NLRC4 Inflammasome Is an Important Regulator of Interleukin-18 Levels in Patients With Acute Coronary SyndromesCLINICAL PERSPECTIVE
Genome-Wide Association Study in the PLATelet inhibition and patient Outcomes Trial (PLATO)
Background—Interleukin 18 (IL-18) promotes atherosclerotic plaque formation and is increased in patients with acute coronary syndromes. However the relative contribution of genetic variants to the IL-18 levels has not been fully determined.
Methods and Results—Baseline plasma IL-18 levels were measured in 16 633 patients with acute coronary syndrome, of whom 9340 had genetic data that passed genotype quality control. A 2-stage genome-wide association study was performed, followed by combined analyses using >10 million genotyped or imputed genetic markers. Single nucleotide polymorphisms at 3 loci (IL18, NLRC4, and MROH6) were identified (P<3.15×10−8) in the discovery cohort (n=3777) and replicated in the remaining patients (n=5563). In the pooled data (discovery+replication cohort), 7 independent associations, in 5 chromosomal regions, were associated with IL-18 levels (minimum P=6.99×10–72). Six single nucleotide polymorphisms are located in predicted promoter regions of which one disrupts a transcription factor binding site. One single nucleotide polymorphism in NLRC4 is a rare missense variant, predicted to be deleterious to the protein. Altogether, the identified genetic variants explained 8% of the total variation in IL-18 levels in the cohort.
Conclusions—Our results show that genetic variants play an important role in determining IL-18 levels in patients with acute coronary syndrome and we have identified genetic variants located in the IL-18 gene (IL18) or close to genes that are involved in procaspase-1 activation (NLRC4 and CARD16, CARD17, and CARD18). These associations also highlight the importance of the NLRC4 inflammasome for IL-18 production in acute coronary syndrome patients.
Acute coronary syndromes (ACSs) are a major cause of emergency hospital care and have a high morbidity and mortality. A large number of biomarkers have been investigated in patients with ACS and found to be related to worse disease outcomes. Interleukin-18 (IL-18), a member of the IL-1 cytokine superfamily, is an interferon-γ–inducing factor1 and is expressed in atherosclerotic plaques.2 IL-18 has been associated with destabilization of plaques leading to subsequent thrombotic events.3 It is known that IL-18 is increased in patients with acute myocardial infarction4 and an association between baseline levels of IL-18 and recurrent cardiovascular events has been reported.5,6
Clinical Perspective on p 506
Experiments in mice have shown that IL-18 promotes atherosclerosis through the release of interferon-γ,7 highlighting the inflammatory context of atherosclerosis. Similarly, expression of plasmid DNA encoding murine IL-18–binding protein has been associated with reduced atherosclerotic plaque progression, resulting in more stable plaques.8 This is consistent with the expression of IL-18 being higher in unstable compared with stable plaques in humans.2 Despite the wide range of data relating IL-18 to clinical and subclinical cardiovascular disease, no prior large-scale genetic study of IL-18 has been performed in patients with ACS.
To gain a better understanding of the molecular mechanisms behind IL-18 release in ACS, we aimed to identify genetic variants that influence IL-18 levels. A genome-wide association study (GWAS) for IL-18 was performed in a large contemporary cohort of all types of ACS. Our results show that genetic variants play an important role in IL-18 variation and highlight the importance of the NLRC4 inflammasome for IL-18 production in patients with ACS.
The PLATelet inhibition and patient Outcomes (PLATO) trial (Clinical Trial Registration: www.clinicaltrials.gov; NCT00391872) was a prospective randomized clinical trial that compared the effect of ticagrelor versus clopidogrel in 18 624 patients with ACS. The design, population, baseline clinical characteristics, and primary results of the PLATO trial have previously been published.9,10 The trial was performed in accordance with the Declaration of Helsinki and approved by regulatory authorities in all participating countries and by participating sites’ institutional review boards. All participants provided written informed consent. Participation in the genetic substudy was voluntary and required an additional consent form at the time of enrollment in the genetic substudy.
