAs are functionally associated with mRNAs or miRNAs. Similar associations have been detected in several diseases such as ischemic cardiomyopathy (23), diabetic cardiomyopathy (24), and heart failure (25), but have not yet been revealed in HCM.

In the present study, a triple network was constructed using human data from the National Center for Biotechnology Information Gene Expression Omnibus (NCBI GEO). Based on the ceRNA mechanism in which lncRNA, miRNA and mRNA form a triplet, where the lncRNA and mRNA share the same miRNA. The present study constructed a lncRNA-miRNA-mRNA network with high reliability, providing a potentially novel understanding of the mechanisms of the development of HCM and potential therapeutic targets.

 Materials and methods <sec> <title>Raw data

NCBI GEO (ncbi.nlm.nih.gov/geo/) is a publicly available genomics database containing data obtained from array- and sequence-based analysis. Users can query and download experimental data and curated gene expression profiles from NCBI-GEO. In the present study, human lncRNA and mRNA expression data were downloaded from dataset GSE68316 in myocardial tissues (21), and human miRNA expression data were downloaded from GSE36946 of surgical myectomy tissues. The age of patients corresponding to the heart samples in GSE68316 ranged from 31-60 years, whereas the age range of the patients in GSE36946 was 9-78 years; therefore, data for heart samples corresponding to patients aged from 31-60 years in GSE36946 were selected. Subsequently, the two datasets were merged to construct a lncRNA-miRNA-mRNA network based on the ceRNA theory. The experimental design and flowchart of the steps performed in the present study are presented in Fig. 1.

 Screening of differentially expressed lncRNAs, miRNAs and mRNAs

To detect differentially expressed lncRNAs, miRNAs and mRNAs between patients with HCM and healthy donors, two-class differential analyses was used. The threshold of dysregulated lncRNAs/mRNAs/miRNAs was a fold change of >2. Student's t-test was performed using SPSS 23.0 (IBM Corp.). P<0.05 was considered to indicate a statistically significant difference.

 Prediction of target lncRNAs and mRNAs of miRNAs

miRNA sequences were downloaded from miRBase (mirbase.org/) and lncRNA sequences were downloaded from the NCBI nucleotide database (ncbi.nlm.nih.gov/nucleotide/). The miRNA targets of lncRNAs were predicted and the minimum free energy of lncRNA-miRNA duplexes was calculated using RNAhybrid (bibiserv.cebitec.uni-bielefeld.de/rnahybrid) (26). miRNA-mRNA interactions were predicted using miRanda (http://www.microrna.org/) and TargetScan (targetscan.org/vert_72/).

 Construction of a lncRNA-miRNA-mRNA network

Pearson's correlation coefficient was used to calculate the correlation between lncRNA and mRNA expression. The pairs of lncRNA-mRNA with r score >0.99 and P score <0.05 were regarded as target pairs. Among all the selected lncRNA-mRNA pairs, if the lncRNA and mRNA were both targeted and were sharing a common miRNA, the lncRNA-miRNA-mRNA ceRNA network was identified as a co-expression competing triplet. Cytoscape software (3.7.1) (27) was used to visualize the lncRNA-miRNA-mRNA network.

 Functional enrichment analysis

To study functional enrichment, Gene Ontology (GO; geneontology.org/) and Kyoto Encyclopedia of Genes and Genomes (KEGG; genome.jp/kegg/) pathway analyses of mRNAs in the lncRNA-miRNA-mRNA network were performed using the Cytoscape plugins BiNGO and Database (http://apps.cytoscape.org). The GO database provides gene ontologies, annotations of genes and gene products based on terms. KEGG is a relational database comprising searchable molecular interaction pathways and reaction networks in metabolism, various cellular processes and multiple human diseases.

 Reconstruction of a key lncRNA-miRNA-mRNA subnetwork

Cytoscape software was used to construct a new subnetwork extracted from all lncRNAs and their related miRNAs and mRNAs in the triple network. The target lncRNAs were identified by calculating the number and node of related lncRNA-miRNA-mRNA triplets,

 Results <sec> <title>Screening for differentially expressed lncRNAs, miRNAs and mRNAs

Human lncRNA/mRNA expression data were obtained from GEO dataset GSE68316 and human miRNA expression data were downloaded from GEO dataset GSE36946. lncRNAs, mRNAs and miRNAs were considered as significantly differentially expressed when the fold change was >2 with P<0.05. A total of 8 miRNAs (Table I), 431 lncRNAs and 609 mRNAs were selected for subsequent analyses. Considering the specific description of miRNAs in different microarrays, the names of selected miRNAs were redefined according to miRBase (Table I). Subsequently, the two datasets were merged to construct a ceRNA based lncRNA-miRNA-mRNA network.

 ceRNA based lncRNA-miRNA-mRNA network

lncRNA-miRNA and miRNA-mRNA pairs were identified according to both base sequence and expression levels. A total of 181 miRNA-lncRNA pairs and 421 miRNA-mRNA pairs were predicted based on the intersecting elements (Fig. 2A and B). In addition, 442 lncRNA-mRNA pairs were further selected according to expression levels and ceRNA score (Fig. 2C). Finally, a network composed of 30 lncRNAs and 94 mRNAs (Fig. 2D), and a ceRNA network composed of 30 lncRNAs, 94 mRNAs and 8 miRNAs were constructed (Fig. 2E).

