Abstract
There is currently no medical therapy to prevent calcific aortic valve stenosis (CAVS). Multi-omics approaches could lead to the identification of novel molecular targets. Here, we perform a genome-wide association study (GWAS) meta-analysis including 14,819 cases among 941,863 participants of European ancestry. We report 32 genomic loci, among which 20 are novel. RNA sequencing of 500 human aortic valves highlights an enrichment in expression regulation at these loci and prioritizes candidate causal genes. Homozygous genotype for a risk variant near TWIST1, a gene involved in endothelial-mesenchymal transition, has a profound impact on aortic valve transcriptomics. We identify five genes outside of GWAS loci by combining a transcriptome-wide association study, colocalization, and Mendelian randomization analyses. Using cross-phenotype and phenome-wide approaches, we highlight the role of circulating lipoproteins, blood pressure and inflammation in the disease process. Our findings pave the way for the development of novel therapies for CAVS.
Similar content being viewed by others
Introduction
Calcific aortic valve stenosis (CAVS) is the most frequent valvular heart disease, affecting up to 2% of the population aged 65 years and older1,2. Progressive fibro-calcific remodeling of aortic valve leaflets leads to the obstruction of blood flow from the left ventricle to the aorta, causing left ventricular remodeling, heart failure, and death3,4. Aortic valve replacement is currently the only effective treatment and there is no pharmacological therapy to prevent CAVS or slow its progression.
In the last decade, several genetic variants associated with CAVS have been identified through genome-wide association studies (GWAS), improving our understanding of the disease pathophysiology. Variants in LPA, associated with lipoprotein (a) levels, and PALMD, a regulator of actin polymerization5,6, remain the most strongly associated with CAVS7,8,9. Large-scale transcriptomic analyses in human aortic valves have confirmed the specific role of PALMD in CAVS8,10, and led to the identification of NAV1 as another contributing gene9. Variants located in TEX41, IL6, ALPL and FADS1/2 have also shown significant association with CAVS in previous GWAS9,11,12. More recently, two large GWAS including 14,451 and 13,765 cases, respectively, led to the identification of 20 additional loci, some of which are associated with blood lipids (CELSR2-SORT1) or obesity (FTO)13,14.
In this work, we perform a GWAS meta-analysis for CAVS in 14,819 cases among 941,863 participants and generate extensive transcriptomic data from 500 human aortic valves to identify novel key molecular drivers. Our objectives were to (1) identify novel genetic loci and candidate causal genes that impact gene expression in the aortic valve and (2) characterize the relationship between the genetic architecture of CAVS and cardiovascular traits.
Results
Identification of CAVS genomic loci
A GWAS meta-analysis for CAVS was performed in 14,819 cases (61% men) and 927,044 controls of European ancestry from six cohorts (Supplementary Fig. 1, Supplementary Fig. 2, and Supplementary Data 1). In total, 17,130,226 variants were included in the meta-analysis, 7,159,971 of which were available in all six cohorts. The genomic inflation factor was 1.099 and the LD score regression intercept was 1.020. Thirty-two loci reached genome-wide significance (P < 5 × 10−8), including 20 novel loci (Table 1, Fig. 1 and Supplementary Data 2). Among the novel associations, variants located in LPL and LDLR, two major regulators of circulating blood lipids, were identified. The 32 lead variants at the genome-wide associated loci had a concordant direction of effect in the recent meta-analysis by Chen et al.14. Thirty showed nominal association with CAVS (P < 0.05), of which 29 remained significant when using a threshold of false discovery rate <5%. In the deCODE cohort (n = 2457 CAVS cases and 349,342 controls)11, 28 out of the 32 lead variants had the same direction of effect (Supplementary Data 3). Variants at seven other loci previously reported in European populations were nominally associated (P < 0.05) with a concordant direction of effect, including six that remained significant when using a threshold of false discovery rate <5% (Supplementary Data 4).
Variant annotation and prioritization
Among variants in linkage disequilibrium with independent significant SNPs, there was a significant enrichment for intronic and exonic functional annotations (Supplementary Data 5). Among the lead SNPs, two were missense variants located in STARD9 (rs28744617) and ASCC2 (rs61736786). Missense variants in ALDH1A2 (rs4646626), MUC4 (rs2246901), and CDAN1 (rs4265781) were in high LD (r2 ≥ 0.8) with a lead SNP. The missense variants in ASCC2, ALDH1A2 and MUC4 had Combined Annotation Dependent Depletion (CADD) scores of 14.52, 17.37, and 21.50, respectively, ranking them in the 5% most deleterious variants in the genome. A nonsense variant in LPL previously associated with circulating triglyceride levels, rs328 (S447X), was in high LD (r2 = 0.958) with the lead variant at this locus (Supplementary Data 6). The 95% credible set, determined using a Bayesian approach implemented in CAVIAR, included five or less variants for five loci (nearest genes: PALMD, PRRX1, TMEM44, LPA, PDE3A and HMGA2). The complete sets for each locus are reported in Supplementary Data 7. Conditional analyses led to the identification of independent genome-wide significant signals at the LPA (n = 3) and IL6 (n = 2) loci (Supplementary Data 8).
Gene mapping
We mapped the protein-coding genes located within 10 kilobases (kb) of a genome-wide significant SNP and the genes located on either side for lead intergenic SNPs, identifying a list of 79 nearby genes (Supplementary Data 2). A gene-based analysis using MAGMA identified 158 genes (117 additional genes) at a false discovery rate threshold <5%, for a total of 196 prioritized genes (Supplementary Data 9).
Expression in human aortic valves
To identify the genes most likely to be involved in CAVS, we completed extensive transcriptomic profiling of the most relevant tissue, the aortic valve. We performed RNA-sequencing on 500 human aortic valve samples of various disease stages (81.6% with CAVS) and valve morphology (44% bicuspid) from patients recruited in our institution to create the QUEBEC-CAVS-RNA dataset (Supplementary Data 10). First, the expression level of the 196 genes prioritized by positional mapping or MAGMA was evaluated. Among the 187 genes with available gene expression quantification in aortic valves, 22 had a median expression level above the 90th percentile of all the protein-coding genes, including seven genes located nearby the meta-analysis lead SNPs: RPLP2, CCND1, ACTR2, TALDO1, PDGFRA, PRRX1, and UQCR10 (Supplementary Data 11).
To identify genes for which there is specific expression in the aortic valve, we compared gene expression with 43 GTEx tissues by calculating expression specificity scores (ESS). Among the genes of interest, nine had an ESS > 0.1 for the aortic valve, which corresponded approximatively to the top 5% of the empirical distribution for all the protein-coding genes in the aortic valve (Supplementary Data 12). Five of them were located nearby the meta-analysis lead SNPs: CHST6, CCND1, ALDH1A2, ADAMTS7 and PRRX1. The maximum ESS was observed in the aortic valve for all five genes except PRRX1, for which there was a slightly higher ESS in fibroblasts in GTEx. Other genes with a high ESS in a relevant tissue included HMGA2 (ESS = 0.80 in fibroblasts), NKX2-5 (ESS = 0.42 in atrial appendage), PDE3A (ESS = 0.14 in tibial artery) and PDGFRA (ESS = 0.13 in fibroblasts) (Supplementary Data 12 and Supplementary Fig. 3).
