Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Jul 6.
Published in final edited form as: Circ Cardiovasc Genet. 2015 Nov 9;8(6):812–822. doi: 10.1161/CIRCGENETICS.115.001145

Calcium Signaling Pathway Genes RUNX2 and CACNA1C Are Associated With Calcific Aortic Valve Disease

Sandra Guauque-Olarte 1, David Messika-Zeitoun 1, Arnaud Droit 1, Maxime Lamontagne 1, Joël Tremblay-Marchand 1, Emilie Lavoie-Charland 1, Nathalie Gaudreault 1, Benoit J Arsenault 1, Marie-Pierre Dubé 1, Jean-Claude Tardif 1, Simon C Body 1, Jonathan G Seidman 1, Catherine Boileau 1, Patrick Mathieu 1, Philippe Pibarot 1, Yohan Bossé 1
PMCID: PMC4934886  NIHMSID: NIHMS797414  PMID: 26553695

Abstract

Background

Calcific aortic valve stenosis (AS) is a life-threatening disease with no medical therapy. The genetic architecture of AS remains elusive. This study combines genome-wide association studies, gene expression, and expression quantitative trait loci mapping in human valve tissues to identify susceptibility genes of AS.

Methods and Results

A meta-analysis was performed combining the results of 2 genome-wide association studies in 474 and 486 cases from Quebec City (Canada) and Paris (France), respectively. Corresponding controls consisted of 2988 and 1864 individuals with European ancestry from the database of genotypes and phenotypes. mRNA expression levels were evaluated in 9 calcified and 8 normal aortic valves by RNA sequencing. The results were integrated with valve expression quantitative trait loci data obtained from 22 AS patients. Twenty-five single-nucleotide polymorphisms had P<5×10−6 in the genome-wide association studies meta-analysis. The calcium signaling pathway was the top gene set enriched for genes mapped to moderately AS-associated single-nucleotide polymorphisms. Genes in this pathway were found differentially expressed in valves with and without AS. Two single-nucleotide polymorphisms located in RUNX2 (runt-related transcription factor 2), encoding an osteogenic transcription factor, demonstrated some association with AS (genome-wide association studies P=5.33×10−5). The mRNA expression levels of RUNX2 were upregulated in calcified valves and associated with eQTL-SNPs. CACNA1C encoding a subunit of a voltage-dependent calcium channel was upregulated in calcified valves. The eQTL-SNP with the most significant association with AS located in CACNA1C was associated with higher expression of the gene.

Conclusions

This integrative genomic study confirmed the role of RUNX2 as a potential driver of AS and identified a new AS susceptibility gene, CACNA1C, belonging to the calcium signaling pathway.

Keywords: aortic valve calcification, aortic valve stenosis, eQTL, gene expression, genome-wide association study, genomics, RNA-Seq

Aortic valve stenosis (AS) is the most frequent acquired heart valve disease in developed countries affecting 2% and 4% of people >65 and 85 years, respectively.1 The actual increase of life expectancy will raise the incidence of AS and amplify this public health problem.2 The risk of death, valve replacement, or progression to heart failure is increased by 80% over a period of 5 years in patients with AS.3 No pharmacological treatments are available, and conventional cardiovascular drugs, such as, statin have failed to prevent the development of AS.4,5 Open heart surgery (aortic valve replacement) and more recently transcatheter aortic valve implantation are the only treatments available for symptomatic and severe cases. AS is the principal indication for valve replacement surgery in North America and Europe.6 In Canada and the United States, the cost of an aortic valve replacement surgery is estimated at $74 602 CAN7 and $190 194 USD,8 respectively. New treatment strategies for this growing disease are urgently needed.

AS is a biologically active process involving pathways and cellular processes, such as inflammation, extracellular matrix remodeling, and calcification.6,9 A genetic component to AS is demonstrated,10,11 and candidate gene studies have identified polymorphisms associated with AS or AS-related phenotypes12 (Table I in the Data Supplement). Rare mutations in NOTCH1 have been associated with bicuspid aortic valve disease and severe valve calcification.13,14 A recent genome-wide association study (GWAS) identified the lipoprotein(a) (LPA) gene associated with aortic valve calcification.15

We performed a GWAS meta-analysis with AS patients from Quebec and France to identify new susceptibility genes for this disease. Gene expression of susceptibility genes was evaluated by RNA sequencing (RNA-Seq) in human aortic valves with and without AS. Finally, the functional meanings of the newly identified single-nucleotide polymorphisms (SNPs) showing some association with AS and that are located in differentially regulated genes were extended by carrying out the first expression quantitative trait loci (eQTL) mapping study in human valve tissues. This integrative genomic study confirmed the role of runt-related transcription factor 2 (RUNX2) as a potential driver of AS development and identified a new AS susceptibility gene, namely CACNA1C, belonging to the calcium signaling pathway.

Materials and Methods

Figure 1 illustrates the overall study design.

Figure 1.

Figure 1

Open in a new tab

Overview of the study design. The flow chart shows the steps of the genome-wide association studies (GWAS) meta-analysis, pathway analysis, gene expression study of aortic valves, and valve expression quantitative trait loci (eQTL) mapping study. DE indicates differentially expressed genes; FC, fold change; GSA-SNP, gene set analysis-single nucleotide polymorphisms; KORA, Kooperative Gesundheitsforschung in der Region Augsburg Study; and PCOS, Polycystic Ovary Syndrome Genetics-NUgene.

Populations

The Quebec cohort consisted of 494 French Canadian patients who underwent aortic valve replacement for severe tricuspid nonrheumatic AS at the Institut universitaire de cardiologie et de pneumologie de Québec (IUCPQ). The France cohort consisted of 505 patients with degenerative AS from the (Aortic Stenosis in Elderly: Determinant of Progression; clinicalTrial.gov number NCT 00338676) and (Genetic of Aortic Valve Stenosis-Clinical and Therapeutic Implications; http://www.clinicaltrial.gov; unique identifier NCT00647088) studies. Inclusion criteria were at least mild AS defined as peak aortic jet velocity (Vpeak) > 2.0 m/s. Some patients from France had bicuspid valve based on echocardiographic examination (n=35+38 possible bicuspid). Exclusion criteria for individual cohorts are provided in Methods in the Data Supplement. A summary of the clinical characteristics of individuals from Quebec and Paris are shown in Table II in the Data Supplement. Written informed consent was obtained from all study participants after institutional review board approval.

