Genetic Association Analyses Highlight IL6, ALPL, and NAV1 As 3 New Susceptibility Genes Underlying Calcific Aortic Valve Stenosis
Abstract
Background:
Calcific aortic valve stenosis (CAVS) is a frequent and life-threatening cardiovascular disease for which there is currently no medical treatment available. To date, only 2 genes, LPA and PALMD, have been identified as causal for CAVS. We aimed to identify additional susceptibility genes for CAVS.
Methods:
A GWAS (genome-wide association study) meta-analysis of 4 cohorts, totaling 5115 cases and 354 072 controls of European descent, was performed. A TWAS (transcriptome-wide association study) was completed to integrate transcriptomic data from 233 human aortic valves. A series of post-GWAS analyses were performed, including fine-mapping, colocalization, phenome-wide association studies, pathway, and tissue enrichment as well as genetic correlation with cardiovascular traits.
Results:
In the GWAS meta-analysis, 4 loci achieved genome-wide significance, including 2 new loci: IL6 (interleukin 6) on 7p15.3 and ALPL (alkaline phosphatase) on 1p36.12. A TWAS integrating gene expression from 233 human aortic valves identified NAV1 (neuron navigator 1) on 1q32.1 as a new candidate causal gene. The CAVS risk alleles were associated with higher mRNA expression of NAV1 in valve tissues. Fine-mapping identified rs1800795 as the most likely causal variant in the IL6 locus. The signal identified colocalizes with the expression of the IL6 RNA antisense in various tissues. Phenome-wide association analyses in the UK Biobank showed colocalized associations between the risk allele at the IL6 lead variant and higher eosinophil count, pulse pressure, systolic blood pressure, and carotid artery procedures, implicating modulation of the IL6 pathways. The risk allele at the NAV1 lead variant colocalized with higher pulse pressure and higher prevalence of carotid artery stenosis. Association results at the genome-wide scale indicated genetic correlation between CAVS, coronary artery disease, and cardiovascular risk factors.
Conclusions:
Our study implicates 3 new genetic loci in CAVS pathogenesis, which constitute novel targets for the development of therapeutic agents.
Introduction
Calcific aortic valve stenosis (CAVS) is characterized by progressive thickening and calcification of the aortic valve leaflets causing blood flow obstruction, heart failure, and death.1 Symptomatic patients with CAVS are treated by surgical aortic valve replacement. CAVS is the second most frequent indication for cardiac surgery after coronary artery bypass grafting.2 No medical treatment is available, and conventional cardiovascular drugs tested in clinical trials of CAVS have failed to slow the progression of the disease and reduce associated adverse events.1,3–5 The pathophysiological mechanisms associated with the development of CAVS remain elusive. Previous epidemiological studies on CAVS described familial patterns compatible with genetic inheritance and indicated that genetics contribute more than environmental factors to disease susceptibility.6,7 Unbiased genomic approaches are just starting to elucidate the genetic determinants of CAVS.8–10 To date, only 3 susceptibility loci have been associated with aortic valve calcification and stenosis in large genomics studies.8–10
The first GWAS (genome-wide association study) on aortic valve calcification identified LPA,9 which encodes for lipoprotein(a). We recently reported a TWAS (transcriptome-wide association study) that led to the identification of PALMD as a candidate target gene for CAVS.8 Another group simultaneously reported associations at the PALMD and TEX41 loci with aortic valve stenosis and bicuspid aortic valve.10 The discovery of new loci to understand the genetic architecture of CAVS will provide invaluable information about key molecular processes and could open novel therapeutic avenues.
To identify additional susceptibility genes for CAVS, we undertook a large GWAS meta-analysis. We then leveraged our aortic valve gene expression dataset to perform a TWAS. We characterized the novel loci identified to investigate the pathophysiological mechanisms involved and evaluated the genetic relationships between CAVS and other cardiovascular diseases.
Methods
A fixed-effects meta-analysis of the GWAS results from the 4 cohorts was performed. Our aortic valve expression quantitative trait loci (eQTL) dataset, which includes 233 individuals, was then used as reference to perform a TWAS. Analyses stratified by sex or aortic valve morphology were performed to further characterize the association of the lead single nucleotide polymorphism (SNP) in each significant locus and CAVS. Fine-mapping, colocalization, and pathway analyses were performed to identify the genes and variants most likely to be causal and the biological mechanisms involved. Association between the lead variant in the newly identified loci and a wide range of phenotypes was verified in the UK Biobank. LD-score regression was used to estimate SNP heritability of CAVS and the genetic correlation with other cardiovascular traits. The complete methods are available in the Data Supplement. The data that support the findings of this study are available from the corresponding authors on reasonable request. The study was approved by the local ethics committees and all patients signed an informed consent for the realization of genetic studies.
Results
GWAS Meta-Analysis
A total of 5115 cases and 354 072 controls of European ancestry were included after quality control filters were applied. Clinical characteristics of the participants for each cohort are available in Table I in the Data Supplement.
The genomic inflation factors in the 4 cohorts were 1.029 (QUEBEC-CAVS), 1.031 (CAVS-France-1), 1.050 (CAVS-France-2), and 1.021 (UK Biobank). A fixed-effects meta-analysis was performed using the summary statistics for the association between ≈8 million SNPs and CAVS risk in each cohort. The quantile-quantile plot indicated no inflation of observed test statistics (Figure I in the Data Supplement). Four loci reached genome-wide significance (PGWAS<5×10−8), including 2 previously described loci, namely LPA on 6q25.3-q26 and PALMD on 1p21.28,9 (Figure 1), and 2 new loci: the interleukin 6 (IL6) gene on 7p15.3 (lead SNP: rs12141569, intronic, P=1.1×10−8) and the alkaline phosphatase (ALPL) gene (lead SNP: rs12141569, intronic, P=3.9×10−8) on 1p36.12 (Figures 1 and 2). The lead SNPs for all loci with PGWAS <1×10−5 are indicated in Table II in the Data Supplement. We observed a similar effect for rs1830321 (located near the TEX41 gene) compared with what was previously reported (OR, 1.12 [95% CI, 1.06–1.18], P=1.57×10−5).