Expression quantitative trait loci
Genome-wide expression quantitative trait loci (eQTL) analyses in 484 human aortic valves identified 4,671,347 significant SNP-gene pairs. Among the 32 meta-analysis lead SNPs, 48 significant SNP-gene pairs were identified in the aortic valve, located in 20 loci (Supplementary Data 13). For these 32 lead SNPs, there was a significantly higher proportion of significant SNP-gene pairs (48/915, 5.25%) compared to all the SNP-gene pairs tested (4,671,347/182,925,823, 2.55%, P = 4.25 × 10−7). A Wilcoxon rank sum test also confirmed an enrichment towards stronger associations for these 915 SNP-gene pairs (P = 1.71 × 10−7) (Supplementary Fig. 4).
In 43 tissues from GTEx, significant eQTLs were found for 25 loci (Supplementary Data 14). The number of significant SNP-gene pairs ranged from one (brain anterior cingulate cortex) to 37 (tibial nerve). Taken together, the most significant eQTL was in the aortic valve for 12 loci, including four genes for which there was no reported eQTL in the GTEx tissues: rs6702619-PALMD (P = 7.1 × 10−119), rs7804522-TWIST1 (P = 5.7 × 10−22), rs1965668-NKX2-5 (P = 7.9 × 10−12) and rs72854462-TEX41 (P = 7.8 × 10−6). For 16 other loci, the most significant eQTL was found in a GTEx tissue. However, these were distributed among ten different tissues, which each had between one and three of the most significant SNP-gene pairs for a given locus (Supplementary Fig. 5). Among these, associations include rs1706003-ATP13A3 in the left ventricle (P = 1.7 × 10−26), rs771264-RNF144A in fibroblasts (P = 1.9 × 10−12) and rs11330858-MECOM in the aorta (P = 7.5 × 10−5).
Transcriptome-wide association study, colocalization and Mendelian randomization
A transcriptome-wide association study (TWAS) was performed to identify genes for which genetically predicted expression in the aortic valve is associated with CAVS. Thirty-five genes reached statistical significance at a false discovery rate <5%, including PALMD and NAV1, identified in a previous TWAS9 (Supplementary Data 15). To obtain further evidence on the potential causal role of these genes in CAVS, colocalization and Mendelian randomization (MR) analyses were performed. After removing genes with a low probability for colocalization of the GWAS and eQTL signals (PP4 < 0.75, n = 20), with no MR instrument (n = 3) or with high heterogeneity of the instrument (n = 2), ten genes remained (PALMD, NAV1, PRRX1, ATP13A3, BCL10, TWIST1, RAD9A, NRBP1, FES and AFAP1). All genes showed a significant association in MR with the inverse-variance weighted and weighted median approaches (PIVW and PWM adjusted for false discovery rate <5%), suggesting a causal association between gene expression in the aortic valve and CAVS. There was no evidence of pleiotropy with the Egger intercept test (Pintercept > 0.05 for all genes). Among the novel genes identified, three were located at a genome-wide significant locus: PRRX1, ATP13A3 and TWIST1, for which gene expression in the aortic valve was positively associated with CAVS (PIVW = 4.5 × 10−17, 9.4 × 10−9, and 5.7 × 10−4, respectively) (Fig. 2). The lead SNPs in the meta-analysis at these three loci were strong eQTLs in the aortic valve (all PeQTL < 1 × 10−15). Five other genes were identified; the association between genetically predicted expression and CAVS was positive for three: BCL10, RAD9A and NRBP1 (PIVW = 0.0066, 5.6 × 10−5, and 9.4 × 10−4, respectively) and negative for two: FES and AFAP1 (PIVW = 7.9 × 10−4 and 5.5 × 10−4, respectively) (Supplementary Figs. 6 and 7).
To evaluate a potential impact of valve morphology on gene expression regulation for these genes, we calculated eQTL separately for tricuspid and bicuspid aortic valves. For all 10 genes identified by the transcriptomic analyses, the effect of the lead variant on gene expression was similar in tricuspid and bicuspid valves (all P < 0.05 for heterogeneity) (Supplementary Data 16).
Differential expression according to genotype at the TWIST1 locus
Considering the potential causal relationship between genetically determined expression of TWIST1 in the aortic valve and CAVS risk, as well as the fact that the lead GWAS SNP, rs7804522, is only associated with gene expression in the aortic valve (valve-specific eQTL), we explored further the impact of this variant on the transcriptomic profile in the aortic valve. After selecting individuals with severe aortic stenosis, we compared gene expression between the 66 individuals homozygous for the risk allele (rs7804522-C) and the 138 individuals homozygous for the other allele. We found 509 differentially expressed genes at a false discovery rate <5%, indicating a profound effect of this genotype on transcriptome-wide gene expression (Fig. 3 and Supplementary Data 17). For comparison, a differential expression analysis performed using the same method for the four other genome-wide significant loci corroborated by transcriptomic analyses in the aortic valve (PALMD, NAV1, PRRX1 and ATP13A3) showed only five or less differentially expressed genes between individuals homozygous for the risk allele and those homozygous for the other allele (Supplementary Data 18). At the TWIST1 locus, a total of 148 genes were up-regulated while 361 genes were down-regulated in the group homozygous for the risk allele (Fig. 3). The dysregulated genes were enriched for several metabolic pathways including neutrophil degranulation, cell cycle, cell division and apoptosis (Fig. 3 and Supplementary Data 19).
Prioritization of causal genes
To prioritize potential causal genes, we combined the evidence from ten different features, including lead variants location and annotation, association at the gene level, expression in the aortic valve, colocalization, TWAS and MR. Taken together, 17 genes located at genome-wide significant loci and eight genes identified by TWAS had four or more features that suggested their implication in CAVS (Fig. 4). Among the genes located at genome-wide significant loci, PRRX1, NAV1, ALDH1A2 and MUC4 had seven or more features while ACTR2, SMAD9, PALMD, ATP13A3, FADS2 and TWIST1 had four or more features including significant associations in the TWAS and MR analyses. Out of the 25 prioritized genes, seven encoded druggable human proteins (MUC4, ALDH1A2, NRBP1, FES, PDGFRA, LPL and NPC1). Existing drugs interacting with these genes according to the drug-gene interaction database are reported in Supplementary Data 20.
Pathway enrichment
A pathway analysis was performed with the Metascape tool by including the genes identified using MAGMA as well as the 35 genes significant in the TWAS analysis. Among the top significantly enriched terms, we identified regulation of interleukin-6 production, embryonic development, regulation of osteoblast differentiation, response to growth factor and plasma lipoprotein assembly, remodeling, and clearance (Supplementary Fig. 8 and Supplementary Data 21).
Cross-phenotype analyses
We used the interactive cross-phenotype analysis of GWAS (iCPAG) database to identify other phenotypes that share genetic associations. After adjustment for multiple testing, there were 35 significantly associated phenotypes (PBonferroni < 0.05), including plasma lipids (lipoprotein (a), LDL-cholesterol, triglycerides, apolipoprotein B, total cholesterol, lipoprotein-associated phospholipase A2), blood pressure traits (pulse, diastolic and systolic blood pressure), other cardiovascular diseases (coronary artery disease, peripheral arterial disease, carotid atherosclerosis, metabolic syndrome), but also aortic root size, coronary artery calcification, bone mineral density, resting heart rate and leukocyte count (Supplementary Data 22).
The association of the lead GWAS and TWAS variants with 44 other cardiovascular traits was then evaluated in UK Biobank (Fig. 5). The CAVS risk allele for several variants was positively associated with circulating lipids such as LDL-cholesterol and apolipoprotein B or blood pressure traits. Notably, 14 variants had a significant association with pulse pressure while 11 were associated with coronary artery disease with a direction of effect concordant with CAVS.