Control cohorts were obtained from the database of genotypes and phenotypes.16 Proper control groups were selected based on genetic ancestry, genotyping data available, and sample size (Data Supplement). For cases from Quebec, the Polycystic Ovary Syndrome Genetics-NUgene (PCOS; phs000368) cohort from United States was selected. This cohort is a control group of individuals recruited at the Northwestern University Medical Center for performing a GWAS on PCOS susceptibility. For cases from France, 1 European cohort was selected, namely the Genetic Epidemiology of Refractive Error in the Kooperative Gesundheitsforschung in der Region Augsburg Study (KORA; phs000303). KORA is a population-based study of adults randomly selected from 430 000 individuals in Augsburg and 16 surrounding counties in Germany. Table III in the Data Supplement presents the characteristics of the database of genotypes and phenotypes cohorts.

Whole-Genome Genotyping, Quality Controls, and Imputation

DNA was extracted from whole blood or buffy coat (Data Supplement). Whole-genome genotyping for the Quebec and France cohorts was performed at the McGill University and Genome Quebec Innovation Center (http://gqinnovationcenter.com) using the Illumina HumanOmniExpress-12v1.0 BeadChip testing 733 202 SNPs. Standard genotyping quality controls were performed independently in the Quebec, France, and database of genotypes and phenotypes cohorts. Details are provided and summarized in Tables IV and V in the Data Supplement. For the Quebec cohort, 639 052 SNPs and 474 patients remained after quality controls. For the France cohort, 654 363 SNPs and 486 samples remained. The PCOS cohort was also genotyped using the Illumina HumanOmniExpress-12v1.0 platform. The KORA cohort was genotyped using the Illumina HumanOmni2.5 platform testing 2 182 680 SNPs. For the PCOS cohort, 645 438 SNPs and 2988 individuals passed quality controls. For the KORA cohort, 1 492 728 SNPs and 1864 individuals passed quality controls. To increase whole-genome coverage, additional SNPs were inferred using the segmented haplotype estimation and imputation tool17/IMPUTE218 pipeline. Only SNPs that passed a certainty score cut off of 0.9 were retained (Data Supplement).

Association Tests and Meta-Analysis

To account for differences in clinical characteristics between AS cases from Quebec and France, genetic association tests were performed independently and then combined into a meta-analysis. Association tests were performed with additive logistic regression models. Seven significant principal components derived from EIGENSTRAT19 were used as covariates to adjust for potential population stratification. After adjustment, the genomic inflation factors for the Quebec-PCOS and the France-KORA comparisons were both 1.0. The meta-analysis was performed using the inverse variance method with fixed- and random-effect models. The meta-analysis included 6 635 034 SNPs after the exclusion of polymorphisms with alternate risk alleles (n=36 590). The significant P value cutoff was set to 5×10−8. Association tests, meta-analysis, and linkage disequilibrium calculation were performed with PLINK.20 Regional plots were created with LocusZoom.21 SNPs with GWAS P<1×10−4 were mapped 1000 kb up- or downstream of genes on the hg19 assembly (UCSC genome browser, www.genome.ucsc.edu) using PLINK to generate a list of possible susceptibility genes (Table VI in the Data Supplement) for further analysis.

Biological, Functional, and Disease Analysis

We performed pathway analysis enrichment of the 6 635 034 SNPs included in the meta-analysis using the default procedure in the gene set analysis-single nucleotide polymorphisms (GSA-SNP) software.22 Instead of relying on individual marker analysis, GSA-SNP tests predefined gene sets or pathways for association with disease and, thus, the contribution of multiple genes. GSA-SNP allowed detecting subtle effects of multiple SNPs in the same gene set that might be missed when assessed individually. The software assigned the most significant SNPs located 10 kb up- and downstream to each gene (ie, padding size of 10 000 bp and k=1) in Kyoto Encyclopedia of Genes and Genomes gene sets. The genes were used to calculate pathway enrichment. The expression levels of genes in significant pathways were verified using the RNA-Seq data. Ingenuity Pathway Analysis system was used to identify the connections among these genes and their associated biological functions. The effects of differentially expressed genes in pathways were predicted with the Ingenuity Pathway Analysis system Molecule Activity Predictor tool.

Ingenuity Pathway Analysis system was used to integrate susceptibility genes in biological processes known to be implicated in AS development and progression.9 A SNP was considered within a gene if it was located in the exon–intron structure or 10 kb up- or downstream of the gene. Genes previously associated with AS were extracted from Phenopedia23 and curated manually (Table I in the Data Supplement).

Gene Expression Analysis in Human Aortic Valves

Aortic valves were collected from patients undergoing cardiac surgery at the IUCPQ. An informed consent was signed before surgery. Gene expression and eQTL mapping in aortic valves were approved by the institutional review board. Gene expression levels were evaluated in human aortic valve tissues (n=20; Table II in the Data Supplement). Stenotic aortic valves (n=10) were tricuspid and have degree 3 of fibro-calcific remodeling.24 To further decrease heterogeneity of expression, only valves taken from white male nonsmoking, nondiabetic subjects and without renal insufficiency (creatinine levels >150 μmol/L) were selected. Exclusion criteria are provided in Methods in the Data Supplement. Normal noncalcified tricuspid valves (n=10) were obtained from patients undergoing heart transplantation. Control valves had no evidence of disease at postexplant examination. Stenotic and control groups were matched for age (±10 years).

RNA was extracted from valve leaflets, and gene expression evaluated by the Illumina HiSeq 2000 platform (Data Supplement). TopHat25 was used to align reads, which were then assembled by Cufflinks.25 Gene expression between the 2 groups of valves were compared with Cuffdiff, DESeq,26 and edgeR.27 Genes with Benjamini–Hochberg adjusted P<0.05 using the 3 softwares were considered as differentially expressed (Data Supplement).

Expression Quantitative Trait Loci

Genome-wide genotyping and gene expression profiling were performed on an independent set of 24 human aortic valves. Tricuspid aortic valves were explanted from 24 white male patients under 67 years of age who underwent aortic valve replacement for severe AS at the IUCPQ from February 2005 to October 2013. Only tricuspid nonrheumatic aortic valve with a Warren score of 324 were selected. Patients who were undergoing concomitant coronary artery bypass grafting were included. Exclusion criteria are provided in Methods in the Data Supplement. The clinical characteristics of these patients are shown in Table VII in the Data Supplement.