 Topological analysis of the HCM-related lncRNA-miRNA-mRNA network

Hub nodes play important roles in biological networks. The topological properties of the HCM-related ceRNA network composed of lncRNA-miRNA-mRNA were first analyzed. The degree (the level of a gene related to its surroundings, where genes with a higher degree is more likely to be the core information exchange of the network), closeness (the length of information transferred from one node to the other) and betweenness (the function and influence of the corresponding gene in the whole network, where the larger the value is, the more capable of the gene is to participate in the communication between other genes) of the network were calculated and the topological features of all the nodes in the network were ranked. The top 20 genes in the network were listed in Table II. Interestingly, 5 lncRNAs appeared in the list. The number of primary lncRNA-miRNA pairs and secondary miRNA-mRNA pairs were calculated as presented in Table III. Among the top 10 lncRNA-miRNA pairs, the 5 lncRNAs identified in the ceRNA network were determined. Therefore, these 5 lncRNAs (ENST00000597346.1, ENST00000458178.1, ENST00000544461.1, ENST00000567093.1 and ENST00000571219.1) not only had higher betweenness and node degree but were also involved in a higher number of lncRNA-miRNA and miRNA-mRNA pairs, which suggested that these 5 lncRNAs may serve a role in the initiation and/or development of HCM.

 Key lncRNA-miRNA-mRNA subnetworks

The 5 key lncRNAs in the ceRNA network were further investigated. The 5 lncRNAs were found to primarily target miR-10a-5p, miR-30c-3p, miR-1247-5p, miR-1268a and miR-144-5p (Table IV). The mRNAs and miRNAs associated with these 5 lncRNAs were identified in the global triple network and new subnetworks were constructed (Table IV). Among the 5 lncRNAs, sequence Basic Local Alignment Search Tool (BLAST) analysis showed that the sequence of lncRNA ENST00000597346.1 is part of ENST00000567093.1, and the sequence of ENST00000544461.1 is also part of ENST00000458178.1. ENST00000567093.1 and ENST00000458178.1 were conserved among human, mouse and rat genomes, whereas ENST00000571219.1 was only detected in the human genome. GO function and KEGG pathway analyses for the lncRNA-related mRNAs were then performed. For ENST00000458178.1, differentially expressed mRNAs were primarily enriched in ‘Nucleobase-containing small molecule interconversion’, ‘Nucleobase-containing small molecule metabolic process’ and ‘Regulation of lipid metabolic process’ (Fig. 3A). Enriched KEGG pathways included the ‘AMPK signaling pathway’, ‘Pantothenate and CoA biosynthesis’ and ‘Arginine biosynthesis’ (Fig. 3B), which shared a similar trend with ENST00000567093.1 (Fig. 3D and E). The subnetworks of ENST00000458178.1 and ENST00000567093.1 are presented in Fig. 3C and F, respectively. The biological processes, KEGG pathways and subnetwork of ENST00000571219.1-related mRNAs are presented in Fig. S1A-C. The HCM pathway was enriched in all 5 lncRNA-related KEGG pathways, indicating the potential roles of these 5 lncRNAs in HCM (Fig. 3B and E; Fig. S1B).

 Module analysis of a HCM-related lncRNA-miRNA network

To further investigate the crosstalk between mRNAs and lncRNAs, a bidirectional hierarchical clustering analysis was performed using gplot in R software (version 3.5.1; http://www/r-project.org/). The heat map showed two modules screened by cluster analysis that were closely related to HCM (Fig. 4A-C). The two modules included two different gene populations. GO enrichment analysis and KEGG analysis of genes were performed in the modules (Fig. 4D-G). In module 1, ‘Positive regulation of adenylate cyclase activity involved in G-protein-coupled receptor signaling pathways (GO: 0010579)’ was significantly and closely related to HCM. KEGG pathway analysis further demonstrated that the ‘AMPC signaling pathway’ was the most significant signaling pathway in HCM. In module 2, ‘Translation’ had the most notable relationship with HCM, and KEGG pathway analysis showed that ‘Ribosome’ served a predominant role in HCM (Fig. 4D-G). Among all the lncRNAs in the two modules, 8 lncRNAs were included in the ceRNA network (ENST00000499521.2, ENST00000458178.1, ENST00000567093.1, ENST00000591866.1, ENST00000367207.3, ENST00000425185.1, ENST00000596350.1 and ENST00000416301.1). According to the results of the current study, ENST00000458178.1 and ENST00000567093.1 not only had higher betweenness and node degree (Table II) but were also involved in a higher number of lncRNA-miRNA and miRNA-mRNA pairs (Table III), which were regarded as potential targets for HCM and were confirmed by module analysis.

 Discussion

HCM is a commonly inherited cardiovascular disease that is the most frequent cause of sudden death in younger individuals and functional disability resulting from heart failure and stroke (28). Although therapeutic advancements have been made in HCM, the treatment of HCM remains limited. Currently, major efforts have been made to detect the underlying mechanisms of HCM (29-31). ncRNAs including miRNAs and lncRNAs, have been identified as novel regulators underlying HCM (32-35). However, current studies of lncRNAs are focused on RNA sequencing or microarray technology (21,22), the functional studies are relative few and complicated compared to miRNAs. Therefore, it is necessary to identify an efficient and accurate way to infer the potential function of lncRNAs with miRNAs and/or mRNAs whose functions have already been annotated. In the present study, interaction data from NCBI GEO (lncRNA-mRNA dataset: GSE68316, miRNA dataset: GSE36946) were used to generate a triple network based on ceRNA theory, which suggested that lncRNAs and mRNAs share the common miRNA in one triplet. The ceRNA network was composed of 30 lncRNA nodes, 94 mRNA nodes and 8 miRNA nodes. Subsequently, hub nodes and the number of relationship pairs were used to perform topological and subnetwork analysis. 2 lncRNAs (ENST00000458178.1 and ENST00000567093.1) were observed to be key topological nodes, whose node degrees and number of lncRNA-miRNA and miRNA-mRNA interaction pairs were higher compared with other lncRNAs, and these 2 lncRNAs were highly conserved in human, mouse and rat genomes. According to the ceRNA network, the 2 lncRNAs primarily targeted miR-10a-5p, miR-30c-3p, miR-1247-5p and miR-1268a.

