两样本孟德尔随机化分析血浆磷脂酰乙醇胺水平与结直肠腺癌发病风险的关系

doi:10.3969/j.issn.1674-8115.2025.05.009

摘要/Abstract

摘要：

目的·采用两样本孟德尔随机化（two-sample Mendelian randomization，TSMR）方法，以遗传变异作为工具变量，评估血浆磷脂酰乙醇胺（phosphatidylethanolamine，PE）水平与结直肠腺癌发病风险之间的因果关系。方法·PE与结直肠癌相关的单核苷酸多态性（single-nucleotide polymorphism，SNP）数据分别来自布里斯托大学MRC综合流行病学小组（The Medical Research Council Integrative Epidemiology Unit，MRC-IEU）和芬兰生物银行。对基于全基因组关联研究（genome-wide association study，GWAS）的所有汇总数据进行二次数据分析，选择与PE密切关联的遗传位点作为工具变量，采用4种孟德尔随机化（Mendelian randomization，MR）方法，包括逆方差加权法（inverse-variance weighted，IVW）、孟德尔随机化Egger回归法（Mendelian randomization-Egger，MR-Egger回归）、加权中位数法（weighted median，WME）和加权众数法（weighted mode，WM）分析评估因果效应。IVW作为主要统计方法，MR-Egger、WME和WM作为辅助方法。采用MR-Egger回归截距和MR-PRESSO检验评估水平多效性全局检验是否违反MR假设。Cochran′s Q检验用于评估异质性。结果·筛选出10个工具变量，Steiger检验显示全部PE相关SNP对结直肠癌因果关系方向一致。10个SNP中rs102275和rs9393903位点与结直肠癌风险正相关，关联度分别为0.45（P=8.01×10^-5）和0.82（P=2.31×10^-2）。4种MR分析的结果一致，均显示血浆PE水平与结直肠腺癌的发病风险存在正向关联。IVW中OR为1.36，95%CI 1.17~1.59，P=7.24×10^-5；MR-Egger回归OR为1.44，95%CI 0.97~2.14，P=1.12×10^-1；WME中OR为1.33，95%CI 1.07~1.65，P=8.81×10^-3；WM中OR为1.41，95%CI 1.12~1.77，P=1.70×10^-2。Cochran′s Q检验结果显示10个SNP对结直肠腺癌估计值无异质性。MR-Egger回归截距和MR-PRESSO检验表明全部SNP无水平多效性。采用留一法在剔除任一个SNP后，总体置信区间重叠，表明结果对单个SNP不敏感，具有较高的稳健性。结论·血浆PE水平与结直肠腺癌风险之间存在因果关联，遗传预测的血浆PE每增加1个标准差，结直肠腺癌发生风险就会增加0.36倍（95%CI 1.17~1.59，P=7.24×10^-5）。

关键词: 结直肠腺癌, 磷脂酰乙醇胺, 两样本孟德尔随机化, 因果关联, 全基因组关联研究

Abstract:

Objective ·To employ the two-sample Mendelian randomization (TSMR) method, using genetic variants as instrumental variables, to investigate the causal relationship between phosphatidylethanolamine (PE) and the risk of colorectal adenocarcinoma. Methods ·The single-nucleotide polymorphism (SNP) data associated with PE and colorectal adenocarcinoma were obtained from the Medical Research Council Integrative Epidemiology Unit (MRC IEU) at the University of Bristol and the Finnish Biobank, respectively. A secondary data analysis was conducted using summary statistics from genome-wide association studies (GWAS), and genetic loci strongly associated with PE were selected as instrumental variables. Four Mendelian randomization (MR) methods, inverse-variance weighted (IVW) method, MR-Egger regression, weighted median (WME) method, and weighted mode (WM) method, were employed to assess the causal effect. The IVW method was used as the primary statistical approach, while MR-Egger, WME, and WM served as supplementary methods. Rigorous assessments for robustness included MR-Egger regression, MR-PRESSO global tests for horizontal pleiotropy, and Cochran′s Q test to evaluate heterogeneity. Results ·Ten instrumental variables were selected, and the Steiger test indicated that all PE-associated SNPs exhibited a consistent direction of causal effect on colorectal cancer. Among the 10 SNPs, rs102275 and rs9393903 showed the strongest positive associations with colorectal adenocarcinoma risk, with effect sizes of 0.45 (P=8.01×10^-5) and 0.82 (P=2.31×10^-2), respectively. Consistent findings from MR analyses demonstrated that PE elevated the risk of colorectal adenocarcinoma across all four methods. In the IVW analysis, the OR was 1.36 (95%CI 1.17‒1.59, P=7.24×10^-5). In the MR-Egger regression, the OR was 1.44 (95%CI 0.97‒2.14, P=1.12×10^-1). In the WEM analysis, the OR was 1.33 (95%CI 1.07‒1.65, P=8.81×10^-3). In the WM analysis, the OR was 1.41 (95%CI 1.12‒1.77, P=1.70×10^-2). Cochran′s Q test revealed no heterogeneity among the effect estimates of the 10 SNPs on colorectal adenocarcinoma. Both MR-Egger regression intercept and MR-PRESSO test indicated no evidence of horizontal pleiotropy among the SNPs. Leave-one-out analysis showed overlapping confidence intervals after excluding any single SNP, indicating that the results were not sensitive to individual SNPs and were highly robust. Conclusions ·There is a causal association between circulating PE levels and the risk of colorectal adenocarcinoma. A genetically predicted increase of one standard deviation in plasma PE levels is associated with a 1.36-fold higher risk of developing colorectal adenocarcinoma (95%CI 1.17‒1.59, P=7.24×10^-5).

Key words: colorectal adenocarcinoma, phosphatidylethanolamine, two-sample mendelian randomization, causal association, genome-wide association study

中图分类号:

R735.3

徐苓, 皇甫昱婵, 沈立松, 马妍慧. 两样本孟德尔随机化分析血浆磷脂酰乙醇胺水平与结直肠腺癌发病风险的关系[J]. 上海交通大学学报（医学版）, 2025, 45(5): 605-613.

XU Ling, HUANGFU Yuchan, SHEN Lisong, MA Yanhui. Causal association between plasma phosphatidylethanolamine and risk of colorectal adenocarcinoma: a two-sample Mendelian randomization study[J]. Journal of Shanghai Jiao Tong University (Medical Science), 2025, 45(5): 605-613.

图/表 6

表1 PE用于TSMR分析的工具变量

Tab 1 Instrumental variables for TSMR analysis of PE

SNP	Chr	Position	Effect allele	Other allele	EAF	β	SE	P value	Nearby gene	R²	F value	SteigerP value
rs102275	11	61790331	C	T	0.35	-0.18	0.01	4.40×10^-66	TMEM258	0.020 9	202.64	8.47×10^-57
rs10781444	9	77534256	C	T	0.80	0.06	0.01	3.10×10^-6	GNA14	0.001 6	14.69	1.48×10^-5
rs1087299	4	184305068	C	T	0.33	0.05	0.01	5.41×10^-6	MYL12BP2	0.001 5	14.17	9.06×10^-6
rs11254897	10	6942964	G	A	0.06	0.09	0.02	6.40×10^-6	LOC105376387	0.001 5	13.79	3.20×10^-5
rs1532085	15	58391167	G	A	0.61	-0.11	0.01	3.01×10^-27	ALDH1A2	0.008 4	79.73	2.58×10^-25
rs17395349	2	132439386	A	G	0.05	-0.11	0.02	2.00×10^-6	GPR39	0.001 6	15.38	7.45×10^-6
rs1800588	15	58431476	T	C	0.22	0.14	0.01	1.10×10^-31	LIPC	0.009 8	94.84	9.27×10^-30
rs75933914	1	51296770	A	G	0.03	0.13	0.03	8.70×10^-6	TTC39A LOC124904177	0.001 4	13.63	4.65×10^-5
rs7655751	4	148712269	T	C	0.21	-0.06	0.01	8.30×10^-7	LOC107986195	0.001 9	16.86	4.04×10^-6
rs9393903	6	11042676	A	G	0.23	0.08	0.01	4.00×10^-12	ELOVL2	0.003 5	33.24	3.13×10^-10