Figure 1. Manhattan plot of the GWAS (genome-wide association study) meta-analysis of 4 cohorts including 5115 cases and 354 072 controls. The y axis represents P in −log10 scale. The horizontal red and blue lines indicate P of 5×10−8 and 1×10−5, respectively. Gene names in black and red correspond to previously known and novel loci for calcific aortic valve stenosis, respectively.
Figure 2. Two new calcific aortic valve stenosis (CAVS)-associated loci.A and B, Regional plots showing the 2 new loci. The y axis shows the P in −log10 scale for SNPs 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. SNPs are plotted based on their chromosomal position on build hg19. C and D, Forest plots showing the effect size of the top CAVS-associated SNP at the 2 new loci in each cohort and meta-analysis. The blue filled squares represent the odds ratio (OR) for each cohort. The horizontal lines represent the 95% CIs of the OR. The gray and the dashed magenta vertical lines represent an OR of 1.0 and the OR of the meta-analysis, respectively.
TWAS Using Gene Expression in the Aortic Valve
We have then integrated the summary statistics from the 4-cohort meta-analysis with our valve eQTL dataset (n=233) to perform a new TWAS on CAVS (Table III in the Data Supplement). PALMD, identified in our previous TWAS,8 was strongly associated with CAVS (PTWAS=3.6×10−17; Figure 3A). One additional gene reached the significance threshold (PTWAS<4.2×10−6), namely neuron navigator 1 (NAV1; PTWAS=2.0×10−6). NAV1 is located on chromosome 1q32.1; the lead GWAS SNP is rs665770, located in intron 3 (PGWAS=1.5×10−6; Figure II in the Data Supplement and Figure 3B). It is also the lead valve eQTL-SNP (PeQTL=5.9×10−11). The risk allele for CAVS (A) is associated with higher mRNA expression levels of NAV1 in valve tissues (Figure 3C). Moreover, Bayesian colocalization test11 indicated a high probability of shared GWAS and NAV1 eQTL signals (posterior probability of shared associations [PP4]=98.3%). Finally, the CAVS risk alleles were consistently associated with increased expression of NAV1 in aortic valve tissues (Figure 3D). The 2 other CAVS risk loci discovered in this study (IL6 and ALPL) did not have significant cis-heritability (part of expression variability explained by SNPs) in our aortic valve dataset and therefore could not be evaluated by TWAS.
Figure 3. Identification of NAV1 by TWAS (transcriptome-wide association study) and causality analysis.A, TWAS in valve tissue with calcific aortic valve stenosis (CAVS). P for gene expression-CAVS associations are on the y axis in −log10 scale. The blue and magenta horizontal lines represent PTWAS of 0.0001 and 4.23×10−6 (Bonferroni), respectively. Each dot represents the association between predicted gene expression and CAVS for a specific probe/transcript. Annotations for the top significant probes/transcripts are indicated. B, GWAS (genome-wide association study) and valve expression quantitative trait loci (eQTL) results surrounding NAV1 on chromosome 1q32.1. The top shows the genetic associations with CAVS. The bottom shows the valve eQTL statistics for the NAV1 gene. The extent of linkage disequilibrium (LD; r2 values) for all SNPs with rs665770 is indicated by colors. The location of genes in this locus is illustrated at the bottom. C, Boxplots of gene expression levels in the valves according to the 3 genotype groups for SNP rs665770. The y axis shows the mRNA expression levels. The x axis represents the 3 genotype groups for rs665770 with the number of individuals in parenthesis. Expression data for the NAV1 probe were not available for 3 individuals (n=230). The risk allele identified in our GWAS meta-analysis is shown in red. Box boundaries, whiskers, and center marks in boxplots represent the first and third quartiles, the most extreme data point which is no more than 1.5× the interquartile range, and the median, respectively. The black bar and the left y axis indicate the variance in NAV1 gene expression explained by rs665770. D, Scatterplot of the 1q32.1 susceptibility locus showing SNP associations with CAVS and NAV1 gene expression in aortic valve tissues. The y axis represents variant association with CAVS (Z score). The x axis shows association with NAV1 gene expression (t statistic). Variants are colored based on the degree of LD (r2) with the top CAVS-associated variant rs665770. The blue line is the regression slope with 95% CI (red lines).
Subgroup Analyses According to Sex and Valve Morphology
Among the lead SNPs in the 5 significant loci, significant heterogeneity was observed for PALMD and NAV1 in sex-stratified analyses. Although the direction of effect was consistent across sex, the effect was stronger in women for both loci (Table IV in the Data Supplement ). Association of the lead SNPs in the 5 significant loci with CAVS was determined according to aortic valve morphology (bicuspid or tricuspid) in the 3 cohorts for which this information was available (QUEBEC-CAVS, CAVS-France-1, and CAVS-France-2). Rs665770 in NAV1 was only associated with tricuspid CAVS (PHet=0.0097; Table 1). The other loci showed overlapping effect sizes for bicuspid and tricuspid CAVS. The association of the 5 loci with tricuspid CAVS was very similar to the results of the meta-analysis including all CAVS cases from the 4 cohorts.
