Abstract
BACKGROUND: Recent research suggests a link between acetyl tributyl citrate (ATBC) exposure and an increased risk of coronary heart disease (CHD). OBJECTIVE: This study investigated the molecular mechanisms underlying ATBC's potential role in CHD pathogenesis. METHODS: Using "Acetyl tributyl citrate" as a search term, relevant targets were retrieved from the ChEMBL database. The standard simplified molecular input line entry system (SMILES) notation of ATBC was submitted to the SwissTargetPrediction database. All the targets obtained were compiled to create a target database for ATBC. Functional enrichment analysis and gene set enrichment analysis (GSEA) were performed to explore the potential pathogenic mechanisms of ATBC. The GSE66360 dataset was used as the training dataset, while GSE48060 and GSE60993 served as validation datasets. A total of 107 combinations of eleven machine learning algorithms, including Random Forest (RF), Elastic Net (Enet), support vector machine (SVM), least absolute shrinkage and selection operator (LASSO) regression, Ridge regression, gradient boosting with component-wise linear model (glmBoost), partial least squares regression for generalized linear model (plsRglm), linear discriminant analysis (LDA), extreme gradient boosting (XGBoost), Naive Bayes, and stepwise generalized linear model (Stepglm), were applied to identify the model with the highest area under the curve (AUC) as the best diagnostic model. Additionally, receiver operating characteristic (ROC) curves were used to identify key hub genes. Single-cell transcriptomic data were employed to locate these hub genes, while molecular docking further validated the binding capacity between ATBC and its hub targets. This included converting the ligand to 3D format, performing molecular docking, and calculating the binding affinity and hydrogen bond formation between the molecules. The binding site with the lowest predicted binding affinity was selected for visualization. RESULT: By integrating ATBC targets with CHD core modules, we identified genes associated with ATBC-induced CHD. Using the RF algorithm, we constructed the optimal diagnostic model and identified key hub genes, including MMP9, NLRP3, and PLAU. These genes were closely associated with glucose and lipid metabolism disorders, induction of estrogen resistance, and vascular inflammation. Furthermore, NLRP3 was predominantly expressed in monocytes, while PLAU showed higher expression in fibroblasts and endothelial cells. The molecular docking results indicated that the calculated predicted binding affinities were all less than or equal to -5.0 kcal/mol. This confirmed the binding affinities of ATBC with MMP9 and PLAU, and supported their involvement in the pathogenesis of coronary heart disease. CONCLUSION: Our study predicted ATBC's potential mechanisms in CHD progression and identified key hub genes, notably MMP9, NLRP3, and PLAU. These findings provide novel molecular targets for future research and highlight the potential health risks of ATBC in everyday applications.