人吸入性损伤与循环炎症蛋白之间因果关系的双样本孟德尔随机化分析

Two-sample Mendelian randomization analysis of the causal relationship between human inhalation injury and circulating inflammatory proteins

目的探究人吸入性损伤与循环炎症蛋白之间的因果关系。方法该研究为基于双样本孟德尔随机化(MR)分析的研究。以吸入性损伤为暴露因素、循环炎症蛋白为结局,从全基因组关联分析数据库中获得吸入性损伤(216993个样本)和91种循环炎症蛋白(14824个样本)的数据,采用双样本MR分析方法进行分析。根据连锁不平衡分析获得与吸入性损伤显著相关的独立位点单核苷酸多态性(SNP)并将其作为工具变量,主要采用逆方差加权(IVW)法进行吸入性损伤与91种循环炎症蛋白之间因果关系的分析,进一步使用加权中位数法、加权模式法、MR-Egger法和简单模式法进行验证。根据前述IVW法分析结果,针对符合假设的吸入性损伤SNP,进行Cochran Q检验评估异质性,进行MR-Egger回归检验、MR-PRESSO离群值检验评估水平多效性,进行留一法分析评估可靠性。结果筛选出6个达到显著阈值(P<5×10^(-5))的SNP作为代表吸入性损伤的工具变量,其F值均>10,提示均为强相关工具变量。基于6个吸入性损伤SNP,IVW法分析显示,吸入性损伤与白细胞介素20(IL-20)、IL-20受体亚基α(IL-20RA)、IL-5、肿瘤坏死因子受体超家族成员9(TNFRSF9)之间均存在显著因果关系(比值比分别为1.01、1.01、1.02、1.01,95%置信区间分别为1.00~1.02、1.00~1.03、1.01~1.03、1.00~1.03,P<0.05)。经加权中位数法和MR-Egger法验证,吸入性损伤与IL-5(比值比分别为1.02、1.03,95%置信区间分别为1.00~1.04、1.01~1.04,P<0.05)、TNFRSF9(比值比分别为1.02、1.03,95%置信区间分别为1.00~1.04、1.01~1.04,P<0.05)之间均存在显著因果关系;经加权模式法和简单模式法验证,吸入性损伤与IL-20、IL-20RA、IL-5和TNFRSF9之间的因果关系不明显(P值均>0.05),仍需以IVW法结果为准。根据前述IVW法分析结果,Cochran Q检验评估显示,与IL-20、IL-20RA、IL-5和TNFRSF9存在显著因果关系的6个吸入性损伤SNP均不存在显著异质性(Q值分别为2.67、5.00、5.17、5.29,P>0.05);MR-Egger回归检验、MR-PRESSO离群值检验评估显示,与IL-20、IL-20RA、IL-5和TNFRSF9存在显著因果关系的6个吸入性损伤SNP均不存在显著水平多效性(截距分别为0.01、<0.01、-0.02、-0.03,RSSobs值分别为3.33、9.00、7.88、7.26,P>0.05);留一法分析显示,吸入性损伤与IL-20、IL-20RA、IL-5和TNFRSF9之间的显著因果关系在逐个剔除6个吸入性损伤SNP后结果稳定可靠。结论通过双样本MR分析,明确吸入性损伤与4种循环炎症蛋白IL-20、IL-20RA、IL-5和TNFRSF9存在显著因果关系,提示发生吸入性损伤后以上4种循环炎症蛋白的生成呈增多趋势。

ObjectiveTo explore the causal relationship between human inhalation injury and circulating inflammatory proteins.MethodsThis research was based on two-sample Mendelian randomization(MR)analysis.With inhalation injury as the exposure factor and circulating inflammatory proteins as the result,data on inhalation injury(216993 samples)and 91 circulating inflammatory proteins(14824 samples)were obtained from the genome-wide association study database,and analysis was conducted by two-sample MR analysis methods.Based on linkage disequilibrium analysis,independent site single nucleotide polymorphisms(SNPs)that were significantly associated with inhalation injury were identified as the instrumental variables.The inverse variance weighted(IVW)method was mainly used to analyze the causal relationship between inhalation injury and 91 circulating inflammatory proteins,which were further verified using the weighted median method,weighted pattern method,MR-Egger method,and simple pattern method.Based on the aforementioned IVW method analysis results,SNPs of inhalation injury conformed to the hypothesis were subjected to Cochran's Q test for heterogeneity assessment,the MR-Egger regression test and MR-PRESSO outlier test for assessment of horizontal pleiotropy,and the leave-one-out method analysis for reliability assessment.ResultsSix SNPs with a significant threshold(P<5×10^(-5))were identified as representative instrumental variables of inhalation injury,with F values greater than 10,indicating strong correlated instrumental variables.Based on the 6 inhalation injury SNPs,the IVW method analysis revealed a significant causal relationship between inhalation injury and interleukin-20(IL-20),IL-20 receptor subunit alpha(IL-20RA),IL-5,and tumor necrosis factor receptor superfamily member 9(TNFRSF9),with odds ratios of 1.01,1.01,1.02,and 1.01,respectively,and 95%confidence intervals of 1.00-1.02,1.00-1.03,1.01-1.03,and 1.00-1.03,respectively,P<0.05.Verification through the weighted median method and MR-Egger method confirmed that the causal relationships between inhalation injury and IL-5(with odds ratios of 1.02 and 1.03,respectively,confidence intervals of 1.00-1.04 and 1.01-1.04,respectively,P<0.05)as well as TNFRSF9(with odds ratios of 1.02 and 1.03,respectively,confidence intervals of 1.00-1.04 and 1.01-1.04,respectively,P<0.05)were statistically significant.Conversely,verification through the weighted pattern method and simple pattern method indicated that the causal relationships between inhalation injury and IL-20,IL-20RA,IL-5,and TNFRSF9 were not statistically significant(with all P values>0.05),thus still needing IVW method results as standards.Based on the aforementioned IVW method analysis results,the Cochran's Q test demonstrated there was no significant heterogeneity in the 6 inhalation injury SNPs that had significant causal relationships with IL-20,IL-20RA,IL-5,and TNFRSF9(with Q values of 2.67,5.00,5.17,and 5.29,respectively,P>0.05);assessments using the MR-Egger regression test along with MR-PRESSO outlier test showed that none of the 6 inhalation injury SNPs that had significant causal relationships with IL-20,IL-20RA,IL-5,and TNFRSF9 had significant horizontal pleiotropy(with intercepts of 0.01,<0.01,-0.02,and-0.03,respectively,RSSobs values of 3.33,9.00,7.88,and 7.26,respectively,P>0.05);the leave-one-out method analysis showed that the significant causal relationship between inhalation injury and IL-20,IL-20RA,IL-5,and TNFRSF9 was stable and reliable after removing the 6 inhalation injury SNPs one by one.ConclusionsThrough two-sample MR analysis,it is clear that there is a significant causal relationship between inhalation injury and four circulating inflammatory proteins,namely IL-20,IL-20RA,IL-5,and TNFRSF9,suggesting the production of the above four circulating inflammatory proteins is in an increasing trend following inhalation injury.