A number of transcriptomic and proteomic profiling analyses of HCM have been previously performed and several dysregulated miRNAs have been identified. For example, Song et al (36) identified 13 dysregulated miRNAs involved in myocardial tissues in patients with HCM. The list of miRNAs in GSE36946 and previous reports on HCM or cardiac hypertrophy were compared. Several miRNAs, including miR-10a, miR-30c and miR-373, were common between the present study and previous studies (37-41). Additionally, studies have shown that miR-10a was downregulated in a transverse abdominal aortic constriction-induced cardiac hypertrophy model and angiotensin II (Ang II)-stimulated cardiomyocytes. Overexpression of miR-10a in Ang II-treated cardiomyocytes ameliorated cell hypertrophy and decreased the expression of natriuretic peptides A and B by directly inhibiting T-box transcription factor TBX5 (37,38). Diabetic cardiomyopathy (DCM) is characterized by endothelial dysfunction, myocyte hypertrophy, necrosis, apoptosis and increased fibrosis deposition (39). A previous study demonstrated that cardiac miR-30c expression was decreased in rats and patients with DCM and in high HG-treated cardiomyocytes (40). Overexpression of miR-30c attenuated HG-induced cardiomyocyte hypertrophy by inhibiting cell division control protein 42 homolog and serine/threonine-protein kinase PAK 1(40). Furthermore, a study reported that concurrent overexpression of miR-30c and miR-181a resulted in a greater decrease in cardiomyocyte hypertrophy and apoptosis via the p53-p21 pathway compared with the overexpression of miR-30c or miR-181a alone, suggesting a synergistic effect of these two miRNAs in DCM-induced cardiac hypertrophy (41). Contrary to our results, miR-373 levels were decreased in the plasma of patients with HCM, and overexpression of miR-373 improved cardiac hypertrophy induced by DCM. The discrepancy in the expression profiles of miRNAs may have arose from differences in tissue samples and models of cardiac hypertrophy. However, although studies have shown that these miRNAs serve important roles in cardiac hypertrophy, the functions and underlying mechanisms of the miRNAs in HCM remains to be elucidated.

In the present study, GO and KEGG pathway analyses were used to assess biological functions that are enriched among differentially expressed coding genes. Owing to similar miRNA targets (miR-10a-5p, hsa-miR-30c-3p and miR-1268a), the significant GO terms of ENST00000567093.1 and ENST00000458178.1 shared common trends involving ‘Nucleobase-containing small molecule interconversion’, ‘Nucleobase-containing small molecule metabolic process’ and ‘Regulation of lipid metabolic process’, and the results were consistent with previous studies on HCM (42,43). Pathway analysis of lncRNA-related mRNAs showed that the HCM pathway was highly enriched. The other top pathways based on KEGG pathway analyses, including ‘AMPK signaling pathway’, ‘Pantothenate and CoA biosynthesis’, ‘Arginine biosynthesis’, ‘Adrenergic signaling in cardiomyocytes’ and ‘Histidine metabolism’ were primarily metabolic pathways that have been shown to serve important roles in HCM (44-47).

Bidirectional hierarchical clustering analysis was performed to investigate the crosstalk between mRNAs and lncRNAs. GO and KEGG pathway analyses of these 2 modules indicated that in module 1, the dysregulated genes were primarily involved in metabolic regulation, including lipid regulation, nucleobase-containing small molecule metabolic process and CoA biosynthesis, which has been discussed earlier. In module 2, the mRNAs were primarily enriched in translational regulation, including translation initiation, termination and elongation, ribosome and RNA degradation. Once the transcriptional process is impaired, the expression levels of numerous proteins are dysregulated, and this may alter normal physiological processes. To the best of our knowledge, there are no previous studies which have addressed translational regulation in HCM, and thus, it would be meaningful to investigate transcriptional dysfunction in HCM. Nevertheless, additional studies are required to fully understand the molecular mechanisms underlying HCM.

Currently, identifying ncRNA-disease associations is playing an increasingly vital role in diagnostic and therapeutic tools for diseases including HCM. However, since the fact that it is expensive and time-consuming via experimental studies to uncover associations between ncRNAs and disease, novel and effective computational models for the identification of ncRNAs associated with HCM or other diseases are being developed. Several novel computational methods have been used to calculate potential ncRNAs-disease association scores (48,49). lncRNAs have emerged as one of the largest and significantly diverse type of RNA family (50,51). The biological role and functions of lncRNAs are diverse and still mostly uncharacterized. Their target-mimetic and sponge/decoy function on miRNAs have been identified. Previous studies demonstrated that lncRNAs can act as miRNA sponges, reducing their regulatory effect on mRNAs (52,53). This function introduces an extra layer of complexity in the miRNA-mRNA interaction network. In addition, while several studies examining. lncRNAs or miRNAs in cardiac hypertrophy have been performed (54-56), to the best of our knowledge, there are relatively fewer studies focusing on a ceRNA network between lncRNAs and miRNAs of HCM (57,58), particularly in human models. In the present study, a ceRNA based lncRNA-miRNA-mRNA network in human HCM models was constructed. A total of 2 lncRNAs (ENST00000458178.1 and ENST00000567093.1) were determined to be key topological nodes and highly conserved in human, mouse and rat genomes. According to the ceRNA network, the 2 lncRNAs primarily targeted miR-10a-5p, miR-30c-3p, miR-1247-5p and miR-1268a, which were highly related to the development of HCM. Therefore, the study provides a framework for constructing powerful computational methods in human HCM models to predict potential lncRNA-miRNA-disease associations and select the most promising lncRNAs/miRNAs related to HCM or other diseases for experimental validation.

