摘要目的:探讨人免疫细胞表型与瘢痕疙瘩之间的因果关系。方法:该研究为基于双样本孟德尔随机化（MR）分析的研究。以人免疫细胞表型为暴露因素，以瘢痕疙瘩为结局，从全基因组关联分析数据库中获取免疫细胞表型（3 757个样本）与瘢痕疙瘩（668个样本）的数据。将与免疫细胞表型显著相关的单核苷酸多态性（SNP）作为工具变量并排除弱工具变量的影响，应用双样本MR分析评估731种人免疫细胞表型与瘢痕疙瘩之间的因果关系。采用逆方差加权（IVW）法推断因果关系，使用MR-Egger法、加权模式法和加权中位数法进行验证。针对符合假设的免疫细胞表型SNP，利用Cochran Q检验评估异质性，利用MR-Egger回归检验及MR-PRESSO离群值检验评估水平多效性。结果:筛选出18 204个达到显著阈值（ P<1×10 -5）的SNP作为代表731种免疫细胞表型的工具变量，这些SNP均不是弱工具变量（ F值均>10）。根据IVW法，确定了21种与瘢痕疙瘩存在潜在因果关系的免疫细胞表型，其中CD62L -单核细胞绝对计数、初始-成熟B细胞上的CD19表达、IgD +B细胞上的CD19表达、浆细胞前体（浆母细胞）上的CD27表达、CD62L +髓系树突状细胞上的CD86表达、自然杀伤T细胞上的CD45表达、CD39 +CD4 +调节性T细胞上的CD25表达、来源于单核髓系的抑制细胞上的CD45表达、效应记忆CD8 +T细胞上的CD8表达、静息CD4 +调节性T细胞上的CD45RA表达与瘢痕疙瘩均呈显著正相关（比值比分别为1.12、1.09、1.08、1.21、1.13、1.12、1.17、1.11、1.10、1.07，95%置信区间分别为1.03~1.23、1.02~1.16、1.01~1.15、1.06~1.38、1.02~1.25、1.01~1.24、1.03~1.33、1.00~1.23、1.00~1.20、1.01~1.13， P<0.05），激活并分泌的CD4 +调节性T细胞绝对计数、未转换记忆B细胞上的CD25表达、浆细胞样树突状细胞绝对计数、来源于单核髓系的抑制细胞上的CD14表达、自然杀伤T细胞上的CD8表达、IgD +CD38 +B细胞上的CD20表达、CD11c +CD62L -单核细胞绝对计数、CD66b ++髓系细胞绝对计数、粒细胞上的CD11c表达、CD14 +CD16 +单核细胞上的CD14表达、中心记忆CD8 +T细胞上的CD3表达与瘢痕疙瘩均呈显著负相关（比值比分别为0.95、0.93、0.93、0.93、0.91、0.89、0.89、0.88、0.87、0.86、0.85，95%置信区间分别为0.90~1.00、0.87~0.99、0.88~0.99、0.87~0.99、0.84~1.00、0.81~0.98、0.81~0.98、0.79~0.99、0.78~0.96、0.75~0.99、0.74~0.96， P<0.05）。经MR-Egger法验证，CD39 +CD4 +调节性T细胞上的CD25表达、CD62L +髓系树突状细胞上的CD86表达、IgD +B细胞上的CD19表达、静息CD4 +调节性T细胞上的CD45RA表达、中心记忆CD8 +T细胞上的CD3表达与瘢痕疙瘩存在潜在因果关系（比值比分别为1.32、1.22、1.11、1.09、0.73，95%置信区间分别为1.03~1.70、1.04~1.44、1.02~1.21、1.01~1.19、0.55~0.95， P<0.05）；经加权中位数法验证，自然杀伤T细胞上的CD45表达、激活并分泌的CD4 +调节性T细胞绝对计数、IgD +CD38 +B细胞上的CD20表达、CD66b ++髓系细胞绝对计数与瘢痕疙瘩存在潜在因果关系（比值比分别为1.15、0.93、0.87、0.83，95%置信区间分别为1.01~1.31、0.86~1.00、0.77~0.98、0.71~0.96， P<0.05），其中IgD +CD38 +B细胞上的CD20表达与瘢痕疙瘩的潜在因果关系另得到加权模式法的验证（比值比为0.86，95%置信区间为0.77~0.97， P<0.05）。根据前述IVW法分析结果，评估显示，与瘢痕疙瘩存在显著因果关系的21种免疫细胞表型的SNP均不存在显著异质性（ P>0.05）或显著水平多效性（ P>0.05）。 结论:从遗传学角度揭示了21种人免疫细胞表型和瘢痕疙瘩之间存在潜在因果关系，其中10种免疫细胞表型可能是瘢痕疙瘩的危险因素，11种免疫细胞表型可能是瘢痕疙瘩的保护因素。

abstractsObjective:To explore the causal relationship between human immune cell phenotypes and keloids.Methods:This study was based on a two-sample Mendelian randomization (MR) analysis. Human immune cell phenotypes were considered as the exposure factors, and keloid was the outcome. Data on immune cell phenotypes (3 757 samples) and keloids (668 samples) were obtained from the genome-wide association study database. Using single nucleotide polymorphisms (SNPs) significantly associated with immune cell phenotypes as instrumental variables with the influence of weak instrumental variables being excluded, two-sample MR analysis was employed to evaluate the causal relationship between 731 human immune cell phenotypes and keloids. The inverse variance weighted (IVW) method was used to infer causal relationships, and the MR-Egger, weighted median, and weighted mode methods were used for validation. For SNPs of immune cell phenotypes meeting the hypothesis, the Cochran Q test was used to assess heterogeneity, and the MR-Egger regression and MR-PRESSO outlier tests were used to evaluate horizontal pleiotropy.Results:A total of 18 204 SNPs meeting the significant threshold ( P<1×10??) were selected as instrumental variables for 731 immune cell phenotypes, and none of these SNPs were weak instrumental variables (with F values all >10). According to the IVW method, 21 immune cell phenotypes were identified with potential causal relationships to keloids, among which the CD62L - monocyte absolute count, CD19 on naive-mature B cell, CD19 on IgD + B cell, CD27 on plasma blast-plasma cell, CD86 on CD62L + myeloid dendritic cell, CD45 on natural killer T cell, CD25 on CD39 + CD4 + regulatory T cell, CD45 on monocytic myeloid-derived suppressor cells, CD8 on effector memory CD8 + T cell, and CD45RA on resting CD4 + regulatory T cell showed significant positive correlations with keloids (with odds ratios of 1.12, 1.09, 1.08, 1.21, 1.13, 1.12, 1.17, 1.11, 1.10, and 1.07, respectively, 95% confidence intervals of 1.03-1.23, 1.02-1.16, 1.01-1.15, 1.06-1.38, 1.02-1.25, 1.01-1.24, 1.03-1.33, 1.00-1.23, 1.00-1.20, and 1.01-1.13, respectively, P<0.05), while the activated and secreted CD4 + regulatory T cell absolute count, CD25 on unswitched memory B cell, plasmacytoid dendritic cell absolute count, CD14 on monocytic myeloid-derived suppressor cells, CD8 on natural killer T cell, CD20 on IgD + CD38 + B cell, CD11c + CD62L - monocyte absolute count, CD66b ++ myeloid cell absolute count, CD11c on granulocytes, CD14 on CD14 + CD16 + monocyte, and CD3 on central memory CD8 + T cell showed significant negative correlations with keloids (with odds ratios of 0.95, 0.93, 0.93, 0.93, 0.91, 0.89, 0.89, 0.88, 0.87, 0.86, and 0.85, respectively, 95% confidence intervals of 0.90-1.00, 0.87-0.99, 0.88-0.99, 0.87-0.99, 0.84-1.00, 0.81-0.98, 0.81-0.98, 0.79-0.99, 0.78-0.96, 0.75-0.99, and 0.74-0.96, respectively, P<0.05). MR-Egger method confirmed the potential causal relationship existing respectively between CD25 on CD39 + CD4 + regulatory T cell, CD86 on CD62L + myeloid dendritic cell, CD19 on IgD + B cell, CD45RA on resting CD4 + regulatory T cell, CD3 on central memory CD8 + T cell and keloids (with odds ratios of 1.32, 1.22, 1.11, 1.09, and 0.73, respectively, 95% confidence intervals of 1.03-1.70, 1.04-1.44, 1.02-1.21, 1.01-1.19, and 0.55-0.95, respectively, P<0.05). The weighted median method confirmed the potential causal relationship existing respectively between CD45 on natural killer T cell, activated and secreted CD4 + regulatory T cells absolute count, CD20 on IgD + CD38 + B cell, CD66b ++ myeloid cell absolute count and keloids (with odds ratios of 1.15, 0.93, 0.87, and 0.83, respectively, 95% confidence intervals of 1.01-1.31, 0.86-1.00, 0.77-0.98, and 0.71-0.96, respectively, P<0.05). Among them, the potential causal relationship between CD20 on IgD + CD38 + B cell and keloids was further verified by the weighted mode method (with odds ratio of 0.86, 95% confidence interval of 0.77-0.97, P<0.05). According to the aforementioned IVW method analysis results, the SNPs associated with the 21 immune cell phenotypes that had a significant causal relationship with keloids showed no significant heterogeneity ( P>0.05) or significant horizontal pleiotropy ( P>0.05). Conclusions:From a genetic perspective, the potential causal relationships between 21 human immune cell phenotypes and keloids have been revealed, of which 10 immune cell phenotypes may be risk factors for keloids, while 11 immune cell phenotypes may act as protective factors for keloids.