| SNP | Nearest Gene | Bicuspid CAVS (n=422) | Tricuspid CAVS (n=2557) | Het | All CAVS (n=5115) | |||
|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | P Value | OR (95% CI) | P Value | P Value | OR (95% CI) | P Value | ||
| rs10455872-G | LPA | 1.25 (0.76–2.04) | 0.38 | 1.49 (1.29–1.72) | 9.78×10−6 | 0.50 | 1.54 (1.40–1.68) | 6.48×10−21 |
| rs7543039-T | PALMD | 1.29 (1.10–1.51) | 0.0014 | 1.20 (1.10–1.30) | 1.89×10−6 | 0.40 | 1.24 (1.18–1.30) | 1.24×10−16 |
| rs2069832-A | IL6 | 1.24 (1.05–1.46) | 0.0094 | 1.27 (1.19–1.36) | 6.71×10−10 | 0.78 | 1.16 (1.10–1.23) | 1.11×10−8 |
| rs12141569-C | ALPL | 1.14 (0.98–1.33) | 0.095 | 1.15 (1.07–1.23) | 0.00036 | 0.97 | 1.15 (1.10–1.21) | 3.93×10−8 |
| rs665770-A | NAV1 | 0.98 (0.84–1.15) | 0.79 | 1.22 (1.15–1.30) | 1.22×10−7 | 0.0097 | 1.13 (1.08–1.19) | 1.46×10−6 |
Identification of Causal Variants
Fine-mapping to localize the potential causal variants at CAVS risk loci was performed using PAINTOR12 with the following annotations: conservation score, fetal DNase I hypersensitive sites (DHS), promoter region, and H3K4me1 peaks. For our previously reported PALMD locus, the lead GWAS SNP, rs6702619, provides a posterior probability for causality (PPC) of 0.98. For the IL6 locus, rs1800795 achieved a similarly strong PPC of 0.96. This SNP is located <1 kb from the lead SNP rs2069832 (LD r2=0.97 in 1000 Genomes Europeans). Using the DeepSEA13 algorithm, rs1800795 had a strong probability of effect on chromatin features in different cell lines, including human cardiac fibroblasts and myocytes (DNase I hypersensitive sites) and osteoblasts (histone marks) (Table V in the Data Supplement). For the ALPL locus, the index SNP, rs12141569, has the top PPC of 0.55, but a second plausible functional variant was identified (rs7547128 with a PPC of 0.21). Finally, 3 SNPs were prioritized for the NAV1 locus, including rs7535989, rs665770 (GWAS lead SNP), and rs665834 with PPCs of 0.54, 0.15, and 0.06, respectively.
eQTL in Genotype-Tissue Expression
CAVS-associated variants were evaluated for eQTL in the Genotype-Tissue Expression (GTEx) data set.14 The IL6 intronic variant rs2069832 is a strong eQTL in many tissues (adipose tissues, arteries, breast, esophagus, heart, lung, and others) for the IL6 RNA antisense (LOC541472; Figure III in the Data Supplement). Higher expression corresponds to higher risk of CAVS, with strong probability for colocalization (PP4>96%). Other eQTLs of note in GTEx include rs12141569-ALPL in fibroblasts (PeQTL=1.1×10−7) and rs665770-NAV1 in heart atrial appendage (PeQTL=1.9×10−8).
Phenome-Wide Association Study
The lead SNPs at the 3 new loci were further investigated using a PheWAS (phenome-wide association study) in 353 378 individuals of European ancestry from the UK Biobank, testing a total of 834 curated phenotypes (Tables VI through VIII in the Data Supplement). The IL6 variant rs2069832 was strongly associated with eosinophil count (P=7.9×10−24) and to a lower extent with leukocyte, lymphocyte, and neutrophil counts, pulse pressure, systolic blood pressure, asthma, and carotid artery procedures (Figure 4A). The signals at the CAVS risk locus colocalized with higher eosinophil (PP4=96.8%) and leukocyte counts (PP4=89.4%), pulse pressure (PP4=97.2%), systolic blood pressure (PP4=93.1%), and carotid artery procedures (PP4=97%); the association with lower risk of asthma had a weaker probability for colocalization (PP4=37.7% and 50.1% for raw and coded questionnaires, respectively). For ALPL, no phenotype was associated with rs12141569 after correction for multiple testing, but a positive nominal association was observed between the CAVS risk allele and bone mineral density (P=4.7×10−4, Figure 4B). For NAV1, the CAVS risk allele for rs665770 was associated with higher pulse pressure (P=9.2×10−12), lower diastolic blood pressure (P=6.5×10−10), as well as higher risk of carotid artery stenosis (P=7.4×10−8) and carotid artery procedures (P=1.7×10−5) (Figure 4C). GWAS signals at 1q32.1-NAV1 colocalized between our CAVS meta-analysis and pulse pressure (PP4=98.4%), diastolic blood pressure (PP4=87.1%), carotid stenosis (PP4=98.5%), and carotid procedures (PP4=97%) in the UK Biobank (Figure IV in the Data Supplement).
Figure 4. PheWAS (phenome-wide association study) for the 3 new calcific aortic valve stenosis (CAVS) risk loci in the UK Biobank.A, rs2069832, lead SNP at the IL6 locus. B, rs12141569, lead SNP at the ALPL locus. C, rs665770, lead SNP at the NAV1 locus. Each triangle represents a different phenotype (n=834). Triangles pointing up and down are positive and negative associations with CAVS risk alleles, respectively. The pink horizontal line represents the threshold for significance after correcting for multiple testing (P=0.05/834=6.0×10−5). The blue horizontal line represents the threshold for nominal significance (P=0.05).
Tissue, Gene-Set, and Pathways Enrichment
Tissue and gene-set enrichment were performed with DEPICT15 using SNPs nominally associated with CAVS (PGWAS<1×10−5). At the tissue level, enrichments with nominal PDEPICT<0.05 were observed for arteries, veins, and blood vessels, but none were significant at a false discovery rate <5%. Similarly, no gene sets were significantly enriched at false discovery rate <5%. Top nominal pathways of interest included increased compact bone thickness, increased heart weight, disorganized myocardium, and increased systemic arterial diastolic blood pressure (Table IX in the Data Supplement). We also tested enrichment of CAVS-associated loci (PGWAS<1×10−5) in DNase I hypersensitive sites in cell lines and tissues from ENCODE and the Roadmap Epigenomics Project using FORGE16 and GARFIELD.17 Leading tissues were heart, blood vessels, and fibroblast, but no overlap reached statistical significance (Figure V in the Data Supplement).