Biomarker Laboratory Analysis
Baseline venous blood samples for biomarker investigations were collected, for all patients, in EDTA tubes within 24 hours of admission, before the administration of study medication. During the index hospitalization, an additional venous blood sample was collected in EDTA tubes for genetic analyses for 10 013 individuals. Additional samples, in a serial sample subcohort, were taken at a second visit at the hospital, ≈1 month (mean, 31.3; SD 8.0 days) after randomization. Venous blood samples were centrifuged and plasma samples were frozen at −70°C until analyzed centrally at the Uppsala Clinical Research Center laboratory, Uppsala, Sweden. IL-18 levels were measured at baseline and at the second visit using human IL-18 ELISA (Medical and Biological Laboratories Co. Ltd, Naka-ku, Nagoya, Japan), with a total precision of 9%.
A 2-stage design was used for the GWAS, with a discovery cohort and a replication cohort, followed by combined analyses of the pooled discovery and replication cohort. The discovery cohort consisted of a random set of 3998 patients genotyped using the Illumina HumanOmni2.5-4v1 (Omni2.5) BeadChip (Illumina, San Diego, CA), and the replication cohort of the remaining 6015 patients, which were genotyped using the Illumina Infinium HumanOmniExpressExome-8v1 BeadChip (Illumina, San Diego, CA). Genotyping was performed according to the manufacturer’s instructions. Analysis of the raw data was performed using Illumina GenomeStudio 2011.1 (Illumina, San Diego, CA), the Illumina’s Infinium assay and project sample generated cluster files.11,12 Quality controls were performed using the whole genome association analysis toolset PLINK v1.0713 (http://pngu.mgh.harvard.edu/purcell/plink) in the discovery and replication cohorts separately (Table I in the Data Supplement).
Imputations of Ungenotyped Single Nucleotide Polymorphisms
Imputations were performed in the discovery and replication cohorts separately (Table I in the Data Supplement) using a prephasing approach14 implemented in SHAPEIT version 215 and IMPUTE2 (version 2.2.2).16 The 1000 Genomes17 phase I integrated variant set (National Center for Biotechnology Information build b37, March 2012, updated August 24, 2012) was accessed from the IMPUTE2 web site and used as reference panel. Postimputation quality control was performed using QCTOOL version 1.3 (http://www.well.ox.ac.uk/≈gav/qctool). Imputed data from the discovery and replication cohorts were merged based on the single nucleotide polymorphism (SNP) position (Table I in the Data Supplement).
Pairwise kinship matrices were calculated for the replication and discovery cohorts separately and together, using genotyped autosomal SNPs (Table I in the Data Supplement) and the ibs function (weighted by the allele frequency), which is implemented in GenABEL.18 The kinship matrix was used to calculate the pairwise distance matrix between individuals followed by classical multidimensional scaling analyses using 10 dimensions.
A large set of clinical variables were tested before association analyses (Table II in the Data Supplement). The selection of covariables for the genetic analyses was performed in 2 steps. First, each variable was tested in the discovery cohort using a simple linear regression model with log-transformed IL-18 as dependent variable. Second, all variables that were significantly explanatory for IL-18 at the significance level of P<0.1 were included in a multiple linear regression model (Table III in the Data Supplement). Only variables with P<0.1 in the multiple linear regression model were included as covariables in the genome-wide association analyses.
In the GWAS, IL-18 was analyzed on log scale (to avoid a heavy-tailed variable) using a linear regression model implemented as the mlreg function in the GenABEL R library.18 To adjust for possible population stratification, all genome-wide association analyses were adjusted for the 4 first genetic principal components.19 GWAS analyses were initially performed in the discovery cohort, using genotyped SNPs only. To adjust for the number of statistical tests performed, a genome-wide significance level, according to the Bonferroni method, was set at P=3.152×10−8 (0.05/1 586 429 SNPs). All SNPs that passed the threshold for significance were taken forward for replication, using the same set of covariates. Because the discovery and replication cohorts were not genotyped using the same array, we used imputed genotypes for the replication where necessary. Imputed SNP data were analyzed using the palinear function in the ProbABEL package.20 The threshold for significance for a successful replication was set, according to the Bonferroni method, to P=0.05 divided by the number of SNPs to be replicated.
Finally, the pooled imputed data from the discovery and replication cohorts were analyzed together. To test for independent genetic effects in the pooled sample, the most significant SNP from the primary analyses was added to the previously used covariables and a new round of GWAS analyses was performed. This process, adding the most significant SNP from the previous round to the covariables and rerunning the analyses, was repeated until no genome-wide significant SNP was detected. In the pooled analyses, we used the same threshold for significance as in the discovery GWAS (P<3.152×10−8). To compare IL-18 levels at baseline and at the second visit (≈1 month later), we tested whether the difference between visits 2 and baseline deviated from zero, using Wilcoxon rank test.
