Genetic variants in PLCB4/PLCB1 as susceptibility loci for coronary artery aneurysm formation in Kawasaki disease in Han Chinese in Taiwan

Kawasaki disease (KD) is an acute, inflammatory, and self-limited vasculitis affecting infants and young children. Coronary artery aneurysm (CAA) formation is the major complication of KD and the leading cause of acquired cardiovascular disease among children. To identify susceptible loci that might predispose patients with KD to CAA formation, a genome-wide association screen was performed in a Taiwanese KD cohort. Patients with both KD and CAA had longer fever duration and delayed intravenous immunoglobulin treatment time. After adjusting for these factors, 100 susceptibility loci were identified. Four genes were identified from a single cluster of 35 using the Ingenuity Pathway Analysis (IPA) Knowledge Base. Silencing KCNQ5, PLCB1, PLCB4, and PLCL1 inhibited the effect of lipopolysaccharide-induced endothelial cell inflammation with varying degrees of proinflammatory cytokine expression. PLCB1 showed the most significant inhibition. Endothelial cell inflammation was also inhibited by using a phospholipase C (PLC) inhibitor. The single nucleotide polymorphism rs6140791 was identified between PLCB4 and PLCB1. Plasma PLC levels were higher in patients with KD and CC+CG rs6140791genotypes, and these genotypes were more prevalent in patients with KD who also had CAA. Our results suggest that polymorphism of the PLCB4/B1 genes might be involved in the CAA pathogenesis of KD.

cations, we performed chi-square tests for allelic and genotypic comparisons under the dominant model in a Han Chinese population with KD residing in Taiwan. Stratification of patients for CAA according to either the right or the left coronary artery identified a dilation diameter ≥ 3 mm in children younger than 5 years of age, or ≥ 4 mm in older children 35,31 . As shown in Table S1, significant clinical factors associated with KD CAA formation were fever duration (p < 0.0001) and time of first IVIG treatment (days after the first day of fever) (p < 0.0001). After adjusting for these potential factors by using multivariate regression analyses, significant associations from genome-wide association tests were observed (Fig. 1). A total of 203 single nucleotide polymorphisms (SNPs) that reached p < 0.005 in the genomewide screen of KD with CAA versus without CAA are shown in Table S2. From these, we selected the top 100 SNPs (that reached p < 0.003) for possible pathway mapping and further functional validation. We explored possible functional relationships between the 100 SNP-related genes by using the Ingenuity Pathway Analysis (IPA) Knowledge Base ( Table 1). The IPA network analysis identified a single cluster of 35 genes that included 26 associated genes discovered in this study (Fig. 2). From among these we selected 4 genes, KCNQ5, PLCB1, PLCB4, and PLCL1, for further functional characterization according to the criteria that at least two SNPs were identified in the GWAS study from the same gene and that their p values reached p < 5 × 10 −4 ( Table 1).
Effect of KCNQ5, PLCB1, PLCB4, and PLCL1 knockdown on IL-1, IL-6, and IL-8 proinflammatory cytokine mRNA expression. EC injury and inflammation are some of the major characteristics of the development of KD 25 . When ECs were stimulated with pathogenic mediators such as Scientific RepoRts | 5:14762 | DOi: 10.1038/srep14762 lipopolysaccharide (LPS), the stimulated cells were shown to trigger inflammatory signals to increase permeability and leukocyte recruitment 32 . In this study, we examined 4 genes, KCNQ5, PLCB1, PLCB4, and PLCL1, for their effect on proinflammatory cytokine mRNA expression by silencing them using siRNA ( Fig. 3 and Supplementary Fig. S1) in an LPS-induced endothelial cell inflammation model. Among them, PLCB1 showed the most significant inhibition of IL-1 beta, IL-6, and IL-8 expression. These results suggest that PLCB1 might regulate endothelial cell inflammation via interference with proinflammatory cytokine expression. This is the first study to report that PLCB1 is a regulator of vascular inflammation and thus its use might be beneficial for many inflammatory diseases associated with endothelial dysfunction.