Accordingly, there was a significant positive genetic correlation between CAVS and lipids, blood pressure traits and coronary artery disease. A positive genetic correlation was also observed for abdominal aorta calcification, ischemic stroke, peripheral artery disease, body-mass index, diabetes and C-reactive protein (Fig. 6 and Supplementary Data 23).
Discussion
We performed a GWAS meta-analysis including 14,819 CAVS cases and identified 32 genome-wide significant loci, 20 of which had not been reported before, including 18 with supportive evidence from another study. Leveraging transcriptomic data obtained from RNA sequencing of 500 human aortic valves, we identified novel candidate causal genes, some of which could constitute therapeutic targets. We report several significant eQTL at genomic loci, many of which are specific to the aortic valve, implicating tissue-specific regulatory mechanisms. We also show robust evidence of genetic relationship between CAVS and several traits, including circulating lipids, blood pressure, atherosclerotic cardiovascular diseases and inflammation.
Combining the largest GWAS meta-analysis to date with robust eQTL for the aortic valve generated from the most comprehensive resource so far led to the identification of target genes residing within genomic risk loci and revealed novel risk loci. A TWAS followed by Bayesian colocalization and MR identified eight genes for which genetically determined expression in the aortic valve is robustly associated with CAVS. Three of them were located at genome-wide significant loci: PRRX1, ATP13A3 and TWIST1. PRRX1 and TWIST1 code for transcription factors involved in epithelial-mesenchymal transition, acting as potent mesenchymal promoters15. They were both identified as key members of fibroblast transcriptional regulatory networks and involved in a positive feedback loop that leads to fibroblast activation16,17. PRRX1, paired related homeobox 1, codes for a DNA-associated protein located in the nucleus that acts as a transcription co-activator. It is involved in embryonic development, where it was shown to impact heart looping and laterality in animal models18. PRRX1 has been shown to orchestrate the transition of stromal fibroblasts to myofibroblasts via TGF-β signaling19, which is a known mechanism taking place in valve interstitial cells in CAVS20. Here, we show that PRRX1 is highly expressed in human aortic valves, that the lead SNP at this locus is a strong eQTL specific to this tissue, and report a potential causal role of local PRRX1 expression in CAVS. TWIST1, coding for twist family bHLH transcription factor 1, plays an important role in endothelial-mesenchymal transition and embryonic development for several organs, including the heart21. TWIST1 is highly expressed in endocardial cushions during early development and is then down-regulated for heart valve remodeling22. Persistent expression in valves has been shown to increase cell proliferation, dysregulation of fibrillary collagen and lead to enlarged hypercellular valve leaflets in a mouse model. In the same study, high TWIST1 expression was observed in human diseased aortic valves23. In accordance with these previous observations, we identified a significant positive association between genetically determined expression of TWIST1 in human aortic valves and CAVS. Moreover, the lead variant at this locus (rs7804522, minor allele frequency of 0.42) was a strong eQTL specific to the aortic valve with no reported association in GTEx, despite high expression levels in other tissues. Such remarkable eQTL specificity for the aortic valve was also observed at the PALMD locus, as reported previously8. We show that the aortic valves of individuals with severe CAVS carrying two risk alleles associated with increased TWIST1 expression have a distinct transcriptomic profile compatible with increased cell proliferation, compared to individuals with CAVS of comparable hemodynamic severity carrying no risk allele. ATP13A3 codes for a P-type ATPase involved in polyamine transport24. It is highly expressed in heart and vascular tissues. Protein-truncating variants in this gene have previously been associated with pulmonary artery hypertension. Functional studies have linked the loss of ATP13A3 to inhibition of proliferation and stimulation of apoptosis in endothelial cells25. In our study, the risk allele at the lead SNP was associated with an increase in ATP13A3 expression in the aortic valve. The same variant was also significantly associated with lower diastolic blood pressure and higher QT interval in previous GWAS26,27.
Five other genes were identified as candidate causal genes based on TWAS, colocalization and MR: BCL10, RAD9A and NRBP1 had a positive association, while FES and AFAP1 had a negative association with CAVS. BCL10 is an immune signaling adaptor mostly expressed in lymphocytes. It has a key role in lymphocyte activation through the nuclear factor-kappa B pathway28. Homozygous loss-of-function mutations in this gene cause severe combined immunodeficiency29. Higher expression of BCL10 in the aortic valve could contribute to the immune-related inflammation observed in CAVS20. RAD9A encodes a checkpoint protein required for cell cycle arrest and DNA damage repair. DNA binding to RAD9A was shown to have a critical role in maintaining cell viability and checkpoint activation in the presence of oxidative stress30. To our knowledge, this gene has not been associated with cardiovascular traits before. NRBP1, nuclear receptor binding protein 1, is involved in endoplasmic reticulum to Golgi vesicle-mediated transport. The CAVS risk allele at the lead GWAS SNP at this locus was strongly associated with higher circulating triglycerides, LDL-cholesterol, apolipoprotein B, C-reactive protein, leucocyte and platelet counts, suggesting potential roles in lipids, inflammation and thrombosis. FES encodes a cytoplasmic protein-kinase involved in the regulation of various cellular functions including cell proliferation and differentiation as well as inflammation. This locus has been associated with CAD in several GWAS; the lead CAD SNP, rs17514846, is in high LD (r2 > 0.8) with our lead CAVS SNP. A recent functional study showed that the risk genotype reduces FES expression in monocytes and increases the number of monocytes in human atherosclerotic plaque. Knockout of FES also promoted migration of monocytes as well as vascular smooth muscle cells and increased atherosclerotic plaque size in a murine model31. A decrease in FES expression in the aortic valve could lead to CAVS through a similar mechanism. Of note, the risk allele at the lead SNP was also strongly and positively associated with systolic, diastolic and pulse pressure. AFAP1 encodes actin filament associated protein 1, a potential modulator of actin filament organization in response to cellular signals. AFAP1 was shown to colocalize and bind directly with actin filaments in fibroblast cells. In the same study, deletion of a leucine zipper motif within AFAP1 impacted cell shape and lamellipodia formation in multiple cell lines32. This locus was recently associated with descending thoracic aorta diameter by GWAS33.
Combining multiple lines of evidence to prioritize genes including transcriptomic characterization of a large number of human aortic valves, we highlight a potential role in CAVS for three other genes at novel genome-wide loci: MUC4, ALDH1A2, and SMAD9. MUC4, or mucin 4, encodes an integral membrane glycoprotein found on the cell surface. This locus was associated with bicuspid aortic valve in a recent GWAS34. In the same study, functional experiments on a zebrafish model showed that knockout or knockdown of MUC4 leads to a temporal delay in cardiac valve development. In our study, genetically determined expression in human aortic valve was positively associated with CAVS in MR analyses, further implicating this gene in disease pathophysiology. ALDH1A2, aldehyde dehydrogenase 1 family member A2, encodes an enzyme that catalyzes the synthesis of retinoic acid from retinaldehyde. Loss-of-function mutations in this gene lead to an autosomal recessive condition presenting with diaphragmatic hernia and cardiovascular defects, including atrial as well as ventricular septal defects and aortic root ectasia35,36. In our study, we report a high expression specificity for the aortic valve and a positive relationship between genetically determined expression and CAVS supported by MR. Of note, a missense variant in high LD with the lead SNP was inversely associated with CAVS risk. SMAD9 encodes a member of the SMAD family, which transduces signals from TGF-beta family members. It was shown to inhibit bone morphogenetic protein (BMP) signaling37. Loss-of-function mutations in mice models were reported to induce thickening and smooth muscle hyperplasia in the pulmonary vasculature, with upregulation of TGF-beta signaling and Prx1 (orthologous to human PRRX1) expression38. A rare missense mutation in humans was associated with unexplained high bone mass39. In our study, the lead GWAS SNP was a strong eQTL in the aortic valve. We found a significant inverse relationship between SMAD9 expression in the aortic valve and CAVS, consistent with the inhibitory effect of this protein on TGF-beta signaling and bone formation.