The top SNPs from each independent genetic association signal were imported as custom tracks into the UCSC (University of California Santa Cruz) genome browser (Human Feb. 2009 GRCh37/hg19 Assembly, Data retrieved April 26, 2013). The location of the associated SNPs was compared with the location of (1) known human protein-coding and non–protein-coding genes taken from the National Center for Biotechnology Information RNA reference sequences collection (RefSeq genes: last updated April 25, 2012), (2) regions where transcription factors bind to DNA as assayed by chromatin immunoprecipitation sequencing (transcription factor chromatin immunoprecipitation sequencing from ENCODE [The Encyclopedia of DNA Elements (March 2012 Freeze)], and (3) chromatin state segmentation for 9 human cell types21,22 using the ENCODE. Sorting Intolerant From Tolerant23 and Polymorphism Phenotyping24 scores were retrieved from Ensembl database (http://www.ensembl.org, accessed August 28, 2013).
The clinical background and baseline characteristics were similar in the discovery (n=3998) and replication (n=6015) cohorts (Tables IV and V in the Data Supplement). A large number of clinical variables are likely to influence IL-18 levels, and including those as covariates (precision variables) in the analyses might increase our power to detect association between SNPs and IL-18. We identified 7 variables (age, sex, body mass index, medical history of diabetes mellitus, peripheral artery disease and chronic renal disease, and lipid-lowering agent at the day of randomization) to be included as covariates in the association analyses (Figure 1; Table III in the Data Supplement). The study includes individuals with different genetic origins (Figure I in the Data Supplement). Therefore, we also included the first 4 principal components as covariates (confounders) in the association analyses, because they might affect both IL-18 levels and SNPs by differences in allele frequencies and IL-18 levels between population groups (Figure 1).
Genome-Wide Association Study
A total of 3982 and 5996 patients passed genotype quality control of which 3777 and 5563 had IL-18 measurements available in the discovery and replication cohort, respectively. In the primary GWAS in the discovery cohort, a total of 110 genome-wide significant SNPs (Table VI in the Data Supplement), located in 3 chromosomal regions, were identified. The largest region on chromosome 2 included 100 significant SNPs (minimum P=6.31×10–31) located within or close to NLRC4. The second largest region on chromosome 11 consisted of 9 SNPs (minimum P=9.36×10–24) located at the IL18 locus. The third region had only 1 genome-wide significant SNP (rs13271361, P=3.08×10–8), which was located within MROH6 on chromosome 8. Of the 110 significant SNPs, all but 6 were genotyped or successfully imputed in the replication cohort, and all these SNPs replicated with P values ranging from 3.50×10–42 to 5.98×10–9 with the same direction of the effect (Table VI in the Data Supplement). The most significant SNP from the discovery cohort (kgp1626348) was located at position 32484955 on chromosomes 2. The overlapping variant in the replication, at this position, is rs67593209 (also known as chr2:32484952:D, or chr2_32484952_32484955) which is a TAAG/− deletion polymorphism. In the replication cohort, this polymorphism replicated with similar effect on IL-18 levels as kgp1626348 in the discovery cohort (Table VI in the Data Supplement).
Pooled analyses of the imputed data from the discovery and replication cohorts resulted in 5 chromosomal regions being significantly associated with IL-18 levels (Figure 2; Table VII in the Data Supplement). In addition to the 3 previous regions, 2 SNPs in a region on chromosome 5 (minimum P<5.00×10–12 for rs17229943) located in RAD17, and 17 SNPs (minimum P=1.38×10–8 for rs1842893) located upstream of caspase recruitment domain 16 (CARD16), CARD17, and CARD18. Including the top SNP from each of the 5 identified regions (rs385076, chr11.112034062=rs34649619, rs17229943 and rs2290414 and rs1842893) as covariables in the conditional analyses resulted in 2 additional signals, a second signal in NLRC4 (minimum P=3.79×10–16 for rs149451729=exm184959) and a second signal in the IL18 region (minimum P=1.66×10–12 for rs11214115). Including also the 2 latter association signals in the conditional analyses did not result in any significant results (Figure 2).