Plasma levels of phospholipase C (PLC) and the effect of a PLC inhibitor on IL-1, IL-6, and IL-8 proinflammatory cytokine mRNA expression. The molecular interactions between PLCB1, PLCB4, and PLC and their effect on proinflammatory cytokine production have been previously investigated [33][34][35][36][37] . Furthermore, as shown in Fig. 2, the PLCB1-PLC interaction might then regulate the NF-κ B complex and affect the production of IL-1, IL-6, and IL-8 proinflammatory cytokines. Plasma PLC concentrations were therefore measured in 96 patients with KD by enzyme-linked immunosorbent assays (ELISAs). Phospholipase C levels in plasma samples from patients with KD were analyzed in relation to PLCB4/B1 (rs6140791) genotypes (Fig. 4A). In patients with KD, those with CC+ CG genotypes appeared to have significantly higher plasma levels of PLC than did those with the GG genotype (CC+ CG: 738.7; GG: 549.7; p = 0.029). To investigate the effects of PLC inhibitor application on the IL-1, IL-6, and IL-8 proinflammatory cytokine production via the NF-κ B complex, the PLC inhibitor (U73122) was used to treat ECs. LPS was then added to induce an EC inflammation model. As shown in Fig. 4B-D, U73122 mediated significant inhibition of IL-1 beta, IL-6, and IL-8 expression. Thus, the PLC inhibitor U73122 might regulate EC inflammation via interference with proinflammatory cytokine expression. PLCB4/B1 polymorphism and CAA severity. From our genome wide analysis, we identified one locus as being associated with CAA, the rs6140791 SNP located between the PLCB4 and PLCB1 genes on chromosome 20p12. As shown in Table 1, the frequencies of the CC+ CG genotypes of this polymorphism were significantly higher in patients with KD with CAA than in those without CAA (OR = 2.861; 95% CI = 1.597-5.125; p = 4.08 × 10 −4 ). The distributions of the CC+ CG genotypes of this SNP were further analyzed according to CAA severity (Fig. 5). As compared with the KD without CAA group, the risk genotypes (CC+ CG) of the significantly associated SNP rs6140791 were found to be enriched in the patients with KD with CAA remission between 2-12 months and also in those with giant CAA.

Discussion
Genome-wide association analysis has enabled the discovery of novel genetic loci that contribute to KD susceptibility [10][11][12][13][14][15] . More recently, genetic susceptibility for the development of CAA formation has also been investigated using candidate gene analyses 11,16,[21][22][23][24][25][26][27][28][29][30] . These studies have provided us with a preliminary genetic architecture of KD as well as of CAA formation therein. In the present work, we identified susceptibility loci that might contribute to CAA formation in Taiwanese patients with KD and identified  one locus (rs6140791), located between the PLCB4 and PLCB1 genes, which showed likely contribution through functional analysis. Individuals carrying the CC+ CG genotypes of this polymorphism had higher levels of plasma PLC; furthermore, we also observed a greater number of individuals with the CC+ CG genotypes in the group of KD patients with CAA than in those without CAA. In addition, we showed that the effect of LPS-induced EC inflammation was highly significantly inhibited by application of siPLCB1 or PLC inhibitor, suggesting that the PLCB4/B1 genes might play a role in the CAA pathogenesis of KD. Genome-wide association studies have been applied in children with KD to identify CAA-associated loci 38,39 . Lin and colleagues performed a genome-wide association study using patients with KD with extremely large aneurysms (diameter > 8 mm; n = 64) and those without CAA (n = 70) 38 . Their work led to the discovery that the genetic variant rs2833195 in the intron of the TIAM1 gene was associated with the development and severity of CAA in KD. Their results implied that the TIAM1 protein might be required for chemokine-induced T-cell migration and the lymphocyte infiltration into the arterial wall during the acute stage of KD. In addition, Kim and colleagues