Several identified loci were associated with circulating blood lipids. While a strong association at the LPA locus has previously been reported7,8,9,11, the signals in LDLR and LPL are novel, further implicating atherogenic lipoproteins in the pathophysiology of CAVS. A higher incidence of CAVS has been observed in patients with familial hypercholesterolemia, most frequently caused by a loss-of-function mutation in the LDLR gene40. The protective allele of the lead variant at the LPL locus was in high LD (r2 = 0.958) with rs328 (S447X), a nonsense gain-of-function variant previously shown to increase lipolytic activity41. It has also consistently been associated with lower triglycerides levels and lower coronary artery disease risk42. Other novel loci associated with blood lipids included 9q34.2 (nearest genes MYMK-SLC2A6), for which the lead SNP was associated with circulating LDL-cholesterol and apolipoprotein B and 15q15.2 (nearest gene STARD9), for which the lead SNP was associated with plasma triglycerides. Cross phenotype analysis with other GWAS and genetic correlation also showed a significant genetic relationship with circulating lipids, consistent with evidence from previous Mendelian randomization studies43,44. All in all, our findings support a significant contribution of circulating blood lipids to the development of CAVS, suggesting a value for lipid-lowering interventions for CAVS prevention.
There were also several loci with a significant association with blood pressure traits, most notably for pulse pressure. In fact, 11 lead SNPs had a significant positive association with pulse pressure in UK Biobank. Among the novel loci, a strong association was observed at 4q12 (nearest gene PDGFRA), 11p15.5 (nearest gene PIDD1), 12p12.2 (nearest gene PDE3A) and 22q12.2 (nearest gene ASCC2), which are all loci previously associated with pulse pressure in a recent GWAS meta-analysis26. An increase in pulse pressure reflects a reduction in arterial compliance related to increased arterial stiffness or calcification45. Of note, missense mutations in PDE3A have been associated with an autosomal dominant form of hypertension. Affected individuals manifest with hyperplastic arterial walls resulting from excessive proliferation of vascular smooth muscle cells, which increases peripheral vascular resistance46,47. Genetic correlation results showed consistent positive relationships between CAVS and pulse pressure as well as aortic calcification, further supporting shared genetic mechanisms with vascular remodeling.
This study has some limitations. First, the cohorts only included individuals of European ancestry, due to a lack of sufficient available data for other ancestries. Further studies are needed in individuals of other ancestry to determine if the findings are generalizable and to identify specific associations in non-European populations. Second, complementary evidence is needed to understand the biological mechanisms by which the identified genes impact CAVS risk. Third, except for QUEBEC-CAVS-1, the cohorts all included CAVS patients with a bicuspid aortic valve, which could partly mediate the risk for some loci. However, the genes with supportive evidence from the transcriptomic analyses had similar genetic regulation in tricuspid and bicuspid aortic valves. Fourth, the directionality and causal nature of the relationships identified in the cross-phenotype analyses remain to be determined.
In conclusion, the integration of the largest GWAS meta-analysis to date for CAVS with a large transcriptomic dataset on human aortic valves allowed the identification of novel genomic loci and several candidate causal genes. We highlight the potential contribution of endothelial-mesenchymal transition, circulating lipoproteins, blood pressure, vascular remodeling and inflammation in the disease process. These findings pave the way for the identification of novel therapeutic targets for CAVS.
Methods
Ethical approval
The QUEBEC-CAVS study was approved by the ethics committee of the Institut universitaire de cardiologie et de pneumologie de Québec-Université Laval. The Norwich Local Research Ethics Committee granted ethical approval for the analysis in the European Prospective Investigation into Cancer and Nutrition [EPIC]-Norfolk study. The analysis in the Estonian Biobank [EstBB] was approved by the Estonian Committee on Bioethics and Human Research. UK Biobank received approval from the British National Health Service, North West - Haydock Research Ethics Committee. The Coordinating Ethics Committee of the Hospital District of Helsinki and Uusimaa (HUS) approved the FinnGen study protocol. Informed consent was obtained for all participants.
Study population
A total of six cohorts were included in the genome-wide association analyses: two case-control cohorts (QUEBEC-CAVS-1 and QUEBEC-CAVS-2) and four population-based cohorts (EPIC-Norfolk, EstBB, UK Biobank and FinnGen) (Supplementary Fig. 1). Only individuals of European ancestry were included, since the sample size was too low for other ancestries. A description of the cohorts including the diagnostic criteria used to define CAVS is available in the Supplementary Information. The analyses in UK Biobank were conducted under data application number 25205.
Genetic association analyses
Genotyping was completed in each cohort using genome-wide arrays (Supplementary Data 24). Samples with low call rate, sex mismatch with self-report, outlier heterozygosity or ancestry outliers were excluded. Related samples were excluded in the cohorts that did not use a method accounting for sample relatedness. Variants with a low call rate or marked deviations from Hardy–Weinberg equilibrium were excluded. Genotypes were imputed using reference panels from Haplotype Reference Consortium, 1000 Genomes, UK10K or a reference panel specific to the population (EstBB and FinnGen) (Supplementary Data 24).
The association between the dosage of each genetic variant and CAVS was evaluated using logistic regression. Covariables included age, sex and ancestry-based principal components (Supplementary Data 24). The summary statistics from each cohort were examined for discrepancies in allele frequencies and inflation factor was calculated. Variants with an imputation quality score <0.3, minor allele frequency <0.001, or minor allele count in cases corrected for the imputation quality <5 (when available) were excluded. An inverse-variance weighted fixed-effect meta-analysis was performed using METAL48. The statistical significance threshold was set at P < 5 × 10−8. Heterogeneity was evaluated using Cochran’s Q-test. Independent significant variants were identified by applying a linkage disequilibrium (LD) threshold of r2 ≥ 0.6 using 1000 Genomes phase 3 European as reference panel to the variants with P < 5 × 10−8. Lead variants were selected by applying a LD threshold of r2 ≥ 0.1. Lead variants located within 500 kilobases (kb) were merged into a single genomic locus.
Replication in other cohorts
We retrieved the association results for the lead variants from a recent meta-analysis by Chen et al.14 (n = 13,765 cases and 640,102 controls). Although this study also included UK Biobank participants, only up to 1675 cases are common with our analysis (representing ≤11.3% of the cases included in our study). We also retrieved the results from the deCODE cohort11 (n = 2457 cases and 349,342 controls, no overlap with our data). For variants with no results available, we used a proxy in LD (r2 > 0.8) using 1000 Genomes phase 3 European as reference panel.