Bioinformatic Analyses of Top SNPs
Bioinformatic analyses were performed using the results from the pooled analyses with imputed data. The most significant SNPs at the IL18 locus (Figure IIA in the Data Supplement) were distributed throughout IL18 and the neighboring genes (TEX12 and BCO2). The most significant SNP (chr11:112034962:I or rs34649619, P=1.17×10–50) was located in a site that is enriched for transcription factor binding (Figure 3; Figure II in the Data Supplement). This variant together with neighboring sites (AGAAGCTT[−/A]ATTTATTTTAC) match the DNA motif for among others: FOXD3 (TTATT), FOXA2 (TATTT), and FOXI1 (TATTT) for the wild-type allele (−). The disruption of a binding site by the alternative (minor) allele agrees with lower levels of IL-18 in patients with the alternative allele (Table). The additional association signal from the conditional analyses was centered upstream and in the beginning of IL18 (Figure II in the Data Supplement). Two of the SNPs (rs360717 and rs360718, of which rs360718 as the most significant site in the conditional analyses, P=2.03–12) were located in a site enriched for transcription factor binding in association with the transcription start site (Figure 3; Figure II in the Data Supplement).
The association signal at the NLRC4 locus consisted of a large number of SNPs in strong linkage disequilibrium (Table VII in the Data Supplement; Figure III in the Data Supplement). This signal included NLRC4, but also DPY13, SPAST, and SLC30A6. The most significant SNP (rs385076, P=6.99×10–72) did not have either a considerably lower P value compared with the others (second most significant P=2.78×10–68) or a greater genetic effect (0.113 U of the log-transformed IL-18 values in nanogram per liter, compared with 0.108 for the second largest). However, rs385076 is the only SNP that is located in a predicted regulatory region. The region around rs385076 has been predicted to be an active promoter in GM12878 (B lymphocyte, lymphoblastoid) cells and a weak promoter in HepG2 (hepatocellular carcinoma) cells and is also located close to the transcription start site. One of the most significant SNPs at NLRC4 (rs212717, P=4.27×10–67) was an expression quantitative trait locus that is associated with the expression of NLRC4 (P=4.037×10–7) in Epstein–Barr virus–transformed lymphoblastoid cell lines25 (data retrieved from eqtl.uchicago.edu, April 16, 2013). The second signal at the NLRC4 locus, which was identified in the conditional analyses, was driven by rs149451729, which is located in exon 6 of NLRC4. This SNP is a rare (minor allele frequency, 0.0037) missense variant (also known as 2357G>T) resulting in the amino acid substitution from glycine to valine at position 786 of the protein (Gly786Val). Both Sorting Intolerant From Tolerant23 and Polymorphism Phenotyping24 predictions (Sorting Intolerant From Tolerant score=0.04; Polymorphism Phenotyping score=0.943) suggest that this amino acid change is deleterious and therefore likely to affect the function of the protein.
The association signal on chromosome 8 consists of 14 SNPs located within GSDMD and MROH6 as well as between the genes (Figure IV in the Data Supplement). Although the most significant SNP (rs2290414, P=2.77×10–16) was located in a nonregulatory and noncoding region of GSDMD, 2 of the top 5 SNPs (rs13254352, P=5.88×10–14 and rs13271361, P=9.49×10–16) were located in predicted regulatory regions within MROH6. The association signal on chromosome 5 consisted of 2 SNPs neither of which was found to be associated in the initial GWAS in the discovery cohort. The most significant SNP (rs17229943, P=7.04×10–12) was located in an intron of RAD17, in a region where no functional element has been assigned.
The second region on chromosome 11 included as set of 18 SNPs with P values ranging from 1.10×10–8 to 3.06×10–8 (Figure 4; Table VII in the Data Supplement). None of the SNPs in this region were identified in the GWAS of the discovery cohort only. All SNPs were located upstream of a cluster of CARD (CARD16, CARD17, and CARD18) and CASP (CASP1, CASP5, CASP4, and CASP12) genes, except for 1 which was located upstream of all but CARD18 (Figure V in the Data Supplement). Some of these SNPs (eg, rs17103763 and rs11226633) were located in predicted regulatory regions close to transcription factor binding sites. In previous studies, SNPs with somewhat lower Ps in our study (eg, rs17103597, P=6.06×10–6) have been identified as expression quantitative trait locus (P=2.81×10–118) for CARD16,26 whereas SNPs (eg, rs1941425) that have previously been identified as expression quantitative trait locus for CASP127 were not associated with IL-18 levels in our study (P>0.05).
