Zou et al. J Transl Med (2019) 17:45 https://doi.org/10.1186/s12967-019-1790-x Journal of Translational Medicine

RESEARCH Open Access Bioinformatic analysis for potential biomarkers and therapeutic targets of atrial fbrillation‑related stroke Rongjun Zou1†, Dingwen Zhang1†, Lei Lv1†, Wanting Shi2, Zijiao Song3, Bin Yi1, Bingjia Lai4, Qian Chen5, Songran Yang6,7* and Ping Hua1*

Abstract Background: Atrial fbrillation (AF) is one of the most prevalent sustained arrhythmias, however, epidemiological data may understate its actual prevalence. Meanwhile, AF is considered to be a major cause of ischemic strokes due to irregular heart-rhythm, coexisting chronic vascular infammation, and renal insufciency, and blood stasis. We studied co-expressed to understand relationships between atrial fbrillation (AF) and stroke and reveal potential biomarkers and therapeutic targets of AF-related stroke. Methods: AF-and stroke-related diferentially expressed genes (DEGs) were identifed via bioinformatic analysis Omnibus (GEO) datasets GSE79768 and GSE58294, respectively. Subsequently, extensive target prediction and network analyses methods were used to assess –protein interaction (PPI) networks, (GO) terms and pathway enrichment for DEGs, and co-expressed DEGs coupled with corresponding predicted miRNAs involved in AF and stroke were assessed as well. Results: We identifed 489, 265, 518, and 592 DEGs in left atrial specimens and cardioembolic stroke blood samples at < 3, 5, and 24 h, respectively. LRRK2, CALM1, CXCR4, TLR4, CTNNB1, and CXCR2 may be implicated in AF and the hub- genes of CD19, FGF9, SOX9, GNGT1, and NOG may be associated with stroke. Finally, co-expressed DEGs of ZNF566, PDZK1IP1, ZFHX3, and PITX2 coupled with corresponding predicted miRNAs, especially miR-27a-3p, miR-27b-3p, and miR-494-3p may be signifcantly associated with AF-related stroke. Conclusion: AF and stroke are related and ZNF566, PDZK1IP1, ZFHX3, and PITX2 genes are signifcantly associated with novel biomarkers involved in AF-related stroke. Keywords: Gene analysis, Biomarkers, Atrial fbrillation-related stroke

Background 80-years-of-age [1, 2]. However, epidemiological data Atrial fbrillation (AF) is one of the most prevalent may understate its actual prevalence, because 40% of sustained arrhythmias, having an age-adjusted hospi- patients are asymptomatic and remain undiagnosed with talization incidence of 1–4% of the general population subclinical AF [3]. Tere is also evidence that patients and an prevalence rising of > 13% for those older than with AF have signifcantly increased cardiovascular- related morbidity, given its association with atrial and ventricular mechanical or electrical failure, structural *Correspondence: [email protected]; [email protected] †Rongjun Zou, Dingwen Zhang and Lei Lv contributed equally to this and hemodynamic alterations, and thromboembolic article events [3]. 1 Department of Cardio‑Vascular Surgery, Sun Yat‑sen Memorial Hospital, Stroke is the leading cause of disability and death and Sun Yat-sen University, Guangzhou 510120, China 6 The Biobank of Sun Yat‑sen Memorial Hospital, Sun Yat-sen University, has an estimated incidence of 3.73 (95% CI 3.51–3.96) Guangzhou 510120, China per 1000 person-years among black- and white- adults in Full list of author information is available at the end of the article an atherosclerosis risk in communities (ARIC) cohort [4].

© The Author(s) 2019. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat​iveco​mmons​.org/licen​ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creat​iveco​mmons​.org/ publi​cdoma​in/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Zou et al. J Transl Med (2019) 17:45 Page 2 of 12

Furthermore, global increases in stroke prevalence plus Methods stroke-related disability and mortality associated with Materials and methods aging will increase [5, 6]. Tus, we may not now know the GSE79768 and GSE58294 datasets were downloaded actual true burden of stroke due to limits in brain imag- from GEO (http://www.ncbi.nlm.nih.gov/geo/) [13] and ing identifcation in < 10 mm small hypointense areas and expression profling arrays were generated using GPL570 silent infarctions for 28% of those patients older than (HG‑U133_Plus_2) Afymetrix Genome U133 65-years-of-age [7]. AF is commonly classifed as parox- Plus 2.0 Array (Afymetrix, Santa Clara, CA). Addition- ysmal, persistent or permanent, or new onset arrhythmia ally, the GSE79768 dataset, including 26 specimens with basing on the present continuous time, which mainly paired left atrial (LA) and right atrial (RA) tissue obtained included that paroxysmal AF was self-terminates within from 13 patients was used to identify diferential LA- 7 days, while persistent AF was lasts longer than 7 days or to-RA gene expression and molecular mechanisms needs cardioversion, and usually has lasted for 3 months for patients with persistent AF or sinus rhythm (SR) [8]. As we all kwon, AF is considering to be a major cause abnormalities and we describe potential mechanisms of of ischemic strokes due to irregular heart-rhythm, coex- AF-related remodeling in the LA and the relationship isting chronic vascular infammation, and renal insuf- between LA arrhythmogenesis and thrombogenesis. In fciency, and blood stasis. According to Rivaroxaban this study, persistent AF patients has lasts continuously Once Daily Oral Direct Factor Xa Inhibition Compared for > 6 months, while the SR patients had no evidence of With Vitamin K Antagonism for Prevention of Stroke AF clinically and any anti-arrhythmic drug history. Blood and Embolism Trial in Atrial Fibrillation (ROCKET-AF) samples of GSE58294 were collected from cardioembolic trial study, Steinberg et al. [9] suggested that the par- stroke (N = 69) and control patents (N = 23) at < 3, 5, and oxysmal AF patients carrying a lower adjusted rate of 24 h. stroke or systemic embolism (adjusted HR: 0.78, 95% CI Data processing 0.61–0.99, P = 0.045), all-cause mortality (adjusted HR: 0.79, 95% CI 0.67–0.94, P = 0.006), and the composite of R packages of “afy”, “afyPLM”, and “limma” (http://www. stroke or systemic embolism or death (adjusted HR: 0.82, bioco​nduct​or.org/packa​ges/relea​se/bioc/html/afy.html), 95% CI 0.71–0.94, P = 0.005) than persistent AF patients provided by a bioconductor project [14], were applied after adjusted efcacy and safety outcomes. According to assess GSE79768 and GSE58294 RAW datasets. We to the Oxford vascular study (OXVASC), nearly 43.9% of used background correction, quantile normalization, ischemic strokes were associated with AF among patients probe summarization and log2‑transformation, to create 80 years-of-age or older who had a threefold increase in a robust multi-array average (RMA), a log-transformed AF in the past 3 decades [10]. However, this assumption perfect match, and a mismatch probe (PM and MM) has been challenged by the atrial fbrillation reduction methods. Te Benjamini‑Hochberg method was used atrial pacing trial (ASSERT) which identifed a tempo- to adjust original p-values, and the false discovery rate ral association between subclinical AF and stroke risk (FDR) procedure was used to calculate fold-changes (FC). among patients with implantable pacemakers and def- Genes expression values of the|log2 FC| > 1and adjusted brillators. Tey reported that only 8% and 16% of patients p < 0.05 were used for fltering AF-DEGs. However, the had an association between pre-detected and post- |log2 FC| > 1.5 and adjusted p < 0.05 were used to identify detected AF within months of stroke or systemic embo- stroke-DEGs, given that blood sample specifcity pointed lism, respectively [11]. Of note, AF is often intermittent to many genes. Additionally, we calculated and made and asymptomatic, and presents as an electromechani- Venn diagrams for co-DEGs for AF- and stroke-DEGs. cal disassociation of atrial fbrillation. Clinically, current Finally, we applied online prediction tools utilizing stroke risk scores and traditional diagnosis with an elec- microRNA Data Integration Portal (mirDIP) (http:// trocardiogram are practical, while the limitation of pre- ophid​.utoro​nto.ca/mirDI​P) [15], miRDB (http://mirdb​ dict stroke risk accurately in individual AF patients was .org/) [16], TargetScan (v7.1; http://www.targe​tscan​.org/ signifcantly identifed, especially in persistent AF which vert_71/) [17], and DIANA Tools (http://diana​.imis.athen​ carrying a higher risk of stroke or systemic embolism a-innov​ation​.gr/Diana​Tools​/) [18], to predict poten- and all-cause mortality [12]. In this study, we identifed tial microRNA targeting. Subsequently, we used the co-expressed diferentially expressed genes (co-DEGs) of mirDIP, miRDB, TargetScan, and Diana Tools software persistent AF and stroke and elucidated molecular mech- to predict which of the selected miRNAs could target co- anisms and pathology of AF-related DEGs (AF-DEGs) DEGs. We determined 5 top candidate miRNAs based on and stroke-related DEGs (stroke-DEGs). Finally, we higher predicted scores for ≥ 3 prediction tools for each provide a bioinformatic analysis of DEGs and predicted co-DEG. microRNAs (miRNAs) for AF patients prone to stroke. Zou et al. J Transl Med (2019) 17:45 Page 3 of 12