Variant and gene prioritization
Variant annotation was performed using ANNOVAR49. Enrichment for functional consequences of variants in LD with independent significant variants (r2 ≥ 0.6) was calculated using Fisher’s exact test with 1000 Genomes phase 3 European as reference panel. Exonic variants with P < 1 × 10−5 were retrieved to identify missense and loss of function variants in high LD (r2 ≥ 0.8) with one of the lead SNPs. A 95% credible set of genetic variants was determined for each genomic locus using CAVIAR, assuming a single causal variant50. Conditional and joint association analysis (COJO) was performed using the --cojo-slct function in GCTA v1.92.3beta3 to identify variants independently associated with CAVS at each genomic locus51. Protein-coding genes located within 10 kilobases (kb) of a genome-wide significant SNP or located on either side of a lead intergenic SNP were identified. We also used Multi-marker Analysis of GenoMic Annotation (MAGMA) v1.08 with default settings to perform a gene analysis for 19,982 protein-coding genes52. Statistical significance was set below a false discovery rate of 5%.
Human aortic valve samples
Aortic valve tissues were collected from patients undergoing cardiac surgery at the Institut universitaire de cardiologie et de pneumologie de Québec – Université Laval between 1998 and 2019 (QUEBEC-CAVS-RNA). A total of 500 samples from patients undergoing aortic valve replacement for CAVS (n = 408), for aortic regurgitation in the absence of CAVS (n = 32) or undergoing heart transplant (n = 60, without aortic valve anomaly) were included. Samples were selected to obtain a representation of CAVS with different levels of severity, valve morphology (tricuspid and bicuspid) and patient sex (women and men) (Supplementary Data 10). The presence of aortic valve stenosis and its severity were evaluated using echocardiography, following recognized valve hemodynamic criteria including aortic valve area, mean and peak gradients53. Genotypes were obtained from blood samples for each participant using the Illumina Global Screening Array, as described in Supplementary Data 24 for the QUEBEC-CAVS-2 cohort.
Bulk RNA-sequencing
Aortic valve leaflets were preserved in RNAlater solution following collection and flash frozen in liquid nitrogen. RNA extraction was performed using a modified TRIzol protocol followed by RNA cleanup using QIAGEN RNeasy columns (QIAGEN, 74104). RNA quality was evaluated using a Bioanalyzer. RNA Integrity Number (RIN) was ≥6.0 for all samples selected for sequencing. Libraries were prepared using poly(A) and NEB Library Prep Kit for Illumina to select stranded mRNA (New England Biolabs, E7420L). Expression was assessed by 150 bp length paired-end RNA-sequencing (RNAseq) on a NovaSeq 6000 instrument (Illumina), aiming for >50 million paired reads per sample. The quality of the data was verified using the FastQC v0.11.5 and MultiQC v1.10 applications54. STAR v2.5.1b was used to align reads of each sample to the UCSC hg19 and hg38 reference genomes. Alignment quality was verified using QualiMap v2.2.155.
Comparison of expression between aortic valve and other tissues
Reads were collapsed to a single transcript model using GENCODE Release 26 on build GRCh38 as reference, to allow direct comparison with the Genotype-Tissue Expression (GTEx) project v856. Read counts and Transcripts Per Million (TPM) values were obtained using RNA-SeQC 257. Genes with a median log2(TPM) value across all samples above the 90th percentile of all protein-coding genes in the aortic valve were identified among the genes prioritized based on their proximity with lead GWAS variants or from the MAGMA analysis. TPM values from 43 non sex-specific tissues in GTEx (Supplementary Data 25) were retrieved. We then calculated expression specificity scores (ESS) for each gene58. Briefly, ESS were obtained by dividing the median log2(TPM) value in a given tissue by the sum of the medians of all tissues, resulting in a score ranging from 0 to 1. An ESS above 0.1 in the aortic valve, indicating that expression represents >10% of the overall expression in all tissues, was selected as a threshold to identify genes among the ones prioritized by positional mapping or MAGMA. This threshold corresponded approximately to the top 5% of the distribution of ESS in the aortic valve for all protein-coding genes.
Expression quantitative trait loci (eQTL) in aortic valve
Further quality control was performed to ensure that the genotypes obtained from blood corresponded to the aortic valve tissue of each participant. Sex mismatches and potential contamination were identified based on XIST and PRKY expression in the aortic valve RNAseq. The mbv function in QTLtools v1.1 was used to match the RNAseq inferred genotypes with the genotypes from blood59. Samples with a duplicated sequence, that did not match blood genotype or with potential contamination (>60% of matching heterozygous genotype with another sample) were excluded, which resulted in 484 samples available for further analyses (81.6% with CAVS).
Reads and TPM were generated as mentioned above using GENCODE Release 41 on build GRCh37. Genes with expression >0.1 TPM in at least 20% of samples and ≥6 reads in at least 20% of samples were kept56. Expression values were normalized between samples using trimmed mean of M-values (TMM) as implemented in edgeR v3.24.360. For each gene, expression was normalized across samples using an inverse normal transformation.
We calculated eQTL for genetic variants located 1 megabase (Mb) up- and downstream of the transcription start site of the genes located on chromosome 1 to 22 (cis-eQTL). Age, sex, smoking status (current or not), the first 60 Probabilistic Estimation of Expression Residuals (PEER)61 factors and the first five ancestry-based principal components were included as covariates. Only variants with a minor allele frequency ≥0.01 and imputation quality score ≥0.3 were considered. Association was evaluated using QTLtools v1.162 with the cis-QTL nominal pass function. We performed cis-QTL permutations using tensorQTL v1.0.7 to generate empirical p-values63. Nominal significance thresholds corresponding to a false discovery rate of <5% were calculated for each gene with QTLtools and applied to identify significant eQTLs. A similar analysis was performed separately for tricuspid (n = 215) and bicuspid aortic valves (n = 211), but using the first 30 PEER factors as covariates considering the smaller sample size.
We retrieved all significant eQTLs for the lead SNPs identified in the CAVS GWAS meta-analysis from our aortic valve dataset. Enrichment for eQTLs was evaluated by comparing the proportion of significant SNP-gene pairs for the lead SNPs to the one for the genome-wide SNPs, using a Pearson’s chi-squared test. The distribution of p-values in each group was compared using a Wilcoxon rank sum test. Significant eQTLs for the lead CAVS GWAS meta-analysis SNPs were also retrieved from 43 non sex-specific tissues in GTEx v8. When the lead SNP was not available, a proxy with r2 > 0.9 was used. Tissue-specific eQTLs were defined as significant SNP-gene pairs that were identified in a single tissue.
Transcriptome-wide association study (TWAS)
Effects of genetic variants on gene expression in the aortic valve were combined with the GWAS meta-analysis results to perform a TWAS using MetaXcan v0.7.4, with the S-PrediXcan extension64. First, gene expression models were developed using the PredictDB pipeline65. Briefly, elastic-net models were trained using nested cross validation from genotype and adjusted expression data. A model was considered significant when the average Pearson correlation between predicted and observed expression was greater than 0.1 and the estimated p-value was less than 0.05. Variants located in a window of +/−1 Mb of the gene of interest were selected. Only protein-coding genes were considered. The statistical significance threshold was set at a false discovery rate <5% to retain genes for further analyses.
Bayesian colocalization
COLOC v3.2.1 was used to evaluate colocalization between eQTL and CAVS risk for the genes identified by TWAS66. Variants located within 1 Mb of the gene were considered. A posterior probability of shared signal (PP4) > 0.75 was considered as indicative of colocalization. The LocusCompareR package v1.0.0 was used to illustrate colocalization events67.