Heritability and Genetic correlation
We performed LD score regression18 to estimate the proportion of phenotypic variance explained by SNPs (SNP heritability) for CAVS and the genetic correlation with other cardiovascular traits, using summary statistics from GWAS meta-analyses obtained from publicly available resources (Table X in the Data Supplement). SNP heritability was estimated at 10.0% using LD score regression. There was significant genetic correlation between CAVS and coronary artery disease (CAD), myocardial infarction, diastolic and systolic blood pressure, body mass index, type 2 diabetes mellitus, and large-artery stroke (Table 2).
| Traits/Diseases | Correlation (Rg) | P Value |
|---|---|---|
| Coronary artery disease | 0.23 | 0.0001 |
| Myocardial infarction | 0.26 | 0.0004 |
| Systolic blood pressure | 0.15 | 7.74×10−8 |
| Diastolic blood pressure | 0.12 | 1.90×10−5 |
| BMI | 0.21 | 6.41×10−6 |
| WHR | 0.20 | 0.0010 |
| WHR adjusted for BMI | 0.08 | 0.1798 |
| Type 2 diabetes mellitus | 0.18 | 6.08×10−6 |
| Total cholesterol | 0.12 | 0.0247 |
| LDL-cholesterol | 0.14 | 0.0125 |
| HDL-cholesterol | −0.14 | 0.0073 |
| Triglyceride | 0.16 | 0.0070 |
| Any ischemic stroke | 0.16 | 0.0011 |
| Large-artery stroke | 0.20 | 2.00×10−5 |
| Cardioembolic stroke | 0.15 | 0.0024 |
| Small-vessel stroke | 0.07 | 0.2159 |
| Chronic kidney disease | 0.10 | 0.1727 |
Discussion
We report the discovery of 3 new CAVS risk loci: IL6, ALPL, and NAV1. The CAVS risk alleles at the NAV1 locus are associated with higher expression of NAV1 in the aortic valve, with a strong probability for colocalization between the GWAS and eQTL signals. For each locus, we identify candidate causal variants with fine-mapping and show relevant associations with other health conditions using a PheWAS approach. At the genome-wide scale, we estimate the proportion of phenotypic variance explained by SNPs at 10% and demonstrate significant genetic correlation between CAVS and multiple cardiovascular traits and diseases.
Although the exact pathophysiological mechanisms remain to be elucidated, the 3 identified genes can be functionally linked to CAVS. Interleukin 6 has been identified previously as a strong promoter of valve interstitial cells mineralization.19 The SNP rs1800795 identified as the most likely causal variant is located in the promoter of IL6. There is conflicting evidence in the literature about the impact of rs1800795 on IL6 transcription, with data suggesting haplotype or tissue-dependent effects.20,21 The RNA antisense upregulated in various tissues by the CAVS risk allele C is strongly upregulated in monocytes during fungal infection.22 It was previously shown to induce IL6 expression in glioma cells by causing the enrichment of histone H3 acetylated at lysine 27 (H3K27Ac) of the IL6 promoter.23 The alternate allele G has been associated with exceptional longevity.24 Moreover, our PheWAS analysis suggests that the identified locus in the IL6 gene has impacts on several health conditions recognized as consequences of the modulation of IL6 signaling pathways. Genetic variants in IL6R, which participates to IL6 signaling, have been mapped to CAD and atopic disorders.25–29 Finally, this locus was recently associated with pulse pressure in a recent GWAS,30 a signal that colocalized with CAVS risk in our analysis using the UK Biobank. Taken together, plausible mechanisms include mineralization and blood pressure modulation, but further work is needed to evaluate the specific role of this locus in CAVS pathogenesis.
The ALPL gene codes for tissue nonspecific alkaline phosphatase (ALPL, also called TNAP), a crucial enzyme involved in the mineralization process,31 which was found to be enriched in calcified aortic valves as compared with noncalcified valves.32 Although our lead variant was not associated with blood alkaline phosphatase levels in previous GWAS,33–35 there is evidence to support a local effect on ALPL activity. In fact, our PheWAS analysis shows a positive effect of the CAVS risk allele on bone mineral density, which was confirmed in a recent GWAS (Z score=5.8, P=9.9×10−9).36
NAV1 was first identified as a human homolog of the Caenorhabditis elegans unc-53 gene. The UNC-53 protein contains 2 putative actin-binding domains and has been shown to co-sediment with actin.37 NAV1 is strongly expressed in the aorta (GTEx Portal) and a short transcript seems to be specific to the human heart.37 The identified locus is a significant eQTL in aortic valve tissue, which colocalizes with CAVS risk and with carotid stenosis risk in the UK Biobank. The risk alleles are associated with an increase in NAV1 expression in the aortic valve and the atrial appendage, suggesting an impact on local transcription as the acting mechanism. This locus was also associated with pulse pressure and diastolic blood pressure in our PheWAS analysis and in previous GWAS.30,38 Pulse pressure and the aortic root morphology, 2 factors associated with blood flow turbulence, have been previously identified as risk factors for CAVS and its progression over time.39,40 These data suggest that this locus could have an effect on vascular tissues, possibly through an interaction with actin fibers.