Pairwise Genetic Interactions
We also tested for pairwise genetic interactions between the top SNPs at each identified loci. One of the strongest associations was the SNPs in NLRC4 gene. NLRC4 is known to interact with other proteins including CASP1 and NOD2. However, we did not see any genetic interaction between the SNP in the CASP1 region and the SNPs in NLRC4 (P>0.05). Similarly, we did not find any significant interaction between any of the top SNPs tested (P>0.05). NOD2 is located on chromosome 16: position 50 727 514 to 50 766 987 (build 37/hg19). In this region (10 kb upstream and 10 kb downstream of NOD2), we had included 192 SNPs in our analyses. Of these, the most significant P value was 0.0117. This was not significant after adjusting for the 192 SNPs tested in the region.
Multiple Regression Analyses
The multivariable model, including the 7 top SNPs from the pooled analyses, resulted in similar β estimates and P values as in the original GWAS (Table). The total variance explained by each of the SNPs range from 0.33% to 3.2% of the variation in IL-18 levels and altogether they account for ≤8% of the total variance in IL-18 levels.
Because our study includes individuals with different genetic backgrounds (Figure I in the Data Supplement), we included the first 4 principal components in all analyses. This resulted in λ values (inflation factor) of 1.03, 1.10, and 1.16 in the discovery, replication, and pooled data sets, respectively (estimates are based on genotyped SNPs only). The λ value for the combined data (discovery+replication cohort) increased slightly (1.16), which can be explained by the strong associations between IL-18 levels and multiple SNPs (Figure VI in the Data Supplement). However, we also performed sensitivity analyses by stratifying the analyses to include only individuals with European, only African, or only Asian descent. The European-only analyses gave similar results to the previous analyses, whereas the results with patients from the other ethnicities were too few (n=53 and n=95 with both genotype and IL-18 levels measured) to give any significant results (Table VIII in the Data Supplement).
Serial IL-18 Measurements and ACS Subtype Analyses
In a subset of the patients with genotype data (n=3080), IL-18 was also measured at a second visit. For these patients, the IL-18 levels were higher (P<2×10–16) at the second visit (mean, 306.0 ng/L; median, 341.3 ng/L) compared with baseline (mean, 258.1 ng/L; median, 231.0 ng/L). The 7 identified SNPs were also associated with IL-18 levels at the second visit. All β estimates were numerically larger at the second visit compared with baseline; however, a test for the paired difference was only significant for 2 SNPs (Table). To evaluate the variable effect of the IL-18–associated genetic variants in different ACS subtypes, additional analyses were performed separately for patients with ST-segment–elevation myocardial infarction (STEMI), non-STEMI and unstable angina. The IL-18 levels were similar between the ACS subtypes with mean (interquartile range) log-transformed IL-18 levels of 5.46 (5.20–5.49) for unstable angina, 5.48 (5.21–5.48) for STEMI and 5.45 (5.18–5.46) for non-STEMI. For the most significant SNP (rs34649619) in the IL-18 region, the effect of the SNP seems to be more pronounced in patients with STEMI as compared with non-STEMI (Table). However, the interaction between the rs34649619 and ACS subtypes was weak (P=0.0397) and was not significant after adjusting for testing 7 SNPs.
We have performed a GWAS to identify genetic variants influencing IL-18 levels. In total, we identified and replicated 3 loci (IL18, NLRC4, and MROH6) of which IL18 and NLRC4 each has 2 independent effects. By combining data from the discovery and replication cohorts, 2 additional association signals were identified. The association between IL18 and the mature form of the protein it codes for is not surprising and has been identified in previous GWAS.28,29 However, the identification of 2 independent variants, located in different regulatory regions, is a novel finding. In addition, we identify that one of these variants interrupts a binding site for several transcription factors including FOXD3, FOXA2, and FOXI1.