Mendelian randomization (MR)
MR was used to evaluate the causal effect of the changes in gene expression level (exposure) on CAVS (outcome) for the genes identified by TWAS. Variants located within 1 Mb of the gene of interest and associated with gene expression with P < 0.0001 were selected. Independent variants were selected using the clump function implemented in plink v1.968, with a threshold of r2 < 0.1 in the European subset of 1000 Genomes (n = 503). Only genes with at least 3 variants available were tested. All the selected variants had an estimated F-statistic >1569. MR was performed by regressing genetic effect estimates for CAVS as determined in the GWAS meta-analysis on genetic effect estimates for expression levels (inverse-variance weighting). Heterogeneity was evaluated using Cochran’s Q statistic, instruments with P < 0.01 were not further considered. Egger MR was performed to determine the presence of unmeasured pleiotropy70. Weighted median estimator71 was used to ensure the robustness of the findings. The significance threshold was set at a false discovery rate <5% considering the number of genes tested. The analyses were performed using the MendelianRandomization v0.4.3 package.
Differential expression analyses
To further explore the effect of the risk genotype at a locus that seemed to involve specific regulation in the aortic valve (near TWIST1), differential expression analyses were performed. Genes with expression >0.1 TPM in at least 20% of samples and ≥6 reads in at least 20% of samples were selected, resulting in a total of 25,390 genes. Read counts were normalized using TMM for differential expression analyses using edgeR60. Samples that did not pass the quality control process (i.e., sex mismatch or mismatch of genotypes with mRNA sequences) were excluded. Among individuals with severe aortic stenosis according to echocardiographic criteria, gene expression in the aortic valve was compared between individuals homozygous for the risk allele rs7804522-C (n = 66) and individuals homozygous for the other allele rs7804522-G (n = 138) using a likelihood ratio test implemented in edgeR60. The model was adjusted for age, sex, smoking, RIN, 3’ bias, alignment rate and mitochondrial gene expression proportion. The significance threshold was set at false discovery rate <5%. A differential expression analysis performed using the same method was performed for the other genome-wide significant loci corroborated by transcriptomic analyses in the aortic valve, for comparison purposes.
Identification of candidate genes from selected features
For the genes located nearby the meta-analysis lead SNPs and the genes identified by TWAS, we counted the number of features that supported their role in CAVS. The following criteria were used: (1) nearest gene to a lead GWAS SNP, (2) lead GWAS SNP is located in an intron of the gene, (3) lead GWAS SNP is in linkage disequilibrium (r2 ≥ 0.8) with a missense or nonsense variant for the gene, (4) gene is significant in the MAGMA analysis, (5) gene has high expression in the aortic valve (above the 90th percentile of all protein-coding genes), (6) gene is specifically expressed in the aortic valve (ESS > 0.1), (7) the lead GWAS SNP is a significant eQTL for the gene in the aortic valve, (8) gene is significant in the TWAS analysis, (9) evidence of colocalization (PP4 > 0.75), (10) evidence of causality between gene expression and CAVS using MR (PIVW < 0.05, PWM < 0.05 and PHet > 0.01).
Drug target analysis
Among the genes with four or more supporting features, we identified those previously reported to encode a druggable human protein72. We then retrieved drugs interacting with these genes (interaction score ≥0.1) from the drug-gene interaction database73 (DGIdb, www.dgidb.org). A short description of the clinical indication for each drug was retrieved from DrugBank (www.drugbank.ca) or PubChem (https://pubchem.ncbi.nlm.nih.gov/).
Pathway enrichment
Pathway analyses were performed to identify enriched terms in Gene Ontology biological processes, KEGG pathways, Reactome gene sets, canonical pathways, and WikiPathways using Metascape74. Terms with a minimum of five overlapping genes and with a p-value < 0.001 obtained from a hypergeometric test are reported. These terms were then hierarchically clustered based on similarities among their gene memberships to identify independent leading terms.
Phenome-wide association studies
For each lead genetic variant identified from the GWAS or transcriptomic analyses in the aortic valve, association with 44 relevant cardiovascular traits and diseases were evaluated in UK Biobank9,10. Phenotypes were obtained from diagnosis in medical records, anthropomorphic measures and laboratory markers (Supplementary Data 26). For each phenotype, an additive logistic (binary phenotype) or linear regression (continuous phenotype) regression was performed in 353,378 unrelated individuals of White-British ancestry with adjustment for age, sex and the first 10 ancestry-based principal components using SNPTEST v2.5.475. The statistical significance threshold was set at P < 0.001 considering multiple hypothesis testing.
Shared genetic association with other traits
The interactive Cross-Phenotype Analysis of GWAS database (iCPAGdb, http://cpag.oit.duke.edu/explore/app/) was used to identify traits with shared genetic association with CAVS76. Traits with P < 0.05 following Bonferroni correction were considered significantly enriched.
Genetic correlation
We performed genetic correlation analyses between CAVS and cardiovascular diseases and risk factors using LD-score regression implemented in ldsc v1.0.077. We used publicly available summary statistics from GWAS meta-analyses for blood lipids, blood pressure, aortic size and calcification, height, body mass index, waist-to-hip ratio, coronary artery disease, stroke, peripheral artery disease, type 2 diabetes, chronic kidney disease, C-reactive protein and bone mineral density (Supplementary Data 23).
Statistical analysis
Analyses were performed using R version 3.5.1 unless otherwise specified.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
Summary statistics of the meta-analysis generated in this study have been deposited in the NHGRI-EBI GWAS catalog under accession code GCST90310293. The RNA sequencing data from the 500 human aortic valves generated in this study have been deposited in dbGaP under accession code phs003541.v1.p1. The data are available under restricted access in accordance with the institutional ethics approval. Access can be obtained for research related to cardiovascular diseases by not-for-profit organizations providing a local institutional review board approval and a letter of collaboration with the study investigators. Requests can be made to the corresponding author who will respond within two weeks. The GWAS summary statistics from FinnGen are available here: https://www.finngen.fi/en/access_results. The summary statistics of the GWAS meta-analysis for CAVS by Chen et al.14 used in this study are available here: https://zenodo.org/records/7829401. The GWAS summary statistics for CAVS from the deCODE cohort used in this study are available here: https://www.decode.com/summarydata/. The GTEx project v8 data used in this study are available here: https://gtexportal.org/home/datasets. Individual-level and genotype data from QUEBEC-CAVS-1 and QUEBEC-CAVS-2 are available under restricted access for legal and ethical reasons. Requests can be made to the corresponding author who will respond within two weeks. Access to individual data from CARTaGENE (https://cartagene.qc.ca/), EPIC-Norfolk (https://www.epic-norfolk.org.uk/), Estonian Biobank (https://genomics.ut.ee/en/content/estonian-biobank), UK Biobank (https://www.ukbiobank.ac.uk/) and FinnGen (https://www.finngen.fi/) is available for registered researchers following the respective application process. Further information on data access is available from the study websites. The drug interaction data used in this study are available from the drug-gene interaction database (www.dgidb.org).
References
Lindman, B. R. et al. Calcific aortic stenosis. Nat. Rev. Dis. Prim. 2, 16006 (2016).
Nkomo, V. T. et al. Burden of valvular heart diseases: a population-based study. Lancet 368, 1005–1011 (2006).
Otto, C. M. & Prendergast, B. Aortic-valve stenosis–from patients at risk to severe valve obstruction. N. Engl. J. Med. 371, 744–756 (2014).
Moncla, L. H. M., Briend, M., Bossé, Y. & Mathieu, P. Calcific aortic valve disease: mechanisms, prevention and treatment. Nat. Rev. Cardiol. 20, 546–559 (2023).
Bossé, Y., Mathieu, P. & Thériault, S. PALMD as a novel target for calcific aortic valve stenosis. Curr. Opin. Cardiol. 34, 105–111 (2019).