The IL6 and ALPL loci identified through GWAS could not be included in the TWAS analysis since they did not show sufficient cis-heritability in the aortic valve dataset. This is consistent with the lower than expected number of TWAS genes in previously reported GWAS loci for other complex diseases.41,42 Gene regulation processes specific to other tissues could be involved, as suggested by the colocalization of the signal at the IL6 locus with expression of the IL6 RNA antisense in multiple other tissues and the significant cis-eQTL in fibroblasts for the ALPL locus. In addition, the presence of context-dependent or cell-type specific eQTLs at these loci cannot be excluded. Although the effect sizes for these 2 new GWAS loci are lower than the ones previously reported (LPA and PALMD), they still likely represent core genes with biological significance in CAVS. Of note, many of the top loci associated with CAD (CDKN2A/B, LDLR, PCSK9, PLPP3, PHACTR1, CXCL12) have been extensively studied and shown to have important biological implications despite having modest effect sizes (OR between 1.05 and 1.25 per risk allele).43–46
The same direction of effect was observed in the 4 cohorts for the 5 loci. There was some variability in the effect sizes between cohorts, which could be explained by differences in the study design and characteristics of the cohort. For example, the majority of the individuals in the control group in QUEBEC-CAVS had CAD. Cases from the UK Biobank were younger and identified using diagnosis and procedure codes. They might therefore have less severe CAVS compared with the 3 other cohorts in which cases were recruited in reference centers performing interventions for CAVS, potentially explaining the lower effect sizes at the IL6 and NAV1 loci in the UK Biobank.
The stronger effect of the lead SNPs in PALMD and NAV1 on CAVS observed in women compared with men could be explained by several factors. Men are inherently more at risk of CAVS in part because of a higher prevalence of risk factors such as hypertension, diabetes mellitus, and hypercholesterolemia. Therefore, more CAVS cases in men could be attributed to other causes whereas the genetic variants could be more crucial in women who have less other contributing factors. A different biological effect of PALMD and NAV1 expression in the aortic valve according to sex is also possible.
The 5 significant loci in the current analysis had similar association when only tricuspid CAVS cases were included, suggesting that their effect is not mainly driven by an increased risk of bicuspid aortic valve. The lead SNP at the PALMD locus showed a similar association with bicuspid and tricuspid CAVS, corroborating previous findings.10
Although not genome-wide significant, the locus located near TEX41 showed similar association than previously reported.10 There was no signal at the GATA4 locus recently associated with bicuspid CAVS47 or NOTCH1, for which rare variants were previously associated with familial forms of CAVS.48
The identification of these 3 new genetic loci for CAVS constitutes an important step in deciphering the pathogenesis of the disease and may open novel therapeutic avenues, which are currently lacking. For example, agents modulating the interleukin 6 signaling pathways, some of which are currently available (IL6R antibodies) while others are in development (IL6 antibodies, humanized sgp130Fc),49 could be considered for further evaluation as potential therapies in CAVS. The safety profile of chronic administration of these agents would however need to be determined first.
Enrichment analyses at the genome-wide scale showed nominal associations with blood vessel tissues and relevant pathways involving vascular, cardiac, and bone morphology. Limited power likely explains the absence of enrichment when using a false discovery rate threshold of 5%. Additional studies including a higher number of cases will be needed to further elucidate CAVS genetic architecture. We identified significant positive genetic correlation with CAD and stroke as well as cardiovascular disease risk factors (blood pressure, diabetes mellitus, and BMI). This is consistent with the frequent co-occurrence of CAD and CAVS and the existence of common risk loci, namely LPA9 and TEX41.10 Of note, the lead variants associated with CAVS risk in IL6 and ALPL were not associated with CAD (P>0.10) in a recent GWAS meta-analysis50 while the candidate variant in NAV1 showed association below the genome-wide significance threshold (P=2.4×10−5, same direction of effect).
Limitations
Although we propose credible mechanisms implicating the identified loci, further functional studies will have to be performed to precise the biology involved. We did not perform downstream analyses on loci that did not pass the selected significance thresholds in the GWAS or TWAS analyses. Further studies are needed to confirm the association between the loci approaching these thresholds and CAVS risk. Power was limited to perform stratified analyses based on subphenotypes such as bicuspid aortic valve (only 2 cohorts included cases characterized as bicuspid, n=422). In the French datasets, the control group consisted of population cohorts for which CAVS status was unknown. However, the prevalence of CAVS in this setting is expected to be low (around 1%) and contamination with CAVS cases would impact the effect of genetic variants towards the null. The study population consisted exclusively of individuals of European ancestry and therefore results cannot be generalized to other ethnic groups. Whether the identified loci are associated with CAVS progression remains to be evaluated.
Conclusions
In conclusion, we describe 3 new risk loci for CAVS (IL6, ALPL, and NAV1), implicating inflammation, mineralization, and blood pressure in CAVS etiology. These genes constitute novel potential therapeutic targets for future studies.
Acknowledgments
We thank the research team at the cardiac surgical database and biobank of the Institut universitaire de cardiologie et de pneumologie de Québec (IUCPQ) for their valuable assistance. This research has been conducted using the UK Biobank Resource. We thank Marie Marrec and Guénola Coste for their contribution to clinical data collection. We are most grateful to the Genomics and Bioinformatics Core Facility of Nantes (GenoBiRD, Biogenouest) for its technical support. We would like to specially thank the team of the Centre d’Investigation Clinique (Xavier Duval), the Centre de Ressources Biologiques (Sarah Tubiana), Christophe Aucan from the Assistance Publique - Hopitaux de Paris, Département de la Recherche Clinique et du Développement (DRCD), and Estelle Marcault from the Unité de Recherche Clinique Paris Nord for their help and support during all these years. The PREGO bio-collection is part of the VaCaRMe program founded by the French Regional Council of Pays-de-la-Loire (RFI VaCaRMe). We wish to thank the Fondation Genavie for its financial support. We would like to specially thank Stephanie Chatel for all her contribution to constitute this bio-collection. We thank the Etablissement Français du Sang and the Centre de Ressources Biologiques of CHU de Nantes, for its support. We thank our numerous collaborators for the recruitment of blood donors and all participating individuals.