表2 单一和组合SNP对结直肠腺癌发病风险的影响

Tab 2 Effect of single and combined SNPs on colorectal adenocarcinoma risk

SNP	β	SE	P value
rs102275	0.45	0.11	8.01×10^-5
rs10781444	0.47	0.44	2.89×10^-1
rs1087299	-0.01	0.44	9.87×10^-1
rs11254897	0.51	0.43	2.42×10^-1
rs1532085	0.13	0.18	4.70×10^-1
rs17395349	-0.41	0.53	4.42×10^-1
rs1800588	0.07	0.16	6.48×10^-1
rs75933914	0.77	0.61	2.01×10^-1
rs7655751	0.48	0.50	3.34×10^-1
rs9393903	0.82	0.36	2.31×10^-2
All IVW	0.31	0.08	7.24×10^-5
All MR-Egger	0.36	0.20	1.12×10^-1

图1 PE相关SNP与结直肠腺癌风险关联的森林图Note： The black dots represent the log ORs for colorectal adenocarcinoma using each PE-associated SNP as an instrumental variable, corresponding to the risk per one standard deviation increase in PE levels. The red dots represent the combined causal estimates, derived using all SNPs as instrumental variables, based on the MR-Egger regression and IVW methods. The horizontal lines represent the 95%CI of the estimates.

Fig 1 Forest plot of the association of PE-associated SNPs with risk of colorectal adenocarcinoma

图2 10个SNP与PE、结直肠腺癌关联的汇总数据估计的散点图Note： Each dot in this graph represents an instrumental variable; the lines at each dot reflect the 95%CI. The x-axis represents the effect of each SNP on the exposure factor (PE), and the y-axis represents the effect on the outcome (colorectal adenocarcinoma). The colored lines represent the Mendelian randomization fit (MR Fit).

Fig 2 Scatter plot of summary data estimates for the associations of 10 SNPs with PE and colorectal adenocarcinoma

表3 TSMR估计血浆PE相关SNP与结直肠腺癌发生风险的因果关系

Tab 3 Causal relationship between SNPs associated with plasma PE and the risk of colorectal adenocarcinoma estimated by TSMR

Outcome	MR method	OR (95%CI)	P value	P value (Cochran′s Q)
Colorectal adenocarcinoma	IVW	1.36 (1.17‒1.59)	7.24×10^-5	0.363
Colorectal adenocarcinoma	MR-Egger	1.44 (0.97‒2.14)	1.12×10^-1	0.283
Colorectal adenocarcinoma	WME	1.33 (1.07‒1.65)	8.81×10^-3	‒
Colorectal adenocarcinoma	WM	1.41 (1.12‒1.77)	1.70×10^-2	‒

图3 敏感性分析结果Note： A. The funnel plot visualized the heterogeneity of each genetic variant. B. Leave-one-out plots for sensitivity analysis of the association between PE and colorectal adenocarcinoma.

