www.nature.com/scientificreports Correction: Author Correction OPEN MicroRNA interactome analysis predicts post-transcriptional regulation of ADRB2 and PPP3R1 Received: 3 April 2017 Accepted: 7 June 2018 in the hypercholesterolemic Published online: 04 July 2018 myocardium Bence Ágg 1,2,3, Tamás Baranyai1, András Makkos1, Borbála Vető4, Nóra Faragó5, Ágnes Zvara5, Zoltán Giricz 1, Dániel V. Veres6, Péter Csermely 6, Tamás Arányi4, László G. Puskás5, Zoltán V. Varga1 & Péter Ferdinandy 1,2,7 Little is known about the molecular mechanism including microRNAs (miRNA) in hypercholesterolemia- induced cardiac dysfunction. We aimed to explore novel hypercholesterolemia-induced pathway alterations in the heart by an unbiased approach based on miRNA omics, target prediction and validation. With miRNA microarray we identifed forty-seven upregulated and ten downregulated miRNAs in hypercholesterolemic rat hearts compared to the normocholesterolemic group. Eleven mRNAs with at least 4 interacting upregulated miRNAs were selected by a network theoretical approach, out of which 3 mRNAs (beta-2 adrenergic receptor [Adrb2], calcineurin B type 1 [Ppp3r1] and calcium/calmodulin-dependent serine protein kinase [Cask]) were validated with qRT-PCR and Western blot. In hypercholesterolemic hearts, the expression of Adrb2 mRNA was signifcantly decreased. ADRB2 and PPP3R1 protein were signifcantly downregulated in hypercholesterolemic hearts. The direct interaction of Adrb2 with upregulated miRNAs was demonstrated by luciferase reporter assay. Gene ontology analysis revealed that the majority of the predicted mRNA changes may contribute to the hypercholesterolemia-induced cardiac dysfunction. In summary, the present unbiased target prediction approach based on global cardiac miRNA expression profling revealed for the frst time in the literature that both the mRNA and protein product of Adrb2 and PPP3R1 protein are decreased in the hypercholesterolemic heart. Although, many guidelines efectively support the treatment of cardiovascular disorders, cardiovascular diseases are still the leading cause of mortality and morbidity in developed countries1. Te high morbidity and mortality of cardiovascular diseases are primarily attributed to the growing prevalence of chronic metabolic diseases, such as diabetes mellitus, hypercholesterolemia, obesity, which are well-established risk factors of numerous cardio- vascular diseases (e.g., chronic heart failure, acute myocardial infarction, etc.)1. Hypercholesterolemia is mainly responsible for the development of atherosclerosis, therefore, hypercholesterolemic patients are exposed to more severe progression of acute and chronic ischemic heart diseases2. Importantly, it has been also demonstrated in preclinical and clinical studies that hypercholesterolemia directly, and independently from the development 1Department of Pharmacology and Pharmacotherapy, Semmelweis University, 1089, Budapest, Hungary. 2Pharmahungary Group, 6722, Szeged, Hungary. 3Heart and Vascular Center, Semmelweis University, 1122, Budapest, Hungary. 4Institute of Enzymology, Research Center for Natural Sciences, Hungarian Academy of Sciences, 1117, Budapest, Hungary. 5Institute of Genetics, Biological Research Center of the Hungarian Academy of Sciences, 6726, Szeged, Hungary. 6Department of Medical Chemistry, Semmelweis University, 1094, Budapest, Hungary. 7Cardiovascular Research Group, Department of Biochemistry, University of Szeged, 6720, Szeged, Hungary. Bence Ágg and Tamás Baranyai contributed equally to this work. Zoltán V. Varga and Péter Ferdinandy jointly supervised this work. Correspondence and requests for materials should be addressed to P.F. (email: peter. [email protected]) SCIENTIFIC REPORTS | (2018) 8:10134 | DOI:10.1038/s41598-018-27740-3 1 www.nature.com/scientificreports/ of atherosclerosis, impairs both systolic and diastolic cardiac function3,4, and exacerbates ischemia/reperfusion injury5 possibly via interfering with endogenous protective pathways6–8. However, the underlying mechanism of impaired myocardial function due to hypercholesterolemia is still unclear. It is known that chronic metabolic derangements (e.g., type 2 diabetes mellitus) modify the microRNA (miRNA) expression pattern of fundamental metabolic and survival mechanisms of the heart9. MiRNAs are con- served, non-coding, about 22 nucleotide long RNA molecules, playing pivotal role in the posttranscriptional regulation of gene expression both in physiological and pathophysiological conditions10, including diseases of the cardiovascular system11–13. Generally, high-throughput miRNA analysis is applied as a screening method to identify single targets to investigate in depth. Similarly, we have previously shown that hypercholesterolemia