Sources of Funding
Dr Theriault holds a Junior 1 Clinical Research Scholar award from the Fonds de Recherche du Québec-Santé (FRQS). Dr Clavel holds a Junior 1 Research Scholar award from the FRQS. Dr Arsenault holds a Junior 2 Research Scholar award from the FRQS. Dr Pibarot holds the Canada Research Chair in Valvular Heart Disease and his research program is supported by a Foundation Scheme Grant from Canadian Institutes of Health Research (Ottawa, ON, Canada). Dr Mathieu holds a FRQS Research Chair on the Pathobiology of Calcific Aortic Valve Disease. Dr Bossé holds a Canada Research Chair in Genomics of Heart and Lung Diseases. This work was supported by the Heart and Stroke Foundation of Canada and the Canadian Institutes of Health Research (MOP—102481, MOP—137058, PJT—153396, PJT—159641] to Dr Bossé. This work was supported by an ANR & FRM grant [13-BSV6-0011, DCV20070409278] to Dr Schott, by a Fédération Française de Cardiologie, a Fondation Coeur et Recherche and an Inserm Translational Research grant to Dr Le Tourneau, a PHRC Interregional (API20-20) grant to Dr Probst, a Connect Talent research chair from Région Pays de la Loire and Nantes Métropole to Dr Capoulade and the French Regional Council of Pays-de-la-Loire (VaCaRMe program) to Dr Redon. The PREGO bio-collection is part of the VaCaRMe program founded by the French Regional Council of Pays-de-la-Loire (RFI VaCaRMe). We wish to thank the Fondation Genavie for its financial support. The D.E.S.I.R. study has been supported by INSERM contracts with CNAMTS, Lilly, Novartis Pharma and Sanofi-Aventis; by INSERM (Réseaux en Santé Publique, Interactions entre les déterminants de la santé), Cohortes Santé TGIR, the Association Diabète Risque Vasculaire, the Fédération Française de Cardiologie, La Fondation de France, ALFEDIAM, Société francophone du diabète, ONIVINS, Abbott, Ardix Medical, Bayer Diagnostics, Becton Dickinson, Cardionics, Merck Santé, Novo Nordisk, Pierre Fabre, Roche, Topcon. The COFRASA (clinicalTrial.gov number NCT 00338676) and GENERAC (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).
Disclosures
None.
Appendix
The D.E.S.I.R. Study Group. INSERM U1018: B. Balkau, P. Ducimetière, E. Eschwège; INSERM U1138: F. Alhenc-Gelas; CHU D’Angers: A. Girault; Bichat Hospital: F. Fumeron, M. Marre, R Roussel; CHU de Rennes: F. Bonnet; CNRS UMR8090, Lille: A. Bonnefond, P. Froguel; Université Paris Descartes, UMR1153, Paris: F. Rancière; Centres d’Examens de Santé: Alençon, Angers, Blois, Caen, Chateauroux, Chartres, Cholet, Le Mans, Orléans, Tours; Institute de Recherche Médecine Générale: J. Cogneau; General practitioners of the Region; Institute inter-Regional pour la Santé: C. Born, E. Caces, M. Cailleau, O Lantieri, J.G. Moreau, F. Rakotozafy, J. Tichet, S. Vol.
Footnotes
References
- 1.
Freeman RV, . Spectrum of calcific aortic valve disease: pathogenesis, disease progression, and treatment strategies.Circulation. 2005; 111:3316–3326. doi: 10.1161/CIRCULATIONAHA.104.486738LinkGoogle Scholar - 2.
Rajamannan NM, . Calcific aortic stenosis: an update.Nat Clin Pract Cardiovasc Med. 2007; 4:254–262. doi: 10.1038/ncpcardio0827CrossrefMedlineGoogle Scholar - 3.
Chan KL, ; ASTRONOMER Investigators. Effect of Lipid lowering with rosuvastatin on progression of aortic stenosis: results of the aortic stenosis progression observation: measuring effects of rosuvastatin (ASTRONOMER) trial.Circulation. 2010; 121:306–314. doi: 10.1161/CIRCULATIONAHA.109.900027LinkGoogle Scholar - 4.
Rossebø AB, ; SEAS Investigators. Intensive lipid lowering with simvastatin and ezetimibe in aortic stenosis.N Engl J Med. 2008; 359:1343–1356. doi: 10.1056/NEJMoa0804602CrossrefMedlineGoogle Scholar - 5.
Cowell SJ, ; 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/NEJMoa043876CrossrefMedlineGoogle Scholar - 6.
Martinsson A, . Familial aggregation of aortic valvular stenosis: a nationwide study of sibling risk.Circ Cardiovasc Genet. 2017; 10:e001742.LinkGoogle Scholar - 7.
Probst V , . Familial aggregation of calcific aortic valve stenosis in the western part of France.Circulation. 2006; 113:856–860. doi: 10.1161/CIRCULATIONAHA.105.569467LinkGoogle Scholar - 8.
Thériault S , . A transcriptome-wide association study identifies PALMD as a susceptibility gene for calcific aortic valve stenosis.Nat Commun. 2018; 9:988. doi: 10.1038/s41467-018-03260-6CrossrefMedlineGoogle Scholar - 9.
Thanassoulis G, ; CHARGE Extracoronary Calcium Working Group. Genetic associations with valvular calcification and aortic stenosis.N Engl J Med. 2013; 368:503–512. doi: 10.1056/NEJMoa1109034CrossrefMedlineGoogle Scholar - 10.
Helgadottir A , . Genome-wide analysis yields new loci associating with aortic valve stenosis.Nat Commun. 2018; 9:987. doi: 10.1038/s41467-018-03252-6CrossrefMedlineGoogle Scholar - 11.