Fig 3 Results of sensitivity analysis

参考文献 47

1	SIEGEL R L, MILLER K D, WAGLE N S, et al. Cancer statistics, 2023[J]. CA Cancer J Clin, 2023, 73(1): 17-48.
2	HOFMANOVÁ J, SLAVÍK J, OVESNÁ P, et al. Phospholipid profiling enables to discriminate tumor- and non-tumor-derived human colon epithelial cells: phospholipidome similarities and differences in colon cancer cell lines and in patient-derived cell samples[J]. PLoS One, 2020, 15(1): e0228010.
3	STOICA C, FERREIRA A K, HANNAN K, et al. Bilayer forming phospholipids as targets for cancer therapy[J]. Int J Mol Sci, 2022, 23(9): 5266.
4	KENNY T C, SCHARENBERG S, ABU-REMAILEH M, et al. Cellular and organismal function of choline metabolism[J]. Nat Metab, 2025, 7(1): 35-52.
5	VAN DER VEEN J N, KENNELLY J P, WAN S, et al. The critical role of phosphatidylcholine and phosphatidylethanolamine metabolism in health and disease[J]. Biochim Biophys Acta Biomembr, 2017, 1859(9 Pt B): 1558-1572.
6	NAUDIN S, SAMPSON J N, MOORE S C, et al. Lipidomics and pancreatic cancer risk in two prospective studies[J]. Eur J Epidemiol, 2023, 38(7): 783-793.
7	WANG Y Z, GUO F, GUO Y K, et al. Untargeted lipidomics reveals specific lipid abnormalities in systemic lupus erythematosus[J]. Clin Exp Rheumatol, 2022, 40(5): 1011-1018.
8	BUTLER F M, UTT J, MATHEW R O, et al. Plasma metabolomics profiles in Black and White participants of the Adventist Health Study-2 cohort[J]. BMC Med, 2023, 21(1): 408.
9	JEZIORNY K, PIETROWSKA K, SIEMINSKA J, et al. Serum metabolomics identified specific lipid compounds which may serve as markers of disease progression in patients with Alström and Bardet-Biedl syndromes[J]. Front Mol Biosci, 2023, 10: 1251905.
10	GESTA S, TSENG Y H, RONALD KAHN C. Developmental origin of fat: tracking obesity to its source[J]. Cell, 2007, 131(2): 242-256.
11	XIE P S, ZHANG H N, WU P F, et al. Three-dimensional mass spectrometry imaging reveals distributions of lipids and the drug metabolite associated with the enhanced growth of colon cancer cell spheroids treated with triclosan[J]. Anal Chem, 2022, 94(40): 13667-13675.
12	ELMALLAH M I Y, ORTEGA-DEBALLON P, HERMITE L, et al. Lipidomic profiling of exosomes from colorectal cancer cells and patients reveals potential biomarkers[J]. Mol Oncol, 2022, 16(14): 2710-2718.
13	MIKA A, PAKIET A, CZUMAJ A, et al. Decreased triacylglycerol content and elevated contents of cell membrane lipids in colorectal cancer tissue: a lipidomic study[J]. J Clin Med, 2020, 9(4): 1095.
14	LIU J, DE VRIES P S, DEL GRECO M F, et al. A multi-omics study of circulating phospholipid markers of blood pressure[J]. Sci Rep, 2022, 12(1): 574.
15	CAO Y C, AI M C, LIU C J. The impact of lipidome on breast cancer: a Mendelian randomization study[J]. Lipids Health Dis, 2024, 23(1): 109.
16	FAN Z. A two-sample Mendelian randomization study of lipidome and lung cancer[J]. J Pharm Biomed Anal, 2025, 252: 116514.
17	CHEN Y Q, TORTA F, KOH H W L, et al. Metabolomics profiling in multi-ancestral individuals with type 2 diabetes in Singapore identified metabolites associated with renal function decline[J]. Diabetologia, 2025, 68(3): 557-575.