DNA was extracted and genotyped using the Illumina Infinium HumanOmni2.5–8 BeadChip testing 2 379 855 SNPs throughout the human genome. Twenty-four patients and 1 474 616 SNPs passed all quality controls and were available for analysis (Data Supplement). Total RNA was extracted, and gene expression was measured on the Illumina HumanHT-12v4 Expression BeadChip. After quality controls, 22 samples were available for analysis (Data Supplement). mRNA expression data were log2 transformed and quantile normalized before analysis using Rv3.0.0 statistical software and the Bioconductor28 package lumi.29 Because of sample size and patient selection, no adjustment was applied to the mRNA expression data sets. mRNA transcripts were selected for valve eQTL analysis if they were within 1 Mb of the top GWAS meta-analysis SNPs. Valve eQTLs were also evaluated in genes differentially expressed between calcified and normal aortic valves in the RNA-Seq experiment and with SNPs with GWAS P<1×10−4. Association testings between genotypes and mRNA expression levels were performed in PLINKv1.07 using the –assoc command. A liberal eQTL P<0.05 was considered significant for integration with RNA-Seq and GWAS results.

Results

GWAS Results

The meta-analysis of the 2 independent GWAS is summarized in Figure 2. No SNP reached genome-wide significance. However, 25 SNPs had P<5×10−6 under the fixed-effects model (Table VIII in the Data Supplement). Using a more liberal significant threshold (P<1×10−4), 646 SNPs were associated with AS and mapped to or near 130 unique genes. Table IX in the Data Supplement shows the genotyping counts for cases and controls in the 2 GWAS. Among those SNPs, rs9371549 was located 3.8 kb downstream of ESR1 (estrogen receptor 1), and rs4708867 mapped 109 kb downstream of LPA (Figure I in the Data Supplement). ESR1 and LPA were previously associated with AS15,30 (Table I in the Data Supplement).

Figure 2.

Figure 2

Open in a new tab

Manhattan plot showing the results of the genome-wide association studies (GWAS) meta-analysis. The y axis represents P in −log10 scale combining the results of the 2 independent GWAS. The horizontal red line indicates the genome-wide significance cutoff (P=5×10−8). The horizontal blue line indicates the P cutoff of 5×10−6.

Considering the modest evidence of association from single marker analysis in the GWAS, we performed gene-set association analysis. GSA-SNP revealed 25 pathways significantly enriched (Benjamini–Hochberg corrected P<0.05) among genes mapped to SNPs with some evidence of association in the meta-analysis. Figure 3 shows the distributions of P values for these pathways. The most significant Kyoto Encyclopedia of Genes and Genomes gene set was the calcium signaling pathway (hsa04020, corrected P=7.27×10−5). The median P for this gene set was 0.031, suggesting that GSA-SNP identified moderate but coordinated association for this group of genes. A schematic representation of the calcium signaling pathway and members of this pathway that are drug targets is illustrated in Figure II in the Data Supplement.

Figure 3.

Figure 3

Open in a new tab

Boxplots of genome-wide association studies (GWAS) P values for significant gene sets. Corrected P values for each pathway are shown in parentheses. The vertical dashed line indicates P=0.01. Each point represents a gene. Please see Data Supplement for more details about gene set analysis single-nucleotide polymorphisms. ECM indicates extracellular matrix; and GnRH, gonadotropin-releasing hormone.

Integration With RNA-Seq

We then extended the functional assessment of SNPs with some evidence of association with AS in this study to gene expression in human aortic valves. Combining the results of 3 RNA-Seq analysis tools, 518 genes were found consistently differentially expressed between calcified and normal aortic valves (Figure III in the Data Supplement). A total of 332 and 186 genes were up- and downregulated in calcified valves, respectively. All genes had absolute fold changes of at least 1.52 and higher. Among the differentially expressed genes, 46 were located 1 Mb up- or downstream of SNPs with P<1×10−4 in the GWAS meta-analysis (Table X in the Data Supplement). Twenty-six of these 46 genes (56.5%) were linked to genes in molecular pathways known to be altered in the development and progression of AS, such as inflammation, lipid deposition and oxidation, angiogenesis, calcification, and extracellular matrix remodeling9 (Figure 4; Table XI in the Data Supplement). Interestingly, 11 genes in the calcium signaling pathway identified above were differentially expressed in calcified compared with non-calcified aortic valves in the RNA-Seq experiment (Figure 5A).

Figure 4.

Figure 4

Open in a new tab

Integration of the genome-wide association studies (GWAS) and RNA-Seq results with known aortic valve stenosis (AS)–related biological processes. Integration of genes differentially expressed between calcified and normal valves and with/near single-nucleotide polymorphisms with P<1×10−4 in biological processes known to be implicated in AS development and progression. Cyan indicates down-regulated gene; Red, upregulated genes. Some genes can belong to >1 biological process. This figure was modified from Bossé et al.9 Name of all genes are provided in Table XI in the Data Supplement.

Figure 5.

Figure 5

Open in a new tab

Alterations of the calcium signaling pathway in aortic valve stenosis (AS). A, Heat map of genes differentially expressed in the RNA-Seq experiment that belong to the calcium signaling pathway. Genes in yellow are upregulated and genes in blue are downregulated in calcified valves (TAVc) compared to normal valves (TAVn). The vertical white line separates the 2 groups. Each row represents a different gene indicated at the right. The calcium signaling pathway genes are ADRB1 (adrenoceptor beta 1), AGTR1 (angiotensin II receptor, type 1), BDKRB2 (bradykinin receptor B2), CACNA1C (calcium channel, voltage-dependent, L type, alpha 1C subunit), CACNA1H (calcium channel, voltage-dependent, T type, alpha 1H subunit), EDNRA (endothelin receptor type A), ERBB3 (v-erb-b2 avian erythroblastic leukemia viral oncogene homolog 3), HRH1 (histamine receptor H1), HTR4 (5-hydroxytryptamine [serotonin] receptor 4, G protein-coupled), P2RX6 (purinergic receptor P2X, ligand-gated ion channel, 6), and TBXA2R (thromboxane A2 receptor). B, Ingenuity Pathway Analysis system (IPA) of the calcium pathway showing only genes that are altered (single-nucleotide polymorphisms with P<0.01 or genes differentially expressed) or regulated as predicted by the molecule activity predictor (MAP). Symbol indicates gene function (see legend). Genes in yellow are upregulated and genes in blue are downregulated in TAVc. Predicted activation and inhibition of genes because of the differential expression in TAVc are shown in orange and violet, respectively. TAVc indicates calcified tricuspid aortic valve; and TAVn, normal tricuspid valve.