Giambartolomei C , . Bayesian test for colocalisation between pairs of genetic association studies using summary statistics.PLoS Genet. 2014; 10:e1004383. doi: 10.1371/journal.pgen.1004383CrossrefMedlineGoogle Scholar - 12.
Kichaev G , . Integrating functional data to prioritize causal variants in statistical fine-mapping studies.PLoS Genet. 2014; 10:e1004722. doi: 10.1371/journal.pgen.1004722CrossrefMedlineGoogle Scholar - 13.
Zhou J, . Predicting effects of noncoding variants with deep learning-based sequence model.Nat Methods. 2015; 12:931–934. doi: 10.1038/nmeth.3547CrossrefMedlineGoogle Scholar - 14.
Battle A, ; GTEx Consortium; Laboratory, Data Analysis &Coordinating Center (LDACC)—Analysis Working Group; Statistical Methods groups—Analysis Working Group; Enhancing GTEx (eGTEx) groups; NIH Common Fund; NIH/NCI; NIH/NHGRI; NIH/NIMH; NIH/NIDA; Biospecimen Collection Source Site—NDRI; Biospecimen Collection Source Site—RPCI; Biospecimen Core Resource—VARI; Brain Bank Repository—University of Miami Brain Endowment Bank; Leidos Biomedical—Project Management; ELSI Study; Genome Browser Data Integration &Visualization—EBI; Genome Browser Data Integration &Visualization—UCSC Genomics Institute, University of California Santa Cruz; Lead analysts:; Laboratory, Data Analysis &Coordinating Center (LDACC):; NIH program management:; Biospecimen collection:; Pathology:; eQTL manuscript working group. Genetic effects on gene expression across human tissues.Nature. 2017; 550:204–213. doi: 10.1038/nature24277CrossrefMedlineGoogle Scholar - 15.
Pers TH, ; Genetic Investigation of ANthropometric Traits (GIANT) Consortium. Biological interpretation of genome-wide association studies using predicted gene functions.Nat Commun. 2015; 6:5890. doi: 10.1038/ncomms6890CrossrefMedlineGoogle Scholar - 16.
Dunham I, . FORGE: a tool to discover cell specific enrichments of GWAS associated SNPs in regulatory regions.F1000Res. 2015.CrossrefGoogle Scholar - 17.
Iotchkova V, ; UK10K Consortium. Discovery and refinement of genetic loci associated with cardiometabolic risk using dense imputation maps.Nat Genet. 2016; 48:1303–1312. doi: 10.1038/ng.3668CrossrefMedlineGoogle Scholar - 18.
Bulik-Sullivan BK , ; Schizophrenia Working Group of the Psychiatric Genomics Consortium. LD Score regression distinguishes confounding from polygenicity in genome-wide association studies.Nat Genet. 2015; 47:291–295. doi: 10.1038/ng.3211CrossrefMedlineGoogle Scholar - 19.
El Husseini D, . 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.2014.02.014CrossrefMedlineGoogle Scholar - 20.
Fishman D, . The effect of novel polymorphisms in the interleukin-6 (IL-6) gene on IL-6 transcription and plasma IL-6 levels, and an association with systemic-onset juvenile chronic arthritis.J Clin Invest. 1998; 102:1369–1376. doi: 10.1172/JCI2629CrossrefMedlineGoogle Scholar - 21.
Terry CF, . Cooperative influence of genetic polymorphisms on Interleukin 6 transcriptional regulation.J Biol Chem. 2000; 275:18138–18144. doi: 10.1074/jbc.M000379200CrossrefMedlineGoogle Scholar - 22.
Riege K , . Massive effect on LncRNAs in human monocytes during fungal and bacterial infections and in response to vitamins A and D.Sci Rep. 2017; 7:40598. doi: 10.1038/srep40598CrossrefMedlineGoogle Scholar - 23.
Wang Y, . AS-IL6 promotes glioma cell invasion by inducing H3K27Ac enrichment at the IL6 promoter and activating IL6 transcription.FEBS Lett. 2016; 590:4586–4593. doi: 10.1002/1873-3468.12485CrossrefMedlineGoogle Scholar - 24.
Revelas M, . Review and meta-analysis of genetic polymorphisms associated with exceptional human longevity.Mech Ageing Dev. 2018; 175:24–34. doi: 10.1016/j.mad.2018.06.002CrossrefMedlineGoogle Scholar - 25.
Cai T , ; VA Million Veteran Program. Association of Interleukin 6 receptor variant with cardiovascular disease effects of Interleukin 6 receptor blocking therapy: a phenome-wide association study.JAMA Cardiol. 2018; 3:849–857. doi: 10.1001/jamacardio.2018.2287CrossrefMedlineGoogle Scholar - 26.
Campbell L, . Risk of adverse events including serious infections in rheumatoid arthritis patients treated with tocilizumab: a systematic literature review and meta-analysis of randomized controlled trials.Rheumatology (Oxford). 2011; 50:552–562. doi: 10.1093/rheumatology/keq343CrossrefMedlineGoogle Scholar - 27.
Swerdlow DI , ; Interleukin-6 Receptor Mendelian Randomisation Analysis (IL6R MR) Consortium. The interleukin-6 receptor as a target for prevention of coronary heart disease: a mendelian randomisation analysis.Lancet. 2012; 379:1214–1224. doi: 10.1016/S0140-6736(12)60110-XCrossrefMedlineGoogle Scholar - 28.
Esparza-Gordillo J, . A functional IL-6 receptor (IL6R) variant is a risk factor for persistent atopic dermatitis.J Allergy Clin Immunol. 2013; 132:371–377. doi: 10.1016/j.jaci.2013.01.057CrossrefMedlineGoogle Scholar - 29.
Rincon M, . Role of IL-6 in asthma and other inflammatory pulmonary diseases.Int J Biol Sci. 2012; 8:1281–1290. doi: 10.7150/ijbs.4874CrossrefMedlineGoogle Scholar - 30.