Identifcation of protein–protein interaction (PPI) networks AF- and stroke-DEGs were confrmed. We found 489 of DEGs DEGs in LA specimens of AF patients compared with PPI networks of AF- and stroke-DEGs were analyzed SR patients, including 428 down-regulated genes and using the search tool for the retrieval of interacting genes 61 up-regulated genes. However, total of 265, 518, and (STRING database, V10.5; http://strin​g‑db.org/) that 592 DEGs were identifed following the time points of predicted protein functional associations and protein– less than 3, 5, and 24 h after stroke, respectively. Here, protein interactions. Subsequently, Cytoscape software we defned 210 co expressed DEGs in the three time (V3.5.1; http://cytos​cape.org/) was applied to visualize points mentioned above as the stroke-DEGs. Heatmaps and analyze biological networks and node degrees, after of AF-DEGs in relation to infammatory and immune downloading analytic results of the STRING database response, ion channels, and cell signaling were con- with a confdence score > 0.4 [19]. ducted for genes expression and these data appear in Fig. 1 and Additional fle 1: S1. Simultaneously, Fig. 2 and Additional fle 2: S2 has shown the genes expres- Functional enrichment analysis sion value in relation to infammatory response, energy Gene Ontology (GO) and Kyoto Encyclopedia of Genes metabolism, ions channel and transportation, and neu- and Genomes (KEGG) pathway enrichment analyses of ronal regulation above the stroke-DEGs. AF- and stroke-DEGs were carried out using the data- base for annotation, visualization and integrated discov- ery bioinformatics resources (DAVID Gene Functional Functional enrichment in Co‑DEGs Classifcation Tool, http://david​.abcc.ncifc​rf.gov/) [20], Figure 3c illustrates expressed AF- and stroke-DEGs and REACTOME databases (v62; http://www.react​ome. and co expressed genes. Interestingly, four co expressed org) [21]. GO terms and KEGG maps of biological func- DEGs, including zinc fnger protein 566 (ZNF566), tions associated with a p < 0.05 was considered to be sig- PDZK1 interacting protein 1(PDZK1IP1), zinc fn- nifcantly enriched. In addition, we presented diferent ger homeobox 3 (ZFHX3), paired-like homeodomain biofunctions of AF- and stroke-DEGs in biological pro- 2 (PITX2), were observed. Te AmiGO database was cesses, molecular functions, and cellular components used to confrm GO term enrichment related to biolog- from DAVID and REACTOME databases, respectively. ical processes, molecular functions, and cellular com- Subsequently, the AmiGO database (v2.0; http://amigo​ ponents and Co-DEGs were associated with various .geneo​ntolo​gy.org/amigo​/) was used to analyze the GO processes as indicated in Table 1. consortium for selected co-DEGs to verify the accuracy and annotate biofunctions of identifed co-DEGs [22]. Using microRNA target prediction, online tools from PPI network analysis and functional GO terms and pathway Diana-miRPath (v3.0; http://www.micro​rna.gr/miRPa​ enrichment analyses thv3) [18] were applied to evaluate interactions between We identifed 256 and 43 nodes from PPI network miRNA previously identifed using prediction tools and of AF- and stroke-DEGs, respectively and these data co-DEGs involved in AF and stroke. appear in Fig. 3. Here, the hub nodes, including leu- cine-rich repeat kinase 2 (LRRK2; degree = 38), calm- Identifcation of co‑DEGs associated with nervous odulin 1 (CALM1; degree = 25), chemokine (C-X-C or cardiovascular diseases motif) 4 (CXCR4; degree = 25), toll-like recep- tor 4 (TLR4; degree 21), catenin (cadherin-associated Te comparative toxicogenomics database (http://ctdba​ = protein), beta 1(CTNNB1; degree 21), and chemokine se.org/) was used to fnd integrated chemical-gene, = (C-X-C motif) receptor 2 (CXCR2; degree 21) are chemical-disease, and gene-disease interactions to gen- = considering as hub-genes in related to AF maintaining. erate expanded networks and predict novel associations However, the hub-genes, involved in CD19 (degree 5), [23]. We used these data to analyze relationships between = fbroblast 9 (FGF9; degree 5), SRY (sex gene products and nervous or cardiovascular diseases. = determining region Y)-box 9 (SOX9; degree 5), gua- Here, relationships between co-DEGs and diseases and = nine nucleotide binding protein (), gamma association or an implied association were identifed. transducing activity polypeptide 1(GNGT1; degree = 4), Results and noggin (NOG; degree = 4), are demonstrated in Identifcation of DEGs stroke-DEGs with a relative higher degree. Using the DAVID database, the top 5 GO terms We identifed 54,674 probes corresponding to 20,484 related biological processes among those genes were genes in GSE79768 and GSE58294 datasets and primarily associated with infammatory response (Fold Zou et al. J Transl Med (2019) 17:45 Page 4 of 12

AF-DEGs

abCellular signaling Ions channel

SR SR AF AF ATP13A3 NPY5R 10 GJA5 10 IGSF6 ANO6 AGTR2 FPR3 9 GPR83 9 TMEM38B IL12RB2 MEF2C 8 IFNGR2 8 CLIC2 UPK1B GSTM2 7 KLRB1 7 CXCR4 CXCR2 NPY1R 6 TSPAN15 SOX18 6 CD160 CALM1 5 KLRD1 ANO1 NPY1R 5 CXCR2 4 CALCRL CHRNB1 CD27 4 CCR2 AF.7 AF.2 AF.6 AF.3 AF.1 AF.5 SR.2 SR.6 SR.3 SR.4 SR.1 SR.5 AF.4 ATP1B4 NPY5R ASIC2 CD24 XCR1 PROK2 UTS2 AF.1 AF.3 AF.4 AF.6 AF.7 AF.2 AF.5 SR.1 SR.5 SR.2 SR.4 SR.3 SR.6

cdInflammatory response Immune response

SR SR AF AF

CXCL11 IL18 PROK2 10 SBSPON 12 AGTR2 IL1B XCR1 9 GZMA MS4A2 IGKC 10 S100A9 8 CD27 CXCR4 TLR4 S100A8 7 ETS1 8 NMI PPBP LYN 6 IGLL1 IL1B CCR2 6 IL18 5 ADAMDEC1 CXCL11 ALOX15 4 MS4A2 PPBP IGSF6 4 PTGS2 HLA DQB2 CXCR2 CD1E CCR2 PRG4 S100A12 SLPI BLNK HLA DRA CD27 FCGR3B FPR3 CTSS TLR4 TRIM22 KIT C1QC AF.1 AF.2 SR.5 SR.4 SR.6 SR.2 SR.1 SR.3 AF.7 AF.4 AF.5 AF.6 AF.3 HLA DMB HLA DPB1 SR.6 SR.3 SR.1 SR.5 SR.2 SR.4 AF.1 AF.4 AF.5 AF.3 AF.6 AF.2 AF.7

