Study On The Genetic Polymorphism To Serum Amyloid A And Its Correlation With Risk Of Hypertension Among Guangxi Han Population

ObjectiveAs the common pathological basis of various cardiovascular and cerebrovascular diseases,hypertension is still one of the diseases with the highest disease burden and all-cause mortality in the world.Serum amyloid A（SAA）is a circulating inflammatory marker in the acute phase,which has been proved to be an independent risk factor for chronic inflammatory diseases.However,studies on serum SAA and hypertension in large sample population are limited in China.Therefore,this study aimed to explore the genetic polymorphism of serum SAA levels and its effect on hypertension and metabolic related traits via conventional Epidemiology method,low coverage whole genome sequence（lc WGS）,genome-wide association study（GWAS）and Mendelian randomization（MR）,in attempt to provide scientific reference on whether SAA can be used as a potential anti-inflammatory therapeutic target for hypertension,along with a new methodological support for lc WGS technology in large-scale population GWAS studies.MethodsPart one:Association of serum amyloid A and risk of hypertensionStudy participants were obtained from the cohort of the National Key Project,and a total of 1779 eligible participants were included by 1:2 propensity score matching,including 652 hypertensive group and 1127 non-hypertensive group.The levels of serum SAA,IL-1β,IL-23,IL-17 and TNF-αwere determined by immunoturbidimetry and ELISA.LASSO regression,condition Logistic regression,interaction analysis,stratified analysis and mediation effects were performed to detect the correlation and dose-response relationships between serum SAA levels and risk of hypertension and metabolic related traits.Part two:Low coverage whole genome sequence effectively detect common and novel vatiantsWe randomly selected 2000 Han Chinese from the intermediate phenotype database to evaluate the effectiveness of lc WGS and study the genetic susceptibility to serum SAA levels.Firstly,240 subjects were selected for ASA chip detection,and further divided into six groups for 0.5X,1X,2X,4X,6X and30X whole genome sequencing,with 40 cases in each groups,respectively.After obtaining the sequencing data,（1）Downsampling cohort:GATK Downsample Sam module was applied to downsample 30X data to 0.5X,1X,2X,4X and 6X data.Then,taking 30X data as“truth data”,we compared the differences of ASA chip and 0.5X～6X sequencing in terms of the number of genetic variants,non-reference concordance（NRC）and imputation r² before（“raw data”）and after（“imputed data”）imputation.（2）Actual sequencing cohort:we compared the differences of ASA chip and 0.5X～6X sequencing in terms of the number of genetic variants,non-reference concordance（NRC）and imputation r².We completed the analyses by GATK4.0.2,PLINK1.9 and R4.0.2 softwares.Part three:Study on genetic susceptibility to serum amyloid A among Guangxi Han populationBased on the evaluation results of lc WGS in the second part,1X sequencing was performed among the remaining 1760 subjects.We conducted linear regression and FUMA GWAS analysis platform to detect the significant SNPs related to serum SAA(P<5×10^-8).Bioinformatics analyses including GO enrichment analysis,KEGG signaling pathway,and Wiki Pathways were also used to determine the potential biological functions of SAA-related locus.We further adopted two-sample Mendelian randomization（MR）method to assess the causal associations between serum SAA and hypertension and metabolic related traits among East Asian and European populations.ResultsPart one:Association of serum amyloid A and risk of hypertension（1）SAA was selected as the most influential inflammatory factor for hypertension by LASSO regression model.Multivariate analyses indicated that,serum SAA levels were positively associated with hypertension（Q5 vs.Q1,OR=1.30,95%CI:1.07-1.66）,SBP（Q5 vs.Q1,β=4.60,95%CI:2.09-7.11）and DBP（Q5 vs.Q1,β=2.73,95%CI:1.04-4.42）levels,with a linear dose-response relationship(P_overall<0.05,P_nonlinear>0.05),after adjusting for potential confounders.Besides,elevated SAA levels increased the risk of people with hypertension to be cardiovascular high-risk population（Q5 vs.Q1,OR=2.30,95%CI:1.38-3.83）,and the trend test was statistically significant(P_trend<0.001).（2）Multiple linear regression analysis showed that,serum SAA was positively correlated with BMI,FBG,TC and TG levels（P<0.01）,after adjusting for potential confounders.These metabolic traits could further partly mediate the association between serum SAA and SBP/DBP levels,with the mediating effects ranging from 7.2%to 33.9%,and BMI had the greatest mediating effect.