DOI: 10.1161/CIRCGENETICS.114.000968

Genetic Variations in miRNA Binding Sites Affect miRNA-Mediated Regulation of Several Genes Associated with Cardiometabolic Phenotypes

Running title: Ghanbari et al.; Cardiometabolic-SNPs in miR target site

Mohsen Ghanbari, MD1,2; Oscar H. Franco, MD, PhD1; Hans de Looper, MSc3; Albert Hofman, MD, PhD1; Stefan Erkeland, PhD3; Abbas Dehghan, ehghan, MD, PhD D1

1

Department a me artme ment ntt ooff Ep E Epidemiology, id demiology, 3Department ooff He H Hematology, matology,, Ca Cancer anc ncer Institute, Erasmus Erasmu u

University i y Medical ity Medical C Center, en ente n er, R Rotterdam, ottte otte terd rdam rd am, th am tthee Ne N Netherlands; etheerllandds;2D Department epaart rtme m nt of me of Ge G Genetics, net etic iccs,, S School chooo of ch medicine, Mashhad University Sciences, m dicine me n ,M asshhhadd U niv ver e sity y ooff Medical Meddicaal Sc Me cieencces, Mashhad, Maashhha had, Iran Ira ran

Correspondence: Abbas Dehghan, MD, PhD Department of Epidemiology Erasmus University Medical Center P.O. Box 2040 3000 CA, Rotterdam the Netherlands Tel: +31-10-70 43997 Fax: +31-10-70 44657 E-mail: [email protected]

Journal Subject Code: [89] Genetics of cardiovascular disease

1 Downloaded from http://circgenetics.ahajournals.org/ at University of Colorado on March 30, 2015

DOI: 10.1161/CIRCGENETICS.114.000968

Abstract: Background - Genome-wide association studies (GWAS) enabled us to discover a large number of variants and genomic loci contributing to cardiovascular and metabolic disorders. However, since the vast majority of the identified variants are thought to merely be proxies for other functional variants, the causal mechanisms remain to be elucidated. We hypothesized that part of the functional variants involved in deregulating cardiometabolic genes are located in microRNA (miRNA) binding sites. Methods and Results - Using the largest GWAS available on glycemic indices, lipid traits, anthropometric measures, blood pressure, coronary artery diseases and type yp 2 diabetes,, we identified 11,067 variants that are associated with cardiometabolic phenotypes. 230 enot en otyp ot ypes yp es. Of tthese, es hesee, 23 he 3 variants are located within miRNA binding sites in the 3’UTR of 155 cardiometabolic-genes. cardiom omet om etab et abol ab olic ol ic--ge ic gene Thirty-seven fulfill functional e oout en ut ooff 230 2330 va variants were found to fulfi fill fi l our predefined cr ll criteria for being funct t in their genomic experimental nom nom mic loci. Ten en variants var aria iaant ntss were w ree ssubsequently we ubssequuentlly selected ub seele lect cted ct ted e for for or exp x erim xp menta tall validation ta valli va lida lida d ti tioon based b on GWAS results, co-expression regulatory miRNAs resu re sult l s, eQTL eQT QTL aanalyses QT nalysses aand nd co o-eexp xpreesssionn of of ttheir heirr hhost osst genes gennes an andd regu g laato gu ory m iR R in relevant tissues. Luciferase reporter assays allele-specific regulation genes t Lucifer ras asee repo port po rter rt er assay ay ys revealed reeve veal aled al ed an al lle l lee-sp sppeccif ific ic reg eg gul u ation of gene hosting the variants byy miRNAs. experiments showed miRN RNA RN As. These T ese co-transfection Th co-transffectiion expe p riiments sh howed d that rs174545 (FADS1:miR-181a-2), iR-1181 iR 81aa 2) 2), ),, rs1059611 rs10 rs 1059 10 5961 59 6111 (LPL:miR-136), 61 (LP PL:m :miR iR-1 iR -1136 36), ), rs13702 rs1 s137 3 022 ((LPL:miR-410), 37 LP PL:m :miR i -4410 iR 10), ) rs1046875 ), rs1104 0 68 6875 75 (FN3KRP:miR-34a), rs7956 (MKRN2:miR-154), rs3217992 (CDKN2B:miR-138-2-3p) and rs11735092 (HSD17B13:miR-375) decrease/or abrogate miRNA-dependent regulation of the genes. Conversely, two variants, rs6857 (PVRL2:miR-320e) and rs907091 (IKZF3:miR326), were shown to enhance the activity of miRNAs on their host genes. Conclusions - We provide evidence for a model in which polymorphisms in miRNA binding sites can both positively and negatively affect miRNA-mediated regulation of cardiometabolic genes.