The strongest genetic effect influencing the IL-18 levels was seen for the NLRC4 locus. A recent GWAS also identified SNPs near NLRC4 to be associated with IL-1ra as well as IL-18.28 However, in our study, we also demonstrate that the association signal for IL-18 reflects 1 regulatory and 1 deleterious variant within NLRC4. The NLRC4 protein belongs to the Nod-like receptors. NLRC4 is an inflammasome that activates the inflammatory cascade in the presence of bacterial molecules. NLRC4 recruits and activates procaspase-1, which in its turn is responsible for the maturation of pro-IL-18.30 In our study, we identify 1 variant that regulates the expression of NLRC4 and another variant that alters the amino acid sequence of the protein resulting in loss of function. It has been suggested that the expression of another Nod-like receptor, NLRP3, is activated in patients with coronary artery disease followed by increase in IL-18 levels.31 Our finding also suggests that the NLRC4 inflammasome is also an important complex for IL-18 maturation in patients with ACS, and thereby regulates the release of IL-18. IL-18 is regarded as a risk factor for ACS, by promoting atherosclerosis7 and, consequently, the NLRC4 inflammasome is also likely to play a role in the pathogenesis of the disease. We suggest that genetic variants that decrease the activity or the expression level of NLRC4 could be protective of atherosclerosis. In our study, we have identified 1 deleterious missense variant that decreases IL-18 levels and 1 regulatory genetic variant that decreases the expression of the NLRC4 resulting in decreased IL-18 levels. Recently, a de novo missense mutation that affects the nucleotide-binding domain of the NLRC4 inflammasome was shown to cause early-onset recurrent fever flares and macrophage activation syndrome.32 In contrast to the deleterious variants that we identified, this de novo missense mutation was associated with spontaneous inflammasome formation and production of IL-18. We can therefore conclude that the importance of the NLRC4 inflammasome in IL-18 production is not restricted to patients with ACS.
The NLRC4 inflammasome consists of several different components of which most have a CARD.33 Therefore, another interesting finding in our data was the association of SNPs upstream of CARD16, CARD17, and CARD18 (also known as COP1, INCA, and ICEBERG). These CARD genes are located close to the gene encoding procaspase-1 (CASP1) and are likely to be the result of an historical duplication event followed by the accumulation of mutations changing the functionality of the proteins.34 These genes encode caspase inhibitors that bind to and interact with procaspase-1, resulting in inhibition of procaspase-1 activation. In agreement with this, CARD16, CARD17, and CARD18 have been shown to prevent at least IL-1β release. However, in this study, we cannot distinguish which gene(s) are regulated by the SNPs identified, and clusters of genes with similar functions often share expression patterns. Although expression data show that our association overlaps with an expression quantitative trait locus for CARD16, it is possible that these variants influence the expression of the whole set of CARD and CASP genes in the region. Two additional association signals are located around the genes MROH6 and RAD17. Neither of these genes has previously been associated with any inflammation or cardiovascular phenotypes.
The data sets we have used are genotyped using 2 different arrays. The replication array includes a large number of SNPs that have been selected to be nonsynonymous coding variants and also includes less common variants. In combination with using a large data set and the latest release of the 1000 genomes reference panel for the imputations, we have also been able to evaluate the effect of rare variants. This has resulted in the identification of a rare nonsynonymous deleterious variant in NLRC4 that is associated with decreased IL-18 levels. It has been suggested that rare variants might have larger effects on complex phenotypes compared with common variants that have traditionally been identified in GWAS.35 The effect of the rare NLRC4 variant (rs149451729) on IL-18 levels was large, with an average decrease of ≈1 SD for individuals being heterozygous (average log-transformed IL-18 [units of ng/L]=5.03,) for the variant compared with homozygous for the common allele (average log-transformed IL-18=5.46). Owing to its frequency, this variant explains <1% of the total variation in IL-18 levels in the sample. Altogether, the identified loci explain ≈8% of the total variance in IL-18, which is more than any of the clinical covariables included in the analyses.
There are some limitations of this study. The study samples represent a global selection of patients, which might give rise to population stratification. To resolve this problem, we have included the first 4 genetic principal components in all analyses. This indicated limited levels of inflation. In addition, we have performed subgroup analyses of the major population groups, which resulted in similar results to the analyses in the total cohort. It is also worth noting that even if we have performed a 2-stage GWAS with one discovery and another replication cohort, some of the results from this study were identified in the combined sample set. These include the independent genetic variants in IL18 and NLRC4, as well as the genetic variants in RAD17 and close to the CARD genes.