Fig. 1 Hierarchical clustering analysis of AF-related diferentially expressed genes: a–d results of hierarchical clustering analysis for DEGs expression in relation to cellular signaling, , infammatory and immune responses. Red, greater expression. Blue, less expression

Enrichment: 4.08; p value: 1.11E−07), immune response Enrichment: 1.75; p-value: 8.88E−04), and MHC class (Fold Enrichment: 3.50; p-value: 2.49E−06), regulation II protein complex (Fold Enrichment: 13.47; p-value: of MAP kinase activity (Fold Enrichment: 9.53; p-value: 0.003) in relation to cellular components. In addition, 1.92E−05), and regulation of NF-kappa B activety (Fold the terms related molecular functions were mainly Enrichment: 16.73; p-value: 1.95E−04). Tere is sig- involved in ion channel binding (Fold Enrichment: 5.06; nifcant correlation in plasma membrane (Fold Enrich- p-value: 0.001), Y receptor activity (Fold ment: 1.60; p-value: 1.37E−06), extracellular region (Fold Enrichment: 23.84; p-value: 0.007), and transmembrane Zou et al. J Transl Med (2019) 17:45 Page 5 of 12

a PPI networks of AF-DEGs

IRF2 HLA-DQB2 HLA-DMB KLRC3 IFI44L UPK1B

TPSAB1 TRIM22 CPA3 RLN2 TGM1 RSAD2 CPA4 HLA-DRA PLA2G4D HLA-DPB1 SPINK5 S100A12 CD160 CALCRL KLRD1 GZMK MNDA ALOX15 GPR83 IL12RB2 TRAT1 HTR2B VNN2 KLRB1 IGLL1 PIGA XCR1 PROK2 BTLA S100A8 MS4A2 S100A9 PLA2G7 UTS2 PPBP NCF2 GZMA IGJ RGS18 NMU PLA2G2A ELAVL2 IGSF6 CD1D GTF2IRD2B MLF1 NPY1R SRGN TNFRSF17 CXCL11 COLQ IL18 APPL2 FCGR3B BLNK SPESP1 EVI2B CD27 CCR2 CTSS MPEG1 AGTR2 FPR3 ALYREF CENPH P2RY13 CD69 TTC37 ELMO1 DDX43 CXCR2 BCL2A1 IL1B P2RY12 GTF2I C1QC CDR1 CENPQ PRRX1 CLPS TLR4 ERCC6L NPY5R ADRA2A JAM3 ZNF385B PILRA IL33 LYN SLPI CXCR4 ATP8B4 PAIP2B GPR160 GPR34 JUND CD300C DUSP5 KITLG FLI1 ZNF41 RNF6 PTPRR IGDCC4 COL6A3 COL15A1 MS4A7 DNAJB4 RAPGEF4 PTGS2 UBE2M PDGFD C8orf4 NTNG1 ETS1 CSF2RAFAR1 COL21A1 ID1 ATP1B4 EIF4E3 UCP1 ASB11 ENPEP MAP3K1 ESRRG PPP1R3A STAP1 SNAI2 PABPC5 GOLPH3 KIT KCTD6

HSPA4L EIF2S3 ACN9 MEF2C DHRS9 PLLP EPHA7 DUSP10 NET1 SOX18 CHIC2 DCK PRKG1 FGF23 SLC28A3 OSBPL11 ALDH1A1 ATP13A3 STYX ATF1 VGLL2 HMGCLL1 FGF12 RELN EPCAM PRR15L ZNF22 SCOC LMAN1 KRT19 ASPNGUCY1B3 LRRK2 MSLN LAIR2 PTGR2 TMX1 GNPNAT1 LRRC39 CAPN6 MYL3 MFN1 CALM1 ADAM10 DACT1 ZNF208 FHL2 ZNF781 ZNF280A SLC8A1 ANO1 TMED2 FLRT3 RSPO4 GNPDA2 KCNIP2 ZNF572 FRZB CTNNB1 ZNF566 ZNF780A CPSF6 MAN1A1 PTN KRT18 NEGR1 ZFP112 ZNF366 ZNF532 COG3 RERGL ZNF217 ZNF146 ACTA1 CNGA1 ZNF506 ABRA ZNF684 CSN3 ZNF238 TTC32 CLIC2 FSCN1 ANO6 PKP2 ARL6 LRRN4 ZNF461 TXNDC16 ZNF501 VAMP3 RAB25 GFM2 RAB8A NRAP RDX TOX2 BBS5 RAB8B SLC26A4 SUMO4 HSF2 LRP2 MRPL44 SLC9A3 DSC3 NEDD1 CCDC41 CNN3 KPNA5 MRPL50 SENP7 USP44

TAF12 CORO1C SCARF1 KRT6A

bcPPI networks of Stroke-DEGsVenn diagrams of DEGs

Stroke after 5h GNRHR2 LPAR4 Stroke after 3h Stroke after 24h VANGL2 ACP5 GNGT1 TAS2R31 24 117 NOG PITX2 USP18 TNNC2 SCN3A GNAZ 16 MSX1 LINGO2 1 172 GRAPL SHC4 FGF9 USP43 AF-DEGs SOX9 FGF20 XRCC2 UBE2S 0 201 SHOX CTGF POLD1 472 187 ESRP1 GRHL2

CD22 KLHL13 BCL9L TENM2 CD27 PRSS55 3 4 19 CD72 FAT3 SPIB JUP TNFRSF25 HLA-DQA1 CD19

TCL1A ABCB4 SLCO1B1 5 0 ZNF566 PDZK1IP1 ZFHX3 4 PITX2

Fig. 2 Hierarchical clustering analysis of stroke-related DEGs: a–c results of hierarchical clustering analysis for DEG expression in relation to energy metabolism, ion channel, infammatory response, and neuronal regulation. Red, greater expression. Blue, less expression. a PPI network of AF-related DEGs; b PPI network of stroke-related DEGs. Red, greater degree. yellow, lesser degree; c Venn diagrams of DEGs Zou et al. J Transl Med (2019) 17:45 Page 6 of 12