Integration With Valve eQTL Mapping

Of the 646 SNPs with GWAS P<1×10−4, 363 were genotyped or imputed in the valve eQTL data set. mRNA transcripts located within 1Mb of these SNPs were selected for eQTL analysis. Association between genotypes and gene expression levels in valve tissues had P<0.05 for 2783 SNP-probe pairs, including 230 unique genes and 223 unique SNPs. These eQTL-SNPs and eQTL-regulated genes were overlaid to genes differentially expressed in the RNA-Seq experiment and SNPs showing some evidence of association with AS in the GWAS. Data integration revealed RUNX2 as a potential driver of AS development. Two SNPs located in intron 1 of this gene, rs114193529 and rs144071310 (r2=1), were associated with AS (GWAS P=5.33×10−5, odds ratio [OR] =3.49, minor allele frequency [MAF] cases=0.04, MAF controls=0.03; Figure 6A and 6B). This effect was mainly driven by the Quebec samples (P=1.59×10−5, OR=6.81, MAF cases=0.039, MAF controls=0.035), but the direction of effect was consistent in samples from France (P=0.15, OR=1.86, MAF cases=0.042, MAF controls=0.033; Figure 6A and 6B). All coding exons, 1000 bp of the promoter and the 3′ and 5′ UTR regions of RUNX2, were sequenced in the Quebec samples that are homozygotes (n=2) and heterozygotes (n=33) for the rs114193529 risk allele. Few SNPs were identified, and none were in linkage disequilibrium with the intronic SNPs moderately associated with AS (Figure IV in the Data Supplement). RUNX2 was differentially expressed between calcified and normal aortic valves (fold change=2.68, P adjusted=1.47×10—; Figure 6C).

Figure 6.

Figure 6

Open in a new tab

RUNX2 is a susceptibility gene of aortic valve stenosis (AS) upregulated in calcified aortic valve and associated with expression quantitative trait loci (eQTL)-single-nucleotide polymorphisms (SNPs). A, Regional plots showing SNP rs114193529 located in RUNX2. The y axis shows the P values in −log10 scale for SNPs 1 Mb up- and downstream of the sentinel SNP (purple dot). The extent of linkage disequilibrium (LD; r2 values) for all SNPs with the sentinel SNP is indicated by colors. The location of genes is shown at the bottom. RUNX2 is framed in red. SNPs are plotted based on their chromosomal position on build hg19. B, Forest plot of overall effect size for rs114193529 in the 2 independent genome-wide association studies (GWAS) and meta-analysis. The black filled squares represent the odds ratio (OR) for each cohort and are proportional to the study size. The horizontal lines represent the 95% confidence intervals of the OR. The dashed grey and the dotted red vertical lines represent an OR of 1.0 and the OR of the meta-analysis, respectively. C, Gene expression levels of RUNX2 in human aortic valves with and without AS measured by RNA-Seq and expressed as mean FPKM (fragments per kilobase of exon per million fragments mapped). Cuffdiff adjusted P value is indicated. The number of valves assessed per group is indicated in parenthesis. Error bars represent the standard error. D, E, and F, Boxplots of mRNA expression levels for RUNX2 in human aortic valves according to genotyping groups for SNPs rs1200428, rs12209785, and rs35565233, respectively. The left y axis presents the z-score of log2-tranformed mRNA expression levels. The grey bar on the right y axis presents the percent variance in expression levels explained by the genotype. The x axis represents the 3 genotyping groups for each SNP, with the number of subjects in parenthesis. TAVc indicates calcified tricuspid aortic valve; and TAVn, normal tricuspid valve.

The 2 SNPs showing some evidence of association with AS in RUNX2 were not genotyped or imputed in the eQTL project. The genotyping information was obtained by sequencing the intronic region containing the 2 SNPs (see Data Supplement), but none of the 22 individuals were carriers of the risk variant. However, 3 other SNPs, rs1200428 (eQTL P=5.54×10−5, GWAS P=0.49), rs12209785 (eQTL P=4.35×10−3, GWAS P=0.41), and rs35565233 (eQTL P=8.83×10−3, GWAS P=0.08) were significantly associated with the mRNA expression of RUNX2 (Figure 6D–6F). These 3 SNPs were independent of the SNPs showing evidence of association with AS (r2<0.015). The minor alleles for rs1200428, rs12209785, and rs35565233 were associated with lower mRNA expression levels of RUNX2 in valve tissues. Although these eQTL-SNPs were not significantly associated with AS, rs35565233 had an OR below 1 in both GWAS and OR of 0.68 (95% confidence interval=0.44–1.05) in the meta-analysis (Figure V in the Data Supplement). In addition, the protective allele T for rs35565233 was associated with lower mRNA expression levels of RUNX2 (Figure V in the Data Supplement), suggesting that the SNP decreases susceptibility to AS through downregulation of RUNX2 in valve tissue.

Four additional eQTL-regulated genes overlap with the differentially expressed genes in the RNA-Seq analysis, namely HSD17B13 (hydroxysteroid [17-beta] dehydrogenase 13), PLAUR (plasminogen activator, urokinase receptor), SLC16A9 (solute carrier family 16, member 9), and SYT12 (synaptotagmin XII). GWAS, RNA-Seq, and eQTL results for these 4 genes are illustrated in Figure VI in the Data Supplement. The observed direction of effect for the valve eQTL-SNPs associated with HSD17B13 expression was concordant with the expected direction of effect based on GWAS and RNA-Seq results.

Valve eQTL data were further studied for genes differentially expressed in the calcium signaling pathway. A total of 460 significant eQTL-SNPs in 11 of these differentially expressed genes (ADRB1, AGTR1, BDKRB2, CACNA1C, CACNA1H, EDNRA, ERBB3, HRH1, HTR4, P2RX6, and TBXA2R) in this pathway were identified. From these eQTL-SNPs, a large majority (n=268) were associated with the expression levels of a single gene, CACNA1C. Rs2239118 located in intron 10 of CACNA1C is the eQTL-SNP with the strongest genetic association signal (GWAS P=4.09×10−4) in the meta-analysis. The risk allele G of rs2239118 was associated with increased mRNA levels of CACNA1C (eQTL P=0.030), and CACNA1C is increased 1.8× in calcified compared with normal aortic valves (P=1.47×10−3; Figure 7). The observed direction of effect of a second eQTL-SNP, rs1034936 (GWAS P=9.38×10−4), was also concordant with the expected direction of effect based on GWAS and RNA-Seq results (Figure VII in the Data Supplement). These results suggest that risk alleles for rs2239118 and rs1034936 mediate their effects by increasing the mRNA expression levels of CACNA1C in valve tissue. Sixteen additional significant eQTL-SNPs with GWAS P<0.01 were also associated with the expression levels of CACNA1C.