Evangelou E, . Genetic analysis of over 1 million people identifies 535 new loci associated with blood pressure traits.Nat Genet. 2018.Google Scholar - 31.
Orimo H . The mechanism of mineralization and the role of alkaline phosphatase in health and disease.J Nippon Med Sch. 2010; 77:4–12.CrossrefMedlineGoogle Scholar - 32.
Schlotter F , . Spatiotemporal multi-omics mapping generates a molecular atlas of the aortic valve and reveals networks driving disease.Circulation. 2018; 138:377–393. doi: 10.1161/CIRCULATIONAHA.117.032291LinkGoogle Scholar - 33.
Verma SS , . Identifying genetic associations with variability in metabolic health and blood count laboratory values: diving into the quantitative traits by leveraging longitudinal data from an EHR.Pac Symp Biocomput. 2017; 22:533–544. doi: 10.1142/9789813207813_0049MedlineGoogle Scholar - 34.
Chambers JC , ; Alcohol Genome-wide Association (AlcGen) Consortium; Diabetes Genetics Replication and Meta-analyses (DIAGRAM+) Study; Genetic Investigation of Anthropometric Traits (GIANT) Consortium; Global Lipids Genetics Consortium; Genetics of Liver Disease (GOLD) Consortium; International Consortium for Blood Pressure (ICBP-GWAS); Meta-analyses of Glucose and Insulin-Related Traits Consortium (MAGIC). Genome-wide association study identifies loci influencing concentrations of liver enzymes in plasma.Nat Genet. 2011; 43:1131–1138. doi: 10.1038/ng.970CrossrefMedlineGoogle Scholar - 35.
Kanai M, . Genetic analysis of quantitative traits in the Japanese population links cell types to complex human diseases.Nat Genet. 2018; 50:390–400. doi: 10.1038/s41588-018-0047-6CrossrefMedlineGoogle Scholar - 36.
Morris JA , ; 23andMe Research Team. An atlas of genetic influences on osteoporosis in humans and mice.Nat Genet. 2019; 51:258–266. doi: 10.1038/s41588-018-0302-xCrossrefMedlineGoogle Scholar - 37.
Maes T, . Neuron navigator: a human gene family with homology to unc-53, a cell guidance gene from Caenorhabditis elegans.Genomics. 2002; 80:21–30. doi: 10.1006/geno.2002.6799CrossrefMedlineGoogle Scholar - 38.
Hoffmann TJ , . Genome-wide association analyses using electronic health records identify new loci influencing blood pressure variation.Nat Genet. 2017; 49:54–64. doi: 10.1038/ng.3715CrossrefMedlineGoogle Scholar - 39.
Capoulade R , . Relationship between proximal aorta morphology and progression rate of aortic stenosis.J Am Soc Echocardiogr. 2018; 31:561.e1–569.e1. doi: 10.1016/j.echo.2017.12.008CrossrefGoogle Scholar - 40.
Rahimi K , . Elevated blood pressure and risk of aortic valve disease: a cohort analysis of 5.4 million UK adults.Eur Heart J. 2018; 39:3596–3603. doi: 10.1093/eurheartj/ehy486CrossrefMedlineGoogle Scholar - 41.
Gusev A , ; Schizophrenia Working Group of the Psychiatric Genomics Consortium. Transcriptome-wide association study of schizophrenia and chromatin activity yields mechanistic disease insights.Nat Genet. 2018; 50:538–548. doi: 10.1038/s41588-018-0092-1CrossrefMedlineGoogle Scholar - 42.
Mancuso N , . Probabilistic fine-mapping of transcriptome-wide association studies.Nat Genet. 2019; 51:675–682. doi: 10.1038/s41588-019-0367-1CrossrefMedlineGoogle Scholar - 43.
Döring Y , . CXCL12 derived from endothelial cells promotes atherosclerosis to drive coronary artery disease.Circulation. 2019; 139:1338–1340. doi: 10.1161/CIRCULATIONAHA.118.037953LinkGoogle Scholar - 44.
Krause MD , . Genetic variant at coronary artery disease and ischemic stroke locus 1p32.2 regulates endothelial responses to hemodynamics.Proc Natl Acad Sci USA. 2018; 115:E11349–E11358. doi: 10.1073/pnas.1810568115CrossrefMedlineGoogle Scholar - 45.
Visel A , . Targeted deletion of the 9p21 non-coding coronary artery disease risk interval in mice.Nature. 2010; 464:409–412. doi: 10.1038/nature08801CrossrefMedlineGoogle Scholar - 46.
Wang X, . Confirmation of causal rs9349379- PHACTR1 expression quantitative trait locus in human-induced pluripotent stem cell endothelial cells.Circ Genom Precis Med. 2018; 11:e002327. doi: 10.1161/CIRCGEN.118.002327LinkGoogle Scholar - 47.
Yang B , . Protein-altering and regulatory genetic variants near GATA4 implicated in bicuspid aortic valve.Nat Commun. 2017; 8:15481. doi: 10.1038/ncomms15481CrossrefMedlineGoogle Scholar - 48.
Garg V , . Mutations in NOTCH1 cause aortic valve disease.Nature. 2005; 437:270–274. doi: 10.1038/nature03940CrossrefMedlineGoogle Scholar - 49.
Rose-John S . The soluble Interleukin 6 receptor: advanced therapeutic options in inflammation.Clin Pharmacol Ther. 2017; 102:591–598. doi: 10.1002/cpt.782CrossrefMedlineGoogle Scholar - 50.
van der Harst P, . Identification of 64 novel genetic loci provides an expanded view on the genetic architecture of coronary artery disease.Circ Res. 2018; 122:433–443. doi: 10.1161/CIRCRESAHA.117.312086LinkGoogle Scholar