found that the genetic variant rs17136627 of the KCNN2 gene was associated with CAA development in KD also by using patients with KD with only very large aneurysms (diameter > 5 mm; n = 17) and those without CAA (n = 123) for their KD CAA genome-wide screen 39 . They suggested that the as the KCNN2 gene is highly expressed in arterial myocytes it might be associated with heart disease such as arrhythmias. However, the biological role of the KCNN2 gene in KD pathogenesis remains to be elucidated. Because of limited sample numbers and lack of functional analyses, the significance of these newly identified CAA-associated loci remain inconclusive. In the present work, we analyzed genome-wide screen data generated from a larger sample size. We also adjusted the potential genetic factors for clinical factors such as fever duration and the time of first IVIG use as these factors have been previously reported to be associated with CAA formation in     KD [40][41][42] . From this analysis, we identified CAA susceptibility loci in a Taiwanese cohort of patients with KD; we further characterized these novel genes, and identified PLCB4 and PLCB1 as important regulators of proinflammatory cytokine expression in vascular ECs. Patients with KD tend to have abnormal serum lipid profiles and are associated with important abnormalities in lipid metabolism [43][44][45][46] . Serum lipid profiles such as triglycerides, total cholesterol, high-density lipoprotein-cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) levels have been used as predictors of atherosclerosis, and have also been associated with cardiovascular diseases. As previously reported, patients with KD and coronary abnormalities have decreased HDL-C levels compared with those without such aneurysms 45,46 . Furthermore, another study reported that the levels of total cholesterol, LDL-C, and apolipoprotein B were higher in patients with KD and coronary abnormalities than in those without such aneurysms 44 . In the present work, we found 100 susceptible loci that were associated with KD CAA. Among them, we observed that a single cluster of 35 genes was associated with the top networks, i.e., embryonic development, nervous system development and function, organ development, and lipid metabolism by using the Ingenuity Pathway Analysis (IPA) Knowledge Base (Table S3). Our KD CAA genome-wide findings suggest that these potential CAA-associated gene variants might affect not only the severity of vasculitis during the acute stage of KD, but also might be related to lipid metabolism after KD illness. Children with these potential gene susceptibility variants might have high lipid profiles and arterial stiffness subsequent to KD, indicating an increased risk for cardiovascular disease.
In the present work, we chose 4 lipid metabolism associated genes, KCNQ5, PLCB1, PLCB4, and PLCL1, from among the highlighted gene cluster for further functional investigation. We used an LPS-induced EC inflammation model and knockdown of KCNQ5, PLCB1, PLCB4, and PLCL1 by siRNA and found that these genes might regulate EC inflammation via interference with proinflammatory cytokine expression. Of these, the PLCB1 gene showed the most significant inhibition. Consistent with this, our results also showed that EC inflammation was inhibited by using a PLC inhibitor as well. PLCB1 is expressed in cardiomyocytes, and overexpression of this protein causes cardiomyocyte hypertrophy. For synovial fibroblasts, macrophages, dendritic cells, and ECs, inflammation induced by thrombin, bradykinin, LPS, porphyromonas gingivalis, or advanced glycation end products has been shown to be affected by PLC inhibitors, suggesting that PLC might also be involved in NF-κ B activation and proinflammatory cytokine release in inflammatory diseases [33][34][35][36][37] . The SNP rs6140791 is located between the PLCB4 and PLCB1 genes. KD patients carrying the CC+ CG genotypes of this polymorphism appeared to have higher plasma PLC levels than did those with the GG genotype. In