（3）Compared with High_DL-c+Low_SAA group,Low_HDL-c+High_SAAgroup had a1.54-fold increased risk of hypertension（95%CI:1.11-2.12）,The trend test was statistically significant(P_trend=0.012).We stratified SAA levels into 80^th（≥5.71mg/L）,no statistical significance was observed between HDL-c and hypertension（P>0.05）.Compared with Low_TG+Low_SAAgroup,High_TG+Low_SAAand High_TG+High_SAA groups increased the risk of hypertension by 1.41 fold（95%CI:1.05-1.89）and 2.01 fold（95%CI:1.49-2.72）,respectively.The trend tests and interaction effects were statistically significant（P<0.05）.（4）Stratified analyses revealed that serum SAA levels were positively associated with hypertension among the participants who were male（Q5 vs.Q1,OR=1.41,95%CI:1.07-1.87）,BMI≥28kg/m²（Q5 vs.Q1,OR=1.58,95%CI:1.17-5.41）,HDL-c<1.04mmol/L（Q5 vs.Q1,OR=2.93,95%CI:1.40-6.11）,18-29years old（Q5 vs.Q1,OR=3.58,95%CI:1.11-11.53）and 30-44 years old（Q5 vs.Q1,OR=2.03,95%CI:1.07-3.85）,and their trend tests were statistical significant(P_trend<0.05).Part two:Low coverage whole genome sequence effectively detect common and novel vatiants（1）“Raw data”of downsampling cohort:Taking 30X data as“truth data”,the overall NRC of ASA chip and 0.5X～6X data were 0.03,0.08,0.18,0.37,0.63and 0.77,respectively.The number and concordance of SNPs,common variants,singletons and novel variants,along with the sensitivity and positive predictive value of variants were increased as the increasing of sequencing depth.（2）“Imputed data”of downsampling cohort:Taking 30X data as“truth data”,the common variants and singletons concordance of ASA chip and 0.5X～6X data were 0.74 and 0.36,respectively,while the overall NRC were 0.28,0.45,0.55,0.67,0.82 and 0.90,respectively.The accuracy of ASA chip and 0.5X～2X sequencing improved dramatically after imputation,but these effects decreased after 4X sequencing.（3）Actual sequencing cohort:0.5X～1X sequencing were cost-effectively outperformed than ASA chip in detecting the number of common and novel variants,singletons,and imputation r²,regardless of impulation or not.Part three:Study on genetic susceptibility to serum amyloid A among Guangxi Han population（1）After strict quality control steps,a total of 1639 Han participants were included for GWAS analysis.There were three genetic predisposition regions were found to be associated with serum SAA levels,accounting for 21.48%of the total variation.The first susceptible region（205.7 kb in length）is located at 11p15(rs11024585,P=1.96×10^-81),which plays a key role in the regulation of baseline SAA levels.The second susceptible region is TENM4 gene(rs11237750,P=1.36×10^-8),where involving in the occurrence of neurological diseases and tumors.The third susceptible region（length 69.4kb）was 17q24(rs73992267,P=2.33×10^-9),which containing the CEP112 gene associated with spermatogenic failure.（2）The results of fine mapping analysis for serum SAA candidated SNPs showed that,SAA1,SAA2,SAA4,HPS5,LDHC,LDHAL6A,TENM4 and CEP112genes were associated with protein synthesis.These genes were up-regulated in liver tissues,and six genes in 11p15 region interacted with protein network.（3）Bioinformatics results demonstrated that SAA-related candidate genes involved in the biological process of acute inflammatory response,cell components of high-density lipoprotein particles and lipid proteins,chemical inducer activity and molecular functions of lactate dehydrogenase activity.Additionally,it also involved in glycolysis and gluconeogenesis metabolism,cysteine and methionine metabolism,propionic acid and pyruvate metabolism,vitamin B12,folic acid and selenium metabolism as well.（4）MR analyses indicated a positive causal relationship of serum SAA and SBP levels among East Asian population（β=0.234,95%CI:0.011-0.457,P=0.040）,while positive causal associations between serum SAA and TC（β=0.012,95%CI:0.003-0.021,P=0.007）and HDL-c（β=0.020,95%CI:0.012-0.028,P=1.78E-6）levels among European population.Sensitivity analyses detected none of horizontal pleiotropic or heterogeneity effects（P>0.05）.Conclusions（1）Serum SAA is positively associated with risk of hypertension,especially in the participants who were male,18-44 years old,BMI≥28kg/m² and HDL-c<1.04mmol/L.Meanwhile,elevated SAA levels could accelerate the risk of people with hypertension to be cardiovascular high-risk population.Besides,BMI,FBG,TC and TG could partly mediate the associations of serum SAA and SBP/DBP levels,suggesting that SAA may influence blood pressure via obesity,glucose and lipid metabolism pathways indirectly.（2）At the same cost-effectiveness with ASA chip,0.5X～1X sequencing combined with impulation are effectively detecting more common and novel variants,singletons,and enhancing the accuracy of identifying true variations.（3）Population heterogeneity exsits in genetic susceptibility to serum SAA,which is manifested in two newly susceptible regions include TENM4 gene and CEP112 gene in the 17q24 region,and serum SAA was positively associated with SBP levels in East Asian population,while positively associated with TC and HDL-c levels in European population.