Key words: cardiovascular disease, Genome Wide Association Study, microRNA, variation, cardiometabolic Phenotypes, GWAS, miRNA binding site, genetic variation, miRNA binding site SNPs

2 Downloaded from http://circgenetics.ahajournals.org/ at University of Colorado on March 30, 2015

DOI: 10.1161/CIRCGENETICS.114.000968

Cardiometabolic disorders including cardiovascular disease (CVD) and a number of metabolic abnormalities that increase CVD risk are major worldwide health problem.1 Over the past few decades, remarkable progress has been made in identifying the genes that are responsible for cardiometabolic disorders. The recent development of novel techniques, including genome-wide association studies (GWAS) has led to the identification of a large number of single-nucleotide polymorphisms (SNPs) and specific genomic loci that contribute to the induction of different cardiometabolic traits and diseases.2-4 A large proportion of the identified SNPs associated with cardiometabolic disorders are located in non-coding regions of the genome nome that are respo responsible poons n for fine-tuning gene expression regulation.2, 3, 5, 6 Many studies thus farr have focused focusedd onn G WAS WA S hits h ts and hi nd functional variants that are aree located in the 3’ 33’-untranslated ’-untranslated region n annotating GWA GWAS regions (3’UTR).6, 7 In Increasing ncreasing evidence evi videnc vi ncce su suggests uggessts tha that hat th the he ge genes enes 33’UTRs ’UT UTRss ccontain onttaiin rregulatory egulat a oryy elem at elements menn that t roles tant rol oles les in in gene gene expression. exp xpre r ssio ionn.77,, 8 Ge io Genetic Gene neti ne ticc va ti vari variations riiat atio ions io o s iin n these th hes esee re region regions onss may affect on affe af ffe f ctt ggene play important expression through g different mecha mechanisms. h nisms. This Thi his includes includ l des altering alteriing recognition recogn g itio i n sites for 88,, 9 microRNAss (miRNA) RNA-binding proteins, polyadenylation machinery. ((miRNA), miR iRNA NA)) RNA R NA bbinding indi di proteins teiin and ndd the thhe pol oll aden de lation latii machiner hi MiRNAs MiRN MiR Mi RN

play an important role in modulating the susceptibility to and onset of complex disorders, therefore, research has focused on identifying polymorphisms in miRNA binding sites.10, 11 MiRNAs are small non-coding RNAs, consisting of 20-24 nucleotides, that regulate numerous biological processes by post-transcriptional regulation of gene expression.12 MiRNAs can repress translation and, to a lesser extent, decrease the stability of mRNAs by interacting with the complementary sequences in the 3’UTR of target genes.12 The seed sequence, which has been defined as nucleotides 2-8 of a miRNA, is essential for target recognition and regulation.12 Genetic variation in either a miRNA's seed sequence or the complementary sequence in the 3’UTR of target gene could adversely affect target recognition, resulting in aberrant gene

