An analysis of metabolic changes and potential biomarkers in ischemic stroke based on untargeted metabolomics
10.3969/j.issn.1672-5921.2025.03.006
- VernacularTitle:基于非靶向代谢组学对缺血性卒中代谢变化及潜在生物标志物的分析研究
- Author:
Yunyu WANG
1
;
Yaqi LI
;
Tian ZHAO
;
Liyuan HAN
;
Yongan LI
;
Qingzeng QIAN
Author Information
1. 063210 唐山,华北理工大学公共卫生学院
- Publication Type:Journal Article
- Keywords:
Ischemic stroke;
Metabolomics;
Serum;
Biomarkers;
Diagnostic model
- From:
Chinese Journal of Cerebrovascular Diseases
2025;22(3):199-209
- CountryChina
- Language:Chinese
-
Abstract:
Objective To investigate serum metabolites and metabolic pathways alterations in patients with ischemic stroke(IS)through metabolomic analysis,and to identify reliable serum metabolic biomarkers for IS diagnosis.Methods This prospective study enrolled patients with IS admitted to the Department of Neurology at Xiangcheng People's Hospital of Suzhou from December 1,2022 to December 31,2023.Age-and sex-matched healthy individuals were recruited as controls during the same period.Baseline characteristics were collected,including age,sex,height,body mass index,and blood pressure.Venous blood samples were obtained after an 8 h fast for biochemical analysis of blood glucose,total bilirubin,serum creatinine,urea nitrogen,total cholesterol,triglycerides,high-density lipoprotein cholesterol,and low-density lipoprotein cholesterol.Serum metabolites of both groups were extracted and analyzed using ultra-high-performance liquid chromatography coupled with tandem mass spectrometry.Metabolomic data were processed using Simca-p software for unsupervised principal component analysis(PCA)and orthogonal partial least squares discriminant analysis(OPLS-DA)to evaluate group separation and experimental stability.Differential metabolites were defined by variable importance in projection(VIP)≥1.0,fold change(FC)≥2.0 or ≤0.5,and P＜0.05.Drug-derived exogenous metabolites were excluded by cross-referencing the Human Metabolome Database(HMDB,https://hmdb.ca/)and PubChem(https://pubchem.ncbi.nlm.nih.gov/).MetaboAnalyst 6.0(http://www.metaboanalyst.ca),a comprehensive web-based tool for metabolomic data analysis,was employed to map differential metabolites to the Kyoto encyclopedia of genes and genomes(KEGG)databased and to perform pathway enrichment analysis.Machine learning models were developed using Python.Least absolute shrinkage and selection operator(LASSO)regression and random forest(RF)algorithms were employed to identify diagnostic biomarkers capable of effectively distinguishing IS patients from controls.Metabolites identified by both methods were integrated into an extreme gradient boosting(XGBoost)model.Model performance was evaluated using receiver operating characteristic(ROC)curves with 5-fold cross-validation and internal validation(70％training,30％validation set).Results A total of 51 IS patients and 51 matched controls were included.(1)A total of 1 255 serum metabolites were identified(964 in positive ion mode,291 in negative ion mode).PC A and OPLS-DA demonstrated distinct metabolic separation between IS patients and controls.In IS group,260 metabolites were upregulated and 337 downregulated in positive ion mode;99 were upregulated and 34downregulated in negative ion mode.(2)Among the 1 255 metabolites,259 were identified as differential metabolites based on the criteria of VIP ≥ 1.0,FC≥2.0 or≤0.5 and P＜0.05.After excluding drug-derived metabolite through referencing HMDB and PubChem databases,a total of 220 endogenous differential metabolites were found to coexist in both positive and negative ion modes.Among them,119 metabolites were up-regulated and 101 were down-regulated in the IS group.The expression of these 220 metabolites showed significant differences between the IS and control groups.(3)KEGG pathway analysis highlighted five dysregulated pathways:upregulation of denovo triacylglycerol biosynthesis,glycerophosphate shuttle,and cardiolipin biosynthesis;downregulation of bile acid biosynthesis and methylhistidine metabolism.(4)LASSO and RF algorithms identified 24 and 30 candidate biomarkers,respectively.Four overlapping metabolites were selected:2-((3R)-3-((3R,5S,7S,9S,10S,13R,14S,17R)-3,7-dihydroxy-10,13-dimethylhexadecahydro-1H-cyclopenta[a]phenanthren-17-yl)butanamido)ethane-1-sulfonic acid(m/z 971.571 29),arginine-conjugated cholic acid(m/z 587.379 21),laccaic acid A(m/z 576.010 93)and NCGC00380235-01_C32H48O9_beta-D-xylopyranoside,3,17-dihydroxyspirosta-5,25(27)-dien-1-yl(m/z 559.326 48).The expression levels of 2-((3R)-3-((3R,5S,7S,9S,10S,13R,14S,17R)-3,7-dihydroxy-10,13-dimethylhexadecahydro-1H-cyclopenta[a]phenanthren-17-yl)butanamido)ethane-1-sulfonic acid(m/z 971.571 29),arginine-conjugated cholic acid(m/z 587.379 21),and laccaic acid A(m/z 576.010 93)were upregulated,while the expression level of NCGC00380235-01_C32H48O9_beta-D-xylopyranoside,3,17-dihydroxyspirosta-5,25(27)-dien-1-yl(m/z 559.326 48)was downregulated.An IS diagnostic model was established based on four metabolic biomarkers using the XGBoost algorithm.The area under the ROC curve was 1.000(95％CI 1.000-1.000)in the training set and 0.988 in the validation set(95％CI 0.963-1.000).Conclusions Patients with IS exhibit significant metabolic disturbance.The four identified biomarkers may serve as potential biomarkers for the effective identification of IS:2-((3 R)-3-((3R,5S,7S,9S,10S,13R,14S,17R)-3,7-dihydroxy-10,13-dimethylhexadecahydro-1H-cyclopenta[a]phenanthren-17-yl)butanamido)ethane-1-sulfonic acid(m/z971.571 29),arginine-conjugated cholic acid(m/z587.379 21),laccaic acid A(m/z 576.010 93),and NCGC00380235-01_C32H48O9_beta-D-xylopyranoside,3,17-dihydroxyspirosta-5,25(27)-dien-1-yl(m/z 559.326 48).