However, there are several disadvantages to using bioinformatics analysis. Firstly, the changes and potential functions of lncRNAs/miRNAs/mRNAs are hypothetical, and need to be further verified in animal or human models of hypertrophic cardiomyopathy. Secondly, a lncRNA-miRNA-mRNA ceRNA network was constructed by combining two different GSE datasets (GSE68313 and GSE36946); there may be missing and overlapping data between different databases. HCM myocardial samples will be collected and a complete lncRNA/miRNA/mRNA profile will be performed in future studies. The functional role of the 2 lncRNAs on HCM and the relationship between lncRNAs and miRNAs detected in the present study will be further explored. The underlying mechanisms can be investigated by western blotting and proteomics studies. A luciferase reporter assay will also be performed to confirm the target genes of miRNAs in future work.

The present study has some limitation. Due to the lack of samples, the changes in lncRNAs, miRNAs and mRNAs were not validated in myocardial tissues from patients with HCM, and thus, there may have been false positives. Additionally, during the process of converting gene IDs from different databases, a number of genes may have been lost, which may have decreased the accuracy of the present results. Finally, the present study was primarily focused on alterations in lncRNAs/miRNAs in HCM samples, whereas the specific functions of ceRNAs in HCM remain unknown, therefore the underlying biological functions and mechanisms warrants further exploration.

In conclusion, a ceRNA based lncRNA-miRNA-mRNA network was constructed in the present study, providing a new strategy for studying HCM and other diseases. Furthermore, lncRNA-miRNA pairs may be regarded as candidate diagnostic biomarkers or potential therapeutic targets for treatment of HCM.

 Supplementary Material lncRNA-miRNA-mRNA subnetwork of lncRNA ENST00000544461. (A) Biological function and (B) pathway analysis of differentially expressed mRNAs related to lncRNA ENST00000544461.1. (C) lncRNA ENST00000544461.1 related miRNA-mRNA subnetwork. Green nodes represent miRNAs and red nodes represent mRNAs. miRNA, microRNA; lncRNA, long non-coding RNA. Acknowledgements

Not applicable.

 Funding

This work was supported by grants from the National Natural Science Foundation of China (grant no. 81700343), Natural Science Foundation of Jiangsu Province (grant no. BK20170296), China Postdoctoral Science Foundation (grant no. 2018M642317), Post-Doctoral Foundation of Jiangsu Province (grant no. 2018K095B), Six Talent Peaks Project of Jiangsu Province (grant nos. WSN-202 and WSW-183), Maternal and Child Health Research Project of Jiangsu Province (grant no. F201803), and Changzhou High-Level Medical Talents Training Project (grant no. 2016ZCL J020).

 Availability of data and materials

The datasets generated and/or analyzed during the current study are available in the GEO repository. GSE36949: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE36949GSE68316: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE68316

 Authors' contributions

FH and JS analyzed, interpreted and ensured the quality of the data. LT conceived and designed the study. LY and XH developed the methodology. LT wrote and reviewed the manuscript. All authors read and approved the final manuscript.

 Ethics approval and consent to participate

Not applicable.