18	HUANG Y, STINSON S E, THODBERG M, et al. Genetic factors shaping the plasma lipidome and the relations to cardiometabolic risk in children and adolescents[J]. EBioMedicine, 2025, 112: 105537.
19	BURGESS S, DAVEY SMITH G, DAVIES N M, et al. Guidelines for performing mendelian randomization investigations: update for summer 2023[J]. Wellcome Open Res, 2023, 4: 186.
20	HEMANI G, ZHENG J, ELSWORTH B, et al. The MR-Base platform supports systematic causal inference across the human phenome[J]. eLife, 2018, 7: e34408.
21	ELSWORTH B, LYON M, ALEXANDER T, et al. The MRC IEU OpenGWAS data infrastructure[DB/OL]. (2020-08-10)[2024-12-26]. https://www.biorxiv.org/content/10.1101/2020.08.10.244293v1.
22	WOOTTON R E, LAWN R B, MILLARD L A C, et al. Evaluation of the causal effects between subjective wellbeing and cardiometabolic health: mendelian randomisation study[J]. BMJ, 2018, 362: k3788.
23	刘明, 高亚, 杨珂璐, 等. 孟德尔随机化研究的报告规范(STROBE-MR)解读[J]. 中国循证医学杂志, 2022, 22(8): 978-987.
	LIU M, GAO Y, YANG K L, et al. Interpretation of STROBE-MR: a statement for strengthening the reporting of observational studies in epidemiology using Mendelian randomization[J]. Chinese Journal of Evidence-Based Medicine, 2022, 22(8): 978-987.
24	BURGESS S, THOMPSON S G, CHD GENETICS COLLABORATION C R P. Avoiding bias from weak instruments in Mendelian randomization studies[J]. Int J Epidemiol, 2011, 40(3): 755-764.
25	BOWDEN J, DAVEY SMITH G, HAYCOCK P C, et al. Consistent estimation in mendelian randomization with some invalid instruments using a weighted Median estimator[J]. Genet Epidemiol, 2016, 40(4): 304-314.
26	DENG L, HUANG-FU Y C, MA Y H. Dietary nutrients involved in one-carbon metabolism and colorectal cancer risk[J]. LabMed Discov, 2024, 1(2): 100022.
27	BIAN X L, LIU R, MENG Y, et al. Lipid metabolism and cancer[J]. J Exp Med, 2021, 218(1): e20201606.
28	PETKEVICIUS K, PALMGREN H, GLOVER M S, et al. TLCD1 and TLCD2 regulate cellular phosphatidylethanolamine composition and promote the progression of non-alcoholic steatohepatitis[J]. Nat Commun, 2022, 13(1): 6020.
29	TAVASOLI M, LAHIRE S, REID T, et al. Genetic diseases of the Kennedy pathways for membrane synthesis[J]. J Biol Chem, 2020, 295(51): 17877-17886.
30	DAS S, CASTILLO C, STEVENS T. Phospholipid remodeling/generation in Giardia: the role of the lands cycle[J]. Trends Parasitol, 2001, 17(7): 316-319.
31	BARTOLACCI C, ANDREANI C, VALE G, et al. Targeting de novo lipogenesis and the Lands cycle induces ferroptosis in KRAS-mutant lung cancer[J]. Nat Commun, 2022, 13(1): 4327.
32	CIKES D, ELSAYAD K, SEZGIN E, et al. PCYT2-regulated lipid biosynthesis is critical to muscle health and ageing[J]. Nat Metab, 2023, 5(3): 495-515.
33	FU G T, GUY C S, CHAPMAN N M, et al. Metabolic control of Tfh cells and humoral immunity by phosphatidylethanolamine[J]. Nature, 2021, 595(7869): 724-729.
34	PING Y, SHAN J Q, QIN H M, et al. PD-1 signaling limits expression of phospholipid phosphatase 1 and promotes intratumoral CD8⁺ T cell ferroptosis[J]. Immunity, 2024, 57(9): 2122-2139.e9.