Stroke-DEGs

abEnergy Metabolism Ions channel Normal Normal Stroke Stroke

AADAT SLC17A6 CNR2 KCNG2 10 AASS 10 KCNJ13 TJP1 TJP1 BOK KCNH8 IL12RB2 8 SLC13A5 8 BAK1 SLC5A10 CCR7 LIM2 CD27 6 SLC5A11 6 IDO1 TMEM109 ACP5 DHRS3 HIP1R 4 NOG 4 TNFRSF25 Stroke15 Stroke16 Stroke8 Stroke19 Stroke17 Stroke18 Stroke5 Stroke14 Stroke9 Stroke23 Stroke13 Stroke12 Stroke11 Stroke21 Stroke3 Stroke2 Stroke7 Stroke1 Stroke4 Stroke10 Stroke20 Stroke22 Normal2 Normal1 Normal1 Normal8 Normal2 Stroke6 Normal1 Normal17 Normal Normal1 Normal2 Normal1 Normal3 Normal2 Normal5 Normal4 Normal9 Normal1 Normal Normal10 Normal15 Normal1 Normal23 Normal14 SCN3A CLIC3 2 HES1 2 CLDN5 6 7 2 2 0 1 9 3 1 8 6 SLC5A2 Stroke10 Stroke15 Stroke20 Str o Stroke12 Stroke13 Stroke14 Stroke16 Stroke21 Stroke2 Stroke22 Stroke3 Stroke1 Stroke11 Stroke23 Strok Normal Normal Normal Normal12 Normal18 Normal16 Normal21 Normal23 Normal14 Stroke6 Stroke17 Stroke4 Normal Normal13 Normal15 Normal Normal Normal10 Normal Normal Normal22 Normal17 Stroke18 Stroke8 Stroke19 Normal20 Normal Normal11 Normal19 Stroke5 k e7 e9 5 8 9 4 1 3 6 7 2

c Inflammatory response d Neuronal regulation Normal Normal Stroke Stroke

MMP26 OPCML 10 OVOL2 LEP 10 CCL20 LHX2 CXCL17 SOX8 CNR2 8 VANGL2 CCR6 CNTNAP2 8 NT5E SEMA6C HES1 6 GRHL2 IL23A TWIST1 6 CD28 MTHFD1 ACP5 LEF1 4 NOG Normal1 Normal6 Normal2 Normal4 Normal1 Normal1 No Normal8 Normal3 Normal18 Normal2 Normal23 Normal11 Normal17 Normal1 No r Normal1 Normal13 Normal5 No r Normal2 Normal14 Stro Stroke18 Normal1 Stroke12 Stroke14 Stroke16 Stroke 5 Stroke21 Stroke19 Stro Stroke 8 Stroke 7 Stroke13 Stroke20 Stroke 2 Stroke22 Stroke 4 Stroke11 Stro Stroke10 Stroke23 CARD11 Stroke 1 Stroke 3 Stroke 9 4 Normal1 Stroke17 Stroke 8 Stroke10 Stroke13 Stroke14 Stroke19 Stroke 6 Stroke 5 Stroke12 Stroke21 Stroke 9 Stroke23 Stroke 1 Stroke11 Stroke 2 Stroke15 Stroke16 Stroke20 Normal2 Normal Normal15 Stroke18 Stroke 7 Stroke 4 Stroke22 Stroke 3 Normal1 Normal2 Normal2 Normal2 Normal5 Normal4 Normal8 Normal1 Normal6 Normal1 Normal1 Normal Normal1 Normal1 Normal16 Normal1 Normal22 Normal Normal1 rmal 7 mal9 mal2 ke 6 ke17 ke15 2 9 3 7 5 2 6 0 2 0 1 9 3 0 0 1 3 4 2 1 8 7 9 Fig. 3 PPI network and Venn diagrams: (1) PPI networks from a and b constructed using STRING database for DEGs (threshold > 0.4). (2) Venn diagrams of c of DEGs related to AF and < 3, 5, and 24 h after stroke, respectively. Co-expressed genes, including ZNF566, PDZK1IP1, ZFHX3, and PITX2, are identifed

Table 1 The Gene Ontology (GO) terms enrichment for the co-expressed genes of the AF-related stroke Gene/product GO class (direct) Evidence Evidence with Reference

PDZK1IP1 Integral component of membrane IEA UniProtKB-KW:KW-0812 GO_REF:0000037 Extracellular exosome IDA PMID:19056867 ZNF566 DNA binding IEA UniProtKB-KW:KW-0238 GO_REF:0000037 Nucleus IEA UniProtKB-SubCell:SL-0191 GO_REF:0000039 Transcription, DNA-templated IEA UniProtKB-KW:KW-0804 GO_REF:0000037 Metal ion binding IEA UniProtKB-KW:KW-0479 GO_REF:0000037 ZFHX3 Negative regulation of transcription from RNA polymerase II promoter IGI UniProtKB:P01104 PMID:10318867 RNA polymerase II proximal promoter sequence-specifc DNA binding IDA PMID:7507206 core promoter sequence-specifc DNA binding ISS UniProtKB:Q61329 GO_REF:0000024 Transcriptional repressor activity, RNA polymerase II proximal promoter IC GO:0000122/GO:0000978 GO_REF:0000036 sequence-specifc DNA binding Protein binding IPI UniProtKB:Q13761 PMID:20599712 Nucleus TAS PMID:1719379 PITX2 Transcription regulatory region sequence-specifc DNA binding IDA PMID:9685346 Transcriptional activator activity, RNA polymerase II proximal promoter IEA UniProtKB:Q9R0W1 GO_REF:0000107 sequence-specifc DNA binding RNA polymerase II transcription coactivator activity IDA PMID:9685346 Branching involved in blood vessel morphogenesis IEA UniProtKB:P97474 GO_REF:0000107 Vasculogenesis IEA UniProtKB:P97474 GO_REF:0000107 In utero embryonic development IEA UniProtKB:P97474 GO_REF:0000107 Neuron migration IEA UniProtKB:P97474 GO_REF:0000107 Extraocular skeletal muscle development IEA UniProtKB:P97474 GO_REF:0000107

ISS sequence similarity evidence used in manual assertion, IGI genetic interaction evidence used in manual assertion, IDA direct assay evidence used in manual assertion, TAS traceable author statement used in manual assertion, IEA evidence used in automatic assertion, IPI physical interaction evidence used in manual assertion Zou et al. J Transl Med (2019) 17:45 Page 7 of 12

abGene Ontology Trems of AF-DEGs from DAVID database Gene Ontology Trems of Stroke-DEGs from DAVID database

binding RNA polyme rase II transcription factor activity ion channel binding nucleic acid binding receptor activity transmembrane receptor protein tyrosine kinase adaptor ISG15 pecific protease activity activity transcriptionfactor activity, sequence pecific DNA binding RAGE receptor binding RNA polyme rase II transcription regulato ry region sequence G protein coupled pu rinergicnucleotide receptor activity specific binding MHC class II receptor activity integral component of plasma memb rane receptor activity Count negative regulation of myoblast differentiation negLog10_pv alue external side of plasma memb rane 5 endocardial cushion mo rphogenesis sarcomere positive regulation of epithelial cell proliferation 10 mast cell granule cartilage condensation 4 MHC class II protein complex 15 in utero embryonic development somatic stem cell population maintenance brush border memb rane 20 3 ER to Golgi transportvesicle membrane transcription from RNA polyme rase II promoter 25 apical plasma memb rane odontogenesis 2 positive regulation of MAP kinase activity embryonic hindlimb mo rphogenesis immune response neural tube closure positive regulation of inflammato ry response negLog10_pvalue embryonic skeletal system development inflammato ry response 5 ossification Count cellular defense response fibroblast signaling path way 4 positive regulation of NF appaB impo rt into nucleus 4 retinal rod cell differentiation leukocyte migration involved in inflammato ry response morphogenesis of a branching epithelium 8 adaptive immune response 3 negative regulation of photoreceptor cell differentiation 12 signaling path way 2 metanephric nephron tu bule formation positive regulation of inter feron gamma production cardiac neural crest cell migration involved in outflow 16 tract mo rphogenesis monovalent inorganic cation transport astrocytefate commitment B cell receptor signaling pathway negative regulation of osteoblast differentiation neural crest cell differentiation BMP signaling path way involved in heart development muscle contraction regulation of ion transmembrane transport response to lipopolysaccharide regulation of transcription from RNA polyme rase II promoter outflow tract mo rphogenesis skeletal muscle cell differentiation cation transport atrial cardiacmuscle tissue mo rphogenesis