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 References MaronBJCaseySAOlivottoISherridMVSemsarianCAutoreCAhmedABorianiGFranciaPWintersSLClinical course and quality of life in high-risk patients with hypertrophic cardiomyopathy and implantable cardioverter-defibrillatorsCirc Arrhythm Electrophysiol11e00582020182962597010.1161/CIRCEP.117.005820 MaronBJRecognition of hypertrophic cardiomyopathy as a contemporary, relatively common, and treatable disease (from the international summit V)Am J Cardiol11373974420142448486410.1016/j.amjcard.2013.11.018 SpudichJAHypertrophic and dilated cardiomyopathy: Four decades of basic research on muscle lead to potential therapeutic approaches to these devastating genetic diseasesBiophys J1061236124920142465549910.1016/j.bpj.2014.02.011 Roma-RodriguesCFernandesARGenetics of hypertrophic cardiomyopathy: Advances and pitfalls in molecular diagnosis and therapyAppl Clin Genet719520820142532841610.2147/TACG.S49126 de GregorioCAndòGRisk of sudden death and outcome in patients with hypertrophic cardiomyopathy with benign presentation and without risk factors: A word of comfort to younger patients?Am J Cardiol11450050120142501766310.1016/j.amjcard.2014.05.005 SpiritoPAutoreCFormisanoFAssenzaGEBiaginiEHaasTSBongioanniSSemsarianCDevotoEMusumeciBRisk of sudden death and outcome in patients with hypertrophic cardiomyopathy with benign presentation and without risk factorsAm J Cardiol1131550155520142463078610.1016/j.amjcard.2014.01.435 ToweECBosJMOmmenSRGershBJAckermanMJGenotype-phenotype correlations in apical variant hypertrophic cardiomyopathyCongenit Heart Dis10E139E14520152549500710.1111/chd.12242 CirinoALHarrisSLakdawalaNKMichelsMOlivottoIDaySMAbramsDJCharronPCaleshuCSemsarianCRole of genetic testing in inherited cardiovascular disease: A reviewJAMA Cardiol21153116020172879314510.1001/jamacardio.2017.2352 Roma-RodriguesCRaposoLRFernandesARMicroRNAs based therapy of hypertrophic cardiomyopathy: The road traveled so farBioMed Res Int201598329020152650485010.1155/2015/983290 SartorioCLLazzeroniDBertoliGCamiciPGTheranostic biomarkers in hypertrophic cardiomyopathy: Insights in a long road aheadFront Biosci (Landmark Ed)2217241749201728410142 LiebetrauCMollmannHDörrOSzardienSTroidlCWillmerMVossSGaedeLRixeJRolfARelease kinetics of circulating muscle-enriched microRNAs in patients undergoing transcoronary ablation of septal hypertrophyJ Am Coll Cardiol6299299820132374777910.1016/j.jacc.2013.05.025 BagnallRDTsoutsmanTShephardRERitchieWSemsarianCGlobal microRNA profiling of the mouse ventricles during development of severe hypertrophic cardiomyopathy and heart failurePLoS One7e4474420122302475810.1371/journal.pone.0044744 SayedDHongCChenIYLypowyJAbdellatifMMicroRNAs play an essential role in the development of cardiac hypertrophyCirc Res10041642420171723497210.1161/01.RES.0000257913.42552.23 NteliosDMeditskouSEfthimiadisGPitsisANikolakakiEGirtovitisFParcharidouDZegkosTKouidouSKarvounisHTzimagiorgisGElevated plasma levels of miR-29a are associated with hemolysis in patients with hypertrophic cardiomyopathyClin Chim Acta47132132620172868421910.1016/j.cca.2017.07.004 FangLEllimsAHMooreXLWhiteDATaylorAJChin-DustingJDartAMCirculating microRNAs as biomarkers for diffuse myocardial fibrosis in patients with hypertrophic cardiomyopathyJ Transl Med1331420152640454010.1186/s12967-015-0672-0 LiXHeXWangHLiMHuangSChenGJingYWangSChenYLiaoWLoss of AZIN2 splice variant facilitates endogenous cardiac regenerationCardiovasc Res1141642165520182958481910.1093/cvr/cvy075 OnoKKuwabaraYHorieTKimuraTLong non-coding RNAs as key regulators of cardiovascular diseasesCirc J821231123620182952698510.1253/circj.CJ-18-0169 KaurHSarmahDSarafJVatsKKaliaKBorahAYavagalDRDaveKRGhoshZBhattacharyaPNoncoding RNAs in ischemic stroke: Time to translateAnn N Y Acad Sci1421193620182968350610.1111/nyas.13612 LiangHPanZZhaoXLiuLSunJSuXXuCZhouYZhaoDXuBLncRNA PFL contributes to cardiac fibrosis by acting as a competing endogenous RNA of let-7dTheranostics81180119420182946400810.7150/thno.20846 HeLChenYHaoSQianJUncovering novel landscape of cardiovascular diseases and therapeutic targets for cardioprotection via long noncoding RNA-miRNA-mRNA axesEpigenomics1066167120182969221910.2217/epi-2017-0176 LiuXMaYYinKLiWChenWZhangYZhuCLiTHanBLiuXLong non-coding and coding RNA profiling using strand-specific RNA-seq in human hypertrophic cardiomyopathySci Data69020193119715510.1038/s41597-019-0094-6 YangWLiYHeFWuHMicroarray profiling of long non-coding RNA (lncRNA) associated with hypertrophic cardiomyopathyBMC Cardiovasc Disord156220152614170110.1186/s12872-015-0056-7 LiangHSuXWuQShanHLvLYuTZhaoXSunJYangRZhangLLncRNA 2810403D21Rik/Mirf promotes ischemic myocardial injury by regulating autophagy through targeting Mir26aAutophagy1211520193151255610.1080/15548627.2019.1659610 FengYXuWZhangWWangWLiuTZhouXLncRNA DCRF regulates cardiomyocyte autophagy by targeting miR-551b-5p in diabetic cardiomyopathyTheranostics94558456620193128577910.7150/thno.31052 ZhangMJiangYGuoXZhangBWuJSunJLiangHShanHZhangYLiuJLong non-coding RNA cardiac hypertrophy-associated regulator governs cardiac hypertrophy via regulating miR-20b and the downstream PTEN/AKT pathwayJ Cell Mol Med237685769820193146563010.1111/jcmm.14641 RehmsmeierMSteffenPHochsmannMGiegerichRFast and effective prediction of microRNA/target duplexesRNA101507151720041538367610.1261/rna.5248604 ShannonPMarkielAOzierOBaligaNSWangJTRamageDAminNSchwikowskiBIdekerTCytoscape: A software environment for integrated models of biomolecular interaction networksGenome Res132498250420031459765810.1101/gr.1239303 MaronBJMaronMSHypertrophic cardiomyopathyLancet38124225520132287447210.1016/S0140-6736(12)60397-3 RakowskiHDetermining hypertrophic cardiomyopathy mortality: Gaining wisdom from knowledge. JAMA Cardiol 2019. LiuFFuJHisDSunCHeGHuRZhangJLiuLPercutaneous intramyocardial septal radiofrequency ablation for interventricular septal reduction: An ovine model with 1-year outcomesCardiology2011020193174766510.1159/000502973 