alters miRNA expression profle signifcantly, and as a consequence, decreased expression of miR-25 induces oxidative/nitrative stress in the myocardium14. A single miRNA is able to regulate even hundreds of mRNAs, and a single mRNA might be targeted by several miRNAs, implying that a slightly changed miRNA expression profle might substantially alter multiple pathways simultaneously thereby markedly infuencing the phenotype of var- ious diseases15,16. Terefore, the systematic interpretation of high-throughput methods might be a powerful tool to understand the underlying complex mechanisms of cardiovascular disorders17. Here we aimed to analyze the miRNA expression profle of hypercholesterolemic rat hearts with compre- hensive bioinformatic methods as recommended in the Position Paper of the European Society of Cardiology Working Group on Cellular Biology of the Heart18,19. We developed an unbiased target prediction approach and subsequently validated changes in the expression of predicted genes, Adrb2 and Ppp3r1, in case of Adrb2 direct interaction with miR-195 and miR-322 was also demonstrated. Te adrenoceptor beta 2 (Adrb2) gene encodes the mRNA and the protein of an adenylate cyclase-activating, G-protein coupled adrenergic receptor denoted by the symbols Adrb2 and ADRB2 respectively. While the tran- scribed mRNA of the protein phosphatase 3, regulatory subunit B, alpha (Ppp3r1) gene is denoted by the same symbol (Ppp3r1), its protein product is the calcineurin B type 1 (CNB1) which is one of the two isoforms of the regulatory subunit of the calcineurin serine/threonine phosphatase heterodimer enzyme. Te mRNA (Cask) and protein product (CASK) of calcium/calmodulin dependent serine protein kinase (Cask) gene was also investi- gated in our study, which could be involved in intracellular calcium mediated signaling pathways20. Furthermore, gene ontology analysis revealed that an altered miRNA fngerprint is associated with impaired cardiac function and an overactivated protein kinase pattern. Results miRNA microarray analysis. Datasets of miRNA microarray analysis of normo- and hypercholesterolemic rats from our previous study14 were further analyzed in the present study. Te miRNA microarray data were deposited in the ArrayExpress database (https://www.ebi.ac.uk/arrayexpress/) in Minimum Information About a Microarray Experiment (MIAME) compliant format with an accession number of E-MTAB-3979. Tree hundred ffy miRNAs were assayed, among which 120 miRNAs were detectable. Fourty seven miRNAs were upregulated, while 10 miRNAs were downregulated in hypercholesterolemic rat hearts as compared to the normocholester- olemic control rat hearts (Fig. 1A). Te expression of 8 altered miRNAs was validated with qRT-PCR (Fig. 1B). miRNA target prediction and validation. Out of the diferentially expressed (47 upregulated and 10 downregulated) miRNAs 43 upregulated and 8 downregulated miRNAs were further investigated. Four upregu- lated and two downregulated miRNAs did not meet our inclusion criteria, i.e. these miRNAs were not involved in any miRNA-target interactions presented in at least two predicted (miRDB, microRNA.org) or one experimen- tally validated (miRTarBase) miRNA-target interaction databases. According to previous studies, the regulation of target mRNAs by several miRNAs may be synergistic exerting a major efect on the bioavailability of the tar- get mRNAs21. To fnd such miRNA target hubs, we constructed a classical miRNA-target network in which the nodes represent miRNAs and putative target mRNAs, while edges symbolize miRNA-target interactions (Fig. 2A; High-resolution network is available as Supplementary Fig. 1). Te total number of putative mRNAs regulated by down- and upregulated miRNAs were 79 and 330, respectively, from which 32 mRNAs were theoretically mod- ulated by both down- and upregulated miRNAs (Fig. 2D). From the indicated 409 mRNAs, 11 mRNA (Adrb2, Cask, Lppr4, Mob4, Myt1l, Ppp3r1, Pth, Ptprz1, Sgk1, Stx1a and Wee1) had at least 4 miRNA-target interactions. Out of the above 11 mRNAs four miRNA targets, i.e, Adrb2, Sgk1, Ppp3r1 and Cask mRNAs have been selected for further validation based on systematical review of the literature (Fig. 2B,C). Te results of the PubMed queries are shown in Supplementary Table 1, also indicating the number of articles found to be relevant by manual cura- tion of the search results. As parathyroid hormone (Pth) has negligible expression in the heart according to the Human Protein Atlas22, literature mining was not performed for Pth. Te expression of Adrb2 mRNA was signifcantly downregulated in hypercholesterolemia compared to the normocholesterolemic group, however, mRNA expression of Ppp3r1 and Cask was not afected (Fig. 3A). Since miRNAs are not necessarily mediating