GOTERM_BPGOTERM_CC GOTERM_MF GOTERM_BPGOTERM_CC GOTERM_MF Gene OntologyTerms Gene OntologyTerms

cdKEGG pathway enrichment from DAVID database Functional enrichment from REACTOME database

PI3K Akt signaling pathway FGFR3b ligand binding and acti vation B cell receptor signaling pathway Signalingby acti vated pointmutants of FGFR1 negLog10_pv alue Cytokine interaction Count Leishmaniasis 3 Conjugation of salicylate with glycine Inflammatory bowel disease (IBD) 1.8 6 Signalingby acti vated pointmutants of FGFR3 Hematopoietic cell lineage 9 Staphylococcus aureus in fection FGFR4 ligand binding and activation 1.6 12 Protein digestion and absorption 15 FGFR2c ligand binding and activation 1.4 Tuberculosis FGFR1c ligand binding and activation Intestinal immune networkfor IgA production negLog10_pvalue Antigen processing and presentation FGFR3c ligand binding and activation Count cGMP PKG signaling pathway FGFR3 ligand binding and activation 3 10 Asthma Defective SLCO1B1 causes hyperbilirubinemia, Rotor type 20 Influenza A (HBLRR) 2 30 Graft versus host disease FGFR1 ligand binding and activation 40 NF kappa B signaling pathway Chemokine receptors bind chemokines Rheumatoid arthritis Inte rleukin 18 signaling Phagosome

AF_PAT HWAY Storke_PAT HWAY REACTOME AF REACTOME Stro ke Fig. 4 GO terms and KEGG pathway enrichment: a, b AF-and stroke-related GO term enrichment for DEGs, respectively. c KEGG pathway of AF- and stroke-related DEGs. d Functional and pathway enrichment of AF-and stroke-related DEGs from REACTOME database. Dot sizes represent counts of enriched DEGs, and dot colors represent negative Log10-p values receptor protein tyrosine kinase adaptor activity (Fold KEGG pathway analysis data appear in Fig. 4c. Te Enrichment: 21.46; p-value: 0.008). With respect to results suggesting that the AF-DEGs were mainly stroke-DEGs, the biological processes terms of regula- enriched in pathways of cytokine–cytokine receptor tion of myoblast diferentiation (Fold Enrichment: 25.39; interaction (p-value: 1.02E−04), cGMP-PKG signal- p-value: 4.89E−04), endocardial cushion morphogenesis ing pathway (p-value: 0.025), antigen processing and (Fold Enrichment: 27.38; p-value: 0.005), positive regula- presentation (p-value: 0.022), and NF-kappa B signal- tion of epithelial cell proliferation (Fold Enrichment: 9.73; ing pathway (p-value: 0.037). However, KEGG terms p-value: 0.008), and fbroblast growth factor receptor included PI3 K-Akt signaling pathway (p-value: 0.017) signaling pathway (Fold Enrichment: 7.12; p-value: 0.018) and B cell receptor signaling pathway (p-value: 0.045) were signifcantly enriched. Similarly, the terms of RNA were enriched in stroke-DEGs. (As shown in Fig. 4c polymerase II transcription factor activity, sequence- and Additional fle 4: S4).GO terms enrichment using specifc DNA binding (Fold Enrichment: 5.15; p-value: the REACTOME database identifed additional asso- 0.006), ISG15-specifc protease activity (Fold Enrich- ciations and these appear in Fig. 4d. Te CTD database ment: 146.79; p-value: 0.013), and nucleic acid binding showed that Co-DEGs targeted several nervous system (Fold Enrichment: 2.09; p-value: 0.015) related molecular and cardiovascular diseases and these data appear in functions were primarily enriched (As shown in Fig. 4 Fig. 5 and Additional fle 5: S5. and Additional fle 3: S3). Zou et al. J Transl Med (2019) 17:45 Page 8 of 12

abThe interaction of ZFHX3 gene and nervous and cardiovascular diseases The interaction of ZNF566 gene and nervous and cardiovascular diseases

Vascular Diseases Nervous diseases Memory Disorders Cardiovascular diseases Neurotoxicity Syndromes Heart Defects, Congenital Learning Disorders Hypertension

Hypertension Psychomotor Agitation Cardiomegaly Atherosclerosis Headache Atrial Fibrillation * Nervous System Diseases 0204060 0510 15 20 Inference Score Inference Score

cdThe interaction of PDZK1IP1 gene and nervous and cardiovascular diseases The interaction of PITX2 gene and nervous and cardiovascular diseases

Neurodegenerative Diseases Neural Tube Defects Memory Disorders Memory Disorders Neurotoxicity Syndromes Learning Disorders

Pericardial Effusion Cardiovascular Diseases Hypertension Heart Diseases Cardiomegaly Hypertension Heart Defects, Congenital * 01020304050 0204060 Inference Score Inference Score Fig. 5 Relationship to nervous system and cardiovascular diseases related to co-expressed genes based on the CTD database. *Direct evidence of marker or mechanism in this disease

Identifcation of functional and pathway enrichment to investigate heart rate variability in LRRK2-associated among predicted miRNAs and Co‑DEGs PD [24]. Tere is evidence that LRRK2 may act as a bio- Prediction analysis using mirDIP, miRDB, TargetS- functional mediator to correlate heart rate variability can, and DIANA bioinformatic tools identifed the top and PD [24]. In a molecular mechanistic study, the neu- 5 selected miRNAs targeting each Co-DEG involved ral protective role for regulating mitochondrial complex in AF-related stroke and these data appear in Table 2. I function and oxidative stress in ischemia/reperfusion Tese data enable us to understand how predicted miR- was identifed [25, 26]. According gene–gene interac- NAs are related to AF-related stroke progress. tion analysis, Timasheva’s group illustrated that the loci of CXCR2 is signifcantly associated with stroke devel- Discussion opment in patients with hypertension [26]. In addition, Predicting AF is needed for stroke prevention but 30% CXCR2 antagonism attenuated neurological defcits and of patients have no signs of AF despite months of con- infarct volumes via decreased cerebral neutrophil infl- tinuous cardiac rhythm monitoring. Tus, cardiovas- tration and peripheral neutrophilia in a hyperlipidemic cular malignant events may be correlated with irregular ­ApoE−/− mice stroke model [27]. CALM1 is recognized and infrequent cardiovascular incidents as well as limi- as a major regulator of cardiac ion-current expression tations in electromechanical indices that should predict and calcium handling, and a key determinant of car- problems with atrial contractility [7, 11, 12]. Estimating diac electrical function [28]. Also, specifc risk alleles markers and associations between atrial dysfunction for CALM1 were identifed as being associated with and embolic stroke are thus of interest and may be novel increased risk of stroke in studies of coronary heart dis- therapeutic targets for primary care. Te infammatory ease [29]. Tus, there may be a relationship between car- and immune response, and ion channel and transporta- diovascular and nervous system disease and they may tion are signifcantly associated with AF recurrence and arise from loci mutations or gene variants. maintenance, as well as the stroke occurrence. Several Additionally, PITX2, of the pituitary homeobox (Pitx) hub-genes involved directly or indirectly that regulate the family, has a critical role in organ morphogenesis and nervous system were found among AF-DEGs. Visanji’s AF maintenance which is related to short stature home- group compared resting electrocardiograms of LRRK2- obox 2 (Shox2) [30]. Pitx2 is expressed in the LA and associated Parkinson’s disease (PD) patients, nonmani- the pulmonary vein, which is considered a substrate festing carriers, noncarriers, and idiopathic PD patients and trigger for AF maintenance respectively. However, Zou et al. J Transl Med (2019) 17:45 Page 9 of 12