Figure 7.

Figure 7

Open in a new tab

CACNA1C a new susceptibility gene of aortic valve stenosis (AS). A, Regional plot showing single-nucleotide polymorphism (SNP) rs2239118 located in CACNA1C. The y axis shows the P values in −log10 scale for SNPs 1 Mb up- and downstream of the sentinel SNP (purple dot). The extent of linkage disequilibrium (LD; r2 values) for all SNPs with the sentinel SNP is indicated by colors. The location of genes is shown at the bottom. CACNA1C is framed in red. SNPs are plotted based on their chromosomal position on build hg19. B, Forest plot of overall effect size for rs2239118 in the 2 independent genome-wide association studies (GWAS) and meta-analysis. The black filled squares represent the odds ratio (OR) for each cohort and are proportional to the study size. The horizontal lines represent the 95% confidence intervals of the OR. The dashed grey and the dotted red vertical lines represent an OR of 1.0 and the OR of the meta-analysis, respectively. C, Boxplot of gene expression levels in human aortic valves for CACNA1C according to genotyping groups of rs2239118. The left y axis presents the z-score of log2-tranformed mRNA expression levels. The grey bar on the right y axis presents the percent variance in gene expression explained by the genotype. Number of patients per genotype is presented in parenthesis. D, Gene expression levels of CACNA1C in human aortic valves measured by RNA-Seq and expressed as mean FPKM (fragments per kilobase of exon per million fragments mapped). Cuffdiff adjusted P value is indicated. The number of valves assessed per group is indicated in parenthesis. Error bars represent the standard error. TAVc indicates calcified tricuspid aortic valve; and TAVn, normal tricuspid aortic valve.

Discussion

We performed a meta-analysis of 2 independent GWAS in white individuals with diagnosis of AS from Quebec and France. By traditional single marker analysis, no SNP reached stringent GWAS significance; however, 25 SNPs located in or near biologically relevant genes were moderately significant (P<5×10−6). Gene-set association analysis was performed to identify coordinated association patterns for genes involved in the same biological pathway. Twenty-five pathways were significantly enriched for genes mapping to moderate disease-associated SNPs. This is consistent with the small to moderate effect sizes observed for individual SNPs in GWAS that collectively, acting on a common pathway, may reveal the underlying biological mechanisms underpinning AS. The most significant gene set was the calcium signaling pathway. Eleven genes in this pathway were differentially expressed in calcified compared with normal aortic valves. Variants showing evidence of association in RUNX2 and CACNA1C were found to regulate mRNA expression levels of their respective gene in valve tissues. This is concordant with the overexpression of RUNX2 and CACNA1C observed in calcified valves and suggest that newly discovered AS-associated variants increase the risk of AS by increasing the mRNA expression levels of RUNX2 and CACNA1C in valve tissue.

In a typical GWAS, each SNP is tested individually for association with the disease or trait of interest. A wealth of empirical data show that this approach is successful to identify some robust genetic variants associated with complex diseases, but the effect sizes of risk variants is small to modest. Many polymorphisms are found in the grey zone of GWA studies, an area where the statistical signals are below the stringent threshold used to control for multiple SNP disease tests, but are likely to contain true positive associations31 with only moderate-effect sizes. Many strategies are being pursued to alleviate the large burden of multiple testing encountered in GWA studies to find true-positive signals in the grey zone.32 In this study, we used gene set association analysis22 and found coordinated association patterns for genes involved in the calcium signaling pathway. Most SNPs located near or within genes in this gene set were not individually associated with AS in our meta-analysis. However, considered together, members of this gene set showed an enrichment of small P values beyond what is expected by chance (41 out of 178 SNPs with P value <0.01; Figure 3). The functional meaning of this discovery was extended to gene expression in aortic valve specimens. Eleven gene members of the calcium signaling pathway were differentially expressed between patients with and without AS. The differential expression of these genes could lead to AS by altering normal homeostasis of calcium and apoptosis or enhancing fibrosis (Figure 5B). Valve eQTL-SNPs in differentially expressed genes were also assessed, leading to the identification of CACNA1C as a new AS susceptibility gene.

CACNA1C was overexpressed in calcified valves compared with normal valve. In addition, its mRNA expression levels were modulated by 2 moderately AS-associated SNPs located in intron 10 of the gene. CACNA1C encodes the α-1C subunit of the voltage-dependent calcium channel.33 Polymorphisms in this gene have been associated with systolic and diastolic blood pressure and hypertension.34 Angiotensin II was shown to upregulate the expression and activity of CACNA1C in cultured arteries.35 Further studies will be needed to elucidate the role of CACNA1C in AS, but a crosstalk with the renin–angiotensin is possible.36,37 Several commercial calcium channel inhibitors/blockers are available for treating hypertension, arrhythmias, and chronic angina pectoris. Whether calcium channel blockers can be used to treat patients with AS is currently unknown. In short terms, functional studies in valve interstitial cells38 and mouse models of AS39 are needed to investigate CACNA1C as a potential target to treat AS.

Twenty-six candidate genes differentially expressed in this study, including RUNX2, were overlaid on known biological processes involved in AS (Figure 4). RUNX2 is an osteogenic transcription factor repressed by the action of Notch1 and the Hairy family of repressors.40 RUNX2 regulates the expression of osteopontin and is upregulated in mouse and rabbit models of AS.14 Circulating and valve tissue levels of RUNX2 significantly correlated with valve stenosis severity.41 In this study, we demonstrated that 2 SNPs in perfect linkage disequilibrium, rs114193529 and rs144071310, located in intron 1 of RUNX2 were moderately associated with AS. Whether one of these variants is functional or in linkage disequilibrium with another rare variant in RUNX2 is unclear from this study. In addition, the mRNA expression of RUNX2 was upregulated in calcified valves and associated with valve eQTL-SNPs. Together, these results further support the role of RUNX2 as a potential driver of AS development.

A recent GWAS reported the association of rs10455872 with aortic valve calcification.15 A subsequent study demonstrated that this polymorphism is associated with circulating lipoprotein(a) levels and the risk of AS.42 This SNP, located in the LPA gene, was not present in our genotyping array and was excluded after imputation based on call rate <90%. In the current GWAS meta-analysis, the SNP rs4708867 located in exon 5 of SLC22A3 and 109 kb of the LPA gene was found modestly associated with AS (Figure I in the Data Supplement). This modest genetic association signal is independent from the previous rs10455872 finding. A large-scale meta-analysis of the LPA locus is needed to understand its contribution in aortic valve calcification and AS.

