Mendelian randomization analysis of the causal relationships between human inflammatory proteins and keloids
-
摘要:
目的 探讨人炎症蛋白与瘢痕疙瘩之间的因果关系。 方法 该研究为基于孟德尔随机化（MR）分析的研究。以人炎症蛋白为暴露因素、瘢痕疙瘩为结局,从全基因组关联分析数据库中获取91种炎症蛋白（14 824个样本）与瘢痕疙瘩（668个样本）的数据。设置显著阈值,筛选与炎症蛋白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。针对单个工具变量的分析,采用Wald比率法；针对多个工具变量的分析,采用逆方差加权（IVW）法为主要方法,加权中位数法、简单模式法、加权模式法和MR-Egger法作为补充方法,行双样本MR分析评估炎症蛋白与瘢痕疙瘩之间的因果关系。采用IVW法、加权中位数法和MR-Egger法对前述双样本MR分析中具有统计学意义的炎症蛋白进行多样本MR（MVMR）分析,验证这些炎症蛋白与瘢痕疙瘩之间的独立因果关系。针对符合假设的炎症蛋白的SNP,利用Cochran Q检验评估异质性,采用MR-Egger回归检验和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75种炎症蛋白符合暴露因素条件,达到显著阈值的SNP数量从1个到7 082个不等（F值均>10）,提示该研究存在弱工具变量偏倚的可能性较小。IVW法分析显示,真核翻译起始因子4E结合蛋白1（4E-BP1）、CD5及骨保护素与瘢痕疙瘩之间均存在显著因果关系（比值比分别为0.50、0.61、0.71,95%置信区间分别为0.32~0.77、0.41~0.89、0.52~0.97,P<0.05）；经加权中位数法验证,CD5与瘢痕疙瘩之间存在显著因果关系（比值比为0.61,95%置信区间为0.38~0.97,P<0.05）；经简单模式法、加权模式法和MR-Egger法验证,CD5和骨保护素与瘢痕疙瘩之间因果关系均不明显（P>0.05）。Wald比率法分析显示,程序性死亡受体配体1（PD-L1）与瘢痕疙瘩之间存在显著因果关系（比值比为1.83,95%置信区间为1.06~3.15,P<0.05）。仍需以IVW法结果为准。经IVW法验证,4E-BP1、CD5、骨保护素和PD-L1与瘢痕疙瘩之间均存在显著因果关系（比值比分别为0.43、0.58、0.70、1.95,95%置信区间分别为0.28~0.67、0.39~0.86、0.51~0.95、1.16~3.27,P<0.05）,经MR-Egger法验证,4E-BP1和CD5与瘢痕疙瘩之间均存在显著因果关系（比值比分别为0.41、0.58,95%置信区间分别为0.22~0.77、0.39~0.88,P<0.05）；经加权中位数法验证,4E-BP1和PD-L1与瘢痕疙瘩之间均存在显著因果关系（比值比分别为0.46、2.06,95%置信区间分别为0.26~0.82、1.11~3.81,P<0.05）。Cochran Q检验评估显示,与瘢痕疙瘩之间存在显著因果关系的CD5和骨保护素的SNP均不存在显著异质性（P>0.05）；MR-Egger回归检验和MR-PRESSO离群值检验评估显示,与瘢痕疙瘩之间存在显著因果关系的CD5和骨保护素的SNP均不存在显著水平多效性（P>0.05）。留一法分析显示,CD5和骨保护素与瘢痕疙瘩之间的显著因果关系在逐个剔除SNP后结果稳定可靠。 结论 双样本MR分析和MVMR分析明确4E-BP1、CD5及骨保护素与瘢痕疙瘩之间存在显著因果关系,并对瘢痕疙瘩有保护作用。 Abstract:Objective To explore the causal relationships between human inflammatory proteins and keloids. Methods This study was based on Mendelian randomization (MR) analysis. Human inflammatory proteins were considered as the exposure factors, and keloid was considered as the outcome. Data on 91 inflammatory proteins (14 824 samples) and keloids (668 samples) were obtained from the genome-wide association study database. A significance threshold was established to discern single nucleotide polymorphisms (SNPs) significantly associated with inflammatory proteins as instrumental variables with the influence of weak instrumental variables being excluded. For the analysis of a single instrumental variable, the Wald ratio method was used; for the analysis of multiple instrumental variables, the inverse variance weighted (IVW) method was used as the primary method, with the weighted median method, simple mode method, weighted mode method, and MR-Egger method as supplementary methods to employ two-sample MR analysis to analyze the causal relationships between inflammatory proteins and keloids. Using the IVW method, weighted median method, and MR-Egger method to employ multi-sample MR (MVMR) analysis to evaluate the statistically significant inflammatory proteins in the above-mentioned two-sample MR analysis, thus validating their independent causal relationships with keloids. For SNPs of inflammatory proteins conformed to the hypothesis, the Cochran Q test was used to