3 Downloaded from http://circgenetics.ahajournals.org/ at University of Colorado on March 30, 2015

DOI: 10.1161/CIRCGENETICS.114.000968

expression.10, 13 Both computational and experimental studies have revealed that variants in miRNA binding sites can modulate the risk for disease induction by disrupting the original recognition site or creating an alternative binding site for miRNAs.14-17 MiRNAs have been shown to regulate the expression of genes that are involved in induction of cardiovascular disease and regulation of various metabolic processes, such as insulin production, glucose homeostasis, lipid metabolism and obesity.18, 19 However, thus far, only a few pathogenic miRNA-related variants have been identified that affect the regulation of genes associated with cardiometabolic phenotypes.10, 11, 13 In the present nt study, we hypoth hypothesized the th that genetic variants within miRNA binding sites could constitute part of the h functional functio ti nall SNPs SN involved inn deregulating deere regu gula gu l tiing la n cardiometabolic cardiometabolic genes. To o ttest est this hypothesi hypothesis, siss, we have investigat si investigated whether variants riants ri ian nts that wer were re previously preeviouusl s y id identified denttiffied d bbyy GW GWAS AS of of cardio-metabolic caarddio-m mettab bolicc phenotypes pheenoty ypes reside in miRNA i NA bbinding iRNA i ding in ng sites. sit i es es. In I addition, add ddit dd itiion, we it we li llinked nkked d several sev ver eral al ppublicly ubbliicl cly l available avaiila av labl blle da ddatasets tasets tase ts and an tools such as GWAS findings, findinggs,, eQTL eQT QT TL da ddata, ta,, miRN miRNA iRN NA andd ge ggene ne exp expression pressiion profiles p offil pr i es and combi combined them with eexperimental perimental rii ntall st studies t di dies tto id identif identify tif if ffunctional nctional tii all variants ariants iant that h can alter lt miRNA-mediated miRNA iRNA iR RNA mediat diat di regulation of their host cardiometabolic-genes.

Methods Retrieval of cardiometabolic-associated variants We used data from the recent GWAS meta-analyses on 17 cardiometabolic traits and diseases to identify genetic variants that are associated with cardiometabolic phenotypes. To increase our statistical power and robustness, we used the largest GWAS available on these phenotypes. The GWAS summary statistics of glycemic homeostasis indices (fasting glucose, serum glucose after 2hr, fasting insulin, fasting pro-insulin, HbA1c, HOMA-B, and HOMA-IR) from the MetaAnalyses of Glucose and Insulin-related traits Consortium (MAGIC),20-24 lipid phenotypes 4 Downloaded from http://circgenetics.ahajournals.org/ at University of Colorado on March 30, 2015

DOI: 10.1161/CIRCGENETICS.114.000968

(plasma concentrations of total cholesterol, low-density lipoprotein cholesterol, high-density lipoprotein cholesterol and triglyceride) from the Global Lipid Genetics Consortium (GLGC),2 anthropometric measures (body mass index and waist to hip ratio) from the Genetic Investigation of ANthropometric Traits (GIANT) consortium,25, 26 systolic and diastolic blood pressure from the Global BPgen consortium,27 coronary artery disease from the CARDIoGRAMplusC4D consortium,4 and type 2 diabetes from the DIAGRAM consortium3 were analyzed. A description of the GWAS meta-analyses on these cardiometabolic phenotypes, the consortia responsible for these analyses and the number of SNPs associated with each trait or disease isease is provided in T Table 1. In total o l 11,067 otal 111 0677 unique 11,0 uni niqque SNPs (in 629 genomic ni genomi micc loci) mi l ci) were foundd tto lo o be associated withh 17 cardiometabolic phenotypes genome-wide 0.8, limit distance 500 kb, and population panel CEU to retrieve their proxy SNPs in the HapMap project. To examine the functional potential of the SNPs, we utilized the HaploReg database (http://www.broad institute.org/ mammals/haploreg/) which predict the effect of a SNP on protein structure, gene regulation, and splicing using data from the ENCODE project.

Results Cardiometabolic-associated taboolic-associ ta ciat ci ated at d vvariants aria ar iant ia nts t that thaat aree llocated th ocat ated d iin nm miRNA iRN RNA A binding bind ndin in ng sites siite tess 1,0067 6 uunique n qu ni q e SN NPs iin n 62 29 ge geno noomi m c lo loci cii w e e id er den e tiifi fied ed d tto o be aassociated ssoc ss ocia oc iateed wi ia w itth 117 7 A total of 11,067 SNPs 629 genomic were identified with cardiometabolic b bolic phenotype phenotypes pess at a sstatistically pe tati ta tist ti stic st i allly ssignificant ic igni ig n fi ni f ca cant nt llevel ev vel e ((p-value p-va valu luee

Genetic Variations in MicroRNA-Binding Sites Affect MicroRNA-Mediated Regulation of Several Genes Associated With Cardio-metabolic Phenotypes.

Genome-wide association studies enabled us to discover a large number of variants and genomic loci contributing to cardiovascular and metabolic disord...
5MB Sizes 0 Downloads 7 Views