One of the most interesting results of our study is that NLRC4 plays an important role in regulation of IL-18 levels in patients with ACS. However, it should be recognized that the lack of association between genetic variants in other genes (eg, NOD2 and CASP1) does not suggest that they are not involved in IL-18 production. One possibility is that there are no genetic variants that influence NOD2 and CASP1 in our study and therefore we do not identify significant associations with genetic variants in these genes. Another limitation with the study is that we cannot evaluate the importance of IL-18 and NLRC4 on the development and progress of ACS. Previous studies in animal models suggest that IL-18 might have a causal effect on arteriosclerotic plaque formation. Recently, Mendelian randomization studies have been used to evaluate the causal effect of biomarkers on risk of disease, using genetic variants as instrumental variables. However, the individual genetic variants identified in our study explains only a small fraction of the variability in IL-18 levels, and therefore a much larger sample size and a long follow-up would be needed to assess the causal effect of NLRC4 genetic variants and IL-18 levels to ACS.
In summary, we have identified common and rare genetic variants that are associated with IL-18 levels. Our results highlight the importance of the NLRC4 inflammasome, and proteins involved in caspase-1 activation, for IL-18 production in patients with ACS. IL-18 has previously been suggested to promote atherosclerotic plaque formation and a better understanding of the mechanism behind the increased IL-18 levels in patients with ACS should be of significant value in understanding the disease progression and for developing more effective treatments.
Genotyping was performed by the SNP & SEQ Technology Platform in Uppsala, which is supported by Uppsala University, Uppsala University Hospital, Science for Life Laboratory (SciLifeLab) Uppsala and the Swedish Research Council (Contracts 80576801 and 70374401). The computations were performed on resources provided by SNIC through Uppsala Multidisciplinary Center for Advanced Computational Science (UPPMAX) under projects p2012253 and p2012095. Ebba Bergman, PhD, at Uppsala Clinical Research Center, Uppsala, Sweden, provided editorial assistance.
Sources of Funding
The PLATelet inhibition and patient Outcomes study was funded by AstraZeneca. The genetic work was also cofunded by the Swedish Heart-Lung Foundation.
Drs Johansson and Eriksson have received institutional research grant from AstraZeneca. Dr Becker has received grants from AstraZeneca been scientific advisory board member for Bayer, Janssen and Regado Biosciences, and safety monitoring committee member for Portola. Dr Storey has received institutional research grants from AstraZeneca, Daiichi Sankyo/Eli Lilly, and Merck; consultancy fees from AstraZeneca, Accumetrics, Correvio, Daiichi Sankyo/Eli Lilly, Merck, Plaque Tec, Roche, The Medicines Company, Regeneron, and Sanofi-Aventis; speakers fees from AstraZeneca, Accumetrics, and Daiichi Sankyo/Eli Lilly; travel support from AstraZencea; consumables from Accumetrics; and honoraria from Medscape. Dr Hagström has received institutional research grant from AstraZeneca; honoraria from Sanofi. Dr Varenhorst has received institutional research grant from AstraZeneca; speaker fees from AstraZeneca and The Medicines Company, and is Advisory Board member for The Medicines Company and AstraZeneca. Dr James has received institutional research grant from AstraZeneca, Terumo Inc, Medtronic, and Vascular Solutions; honoraria from The Medicines Company and AstraZeneca; and consultant/advisory board fees from AstraZeneca, Dachii Sanchio, Janssen, Medtronic, and Sanofi. Dr Katus has received honoraria from AstraZeneca, Eli Lilly, GlaxoSmithKline, Roche, and Bayer; and holds a Troponin T test Invention patent jointly with Roche and receives royalties for this patent. Dr Steg has received personal fees and nonfinancial support from AstraZeneca; personal fees from Amarin, Bayer, Boehringer-Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo, Eli Lilly, Merck Sharpe & Dohme, Novartis, Otsuka, Pfizer, Roche, Medtronic, Vivus, The Medicines Company, Sanofi, Servier, and GlaxoSmithKline for steering committees, data monitoring committees, event committees, and consulting activities; Dr Steg’s institution receives research grants from Sanofi and Servier and he is a stockholder in Aterovax. Dr Wallentin has received research grants from AstraZeneca, Merck & Co, Boehringer-Ingelheim, Bristol-Myers Squibb/Pfizer, GlaxoSmithKline; consultancy fees from Abbott, Merck & Co, Regado Biosciences, Athera Biotechnologies, Boehringer-Ingelheim, AstraZeneca, GlaxoSmithKline, and Bristol-Myers Squibb/Pfizer; lecture fees from AstraZeneca, Boehringer-Ingelheim, Bristol-Myers Squibb/Pfizer, and GlaxoSmithKline. Dr Wallentin has also received honoraria from Boehringer-Ingelheim, AstraZeneca, Bristol-Myers Squibb/Pfizer, GlaxoSmithKline; travel support from AstraZeneca, Bristol-Myers Squibb/Pfizer, and GlaxoSmithKline. Dr Siegbahn has received institutional research grants from AstraZeneca, Boehringer-Ingelheim, and Bristol-Myers Squibb. Dr Himmelmann is an employee of AstraZeneca. Dr Barratt is an employee of AstraZeneca and has stock and stock options in AstraZeneca. The other authors report no conflicts.