assess heterogeneity, the MR-Egger regression test and MR-PRESSO outlier test were used to evaluate horizontal pleiotropy, and the leave-one-out analysis was performed to assess reliability. Results Seventy-five inflammatory proteins met the exposure factor criteria, with the number of SNPs reaching a significance threshold ranging from 1 to 7 082 (with F values all >10), indicating minimal potential for weak instrumental variable bias in this study. The IVW method analysis revealed significant causal relationships between eukaryotic translation initiation factor 4E-binding protein 1 (4E-BP1), CD5, and osteoprotegerin and keloids (with odds ratios of 0.50, 0.61, and 0.71, respectively, 95% confidence intervals of 0.32-0.77, 0.41-0.89, and 0.52-0.97, respectively, P<0.05); the weighted median method confirmed a significant causal relationship between CD5 and keloids (with odds ratio of 0.61, 95% confidence interval of 0.38-0.97, P<0.05); the simple mode method, weighted mode method, and MR-Egger method confirmed no significant causal relationships between CD5 and osteoprotegerin and keloids (P>0.05). The Wald ratio method analysis revealed a significant causal relationship between programmed death-ligand 1 (PD-L1) and keloids (with odds ratio of 1.83, 95% confidence interval of 1.06-3.15, P<0.05). Thus IVW method results were considered as the standard. The IVW method analysis confirmed that 4E-BP1, CD5, osteoprotegerin, and PD-L1 maintained significant causal relationships with keloids (with odds ratios of 0.43, 0.58, 0.70, and 1.95, respectively, 95% confidence intervals of 0.28-0.67, 0.39-0.86, 0.51-0.95, and 1.16-3.27, respectively, P<0.05). The MR-Egger method confirmed significant causal relationships between 4E-BP1 and CD5 and keloids (with odds ratios of 0.41 and 0.58, respectively, 95% confidence intervals of 0.22-0.77 and 0.39-0.88, respectively, P<0.05). The weighted median method confirmed significant causal relationships between 4E-BP1 and PD-L1 and keloids (with odds ratios of 0.46 and 2.06, respectively, 95% confidence intervals of 0.26-0.82 and 1.11-3.81, respectively, P<0.05). The Cochran Q test assessment indicated no significant heterogeneity in the SNPs of CD5 and osteoprotegerin that had significant causal relationships with keloids (P>0.05). The MR-Egger regression test and MR-PRESSO outlier test showed no significant horizontal pleiotropy in the SNPs of CD5 and osteoprotegerin that had significant causal relationships with keloids (P>0.05). The leave-one-out analysis confirmed that the significant causal relationships between CD5 and osteoprotegerin and keloids remained stable after sequentially removing individual SNP. Conclusions Two-sample MR analysis and MVMR analysis confirmed significant causal relationships between 4E-BP1, CD5, and osteoprotegerin and keloids, all of which are protective factors for keloids. -
Key words:
- Keloid /
- Mendelian randomization analysis /
- Databases, genetic /
- Causality /
- Inflammatory proteins
-
参考文献