Table 2 The Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways enrichment among predicted miRNAs and Co-DEGs Genes Predicted miRNAs Category P value

PDZK1IP1 hsa-miR-1296-5p KEGG pathway Hypertrophic cardiomyopathy (HCM) 0.047 hsa-miR-27a-3p NF-kappa B signaling pathway 0.022 hsa-miR-27b-3p Toll-like receptor signaling pathway 0.017 hsa-miR-6895-5p GO terms Negative regulation of energy homeostasis 4E 04 − hsa-miR-4725-3p Calcium ion-dependent exocytosis of 0.001 Neurotransmitter receptor activity 0.002 Regulation of leukocyte migration 0.003 Smooth muscle hyperplasia 0.011 Immunoglobulin mediated immune response 0.034 ZNF566 hsa-miR-216b-5p KEGG pathway Valine, leucine and isoleucine biosynthesis 0.320 hsa-miR-1277-5p Cardiac muscle contraction 0.411 hsa-miR-6783-5p GO terms Regulation of acute infammatory response 0.007 hsa-miR-369-3p Commissural neuron diferentiation in spinal cord 0.019 hsa-miR-6778-3p Cardiac vascular smooth muscle cell diferentiation 0.020 MHC protein binding 0.035 ATP-activated inward rectifer activity 0.035 Mitochondrial translation 0.041 PITX2 hsa-miR-377-3p KEGG pathway TGF-beta signaling pathway 3.00E 03 − hsa-miR-141-3p Phosphatidylinositol signaling system 0.124 hsa-miR-5692b GO terms Vascular smooth muscle cell diferentiation 0.002 hsa-miR-4789-5p Cell proliferation involved in outfow tract morphogenesis 0.002 hsa-miR-494-3p Cardiac neural crest cell migration involved in outfow tract 0.002 morphogenesis Pulmonary myocardium development 0.002 Atrial cardiac muscle tissue morphogenesis 0.002 Ventricular cardiac muscle cell development 0.003 Cardiac muscle cell diferentiation 0.006 Neuron migration 0.014 Neuron diferentiation 0.014 ZFHX3 hsa-miR-494-3p KEGG pathway Regulating pluripotency of stem cells 3E 07 − hsa-miR-758-3p GO terms Positive regulation of myoblast diferentiation 0.023 hsa-miR-27a-3p Negative regulation of myoblast diferentiation 0.023 hsa-miR-27b-3p Regulation of neuron diferentiation 0.023 hsa-miR-493-5p Muscle organ development 0.041

several experimental data indicate a trend that PITX2 association study using clinical samples from paroxys- gene expression is silenced during aging in LA sam- mal or persistent AF patients, ZFHX3 was signifcantly ples, suggesting genetic evidence for gene silencing associated with LA enlargement and persistent AF for increased AF susceptibility [30, 31]. Ten, miRNAs and subsequently with ablation outcomes [33]. Cor- function analysis and a genomic approach showed that respondingly, Choi’s group found a signifcant asso- miR-17-92 and miR-106b-25 were associated with Pitx2 ciation between top susceptibility loci ( expression regulation and are implicated in human 4q25 [PITX2], 16q22 [ZFHX3]) and AF recurrence after AF susceptibility [31]. To reveal relationships between ablation in a Korean population, despite no top single genetic variants and the risk of ischemic stroke, Malik’s nucleotide polymorphisms (SNPs) that predicted clini- group studied PITX2 and ZFHX3 genes and found a cal recurrence after catheter ablation [34]. A regulatory signifcant association with cardioembolic stroke (CE) role for PDZK1IP1 (MAP17) in reactive oxygen species in a meta-analysis [31, 32]. Similarly, in a genome-wide production has been confrmed and is considered as a Zou et al. J Transl Med (2019) 17:45 Page 10 of 12

marker for increased oxidative stress and may be a new chip may be better for accessing the risk of AF-related therapeutic target [35]. and recent research suggests stroke. Second, validation should be carried out both a potential role for ions channels regulation, linked to in vitro, in vivo and clinical trials. However, as of now the the ­Na+/H+ exchanger 3 and A-kinase anchor protein techniques of in vivo or in vitro models for AF and stroke 2/protein kinase A pathway [36]. However, ZNF566 was immature. And the larger, prospective clinical stud- plays a central role in heart regeneration and repair, and ies may be better to validate our results to some extent. endocardial and epicardial epithelial to mesenchymal transitions [37, 38]. Research suggests potential benefcial efects of miRNA Additional fles transformation therapy vectored by adenovirus, plasmid, and lentivirus for AF therapy [39]. We found that miR- Additional fle 1: S1. Diferentially expressed genes involved in AF 27a-3p, miR-27b-3p, and miR-494-3p were co-DEGs and samples. may be potential biomarkers of AF-related stroke. Inter- Additional fle 2: S2. Diferentially expressed genes involved in stroke. estingly, Vegter’s group compared heart failure-specifc Additional fle 3: S3. Gene Ontology (GO) terms enrichment analysis of circulating miRNAs in 114 heart failure patients with/ AF- and stroke-related diferentially expressed genes. without diferent manifestations of atherosclerotic dis- Additional fle 4: S4. Kyoto Encyclopedia of Genes and Genomes (KEGG) ease, and reported that miR-18a-5p, miR-27a-3p, miR- pathway enrichment analysis among the AF- and stroke-related diferen- 199a-3p, miR-223-3p and miR-652-3p abundance were tially expressed genes. associated with atherosclerosis and cardiovascular- Additional fle 5: S5. Relationship between diseases and co-expressed genes. related rehospitalizations [40]. Similarly, Marques and colleagues found that several miRNAs involved in let- 7b-5p, let-7c-5p, let-7e-5p, miR-122-5p, and miR-21-5p, Abbreviations AF: atrial fbrillation; DEGs: diferentially expressed genes; GEO: gene expres- and absorbed miR-16-5p, miR-17-5p, miR-27a-3p, and sion omnibus; PPI: protein–protein interaction; GO: Gene Ontology; ARIC: miR-27b-3p are target pathways related to heart failure atherosclerosis risk in communities; OXVASC: the Oxford vascular study; and considered to be potential biomarkers [41]. In con- ASSERT: atrial fbrillation reduction atrial pacing trial; co-DEGs: co-expressed diferentially expressed genes; AF-DEGs: AF-related diferentially expressed trast, expression of miR-27b-3p is signifcantly related genes; stroke-DEGs: stroke-related diferentially expressed genes; miRNA: to embryonic myogenesis and protein synthesis but microRNA; LA: left atrial; RA: right atrial; SR: sinus rhythm; RMA: robust multi- miR-494-3p expression is associated with cerebral blood array average; PM: perfect match; MM: mismatch probe; FDR: false discovery rate; FC: fold-changes; mirDIP: microRNA Data Integration Portal; KEGG: Kyoto supply and functional recovery in a rat stroke model Encyclopedia of Genes and Genomes; PD: Parkinson’s disease; Pitx: pituitary according to cerebral cortical miRNA profle changes [42, homeobox; Shox2: short stature homeobox 2; CE: cardioembolic stroke; SNPs: 43]. single nucleotide polymorphisms. Authors’ contributions Conclusion RZ, DZ, and LL: takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation, drafting Te hub-genes of LRRK2, CALM1, CXCR4, TLR4, the article. WS, ZS, BY, BL and QC: takes responsibility for statistical analyses, CTNNB1, CXCR2, KIT, and IL1B may be associated with and interpretation of data. SY and PH: takes responsibility for full text evalua- tion and guidance, fnal approval of the version to be submitted. All authors AF recurrence and maintenance and CD19, FGF9, SOX9, read and approved the fnal manuscript. GNGT1, and NOG may be associated with stroke. Addi- tionally, co-DEGs of ZNF566, PDZK1IP1, ZFHX3, and Author details 1 Department of Cardio‑Vascular Surgery, Sun Yat‑sen Memorial Hospital, Sun PITX2 link AF and stroke. Finally, the top 5 miRNAs for Yat-sen University, Guangzhou 510120, China. 2 Department of Gastroenterol- each co-DEGs may be potential biomarkers or therapeu- ogy, Fifth Afliated Hospital, Sun Yat-sen University, Zhuhai 519000, China. 3 tic targets for AF-stroke, especially miR-27a-3p, miR- Department of Endocrinology, Sun Yat‑sen Memorial Hospital, Sun Yat-sen University, Guangzhou 510120, China. 4 Department of Radiology, Sun Yat‑sen 27b-3p, and miR-494-3p. Tus, there is an association Memorial Hospital, Sun Yat-sen University, Guangzhou 510120, China. 5 The between AF and stroke, and expression of ZNF566, PDZ- Second Department of Orthopedics, Afliated Hospital of Beihua University, 6 K1IP1, ZFHX3, and PITX2 genes favor AF-related stroke. Jilin 132011, China. The Biobank of Sun Yat‑sen Memorial Hospital, Sun Yat-sen University, Guangzhou 510120, China. 7 Guangdong Province Key Laboratory of Brain Function and Disease, Zhongshan School of Medicine, Sun Limitation Yat-sen University, Guangzhou 510080, China. Several limitations still detected in our study. First, this Acknowledgements study is a microarray analysis that all the results based on Not applicable. gene expression value. However, owing to gene expres- Competing interests sion may be not directly equivalent to protein expression, The authors declare that they have no competing interests. the biomarkers of this study should consider as gene, not in protein. In application, assay of PCR and microarray Zou et al. J Transl Med (2019) 17:45 Page 11 of 12