This study has limitations. This study represents the largest collection of well-phenotyped patients for AS, but the sample size is still relatively small for a GWAS. Heterogeneity, because of differences in clinical characteristics of cases from Quebec and France, may attenuate genetic associations in the meta-analysis. Patients from Quebec were all cases with severe AS undergoing aortic valve replacement. In contrast, AS cases from France were identified by echocardiographic examination and had at least mild AS defined as an aortic jet velocity >2.0 m/s. The proportion of severe AS defined by aortic jet velocity >4.0 m/s in France samples was 54.2%. More DNA collections of case–control cohorts well-characterized for AS are needed to carry genetic analyses in more strictly defined phenotype. In addition, controls for the GWAS were from the database of genotypes and phenotypes. Accordingly, clinical data for controls were not specifically collected to study AS, and differences in clinical characteristics were observed between cases and controls. For example, controls that were compared with cases from Quebec were taken from the PCOS cohort, which is a cohort of women only. The selection of this cohort of controls implies that sex does not influence allele frequencies between cases and controls. Controls were also younger. A proportion of these controls is likely to later develop AS, causing a misclassification of a small subset of controls. However, this is not likely to be a major bias as the prevalence of all valve diseases is estimated at 2.5%.2 It was demonstrated that if 5% of controls would meet the definition of cases, thus misclassified, the loss of power would be approximately the same as that because of a reduction of the sample size by 10%.43 Further work is needed to determine whether the identified eQTL-SNPs moderately associated with AS are tagging nongenotyped/imputed functional variants. Although the sample size to detect valve eQTLs is small, the eQTL mapping study was conducted in the most relevant tissue to study AS. eQTLs are known to be cell type–specific and tissue-specific.44,45 Human aortic valve specimens are challenging to obtain, and here we reported the first valve eQTL mapping study. To alleviate the power limitation, we have restricted the eQTL analysis to SNPs showing some evidence of association with AS and genes differentially expressed in calcified compared with normal valves. However, we cannot preclude false-positive results, and a valve eQTL study of larger sample size is needed.

Several genes implicated in AS development and progression have been documented.9 However, the key molecules that will instigate the development of drugs to treat AS are still unknown. This GWAS confirmed that AS is not because of one causal gene but the action of several genes with subtle individual risk effects that are acting in concert to modulate higher-order biological processes leading to AS. One of the most important discoveries of this study is the identification of new plausible candidate genes moderately associated with AS and differentially expressed in calcified compared with normal aortic valves. Gene-set association analysis coupled with RNA-Seq and valve eQTL data revealed CACNA1C as a new gene in the calcium signaling pathway associated with AS and confirmed the role of RUNX2 as a potential driver of AS development.

Supplementary Material

CLINICAL PERSPECTIVE.

Calcific aortic valve stenosis (AS) is the most common valvular heart disease (2%; >65 years old). The only treatments available for symptomatic patients are aortic valve replacement surgery and transcatheter aortic valve implantation. With the current shift toward an older population, AS is becoming a major societal and economic burden. Conventional cardiovascular drugs, such as statin, have failed to prevent the development of AS. It is thus urgent to identify new therapeutic targets for this disease. This article describes the first genome-wide association study on AS. The genome-wide association study was performed with ≈ 1000 AS cases collected in Quebec City (Canada) and Paris (France). We then extended our findings to gene expression measured by RNA sequencing in human aortic valves. A second set of valves were also evaluated using genome-wide genotyping and gene expression to identify genetic variants influencing gene expression in valve tissues. Briefly, we found new genes involved in the calcium signaling pathway that are associated with AS and also differentially expressed in aortic valves of patients with AS compared with normal aortic valves obtained from heart transplantation. We also found AS-associated genetic variants that are associated with the expression of specific genes in valve tissues. The identification of AS-associated genetic variants that are also associated with gene expression in the most disease-relevant tissue is an important step forward to understand the heritable risk of AS and elucidate new molecular targets to treat this frequent and life-threatening cardiovascular disease.

Acknowledgments

We thank the research team at the cardiac surgical database and bio-bank of the Institut universitaire de cardiologie et de pneumologie de Québec (IUCPQ) for their valuable assistance. We also thank the team of the Centre d’Investigation Clinique and the Centre de Ressources Biologiques of Bichat Hospital for their help, as well as the clinical and research staff at the Montreal Heart Institute. We thank the patients at all sites for participating in this study. Complete acknowledgments for the dbGaP cohorts are available in Data Supplement.

Sources of Funding

This study was supported by the Heart and Stroke Foundation of Canada, the Institut universitaire de cardiologie et de pneumologie de Québec (IUCPQ) Foundation, and the Canadian Institutes of Health Research grants MOP102481 (Dr Bossé), MOP137058 (Dr Bossé), MOP79342 (Drs Pibarot and Mathieu), MOP245048 (Drs Mathieu and Pibarot). The COFRASA (Aortic Stenosis in Elderly: Determinant of Progression; clinicalTrial.gov number NCT 00338676) and GENERAC (Genetic of Aortic Valve Stenosis-Clinical and Therapeutic Implications; clinicalTrial.gov number NCT00647088) studies are supported by grants from the Assistance Publique-Hôpitaux de Paris (PHRC National 2005 and 2010 and PHRC regional 2007 [Dr Messika-Zeitoun]). Dr Arsenault was a recipient of a postdoctoral fellowship of the Canadian Institutes of Health Research and is now recipient of a Junior 1 Research Scholar award from the Fonds de recherche Québec-Santé (FRQS). Dr Tardif holds the Canada Research Chair in translational and personalized medicine, and the University of Montreal endowed research chair in atherosclerosis. M. Lamontagne is the recipient of a doctoral studentship from the FRQS and was a recipient of doctoral studentships from the Centre de recherche de l’IUCPQ and the Faculty of medicine of Laval University. Dr Mathieu is a research scholar from the FRQS. Dr Pibarot holds the Canada Research Chair in Valvular Heart Diseases. S.C. Body received funding from the National Institutes of Health (NIH; R01HL114823). S. Guauque-Olarte was a recipient of a studentship from the International Chair on Cardiometabolic Risk and a doctoral scholarship from the Centre de recherche de l’IUCPQ. Y. Bossé was a research scholar from the Heart and Stroke Foundation of Canada and is now recipient of a Junior 2 Research Scholar award from the FRQS.

Footnotes

Disclosures

None.