Chignon, A. et al. Enhancer-associated aortic valve stenosis risk locus 1p21.2 alters NFATC2 binding site and promotes fibrogenesis. iScience 24, 102241 (2021).
Thanassoulis, G. et al. Genetic associations with valvular calcification and aortic stenosis. N. Engl. J. Med. 368, 503–512 (2013).
Thériault, S. et al. A transcriptome-wide association study identifies PALMD as a susceptibility gene for calcific aortic valve stenosis. Nat. Commun. 9, 988 (2018).
Thériault, S. et al. Genetic association analyses highlight IL6, ALPL, and NAV1 as three new susceptibility genes underlying calcific aortic valve stenosis. Circ. Genom. Precis. Med. 12, 002617 (2019).
Li, Z. et al. Phenome-wide analyses establish a specific association between aortic valve PALMD expression and calcific aortic valve stenosis. Commun. Biol. 3, 477 (2020).
Helgadottir, A. et al. Genome-wide analysis yields new loci associating with aortic valve stenosis. Nat. Commun. 9, 987 (2018).
Chen, H. Y. et al. Association of FADS1/2 Locus Variants and Polyunsaturated Fatty Acids With Aortic Stenosis. JAMA Cardiol. 5, 694–702 (2020).
Small, A. M. et al. Multiancestry Genome-Wide Association Study of Aortic Stenosis Identifies Multiple Novel Loci in the Million Veteran Program. Circulation 147, 942–955 (2023).
Chen, H. Y. et al. Dyslipidemia, inflammation, calcification, and adiposity in aortic stenosis: a genome-wide study. Eur. Heart J. 44, 1927–1939 (2023).
Nieto, M. A., Huang, R. Y. J., Jackson, R. A. & Thiery, J. P. EMT: 2016. Cell 166, 21–45 (2016).
Tomaru, Y. et al. A transient disruption of fibroblastic transcriptional regulatory network facilitates trans-differentiation. Nucleic Acids Res. 42, 8905–8913 (2014).
Yeo, S. Y. et al. A positive feedback loop bi-stably activates fibroblasts. Nat. Commun. 9, 3016 (2018).
Ocaña, O. H. et al. A right-handed signalling pathway drives heart looping in vertebrates. Nature 549, 86–90 (2017).
Lee, K. W. et al. PRRX1 is a master transcription factor of stromal fibroblasts for myofibroblastic lineage progression. Nat. Commun. 13, 2793 (2022).
Driscoll, K., Cruz, A. D. & Butcher, J. T. Inflammatory and Biomechanical Drivers of Endothelial-Interstitial Interactions in Calcific Aortic Valve Disease. Circ. Res. 128, 1344–1370 (2021).
Lim, J. & Thiery, J. P. Epithelial-mesenchymal transitions: insights from development. Development 139, 3471–3486 (2012).
Chakraborty, S., Cheek, J., Sakthivel, B., Aronow, B. J. & Yutzey, K. E. Shared gene expression profiles in developing heart valves and osteoblast progenitor cells. Physiol. Genomics 35, 75–85 (2008).
Chakraborty, S. et al. Twist1 promotes heart valve cell proliferation and extracellular matrix gene expression during development in vivo and is expressed in human diseased aortic valves. Dev. Biol. 347, 167–179 (2010).
Hamouda, N. N. et al. ATP13A3 is a major component of the enigmatic mammalian polyamine transport system. J. Biol. Chem. 296, 100182 (2021).
Gräf, S. et al. Identification of rare sequence variation underlying heritable pulmonary arterial hypertension. Nat. Commun. 9, 1416 (2018).
Evangelou, E. et al. Genetic analysis of over 1 million people identifies 535 new loci associated with blood pressure traits. Nat. Genet. 50, 1412–1425 (2018).
Young, W. J. et al. Genetic analyses of the electrocardiographic QT interval and its components identify additional loci and pathways. Nat. Commun. 13, 5144 (2022).
Thome, M. CARMA1, BCL-10 and MALT1 in lymphocyte development and activation. Nat. Rev. Immunol. 4, 348–359 (2004).
Al-Tamemi, S. et al. BCL10 loss-of-function novel mutation leading to atypical severe combined immunodeficiency. Clin. Immunol. 241, 109067 (2022).
Hwang, B. J. et al. DNA binding by the Rad9A subunit of the Rad9-Rad1-Hus1 complex. PLoS One 17, e0272645 (2022).
Karamanavi, E. et al. The FES Gene at the 15q26 Coronary-Artery-Disease Locus Inhibits Atherosclerosis. Circ. Res. 131, 1004–1017 (2022).
Qian, Y., Baisden, J. M., Zot, H. G., Van Winkle, W. B. & Flynn, D. C. The carboxy terminus of AFAP-110 modulates direct interactions with actin filaments and regulates its ability to alter actin filament integrity and induce lamellipodia formation. Exp. Cell Res. 255, 102–113 (2000).
Pirruccello, J. P. et al. Deep learning enables genetic analysis of the human thoracic aorta. Nat. Genet. 54, 40–51 (2022).
Gehlen, J. et al. Elucidation of the genetic causes of bicuspid aortic valve disease. Cardiovasc. Res. 119, 857–866 (2023).
Beecroft, S. J. et al. Biallelic hypomorphic variants in ALDH1A2 cause a novel lethal human multiple congenital anomaly syndrome encompassing diaphragmatic, pulmonary, and cardiovascular defects. Hum. Mutat. 42, 506–519 (2021).
Leon, E., Nde, C., Ray, R. S., Preciado, D. & Zohn, I. E. ALDH1A2-related disorder: A new genetic syndrome due to alteration of the retinoic acid pathway. Am. J. Med. Genet. A 191, 90–99 (2023).
Tsukamoto, S. et al. Smad9 is a new type of transcriptional regulator in bone morphogenetic protein signaling. Sci. Rep. 4, 7596 (2014).
Huang, Z., Wang, D., Ihida-Stansbury, K., Jones, P. L. & Martin, J. F. Defective pulmonary vascular remodeling in Smad8 mutant mice. Hum. Mol. Genet. 18, 2791–2801 (2009).
Gregson, C. L. et al. A Rare Mutation in SMAD9 Associated With High Bone Mass Identifies the SMAD-Dependent BMP Signaling Pathway as a Potential Anabolic Target for Osteoporosis. J. Bone Min. Res. 35, 92–105 (2020).
Mundal, L. J. et al. Association of Low-Density Lipoprotein Cholesterol With Risk of Aortic Valve Stenosis in Familial Hypercholesterolemia. JAMA Cardiol. 4, 1156–1159 (2019).
Rip, J. et al. Lipoprotein lipase S447X: a naturally occurring gain-of-function mutation. Arterioscler Thromb. Vasc. Biol. 26, 1236–1245 (2006).
Sagoo, G. S. et al. Seven lipoprotein lipase gene polymorphisms, lipid fractions, and coronary disease: a HuGE association review and meta-analysis. Am. J. Epidemiol. 168, 1233–1246 (2008).
Nazarzadeh, M. et al. Plasma lipids and risk of aortic valve stenosis: a Mendelian randomization study. Eur. Heart J. 41, 3913–3920 (2020).
Kaltoft, M., Langsted, A. & Nordestgaard, B. G. Triglycerides and remnant cholesterol associated with risk of aortic valve stenosis: Mendelian randomization in the Copenhagen General Population Study. Eur. Heart J. 41, 2288–2299 (2020).
Dart, A. M. & Kingwell, B. A. Pulse pressure–a review of mechanisms and clinical relevance. J. Am. Coll. Cardiol. 37, 975–984 (2001).