35	LAINS I, ZHU S J, HAN X K, et al. Genomic-metabolomic associations support the role of LIPC and glycerophospholipids in age-related macular degeneration[J]. Ophthalmol Sci, 2021, 1(1): 100017.
36	MANOLIO T A. Cohort studies and the genetics of complex disease[J]. Nat Genet, 2009, 41(1): 5-6.
37	GUERRA R, WANG J, GRUNDY S M, et al. A hepatic lipase (LIPC) allele associated with high plasma concentrations of high density lipoprotein cholesterol[J]. Proc Natl Acad Sci USA, 1997, 94(9): 4532-4537.
38	KANG Y P, YOON J H, LONG N P, et al. Spheroid-induced epithelial-mesenchymal transition provokes global alterations of breast cancer lipidome: a multi-layered omics analysis[J]. Front Oncol, 2019, 9: 145.
39	JEONG D, HAM J, KIM H W, et al. ELOVL2: a novel tumor suppressor attenuating tamoxifen resistance in breast cancer[J]. Am J Cancer Res, 2021, 11(6): 2568-2589.
40	DE ANTUENO R J, KNICKLE L C, SMITH H, et al. Activity of human Delta5 and Delta6 desaturases on multiple n-3 and n-6 polyunsaturated fatty acids[J]. FEBS Lett, 2001, 509(1): 77-80.
41	COLTELL O, SORLÍ J V, ASENSIO E M, et al. Genome-wide association study for serum omega-3 and omega-6 polyunsaturated fatty acids: exploratory analysis of the sex-specific effects and dietary modulation in Mediterranean subjects with metabolic syndrome[J]. Nutrients, 2020, 12(2): 310.
42	SAJUTHI S P, SHARMA N K, COMEAU M E, et al. Genetic regulation of adipose tissue transcript expression is involved in modulating serum triglyceride and HDL-cholesterol[J]. Gene, 2017, 632: 50-58.
43	KOLODZIEJSKI P A, PRUSZYNSKA-OSZMALEK E, SASSEK M, et al. Changes in obestatin gene and GPR39 receptor expression in peripheral tissues of rat models of obesity, type 1 and type 2 diabetes[J]. J Diabetes, 2017, 9(4): 353-361.
44	DOBOSZEWSKA U, MARET W, WLAŹ P. GPR39 deorphanization: the long and winding road to eicosanoids and a crosstalk between GPR39 and hedgehog signaling in angiogenesis[J]. Proc Natl Acad Sci USA, 2023, 120(28): e2308227120.
45	PARK R, LEE S, CHIN H, et al. Tumor-promoting role of GNA14 in colon cancer development[J]. Cancers (Basel), 2023, 15(18): 4572.
46	FAN H, WANG R, WEN B, et al. Biomarkers and potential therapeutic targets driving progression of non-alcoholic steatohepatitis to hepatocellular carcinoma predicted through transcriptomic analysis[J]. Front Immunol, 2024, 15: 1502263.
47	SKRIVANKOVA V W, RICHMOND R C, WOOLF B A R, et al. Strengthening the reporting of observational studies in epidemiology using mendelian randomization: the STROBE-MR statement[J]. JAMA, 2021, 326(16): 1614-1621.

[1]	韩龙传, 李悦, 邹智慧, 罗静, 李若伊, 张颖婷, 唐欣欣, 田丽红, 陆宇恒, 黄莺, 贺明, 付寅坤. 磷脂酰乙醇胺引起内质网应激促进巨噬细胞衰老及肝损伤[J]. 上海交通大学学报（医学版）, 2025, 45(6): 693-704.
[2]	徐宁, 周甜, 侯国俊, 沈南, 唐元家. CRISPR/Cas9介导的同源重组在非编码区疾病相关位点功能研究中的应用[J]. 上海交通大学学报(医学版), 2021, 41(1): 8-15.
[3]	高润颖，吴轲，祝捷，蔡美琴 . 不同泌乳期人乳磷脂成分的研究#br#[J]. 上海交通大学学报（医学版）, 2017, 37(8): 1151-.
[4]	宋立伟，王文越，邹秀群，等. 结直肠腺癌中特异蛋白1表达与临床病理指标的关系[J]. 上海交通大学学报（医学版）, 2014, 34(7): 973-.
[5]	吴琪俊, 徐剑锋, 施榕. 前列腺癌易感基因及遗传流行病学研究进展[J]. , 2011, 31(5): 672-.