References

  • 1.Carabello BA, Paulus WJ. Aortic stenosis. Lancet. 2009;373:956–966. doi: 10.1016/S0140-6736(09)60211-7. [DOI] [PubMed] [Google Scholar]
  • 2.Nkomo VT, Gardin JM, Skelton TN, Gottdiener JS, Scott CG, Enriquez-Sarano M. Burden of valvular heart diseases: a population-based study. Lancet. 2006;368:1005–1011. doi: 10.1016/S0140-6736(06)69208-8. [DOI] [PubMed] [Google Scholar]
  • 3.Otto CM, Burwash IG, Legget ME, Munt BI, Fujioka M, Healy NL, et al. Prospective study of asymptomatic valvular aortic stenosis. Clinical, echocardiographic, and exercise predictors of outcome. Circulation. 1997;95:2262–2270. doi: 10.1161/01.cir.95.9.2262. [DOI] [PubMed] [Google Scholar]
  • 4.Cowell SJ, Newby DE, Prescott RJ, Bloomfield P, Reid J, Northridge DB, et al. Scottish Aortic Stenosis and Lipid Lowering Trial, Impact on Regression (SALTIRE) Investigators. A randomized trial of intensive lipid-lowering therapy in calcific aortic stenosis. N Engl J Med. 2005;352:2389–2397. doi: 10.1056/NEJMoa043876. [DOI] [PubMed] [Google Scholar]
  • 5.Rossebø AB, Pedersen TR, Boman K, Brudi P, Chambers JB, Egstrup K, et al. SEAS Investigators. Intensive lipid lowering with simvastatin and ezetimibe in aortic stenosis. N Engl J Med. 2008;359:1343–1356. doi: 10.1056/NEJMoa0804602. [DOI] [PubMed] [Google Scholar]
  • 6.Rajamannan NM, Bonow RO, Rahimtoola SH. Calcific aortic stenosis: an update. Nat Clin Pract Cardiovasc Med. 2007;4:254–262. doi: 10.1038/ncpcardio0827. [DOI] [PubMed] [Google Scholar]
  • 7.Sehatzadeh S, Doble B, Xie F, Blackhouse G, Campbell K, Kaulback K, et al. Transcatheter aortic valve implantation (TAVI) for treatment of aortic valve stenosis: an evidence update. Ont Health Technol Assess Ser. 2013;13:1–40. [PMC free article] [PubMed] [Google Scholar]
  • 8.Mozaffarian D, Benjamin EJ, Go AS, Arnett DK, Blaha MJ, Cushman M, et al. American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics–2015 update: a report from the American Heart Association. Circulation. 2015;131:e29–322. doi: 10.1161/CIR.0000000000000152. [DOI] [PubMed] [Google Scholar]
  • 9.Bossé Y, Miqdad A, Fournier D, Pépin A, Pibarot P, Mathieu P. Refining molecular pathways leading to calcific aortic valve stenosis by studying gene expression profile of normal and calcified stenotic human aortic valves. Circ Cardiovasc Genet. 2009;2:489–498. doi: 10.1161/CIRCGENETICS.108.820795. [DOI] [PubMed] [Google Scholar]
  • 10.Horne BD, Camp NJ, Muhlestein JB, Cannon-Albright LA. Evidence for a heritable component in death resulting from aortic and mitral valve diseases. Circulation. 2004;110:3143–3148. doi: 10.1161/01.CIR.0000147189.85636.C3. [DOI] [PubMed] [Google Scholar]
  • 11.Probst V, Le Scouarnec S, Legendre A, Jousseaume V, Jaafar P, Nguyen JM, et al. Familial aggregation of calcific aortic valve stenosis in the western part of France. Circulation. 2006;113:856–860. doi: 10.1161/CIRCULATIONAHA.105.569467. [DOI] [PubMed] [Google Scholar]
  • 12.Gaudreault N, Ducharme V, Lamontagne M, Guauque-Olarte S, Mathieu P, Pibarot P, et al. Replication of genetic association studies in aortic stenosis in adults. Am J Cardiol. 2011;108:1305–1310. doi: 10.1016/j.amjcard.2011.06.050. [DOI] [PubMed] [Google Scholar]
  • 13.Ducharme V, Guauque-Olarte S, Gaudreault N, Pibarot P, Mathieu P, Bossé Y. NOTCH1 genetic variants in patients with tricuspid calcific aortic valve stenosis. J Heart Valve Dis. 2013;22:142–149. [PubMed] [Google Scholar]
  • 14.Garg V, Muth AN, Ransom JF, Schluterman MK, Barnes R, King IN, et al. Mutations in NOTCH1 cause aortic valve disease. Nature. 2005;437:270–274. doi: 10.1038/nature03940. [DOI] [PubMed] [Google Scholar]
  • 15.Thanassoulis G, Campbell CY, Owens DS, Smith JG, Smith AV, Peloso GM, et al. CHARGE Extracoronary Calcium Working Group. Genetic associations with valvular calcification and aortic stenosis. N Engl J Med. 2013;368:503–512. doi: 10.1056/NEJMoa1109034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Mailman MD, Feolo M, Jin Y, Kimura M, Tryka K, Bagoutdinov R, et al. The NCBI dbGaP database of genotypes and phenotypes. Nat Genet. 2007;39:1181–1186. doi: 10.1038/ng1007-1181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Delaneau O, Marchini J, Zagury JF. A linear complexity phasing method for thousands of genomes. Nat Methods. 2012;9:179–181. doi: 10.1038/nmeth.1785. [DOI] [PubMed] [Google Scholar]
  • 18.Howie BN, Donnelly P, Marchini J. A flexible and accurate genotype imputation method for the next generation of genome-wide association studies. PLoS Genet. 2009;5:e1000529. doi: 10.1371/journal.pgen.1000529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Patterson N, Price AL, Reich D. Population structure and eigenanalysis. PLoS Genet. 2006;2:e190. doi: 10.1371/journal.pgen.0020190. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007;81:559–575. doi: 10.1086/519795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Pruim RJ, Welch RP, Sanna S, Teslovich TM, Chines PS, Gliedt TP, et al. LocusZoom: regional visualization of genome-wide association scan results. Bioinformatics. 2010;26:2336–2337. doi: 10.1093/bioinformatics/btq419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Nam D, Kim J, Kim SY, Kim S. GSA-SNP: a general approach for gene set analysis of polymorphisms. Nucleic Acids Res. 2010;38(Web Server issue):W749–W754. doi: 10.1093/nar/gkq428. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Yu W, Clyne M, Khoury MJ, Gwinn M. Phenopedia and Genopedia: disease-centered and gene-centered views of the evolving knowledge of human genetic associations. Bioinformatics. 2010;26:145–146. doi: 10.1093/bioinformatics/btp618. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Warren BA, Yong JL. Calcification of the aortic valve: its progression and grading. Pathology. 1997;29:360–368. doi: 10.1080/00313029700169315. [DOI] [PubMed] [Google Scholar]