BollGRowinEJMaronBJWangWRastegarHMaronMSEfficacy of combined cox-maze IV and ventricular septal myectomy for treatment of atrial fibrillation in patients with obstructive hypertrophic cardiomyopathyAm J Cardiol12512012620203172726110.1016/j.amjcard.2019.09.029 KitowJDerdaAABeermannJKumarswarmyRPfanneAFendrichJLorenzenJMXiaoKBavendiekUBauersachsJThumTMitochondrial long noncoding RNAs as blood based biomarkers for cardiac remodeling in patients with hypertrophic cardiomyopathyAm J Physiol Heart Circ Physiol311H707H71220162742298410.1152/ajpheart.00194.2016 GómezJLorcaRRegueroJRMartinMMorisCAlonsoBIglesiasSDiaz-MolinaBAvanzasPCotoEGenetic variation at the long noncoding RNA H19 gene is associated with the risk of hypertrophic cardiomyopathyEpigenomics1086587320182996222510.2217/epi-2017-0175 ShiHLiJSongQChengLSunHFanWLiJWangZZhangGSystematic identification and analysis of dysregulated miRNA and transcription factor feed-forward loops in hypertrophic cardiomyopathyJ Cell Mol Med2330631620193033890510.1111/jcmm.13928 LiMChenXChenLChenKZhouJSongJMiR-1-3p that correlates with left ventricular function of HCM can serve as a potential target and differentiate HCM from DCMJ Trans Med1616120182988565210.1186/s12967-018-1534-3 SongLSuMWangSZouYWangXWangYCuiHZhaoPHuiRWangJMiR-451 is decreased in hypertrophic cardiomyopathy and regulates autophagy by targeting TSC1J Cell Mol Med182266227420142520990010.1111/jcmm.12380 WangDZhaiGJiYJingHmicroRNA-10a targets T-box 5 to inhibit the development of cardiac hypertrophyInt Heart J5810010620172810087310.1536/ihj.16-020 WangFYangXYZhaoJYYuLWZhangPDuanWYChongMGuiYHmiR-10a and miR-10b target the 3'-untranslated region of TBX5 to repress its expressionPediatr Cardiol351072107920142471497910.1007/s00246-014-0901-y JiaGDeMarcoVGSowersJRInsulin resistance and hyperinsulinaemia in diabetic cardiomyopathyNat Rev Endocrinol1214415320162667880910.1038/nrendo.2015.216 RautSKKumarASinghGBNaharUSharmaVMittalASharmaRKhullarMmiR-30c mediates upregulation of Cdc42 and Pak1 in diabetic cardiomyopathyCardiovasc Ther33899720152578119010.1111/1755-5922.12113 RautSKSinghGBRastogiBSaikiaUNMittalADograNSinghSPrasadRKhullarMmiR-30c and miR-181a synergistically modulate p53-p21 pathway in diabetes induced cardiac hypertrophyMol Cell Biochem41719120320162722173810.1007/s11010-016-2729-7 LiuXYeBMillerSYuanHZhangHTianLNieJImaeRAraiHLiYAblation of ALCAT1 mitigates hypertrophic cardiomyopathy through effects on oxidative stress and mitophagyMol Cell Biol324493450420122294950310.1128/MCB.01092-12 KrishnanJSuterMWindakRKrebsTFelleyAMontessuitCTokarska-SchlattnerMAasumEBogdanovaAPerriardEActivation of a HIF1alpha-PPARgamma axis underlies the integration of glycolytic and lipid anabolic pathways in pathologic cardiac hypertrophyCell Metab951252420091949090610.1016/j.cmet.2009.05.005 MurphyRTMogensenJMcGarryKBahlAEvansAOsmanESyrrisPGormanGFarrellMHoltonJLAdenosine monophosphate-activated protein kinase disease mimicks hypertrophic cardiomyopathy and wolff-parkinson-white syndrome: Natural historyJ Am Coll Cardiol4592293020051576683010.1016/j.jacc.2004.11.053 BauersachsJStörkSKungMWallerCFidlerFHoyerCFrantzSWeidemannFErtlGAngermannCEHMG CoA reductase inhibition and left ventricular mass in hypertrophic cardiomyopathy: A randomized placebo-controlled pilot studyEur J Clin Invest3785285920071797378110.1111/j.1365-2362.2007.01877.x RaniDSNallariPPriyamvadaSNarasimhanCSinghLThangarajKHigh prevalence of arginine to glutamine substitution at 98, 141 and 162 positions in troponin I (TNNI3) associated with hypertrophic cardiomyopathy among IndiansBMC Med Genet136920122287677710.1186/1471-2350-13-69 JimenezJTardiffJCAbnormal heart rate regulation in murine hearts with familial hypertrophic cardiomyopathy-related cardiac troponin T mutationsAm J Physiol Heart Circ Physiol300H627H63520112113147510.1152/ajpheart.00247.2010 ChenXXieDZhaoQYouZHMicroRNAs and complex diseases: From experimental results to computational modelsBrief Bioinform2051553920192904568510.1093/bib/bbx130 ChenXWangLQuJGuanNNLiJQPredicting miRNA-disease association based on inductive matrix completionBioinformatics344256426520182993922710.1093/bioinformatics/bty503 LiYZhangJHuoCDingNLiJXiaoJLinXCaiBZhangYXuJDynamic organization of lncRNA and circular RNA regulators collectively controlled cardiac differentiation in humansEBioMedicine2413714620172903760710.1016/j.ebiom.2017.09.015 LiZZhangYDingNZhaoYYeZShenLYiHZhuYInhibition of lncRNA XIST improves myocardial I/R injury by targeting miR-133a through inhibition of autophagy and regulation of SOCS2Mol Ther Nucleic Acids1876477320193173455710.1016/j.omtn.2019.10.004 TayYRinnJPandolfiPPThe multilayered complexity of ceRNA crosstalk and competitionNature50534435220142442963310.1038/nature12986 DykesIMEmanueliCTranscriptional and post-transcriptional gene regulation by long non-coding RNAGenomics Proteomics Bioinformatics1517718620172852910010.1016/j.gpb.2016.12.005 OhJGWatanabeSLeeAGorskiPALeePJeongDLiangLLiangYBaccariniASahooSmiR-146a suppresses SUMO1 expression and induces cardiac dysfunction in maladaptive hypertrophyCirc Res12367368520183035523310.1161/CIRCRESAHA.118.312751 WangKLongBLiuFWangJXLiuCYZhaoBZhouLYSunTWangMYuTA circular RNA protects the heart from pathological hypertrophy and heart failure by targeting miR-223Eur Heart J372602261120162680213210.1093/eurheartj/ehv713 BangCBatkaiSDangwalSGuptaSKFoinquinosAHolzmannAJustARemkeJZimmerKZeugACardiac fibroblast-derived microRNA passenger strand-enriched exosomes mediate cardiomyocyte hypertrophyJ Clin Invest1242136214620142474314510.1172/JCI70577 YuanYWangJChenQWuQDengWZhouHShenDLong non-coding RNA cytoskeleton regulator RNA (CYTOR) modulates pathological cardiac hypertrophy through miR-155-mediated IKKi signalingBiochim Biophys Acta Mol Basis Dis18651421142720193079486610.1016/j.bbadis.2019.02.014 LiJWuZZhengDSunYWangSYanYBioinformatics analysis of the regulatory lncRNAmiRNAmRNA network and drug prediction in patients with hypertrophic cardiomyopathyMol Med Rep2054955820193118054010.3892/mmr.2019.10289