addition, the KD with CAA group tended to have more individuals with CC+ CG genotypes than did the group without CAA. The CC+ CG genotypes were further enriched in the patients with KD who exhibited CAA remission between 2-12 months and also in the patients with KD and giant CAA, suggesting that individuals carrying CC+ CG Figure 5. Distribution of the rs6140791 risk genotype frequency in the PLCB4/B1 gene according to CAA severity. CAA was classified as an > 3 or > 4 mm increase in coronary artery diameter in children under or over 5 years of age during the first 2 months of KD diagnosis, respectively 31,55 . CAA severity grade: KD without CAA indicates patients with no CAA complications; KD with CAA (remission between 2-6 months) indicates patients with CAA, but who showed remission between 2-6 months after KD illness; KD with CAA (remission between 6-12 months) indicates patients with CAA, but who showed remission between 6-12 months after KD illness; KD with CAA (giant CAA) indicates patients with giant CAA (≥ 8 mm) or severe stenosis or occlusion. The genotypes of rs6140791 are shown according to CAA severity, and the number of patients in each category are indicated. genotypes might have higher expression levels of PLC. In turn, people with higher levels of PLC might have more severe inflammations and more potential to develop CAA in KD. This is the first study to report that PLCB1 is a regulator of vascular inflammation and to suggest that its use might therefore be beneficial for the management of inflammatory diseases associated with endothelial dysfunction.
KD is the leading cause of acquired cardiovascular diseases among children with a possible underlying pathology of immune-mediated vasculitis 1,47 . Endothelial cell injury and inflammation play important roles in the development of KD as well as in CAA formation 25 . When ECs were induced by pathogenic mediators including LPS, the cells were shown to activate inflammatory signals to increase permeability and leukocyte recruitment 32 . In this study, functional investigation suggested that RNA silencing of PLC genes or pharmacologic inhibition of PLC downregulated EC inflammation. The PLC gene, PLCB1, consists of 36 exons and localizes to 20p12. PLCB1 codes for the PLC beta 1 protein, which represents one of the four PLC beta isoforms (PLC beta 1-4), has two transcript variants comprised of 1216 and 1173 residues that produce different proteins. PLC beta plays important roles in cardiomyocytes, synovial fibroblasts, macrophages, dendritic cells, and endothelial cells [33][34][35][36][37]48,49 . Furthermore, genetic variants in PLCB4/PLC1 have been significantly associated with several phenotypic traits including levels of apolipoprotein B, cholesterol and HDL, along with body weight, body mass index and stroke (Table S4). For cardiomyocytes, PLC beta is the immediate downstream target of Gα q, a heterotrimeric G protein that regulates its activation, and hydrolyzes the plasma membrane phosphatidylinositol 4,5-bisphosphate (PIP2) protein to produce the second messengers inositol 1,4,5-triphosphate (IP3; a regulator of the intracellular calcium response) and diacylglycerol (DAG, an activator of protein kinase C subtypes) 50 . Studies have reported that PLC beta might transmit the cardiac hypertrophy signal initiated by Gα q 48,51 . Our findings are also in agreement with these studies. Our results demonstrated that the effect of LPS-induced EC inflammation was inhibited with the most significant inhibition by using siPLCB1 or a PLC inhibitor, suggesting that PLCB1 might be involved in EC inflammation, a possibility that has important implications for understanding the pathogenesis of immune-mediated vascular diseases.
In conclusion, we have identified a gene locus that is associated with a genetic predisposition to the development of CAA in KD in Taiwanese children of Han Chinese ethnic background. Functional analysis also supports the possibility that PLCB1 might predispose patients with KD to CAA formation.