  • 25.Trapnell C, Hendrickson DG, Sauvageau M, Goff L, Rinn JL, Pachter L. Differential analysis of gene regulation at transcript resolution with RNA-seq. Nat Biotechnol. 2013;31:46–53. doi: 10.1038/nbt.2450. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010;11:R106. doi: 10.1186/gb-2010-11-10-r106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010;26:139–140. doi: 10.1093/bioinformatics/btp616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S, et al. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 2004;5:R80. doi: 10.1186/gb-2004-5-10-r80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Du P, Kibbe WA, Lin SM. lumi: a pipeline for processing Illumina micro-array. Bioinformatics. 2008;24:1547–1548. doi: 10.1093/bioinformatics/btn224. [DOI] [PubMed] [Google Scholar]
  • 30.Nordström P, Glader CA, Dahlén G, Birgander LS, Lorentzon R, Waldenström A, et al. Oestrogen receptor alpha gene polymorphism is related to aortic valve sclerosis in postmenopausal women. J Intern Med. 2003;254:140–146. doi: 10.1046/j.1365-2796.2003.01179.x. [DOI] [PubMed] [Google Scholar]
  • 31.Panagiotou OA, Ioannidis JP Genome-Wide Significance Project. What should the genome-wide significance threshold be? Empirical replication of borderline genetic associations. Int J Epidemiol. 2012;41:273–286. doi: 10.1093/ije/dyr178. [DOI] [PubMed] [Google Scholar]
  • 32.Manolio TA. Bringing genome-wide association findings into clinical use. Nat Rev Genet. 2013;14:549–558. doi: 10.1038/nrg3523. [DOI] [PubMed] [Google Scholar]
  • 33.Hennessey JA, Boczek NJ, Jiang YH, Miller JD, Patrick W, Pfeiffer R, et al. A CACNA1C variant associated with reduced voltage-dependent inactivation, increased CaV1.2 channel window current, and arrhythmogenesis. PLoS One. 2014;9:e106982. doi: 10.1371/journal.pone.0106982. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Johnson AD, Newton-Cheh C, Chasman DI, Ehret GB, Johnson T, Rose L, et al. Cohorts for Heart and Aging Research in Genomic Epidemiology Consortium; Global BPgen Consortium; Women’s Genome Health Study. Association of hypertension drug target genes with blood pressure and hypertension in 86,588 individuals. Hypertension. 2011;57:903–910. doi: 10.1161/HYPERTENSIONAHA.110.158667. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Wang W, Pang L, Palade P. Angiotensin II upregulates Ca(V)1.2 protein expression in cultured arteries via endothelial H(2)O(2) production. J Vasc Res. 2011;48:67–78. doi: 10.1159/000318776. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Helske S, Lindstedt KA, Laine M, Mäyränpää M, Werkkala K, Lommi J, et al. Induction of local angiotensin II-producing systems in stenotic aortic valves. J Am Coll Cardiol. 2004;44:1859–1866. doi: 10.1016/j.jacc.2004.07.054. [DOI] [PubMed] [Google Scholar]
  • 37.O’Brien KD, Shavelle DM, Caulfield MT, McDonald TO, Olin-Lewis K, Otto CM, et al. Association of angiotensin-converting enzyme with low-density lipoprotein in aortic valvular lesions and in human plasma. Circulation. 2002;106:2224–2230. doi: 10.1161/01.cir.0000035655.45453.d2. [DOI] [PubMed] [Google Scholar]
  • 38.El Husseini D, Boulanger MC, Mahmut A, Bouchareb R, Laflamme MH, Fournier D, et al. P2Y2 receptor represses IL-6 expression by valve interstitial cells through Akt: implication for calcific aortic valve disease. J Mol Cell Cardiol. 2014;72:146–156. doi: 10.1016/j.yjmcc.2014.02.014. [DOI] [PubMed] [Google Scholar]
  • 39.Le Quang K, Bouchareb R, Lachance D, Laplante MA, El Husseini D, Boulanger MC, et al. Early development of calcific aortic valve disease and left ventricular hypertrophy in a mouse model of combined dyslipidemia and type 2 diabetes mellitus. Arterioscler Thromb Vasc Biol. 2014;34:2283–2291. doi: 10.1161/ATVBAHA.114.304205. [DOI] [PubMed] [Google Scholar]
  • 40.Towler DA. Molecular and cellular aspects of calcific aortic valve disease. Circ Res. 2013;113:198–208. doi: 10.1161/CIRCRESAHA.113.300155. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Nagy E, Eriksson P, Yousry M, Caidahl K, Ingelsson E, Hansson GK, et al. Valvular osteoclasts in calcification and aortic valve stenosis severity. Int J Cardiol. 2013;168:2264–2271. doi: 10.1016/j.ijcard.2013.01.207. [DOI] [PubMed] [Google Scholar]
  • 42.Arsenault BJ, Boekholdt SM, Dubé MP, Rhéaume E, Wareham NJ, Khaw KT, et al. Lipoprotein(a) levels, genotype, and incident aortic valve stenosis: a prospective Mendelian randomization study and replication in a case-control cohort. Circ Cardiovasc Genet. 2014;7:304–310. doi: 10.1161/CIRCGENETICS.113.000400. [DOI] [PubMed] [Google Scholar]
  • 43.Colhoun HM, McKeigue PM, Davey Smith G. Problems of reporting genetic associations with complex outcomes. Lancet. 2003;361:865–872. doi: 10.1016/s0140-6736(03)12715-8. [DOI] [PubMed] [Google Scholar]
  • 44.Dimas AS, Deutsch S, Stranger BE, Montgomery SB, Borel C, Attar-Cohen H, et al. Common regulatory variation impacts gene expression in a cell type-dependent manner. Science. 2009;325:1246–1250. doi: 10.1126/science.1174148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Fu J, Wolfs MG, Deelen P, Westra HJ, Fehrmann RS, Te Meerman GJ, et al. Unraveling the regulatory mechanisms underlying tissue-dependent genetic variation of gene expression. PLoS Genet. 2012;8:e1002431. doi: 10.1371/journal.pgen.1002431. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

NIHMS797414-supplement-supplement_1.pdf (988.6KB, pdf)

RESOURCES