Schematic depiction of the experimental design and flowchart of the steps performed in the present study.

View of the lncRNA-miRNA-mRNA network. (A) A total of 181 miRNA-lncRNA pairs and (B) 421 miRNA-mRNA pairs were identified according to both base sequence and expression level. (C) A total of 442 lncRNA-mRNA pairs were selected according to the ceRNA score and expression level. lncRNA, long non-coding RNA; miRNA, microRNA; ceRNA, competing endogenous RNA. View of the lncRNA-miRNA-mRNA network. (D) View of the lncRNA-mRNA network. The squares represent mRNAs and the circles represent lncRNAs. (E) View of the lncRNA-mRNA-miRNA network. Red nodes represent mRNAs, green nodes represent lncRNAs and blue nodes represent miRNAs. lncRNA, long non-coding RNA; miRNA, microRNA; ceRNA, competing endogenous RNA.

Functional analysis of differentially expressed mRNAs related to lncRNAs ENST00000597346.1 and ENST00000458178.1. (A) Biological processes and (B) KEGG analysis of differentially expressed mRNAs related to lncRNA ENST00000597346.1. lncRNA, long non-coding RNA; miRNA, microRNA; KEGG, Kyoto Encyclopedia of Genes and Genomes. Functional analysis of differentially expressed mRNAs related to lncRNAs ENST00000597346.1 and ENST00000458178.1. (C) Subnetwork of lncRNA ENST00000597346.1 and ENST00000458178.1. (D) lncRNA ENST00000458178.1-related biological processes and (E) KEGG analysis of differentially expressed mRNAs. lncRNA, long non-coding RNA; miRNA, microRNA; KEGG, Kyoto Encyclopedia of Genes and Genomes. Functional analysis of differentially expressed mRNAs related to lncRNAs ENST00000597346.1 and ENST00000458178.1. (F) lncRNA ENST00000458178.1-related miRNA-mRNA subnetwork. Green nodes represent miRNAs and red nodes represent mRNAs. lncRNA, long non-coding RNA; miRNA, microRNA; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Module analysis of the HCM-related lncRNA-miRNA network. (A) Two modules were determined to be highly related to HCM. lncRNA-mRNA networks in (B) module 1. HCM, hypertrophic cardiomyopathy; lncRNA, long non-coding RNA; miRNA, microRNA; KEGG, Kyoto Encyclopedia of Genes and Genomes. Module analysis of the HCM-related lncRNA-miRNA network. lncRNA-mRNA networks in (C) module 2. Red nodes represent lncRNAs and green nodes represent mRNAs. (D) Biological processes and (E) KEGG analysis of differentially expressed mRNAs related to module 1. HCM, hypertrophic cardiomyopathy; lncRNA, long non-coding RNA; miRNA, microRNA; KEGG, Kyoto Encyclopedia of Genes and Genomes. Module analysis of the HCM-related lncRNA-miRNA network. (F) Biological processes and (G) KEGG analysis of differentially expressed mRNAs related to module 2. HCM, hypertrophic cardiomyopathy; lncRNA, long non-coding RNA; miRNA, microRNA; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Selected differently expressed miRNAs in samples of hypertrophic cardiomyopathy patients and healthy controls.