Methods
Study population. All experiments were performed in accordance with the relevant guidelines and regulations. This study was approved by the Human Studies Committee of China Medical University Hospital. Written informed consent was obtained from the participants or their parent or legal guardian. Two hundred and sixty-two unrelated patients fulfilling the KD diagnostic criteria were identified and enrolled from the Children's Hospital of China Medical University in Taichung, Taiwan 25,42,[52][53][54] . There were 76 patients with KD with CAA complications and 186 patients with KD without CAA complications, with an average age at onset of 1.86 ± 1.78 and 1.70 ± 1.51 years old, respectively (Table S1). All patients were diagnosed according to KD criteria 55,56 , including fever lasting 5 days or more and at least 4 of the following symptoms: (1) changes in extremities (e.g., erythema, edema, or desquamation), (2) bilateral conjunctivitis, (3) polymorphous rash, (4) cervical lymphadenopathy, and (5) changes in the lips or oral cavity (e.g. pharyngeal erythema, dry/fissured or swollen lips, or "strawberry tongue"). All patients with KD received IVIG treatment and had regular echocardiographic examinations during the 1-year follow-up period. Echocardiographic examinations were completed during the acute stage, at 2 and 6 months after onset, and once per year thereafter. CAA was identified when either the right or the left coronary artery showed a dilated diameter ≥ 3 mm in children younger than 5 years of age, or ≥ 4 mm in older children 31,55 , regardless of whether the CAA went into remission between 2-12 months after KD illness. Only Han Chinese individuals, who account for 98% of Taiwanese residents, were recruited. The ethnic background was assigned based on the results of self-reported questionnaires. The characteristics and clinical profiles of patients with KD enrolled in the study are summarized in Table S1. Statistically significant differences were observed for fever duration (p < 0.0001), first time of IVIG treatment (days after the first day of fever) (p < 0.0001), and efficacy of the first IVIG treatment (p = 0.002). No significant differences were observed in our study for laboratory test results obtained at the acute stage within 24 h before IVIG treatment and for KD status within 3-7 days after IVIG treatment.
Genotyping and quality control. Genomic DNA was extracted from patient blood according to standard protocols by using the Qiagen Genomic DNA Isolation Kit (Qiagen, Taichung, Taiwan). For the genome-wide screen, each sample was genotyped by the National Genotyping Center at Academia Sinica (Taipei, Taiwan) using the Affymetrix Genome-Wide Human SNP Array 6.0, which contains a total of 906,600 SNPs, according to the manufacturer's procedure. Genotype data were quality controlled and SNPs were excluded for further analysis if (1) only one allele appeared in patients; (2) the total call rate was less than 95%; (3) the total minor allele frequency was less than 5%; or (4) the SNP significantly departed from Hardy-Weinberg equilibrium proportions (p < 0.05).
Statistical analysis. Data were expressed as indicated for continuous variables (Table S1). Chi-squared tests were used to identify differences in categorical variables, and ORs and 95% CIs were calculated for the factors under consideration. Genotypes were obtained by direct counting followed by forward stepwise multivariate regression analyses for adjusting clinical factors (fever duration and first time of IVIG use) ( Table 1). All statistical analyses were performed using SPSS v12.0 for Windows (IBM, Armonk, NY, USA). Endothelial cell inflammation assay. To measure endothelial cell inflammation, HUVECs were aliquoted in 6-well plates. Cells were transfected with the desired siRNAs or siNCs as controls using Lipofectamine 2000 (Invitrogen). The transfected cells were then treated with 50 μ g/mL LPS for another 24 h. Cellular RNA extraction and real-time reverse transcription (RT)-PCR analyses were performed as described below.
In vivo and in vitro PLC analysis. Serum PLC levels in patients with KD were assessed using a Human Phospholipase C (PLC) Enzyme-linked Immunosorbent Assay (ELISA) kit (Cat. No: E0788Hu) obtained from Shanghai Crystal day Biotech Co., Ltd. (Shanghai, China) according to manufacturer instruction (http://www.bt-laboratory.com/admin/upload/201212139414181958.pdf). For determination of the in vitro effects of the PLC inhibitor U73122 (Cat. No: U 6756, Sigma), HUVEC cells were treated with10 μ M for 24 h at 37 °C followed by 50 μ g/mL LPS for another 24 h., followed by qRT-PCR analysis.