(40) [1] LimandjajaGC, NiessenFB, ScheperRJ, et al. The keloid disorder: heterogeneity, histopathology, mechanisms and models[J]. Front Cell Dev Biol,2020,8:360.DOI: 10.3389/fcell.2020.00360. [2] 穆泽兰, 滕泳翔, 张键, 等. 代谢重塑及蛋白酰化修饰在瘢痕疙瘩中的作用研究进展[J]. 中华烧伤与创面修复杂志, 2024, 40(6): 589-593. DOI: 10.3760/cma.j.cn501225-20231207-00229. [3] 贾坤朋,周婧.瘢痕疙瘩发病机制的研究进展[J].皮肤性病诊疗学杂志,2023,30(2):176-180.DOI: 10.3969/j.issn.1674-8468.2023.02.016. [4] JinJ, WangK, LuC, et al. NEDD4L inhibits the proliferation and migration of keloid fibroblasts by regulating YY1 ubiquitination-mediated glycolytic metabolic reprogramming[J]. Exp Dermatol,2024,33(11):e70008.DOI: 10.1111/exd.70008. [5] TangM, BianW, ChengL, et al. Ginsenoside Rg3 inhibits keloid fibroblast proliferation, angiogenesis and collagen synthesis in vitro via the TGF-β/Smad and ERK signaling pathways[J]. Int J Mol Med,2018,41(3):1487-1499.DOI: 10.3892/ijmm.2018.3362. [6] XiaoY, FangH, ZhuY, et al. Multifunctional cationic hyperbranched polyaminoglycosides that target multiple mediators for severe abdominal trauma management[J]. Adv Sci (Weinh),2024,11(1):e2305273.DOI: 10.1002/advs.202305273. [7] NiewczasMA, PavkovME, SkupienJ, et al. A signature of circulating inflammatory proteins and development of end-stage renal disease in diabetes[J]. Nat Med,2019,25(5):805-813.DOI: 10.1038/s41591-019-0415-5. [8] ZhanH, LiH, LiuC,et al. Association of circulating vascular endothelial growth factor levels with autoimmune diseases: a systematic review and meta-analysis[J]. Front Immunol,2021,12:674343.DOI: 10.3389/fimmu.2021.674343. [9] HuangH, FuZ, YangM, et al. Levels of 91 circulating inflammatory proteins and risk of lumbar spine and pelvic fractures and peripheral ligament injuries: a two-sample mendelian randomization study[J]. J Orthop Surg Res,2024,19(1):161.DOI: 10.1186/s13018-024-04637-8. [10] BurgessS, DanielRM, ButterworthAS, et al. Network Mendelian randomization: using genetic variants as instrumental variables to investigate mediation in causal pathways[J]. Int J Epidemiol,2015,44(2):484-495.DOI: 10.1093/ije/dyu176. [11] LarssonSC, ButterworthAS, BurgessS. Mendelian randomization for cardiovascular diseases: principles and applications[J]. Eur Heart J,2023,44(47):4913-4924.DOI: 10.1093/eurheartj/ehad736. [12] KintuC, SoremekunO, KamizaAB,et al. The causal effects of lipid traits on kidney function in Africans: bidirectional and multivariable Mendelian-randomization study[J]. EBioMedicine,2023,90:104537.DOI: 10.1016/j.ebiom.2023.104537. [13] WoolfB, RajasundaramS, CronjéHT, et al. A drug target for erectile dysfunction to help improve fertility, sexual activity, and wellbeing: mendelian randomisation study[J]. BMJ,2023,383:e076197.DOI: 10.1136/bmj-2023-076197. [14] DaviesNM, HolmesMV, Davey SmithG. Reading Mendelian randomisation studies: a guide, glossary, and checklist for clinicians[J]. BMJ,2018,362:k601.DOI: 10.1136/bmj.k601. [15] ZhaoJH, StaceyD, ErikssonN, et al. Genetics of circulating inflammatory proteins identifies drivers of immune-mediated disease risk and therapeutic targets[J]. Nat Immunol,2023,24(9):1540-1551.DOI: 10.1038/s41590-023-01588-w. [16] SakaueS, KanaiM, TanigawaY, et al. A cross-population atlas of genetic associations for 220 human phenotypes[J]. Nat Genet,2021,53(10):1415-1424.DOI: 10.1038/s41588-021-00931-x. [17] 1000 Genomes Project Consortium, AutonA, BrooksLD, et al. A global reference for human genetic variation[J].Nature,2015,526(7571):68-74.DOI: 10.1038/nature15393. [18] 代站站, 朱沁, 佟希睿, 等. 