miRNA Redefined name P-value Fold-change Expression
hsa-miR-514 hsa-miR-514a-3p 1.72x10-6 3.069 Upregulated
hsa-miR-373 hsa-miR-373-3p 2.33x10-6 2.577 Upregulated
hsa-miR-30c-1a hsa-miR-30c-3p 2.49x10-3 2.451 Downregulated
hsa-miR-144a hsa-miR-144-5p 3.38x10-6 2.398 Downregulated
hsa-miR-1247 hsa-miR-1247-5p 1.80x10-3 2.290 Downregulated
hsa-miR-10aa hsa-miR-10a-3p 6.63x10-6 2.286 Downregulated
hsa-miR-10a hsa-miR-10a-5p 3.18x10-6 2.101 Downregulated
hsa-miR-1268 hsa-miR-1268a 2.30x10-6 2.096 Downregulated

aOriginal miRNA name.

Top 20 genes in degree, betweenness and closeness.

Betweenness Closeness Degree Pagerank Name Gene type
2950 0.439597 60 0.107238 hsa-miR-514 miRNA
2950 0.439597 60 0.107238 hsa-miR-373 miRNA
1462.053 0.314904 46 0.06756 hsa-miR-30c-1a miRNA
2786.915 0.41195 34 0.049831 hsa-miR-1268 miRNA
568.3588 0.293722 31 0.04458 hsa-miR-10a miRNA
2332.149 0.40184 30 0.0444 hsa-miR-144a miRNA
184.1859 0.274059 15 0.022549 hsa-miR-10aa miRNA
147.3383 0.272917 14 0.020977 hsa-miR-1247 miRNA
132.3279 0.348404 6 0.008662 ENST00000597346.1 lncRNA
92.107 0.346561 5 0.007389 AAK1 mRNA
92.107 0.346561 5 0.007389 ENST00000458178.1 lncRNA
54.59979 0.337629 4 0.006136 ENST00000544461.1 lncRNA
4273.936 0.458042 4 0.006679 RELL1 mRNA
58.5175 0.344737 4 0.006111 SCN5A mRNA
49.93945 0.344737 4 0.006131 FNDC1 mRNA
49.93945 0.344737 4 0.006131 FAM123C mRNA
48.78154 0.341146 4 0.006126 ENST00000567093.1 lncRNA
48.78154 0.341146 4 0.006126 TNS1 mRNA
35.3917 0.332487 3 0.004916 ENST00000571219.1 lncRNA
28.90726 0.339378 3 0.004853 ARGFX mRNA

aOriginal miRNA name. miRNA, microRNA; lncRNA, long non-coding RNA; AAK-1, AP2-associated protein kinase 1; RELL1, RELT-like protein 1; SCN5A, sodium channel protein 5 subunit α; FNDC1, fibronectin type III domain-containing protein 1; FAM123C, APC membrane recruitment protein 3; TNS1, tensin-1; ARGFX, arginine-fifty homeobox.

Number of lncRNA-miRNA and miRNA-mRNA pairs.

lncRNA lncRNA-miRNA pairs miRNA-mRNA pairs Total number
ENST00000597346.1 6 58 64
ENST00000458178.1 5 58 63
ENST00000567093.1 4 53 57
ENST00000544461.1 4 44 48
ENST00000570167.1 3 51 54
ENST00000506222.2 3 50 53
ENST00000571219.1 3 44 47
ENST00000499521.2 2 238 240
ENST00000602172.1 2 238 240

miRNA, microRNA; lncRNA, long non-coding RNA.

lncRNA-miRNA-mRNA networks of the 5 selected lncRNAs.

lncRNA ceRNA score miRNA mRNA
ENST00000567093.1 1 hsa-miR-10a; hsa-miR-30c-1a; hsa-miR-1268; hsa-miR-1247 TNS1
ENST00000458178.1 0.687 hsa-miR-10a; hsa-miR-30c-1a; hsa-miR-1268 ANK1
ENST00000544461.1 0.666 hsa-miR-10a; hsa-miR-144a; hsa-miR-1268 KCNN2, SCN5A
    hsa-miR-10a; hsa-miR-1268; hsa-miR-1247 TNS1
ENST00000597346.1 0.666 hsa-miR-10a; hsa-miR-30c-1a; hsa-miR-1268 ANK1
ENST00000571219.1 0.666 hsa-miR-30c-1a; hsa-miR-1247 FAM22F, UROC1, AK1, CRYGN, PPP1R26, TNS1
    hsa-miR-144a; hsa-miR-30c-1a RFT1, FAM196B, CACNB2, EPB49, RGPD1, GRIK2, SCN5A, AAK1

aOriginal miRNA name. miRNA, microRNA; lncRNA, long non-coding RNA; ceRNA, competitive endogenous RNA; TNS1, tensin-1; ANK1, ankyrin-1; KCNN2, small conductance calcium-activated potassium channel protein 2; SCN5A, sodium channel protein 5 subunit α; FAM22F protein; UROC1, urocanate hydratase; AK1, adenylate kinase isoenzyme 1; CRYGN, γ-crystallin N; PP1R26, Protein phosphatase 1 regulatory subunit 26; RFT1, solute carrier family 52, riboflavin transporter, member 1; FAM196B, protein INSYN2B; CACNB2, voltage-dependent L-type calcium channel subunit β-2; EPB49, dematin; RGPD1, RANBP2-like and GRIP domain-containing protein 1; GRIK2, glutamate receptor ionotropic, kainite 2; AAK-1, AP2-associated protein kinase 1.