Ercu, M. et al. Phosphodiesterase 3A and Arterial Hypertension. Circulation 142, 133–149 (2020).
Maass, P. G. et al. PDE3A mutations cause autosomal dominant hypertension with brachydactyly. Nat. Genet. 47, 647–653 (2015).
Willer, C. J., Li, Y. & Abecasis, G. R. METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics 26, 2190–2191 (2010).
Wang, K., Li, M. & Hakonarson, H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 38, e164 (2010).
Hormozdiari, F., Kostem, E., Kang, E. Y., Pasaniuc, B. & Eskin, E. Identifying causal variants at loci with multiple signals of association. Genetics 198, 497–508 (2014).
Yang, J. et al. Conditional and joint multiple-SNP analysis of GWAS summary statistics identifies additional variants influencing complex traits. Nat. Genet. 44, 369–375 (2012).
de Leeuw, C. A., Mooij, J. M., Heskes, T. & Posthuma, D. MAGMA: generalized gene-set analysis of GWAS data. PLoS Comput. Biol. 11, e1004219 (2015).
Otto, C. M. et al. 2020 ACC/AHA guideline for the management of patients with valvular heart disease: A report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. J. Thorac. Cardiovasc. Surg. 162, e183–e353 (2021).
Ewels, P., Magnusson, M., Lundin, S. & Käller, M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics 32, 3047–3048 (2016).
Okonechnikov, K., Conesa, A. & García-Alcalde, F. Qualimap 2: advanced multi-sample quality control for high-throughput sequencing data. Bioinformatics 32, 292–294 (2016).
The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369, 1318–1330 (2020).
DeLuca, D. S. et al. RNA-SeQC: RNA-seq metrics for quality control and process optimization. Bioinformatics 28, 1530–1532 (2012).
Torres, J. M. et al. A Multi-omic Integrative Scheme Characterizes Tissues of Action at Loci Associated with Type 2 Diabetes. Am. J. Hum. Genet. 107, 1011–1028 (2020).
Fort, A. et al. MBV: a method to solve sample mislabeling and detect technical bias in large combined genotype and sequencing assay datasets. Bioinformatics 33, 1895–1897 (2017).
Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010).
Stegle, O., Parts, L., Durbin, R. & Winn, J. A Bayesian framework to account for complex non-genetic factors in gene expression levels greatly increases power in eQTL studies. PLoS Comput. Biol. 6, e1000770 (2010).
Delaneau, O. et al. A complete tool set for molecular QTL discovery and analysis. Nat. Commun. 8, 15452 (2017).
Taylor-Weiner, A. et al. Scaling computational genomics to millions of individuals with GPUs. Genome Biol. 20, 228 (2019).
Barbeira, A. N. et al. Exploring the phenotypic consequences of tissue specific gene expression variation inferred from GWAS summary statistics. Nat. Commun. 9, 1825 (2018).
Gamazon, E. R. et al. A gene-based association method for mapping traits using reference transcriptome data. Nat. Genet. 47, 1091–1098 (2015).
Giambartolomei, C. et al. Bayesian test for colocalisation between pairs of genetic association studies using summary statistics. PLoS Genet. 10, e1004383 (2014).
Liu, B., Gloudemans, M. J., Rao, A. S., Ingelsson, E. & Montgomery, S. B. Abundant associations with gene expression complicate GWAS follow-up. Nat. Genet. 51, 768–769 (2019).
Chang, C. C. et al. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience 4, 7 (2015).
Bowden, J. et al. Improving the accuracy of two-sample summary-data Mendelian randomization: moving beyond the NOME assumption. Int J. Epidemiol. 48, 728–742 (2019).
Bowden, J., Davey Smith, G. & Burgess, S. Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int. J. Epidemiol. 44, 512–525 (2015).
Bowden, J., Davey Smith, G., Haycock, P. C. & Burgess, S. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genet. Epidemiol. 40, 304–314 (2016).
Finan, C. et al. The druggable genome and support for target identification and validation in drug development. Sci. Transl. Med. 9, eaag1166 (2017).
Cotto, K. C. et al. DGIdb 3.0: a redesign and expansion of the drug-gene interaction database. Nucleic Acids Res. 46, D1068–D1073 (2018).
Zhou, Y. et al. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat. Commun. 10, 1523 (2019).
Marchini, J., Howie, B., Myers, S., McVean, G. & Donnelly, P. A new multipoint method for genome-wide association studies by imputation of genotypes. Nat. Genet. 39, 906–913 (2007).
Wang, L. et al. An atlas connecting shared genetic architecture of human diseases and molecular phenotypes provides insight into COVID-19 susceptibility. Genome Med. 13, 83 (2021).
Bulik-Sullivan, B. et al. An atlas of genetic correlations across human diseases and traits. Nat. Genet. 47, 1236–1241 (2015).
Acknowledgements
QUEBEC-CAVS was supported by the Heart and Stroke Foundation of Canada (G-19-0026386), the Canadian Institutes of Health Research (PJT – 153396, PJT – 162344) and the Fonds de recherche du Québec – Santé (278277). Y.B. holds a Canada Research Chair in Genomics of Heart and Lung Diseases. The EPIC-Norfolk study has received funding from the Medical Research Council (MR/N003284/1 MC-UU_12015/1 and MC_UU_00006/1) and Cancer Research UK (C864/A14136). The genetics work in the EPIC-Norfolk study was funded by the Medical Research Council (MC_PC_13048). The work of Estonian Genome Center, Univ. of Tartu was funded by the European Union through Horizon 2020 research and innovation programme under grants no. 810645 and 894987, through the European Regional Development Fund projects GENTRANSMED (2014-2020.4.01.15-0012), MOBERA5 (Norface Network project no 462.16.107), MOBEC008, MOBERA21 and Estonian Research Council Grants PUT1660 and PRG1291. We thank the research team at the cardiac surgical database and biobank of the Institut universitaire de cardiologie et de pneumologie de Québec-Université Laval for their valuable assistance and all the participants. We thank the CARTaGENE study team and participants. We are grateful to all the participants who have been part of the EPIC-Norfolk study and to the many members of the study teams at the University of Cambridge who have enabled this research. We want to acknowledge the participants and investigators of Estonian biobank for their contribution. This work was carried out in part in the High Performance Computing Center of the University of Tartu. We thank all the UK Biobank participants and administration team. We want to acknowledge the participants and investigators of the FinnGen study.
Author information
Authors and Affiliations
Consortia
Contributions
S.T., S.M.B., N.J.W., T.E. and Y.B. supervised the cohort studies. H.D.M., D.K.B., N.G., L.F., D.A., M.D., F.D., M.A.C. and P.P. contributed to data collection. S.T., Z.L., E.A. and J.L. performed the GWAS. S.T. and Z.L. performed the meta-analysis, eQTL, TWAS, colocalization, MR, pheWAS and genetic correlation analyses. S.T., U.H. and P.Z. performed the differential expression analyses. M.B. performed the iCPAG analysis. S.T., B.J.A., P.M. and Y.B. designed the study. S.T. drafted the manuscript. P.M. and Y.B. edited the manuscript. All authors revised the manuscript prior to submission.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
Nature Communications thanks Nirmal Vadgama and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Thériault, S., Li, Z., Abner, E. et al. Integrative genomic analyses identify candidate causal genes for calcific aortic valve stenosis involving tissue-specific regulation. Nat Commun 15, 2407 (2024). https://doi.org/10.1038/s41467-024-46639-4
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41467-024-46639-4
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.