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[J]. 中华烧伤与创面修复杂志, 2024, 40(11): 1043-1051. DOI: 10.3760/cma.j.cn501225-20240429-00155. [19] LevinMG, JudyR, GillD, et al. Genetics of height and risk of atrial fibrillation: a Mendelian randomization study[J]. PLoS Med,2020,17(10):e1003288.DOI: 10.1371/journal.pmed.1003288. [20] DengY, HuangJ, WongMCS. Associations between six dietary habits and risk of hepatocellular carcinoma: a Mendelian randomization study[J]. Hepatol Commun,2022,6(8):2147-2154.DOI: 10.1002/hep4.1960. [21] BurgessS, ButterworthA, ThompsonSG. Mendelian randomization analysis with multiple genetic variants using summarized data[J]. Genet Epidemiol,2013,37(7):658-665.DOI: 10.1002/gepi.21758. [22] BowdenJ, HemaniG, Davey SmithG. Invited commentary:detecting individual and global horizontal pleiotropy in Mendelian randomization-a job for the humble heterogeneity statistic?[J]. Am J Epidemiol, 2018, 187(12):2681-2685.DOI: 10.1093/aje/kwy185. [23] SandersonE, SpillerW, BowdenJ. Testing and correcting for weak and pleiotropic instruments in two-sample multivariable Mendelian randomization[J]. Stat Med,2021,40(25):5434-5452.DOI: 10.1002/sim.9133. [24] HigginsJP, ThompsonSG, DeeksJJ, et al. Measuring inconsistency in meta-analyses[J]. BMJ,2003,327(7414):557-560.DOI: 10.1136/bmj.327.7414.557. [25] 甘文军, 王婧薷, 何佳, 等. 人免疫细胞表型与瘢痕疙瘩之间因果关系的双样本孟德尔随机化分析[J]. 中华烧伤与创面修复杂志, 2025, 41(1): 84-93. DOI: 10.3760/cma.j.cn501225-20231130-00219. [26] VerbanckM, ChenCY, NealeB, et al. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases[J]. Nat Genet,2018,50(5):693-698.DOI: 10.1038/s41588-018-0099-7. [27] 陈文涛, 王小祥, 郑文炼, 等. 基于双样本孟德尔随机化分析探索人肠道菌群与增生性瘢痕之间的因果关系[J]. 中华烧伤与创面修复杂志, 2024, 40(4): 333-341. DOI: 10.3760/cma.j.cn501225-20231129-00215. [28] OgawaR. Keloid and hypertrophic scars are the result of chronic inflammation in the reticular dermis[J]. Int J Mol Sci,2017,18(3):606.DOI: 10.3390/ijms18030606. [29] FerenceBA, HolmesMV, SmithGD. Using Mendelian randomization to improve the design of randomized trials[J]. Cold Spring Harb Perspect Med,2021,11(7):a040980.DOI: 10.1101/cshperspect.a040980. [30] HeY, JiQ, WuZ, et al. 4E-BP1 counteracts human mesenchymal stem cell senescence via maintaining mitochondrial homeostasis[J]. Protein Cell,2023,14(3):202-216.DOI: 10.1093/procel/pwac037. [31] ArmengolG, RojoF, CastellvíJ, et al. 4E-binding protein 1: a key molecular "funnel factor" in human cancer with clinical implications[J]. Cancer Res,2007,67(16):7551-7555.DOI: 10.1158/0008-5472.CAN-07-0881. [32] DongZ, OjhaA, BarlowL, et al. The eIF3a translational control axis in the Wnt/β-catenin signaling pathway and colon tumorigenesis[J]. Cancer Lett,2024,605:217303.DOI: 10.1016/j.canlet.2024.217303. [33] LeeCC, TsaiCH, ChenCH, et al. An updated review of the immunological mechanisms of keloid scars[J]. Front Immunol,2023,14:1117630.DOI: 10.3389/fimmu.2023.1117630. [34] 高景,焦虎,曹蕊,等.