Availability of data and materials embolic vascular events: a population-based study. Circulation. GSE79768 and GSE58294 datasets were downloaded from GEO (http://www. 2014;130(15):1236–44. ncbi.nlm.nih.gov/geo/)) [11] and expression profling arrays were generated 11. Brambatti M, Connolly SJ, Gold MR, Morillo CA, Capucci A, Muto C, Lau using GPL570 (HG‑U133_Plus_2) Afymetrix U133 Plus 2.0 CP, Van Gelder IC, Hohnloser SH, Carlson M, et al. Temporal relationship Array (Afymetrix, Santa Clara, CA). R packages of “afy”, “afyPLM”, and “limma” between subclinical atrial fbrillation and embolic events. Circulation. (http://www.bioco​nduct​or.org/packa​ges/relea​se/bioc/html/afy.html), 2014;129(21):2094–9. provided by a bioconductor project [12], were applied to assess GSE79768 12. Calenda BW, Fuster V, Halperin JL, Granger CB. Stroke risk assessment and GSE58294 RAW datasets. We applied online prediction tools utilizing in atrial fbrillation: risk factors and markers of atrial myopathy. Nat Rev microRNA Data Integration Portal (mirDIP) (http://ophid​.utoro​nto.ca/mirDI​P)) Cardiol. 2016;13(9):549–59. [13], miRDB (http://mirdb​.org)) [14], TargetScan (v7.1; http://www.targe​tscan​ 13. Barrett T, Wilhite SE, Ledoux P, Evangelista C, Kim IF, Tomashevsky M, Mar- .org/vert_71/) [15], and DIANA Tools (http://diana​.imis.athen​a-innov​ation​.gr/ shall KA, Phillippy KH, Sherman PM, Holko M, et al. NCBI GEO: archive for Diana​Tools​/) [16], to predict potential microRNA targeting. functional genomics data sets–update. Nucleic Acids Res. 2013;41(Data- base issue):D991–5. Consent for publication 14. Chen Z, McGee M, Liu Q, Scheuermann RH. A distribution free summa- Not applicable. rization method for Afymetrix GeneChip arrays. Bioinformatics (Oxford, England). 2007;23(3):321–7. Ethics approval and consent to participate 15. Pio G, Malerba D, D’Elia D, Ceci M. Integrating microRNA target predic- Not applicable. tions for the discovery of gene regulatory networks: a semi-supervised ensemble learning approach. BMC Bioinformatics. 2014;15(Suppl 1):S4. Funding 16. Wong N, Wang X. miRDB: an online resource for microRNA target predic- This work was supported by the National Natural Science Foundation of tion and functional annotations. Nucleic Acids Res. 2015;43(Database China (NSFC No. 81771165), the Natural Science Foundation Project in issue):D146–52. Guangdong province, China (Grant Nos. 2016A030313295; 2018A030313172) 17. Agarwal V, Bell GW. Nam JW. Bartel DP: Predicting efective microRNA and Guangzhou Science and Technology project of Major Special Research target sites in mammalian mRNAs. eLife; 2015. p. 4. Topics on International Collaborative Innovation (Grant Nos. 201704030032; 18. Paraskevopoulou MD, Georgakilas G, Kostoulas N, Vlachos IS, Vergoulis 201807010010). T, Reczko M, Filippidis C, Dalamagas T, Hatzigeorgiou AG. DIANA-microT web server v.50: service integration into miRNA functional analysis work- fows. Nucleic Acids Res. 2013;41(Web Server issue):W169–73. Publisher’s Note 19. Szklarczyk D, Franceschini A, Wyder S, Forslund K, Heller D, Huerta-Cepas Springer Nature remains neutral with regard to jurisdictional claims in pub- J, Simonovic M, Roth A, Santos A, Tsafou KP, et al. STRING v10: protein-pro- lished maps and institutional afliations. tein interaction networks, integrated over the tree of life. Nucleic Acids Res. 2015;43(Database issue):D447–52. Received: 20 October 2018 Accepted: 4 February 2019 20. da Huang W, Sherman BT, Lempicki RA. Systematic and integrative analy- sis of large gene lists using DAVID bioinformatics resources. Nat Protoc. 2009;4(1):44–57. 21. Croft D, O’Kelly G, Wu G, Haw R, Gillespie M, Matthews L, Caudy M, Garapati P, Gopinath G, Jassal B, et al. Reactome: a database of reactions, References pathways and biological processes. Nucleic Acids Res. 2011;39(Database 1. Rahman F, Kwan GF, Benjamin EJ. Global epidemiology of atrial fbrilla- issue):D691–7. tion. Nat Rev Cardiol. 2014;11(11):639–54. 22. Munoz-Torres M, Carbon S. Get go! retrieving go data using AmiGO, 2. Freeman JV, Wang Y, Akar J, Desai N, Krumholz H. National trends in QuickGO, API, Files, and Tools. Methods Mol Biol. 2017;1446:149–60. atrial fbrillation hospitalization, readmission, and mortality for medicare 23. Davis AP, Grondin CJ, Johnson RJ, Sciaky D, King BL, McMorran R, Wiegers benefciaries, 1999–2013. Circulation. 2017;135(13):1227–39. J, Wiegers TC, Mattingly CJ. The comparative toxicogenomics database: 3. Thijs V. Atrial fbrillation detection: fshing for an irregular heartbeat update 2017. Nucleic Acids Res. 2017;45(D1):D972–d978. before and after stroke. Stroke. 2017;48(10):2671–7. 24. Visanji NP, Bhudhikanok GS, Mestre TA, Ghate T, Udupa K, AlDakheel A, 4. Koton S, Schneider AL, Rosamond WD, Shahar E, Sang Y, Gottesman RF, Connolly BS, Gasca-Salas C, Kern DS, Jain J, et al. Heart rate variability in Coresh J. Stroke incidence and mortality trends in US communities, 1987 leucine-rich repeat kinase 2-associated Parkinson’s disease. Mov Disord. to 2011. JAMA. 2014;312(3):259–68. 2017;32(4):610–4. 5. Naghavi M, Abajobir AA, Abbafati C, Abbas KM, Abd-Allah F, Abera SF, 25. Liu J, Li T, Thomas JM, Pei Z, Jiang H, Engelender S, Ross CA, Smith WW. Aboyans V, Adetokunboh O, Afshin A, Agrawal A, Ahmadi A. Global, Synphilin-1 attenuates mutant LRRK2-induced neurodegeneration in regional, and national age-sex specifc mortality for 264 causes of death, Parkinson’s disease models. Hum Mol Genet. 2016;25(4):672–80. 1980-2016: a systematic analysis for the Global Burden of Disease Study 26. Timasheva YR, Nasibullin TR, Mustafna OE. The CXCR2 Gene Polymor- 2016. Lancet (London, England). 2017;390(10100):1151–210. phism Is Associated with Stroke in Patients with Essential Hypertension. 6. Krishnamurthi RV, Moran AE, Feigin VL, Barker-Collo S, Norrving B, Mensah Cerebrovasc Dis Extra. 2015;5(3):124–31. GA, Taylor S, Naghavi M, Forouzanfar MH, Nguyen G, et al. Stroke preva- 27. Herz J, Sabellek P, Lane TE, Gunzer M, Hermann DM, Doeppner TR. Role of lence, mortality and disability-adjusted life years in adults aged 20–64 neutrophils in exacerbation of brain injury after focal cerebral ischemia in years in 1990–2013: data from the global burden of disease 2013 Study. hyperlipidemic mice. Stroke. 2015;46(10):2916–25. Neuroepidemiology. 2015;45(3):190–202. 28. Gaborit N, Le Bouter S, Szuts V, Varro A, Escande D, Nattel S, Demolombe 7. Boehme AK, Esenwa C, Elkind MS. Stroke risk factors, genetics, and pre- S. Regional and tissue specifc transcript signatures of ion channel genes vention. Circ Res. 2017;120(3):472–95. in the non-diseased human heart. J Physiol. 2007;582(Pt 2):675–93. 8. Lip GY, Fauchier L, Freedman SB, Van Gelder I, Natale A, Gianni C, Nattel S, 29. Luke MM, O’Meara ES, Rowland CM, Shifman D, Bare LA, Arellano AR, Potpara T, Rienstra M, Tse HF, et al. Atrial fbrillation. Nat Rev Dis Primers. Longstreth WT Jr, Lumley T, Rice K, Tracy RP, et al. Gene variants associ- 2016;2:16016. ated with ischemic stroke: the cardiovascular health study. Stroke. 9. Steinberg BA, Hellkamp AS, Lokhnygina Y, Patel MR, Breithardt G, Hankey 2009;40(2):363–8. GJ, Becker RC, Singer DE, Halperin JL, Hacke W, et al. Higher risk of death 30. Wang J, Klysik E, Sood S, Johnson RL, Wehrens XH, Martin JF. Pitx2 and stroke in patients with persistent vs. paroxysmal atrial fbrillation: prevents susceptibility to atrial arrhythmias by inhibiting left-sided pace- results from the ROCKET-AF Trial. Eur Heart J. 2015;36(5):288–96. maker specifcation. Proc Natl Acad Sci U S A. 2010;107(21):9753–8. 10. Yiin GS, Howard DP, Paul NL, Li L, Luengo-Fernandez R, Bull LM, Welch 31. Wang J, Bai Y, Li N, Ye W, Zhang M, Greene SB, Tao Y, Chen Y, Wehrens XH, SJ, Gutnikov SA, Mehta Z, Rothwell PM. Age-specifc incidence, out- Martin JF. Pitx2-microRNA pathway that delimits sinoatrial node develop- come, cost, and projected future burden of atrial fbrillation-related ment and inhibits predisposition to atrial fbrillation. Proc Natl Acad Sci U S A. 2014;111(25):9181–6. Zou et al. J Transl Med (2019) 17:45 Page 12 of 12