From the Uppsala Clinical Research Center (Å.J., N.E., E.H., C.V., S.K.J., L.W., A.S.), Department of Immunology, Genetics, and Pathology (Å.J.), Department of Medical Sciences, Cardiology (E.H., C.V., S.K.J., L.W.), Department of Medical Sciences, Molecular Medicine, Science for Life Laboratory (T.A., A.-C.S.), and Department of Medical Sciences, Clinical Chemistry (A.S.), Uppsala University, Uppsala, Sweden; Division of Cardiovascular Health and Disease, Heart, Lung and Vascular Institute, Academic Health Center, Cincinnati, OH (R.C.B.); Department of Cardiovascular Science, University of Sheffield, Sheffield, United Kingdom (R.F.S.); AstraZeneca Research and Development, Mölndal, Sweden (A.H.); Duke Clinical Research Institute, Duke University Medical Center, Durham, NC (E.H.); AstraZeneca R&D, Alderley Park, Cheshire, United Kingdom (B.J.B.); Medizinishe Klinik, Universitätsklinikum Heidelberg, Heidelberg, Germany (H.A.K.); INSERM-Unité 1148, Paris, France (P.G.S.); Assistance Publique- Hôpitaux de Paris, Département Hospitalo-Universitaire FIRE, Hôpital Bichat, Paris, France (P.G.S.); Université Paris-Diderot, Sorbonne-Paris Cité, Paris, France (P.G.S.); and NHLI Imperial College, ICMS, Royal Brompton Hospital, London, United Kingdom (P.G.S.).
The Data Supplement is available at http://circgenetics.ahajournals.org/lookup/suppl/doi:10.1161/CIRCGENETICS.114.000724/-/DC1.
- Received June 2, 2014.
- Accepted February 16, 2015.
- © 2015 American Heart Association, Inc.
- Mallat Z,
- Corbaz A,
- Scoazec A,
- Besnard S,
- Lesèche G,
- Chvatchko Y,
- et al
- Seta Y,
- Kanda T,
- Tanaka T,
- Arai M,
- Sekiguchi K,
- Yokoyama T,
- et al
- Hartford M,
- Wiklund O,
- Hultén LM,
- Persson A,
- Karlsson T,
- Herlitz J,
- et al
- James S,
- Akerblom A,
- Cannon CP,
- Emanuelsson H,
- Husted S,
- Katus H,
- et al
- Aulchenko YS,
- Ripke S,
- Isaacs A,
- van Duijn CM.
- Ng PC,
- Henikoff S.
- He M,
- Cornelis MC,
- Kraft P,
- van Dam RM,
- Sun Q,
- Laurie CC,
- et al
- Lamkanfi M,
- Denecker G,
- Kalai M,
- D’hondt K,
- Meeus A,
- Declercq W,
- et al
Acute coronary syndromes (ACS) are a major cause of emergency hospital care and have a high morbidity and mortality. Several biomarkers are known to be increased in patients with ACS, and these biomarkers are often associated with poor disease prognosis. However, biomarkers differ naturally between individuals due factors such as age and smoking, but also genetics. Therefore, the underlying genetic variants might be confounders in relation to clinical cutoffs and when a disease therapy is chosen. One step toward personalized clinical cutoffs is the investigation of genetic variants that influence the variation in biomarker levels. One biomarker that has been associated with ACS is interleukin 18 (IL-18). IL-18 is a proinflammatory cytokine that is known to promote atherosclerosis through the release of interferon-γ. In our study, we have identified genetic variants that dramatically influence IL-18 levels in patients with ACS. Besides identifying regulatory genetic variants in the gene coding for IL-18, we identify that the NLRC4 inflammasome is an important regulator of IL-18 levels in the patients, and probably plays a role in the arteriosclerotic plaque formation. Our results show that the IL-18 levels differ in the patients because of genes involved in the inflammatory cascade. This information is importation for further understanding of the inflammatory context of ACS.