雷帕霉素诱导瘢痕疙瘩成纤维细胞自噬的作用及机制[J].中华整形外科杂志,2016,32(3):208-214.DOI: 10.3760/cma.j.issn.1009-4598.2016.03.012. [35] VoisinneG, Gonzalez de PeredoA, RoncagalliR. CD5, an undercover regulator of TCR signaling[J]. Front Immunol,2018,9:2900.DOI: 10.3389/fimmu.2018.02900. [36] BagabirR, ByersRJ, ChaudhryIH, et al. Site-specific immunophenotyping of keloid disease demonstrates immune upregulation and the presence of lymphoid aggregates[J]. Br J Dermatol,2012,167(5):1053-1066.DOI: 10.1111/j.1365-2133.2012.11190.x. [37] DongC. Cytokine regulation and function in T cells[J]. Annu Rev Immunol,2021,39:51-76.DOI: 10.1146/annurev-immunol-061020-053702. [38] MuraoN, SeinoK, HayashiT, et al. Treg-enriched CD4+ T cells attenuate collagen synthesis in keloid fibroblasts[J]. Exp Dermatol,2014,23(4):266-271.DOI: 10.1111/exd.12368. [39] RochetteL, MelouxA, RigalE, et al. The role of osteoprotegerin and its ligands in vascular function[J]. Int J Mol Sci,2019,20(3):705.DOI: 10.3390/ijms20030705. [40] OnoT, HayashiM, SasakiF, et al. RANKL biology: bone metabolism, the immune system, and beyond[J]. Inflamm Regen,2020,40:2.DOI: 10.1186/s41232-019-0111-3. -
图 1 人炎症蛋白与瘢痕疙瘩的孟德尔随机化分析核心假设与分析流程图
注：SNP为单核苷酸多态性；图中虚线表示2个因子之间不存在关联性,实线表示2个因子之间存在关联性；假设1为关联性假设,假设2为独立性假设,假设3为排他性假设
图 2 留一法分析与瘢痕疙瘩之间存在显著因果关系的2种人炎症蛋白SNP的可靠性。2A.CD5；2B.骨保护素
注：SNP为单核苷酸多态性；每个SNP对应的线段为误差线,线段中的圆点代表β值,均在0的左侧,提示结果可靠
Table 1. 4种人炎症蛋白与瘢痕疙瘩之间因果关系的双样本MR分析结果
炎症蛋白与分析方法 SNP数（个） 比值比 95%置信区间 P值 4E-BP1 2 IVW法 0.50 0.32~0.77 0.002 CD5 6 MR-Egger法 0.77 0.07~8.18 0.836 加权中位数法 0.61 0.38~0.97 0.037 IVW法 0.61 0.41~0.89 0.010 简单模式法 0.52 0.28~0.98 0.099 加权模式法 0.70 0.39~1.26 0.292 骨保护素 4 MR-Egger法 0.49 0.04~6.48 0.643 加权中位数法 0.74 0.52~1.05 0.090 IVW法 0.71 0.52~0.97 0.032 简单模式法 0.61 0.35~1.08 0.190 加权模式法 0.76 0.53~1.08 0.224 PD-L1 1 Wald比率法 1.83 1.06~3.15 0.030 注：MR为孟德尔随机化,4E-BP1为真核翻译起始因子4E结合蛋白1,IVW为逆方差加权,PD-L1为程序性死亡受体配体1,SNP为单核苷酸多态性 
下载: 导出CSV
Table 2. 4种人炎症蛋白与瘢痕疙瘩之间因果关系的多样本MR分析结果
炎症蛋白与分析方法 比值比 95%置信区间 P值 F值 4E-BP1 12.26 IVW法 0.43 0.28~0.67 0.001 MR-Egger法 0.41 0.22~0.77 0.006 加权中位数法 0.46 0.26~0.82 0.008 CD5 15.46 IVW法 0.58 0.39~0.86 0.006 MR-Egger法 0.58 0.39~0.88 0.010 加权中位数法 0.63 0.37~1.10 0.105 骨保护素 24.61 IVW法 0.70 0.51~0.95 0.020 MR-Egger法 0.72 0.49~1.06 0.092 加权中位数法 0.85 0.53~1.35 0.488 PD-L1 8.87 IVW法 1.95 1.16~3.27 0.012 MR-Egger法 1.85 0.93~3.67 0.078 加权中位数法 2.06 1.11~3.81 0.021 注：MR为孟德尔随机化,4E-BP1为真核翻译起始因子4E结合蛋白1,IVW为逆方差加权,PD-L1为程序性死亡受体配体1
下载: 导出CSV
Table 3. 与瘢痕疙瘩之间存在显著因果关系的2种人炎症蛋白SNP的异质性与水平多效性分析结果
炎症蛋白与分析方法 SNP数（个） Cochran Q检验 MR-Egger回归检验 MR-PRESSO离群值检验 Q值 P值 截距 P值 RSSobs P值 CD5 6 IVW法 1.13 0.951 — — — — MR-Egger法 1.09 0.895 -0.03 0.856 1.78 0.953 骨保护素 4 IVW法 2.45 0.484 — — — — MR-Egger法 2.35 0.308 0.06 0.801 5.78 0.601 注：SNP为单核苷酸多态性,IVW为逆方差加权,MR为孟德尔随机化；Cochran Q检验评估异质性,另2种检验评估水平多效性；“—”表示无此项
下载: 导出CSV
-
图(3) / 表(3)
计量
- 文章访问数: 73
- HTML全文浏览量: 23
- PDF下载量: 6
- 被引次数: 0