32. Malik R, Traylor M, Pulit SL, Bevan S, Hopewell JC, Holliday EG, Zhao 38. von Gise A, Pu WT. Endocardial and epicardial epithelial to mes- W, Abrantes P, Amouyel P, Attia JR, et al. Low-frequency and common enchymal transitions in heart development and disease. Circ Res. genetic variation in ischemic stroke: the METASTROKE collaboration. 2012;110(12):1628–45. Neurology. 2016;86(13):1217–26. 39. Friedman RC, Farh KK, Burge CB, Bartel DP. Most mammalian mRNAs are 33. Husser D, Buttner P, Ueberham L, Dinov B, Sommer P, Arya A, Hindricks conserved targets of microRNAs. Genome Res. 2009;19(1):92–105. G, Bollmann A. Association of atrial fbrillation susceptibility genes, atrial 40. Vegter EL, Ovchinnikova ES, van Veldhuisen DJ, Jaarsma T, Berezikov E, fbrillation phenotypes and response to catheter ablation: a gene-based van der Meer P, Voors AA. Low circulating microRNA levels in heart failure analysis of GWAS data. J Transl Med. 2017;15(1):71. patients are associated with atherosclerotic disease and cardiovascular- 34. Choi EK, Park JH, Lee JY, Nam CM, Hwang MK, Uhm JS, Joung B, Ko YG, related rehospitalizations. Clin Res Cardiol. 2017;106(8):598–609. Lee MH, Lubitz SA, et al. Korean atrial fbrillation (AF) network: genetic 41. Marques FZ, Vizi D, Khammy O, Mariani JA, Kaye DM. The transcardiac variants for AF do not predict ablation success. J Am Heart Assoc. gradient of cardio-microRNAs in the failing heart. Eur J Heart Fail. 2015;4(8):e002046. 2016;18(8):1000–8. 35. Carnero A. MAP17 and the double-edged sword of ROS. Biochem Bio- 42. Lie S, Morrison JL, Williams-Wyss O, Suter CM, Humphreys DT, Ozanne SE, phys Acta. 2012;1826(1):44–52. Zhang S, MacLaughlin SM, Kleemann DO, Walker SK, et al. Impact of peri- 36. Coady MJ, El Tarazi A, Santer R, Bissonnette P, Sasseville LJ, Calado J, conceptional and preimplantation undernutrition on factors regulating Lussier Y, Dumayne C, Bichet DG, Lapointe JY. MAP17 Is a necessary myogenesis and protein synthesis in muscle of singleton and twin fetal activator of renal Na /glucose cotransporter SGLT2. J Am Soc Nephrol. sheep. Physiol Rep. 2015;3:8. 2017;28(1):85–93. + 43. Zheng HZ, Jiang W, Zhao XF, Du J, Liu PG, Chang LD, Li WB, Hu HT, Shi XM. 37. Xin M, Olson EN, Bassel-Duby R. Mending broken hearts: cardiac develop- Electroacupuncture induces acute changes in cerebral cortical miRNA ment as a basis for adult heart regeneration and repair. Nat Rev Mol Cell profle, improves cerebral blood fow and alleviates neurological defcits Biol. 2013;14(8):529–41. in a rat model of stroke. Neural Regen Res. 2016;11(12):1940–50.

Ready to submit your research ? Choose BMC and benefit from:

• fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year

At BMC, research is always in progress.

Learn more biomedcentral.com/submissions