Xing Liu,1,* Yuanyuan Zhang,1,* Yan Wang,2,3,* Yanfeng Xu,2,4,5 Wenhao Xia,2,5– 7 Ruiming Liu,4,5 Shiyue Xu2,5,6 

1Department of Cardiology, The Third Affiliated Hospital, Sun Yat-sen University, Guangzhou, Guangdong, People’s Republic of China; 2Department of Hypertension and Vascular Disease, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, Guangdong, People’s Republic of China; 3Health Management Center, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, Guangdong, People’s Republic of China; 4Laboratory of General Surgery, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, Guangdong, People’s Republic of China; 5National - Guangdong Joint Engineering Laboratory for Diagnosis and Treatment of Vascular Diseases, Sun Yat-sen University, Guangzhou, Guangdong, People’s Republic of China; 6NHC Key Laboratory of Assisted Circulation, Sun Yat-sen University, Guangzhou, Guangdong, People’s Republic of China; 7Department of Cardiovascular Medicine, Guangxi Hospital Division of The First Affiliated Hospital of Sun Yat-sen University, Nanning, Guangxi, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Shiyue Xu, Department of Hypertension and Vascular Disease, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, 510080, People’s Republic of China, Email xushy25@mail.sysu.edu.cn Ruiming Liu, Laboratory of General Surgery, The First Affiliated Hospital, Sun Yat-sen University, Guangzhou, 510080, People’s Republic of China, Email liuruim@mail2.sysu.edu.cn

Purpose: Inflammatory activation of immune cells plays a pivotal role in the development of coronary artery diseases (CAD). This study aims to investigate the immune responses of peripheral blood mononuclear cells (PBMCs) in CAD and to identify novel signature genes and biomarkers using machine learning algorithms.
Methods: The GSE113079 dataset was analyzed and differentially expressed genes (DEGs) were identified between CAD and normal samples. The intersection of DEGs with inflammation-related genes was used to identify the differentially expressed inflammation-related genes (DIRGs). Then, the receiver operating characteristic (ROC) curves were plotted for each DIRG, and those with an area under the curve (AUC) greater than 0.9 were selected for subsequent analysis. Furthermore, machine learning algorithms were employed to identify biomarkers. A nomogram was developed based on these biomarkers. The CIBERSORT algorithm and Wilcoxon test method were used to analyze the differences in immune cells between the CAD and normal samples. The identified biomarkers were validated in PBMCs from patients with CAD and in atherosclerotic aortas from ApoE−/− mice.
Results: A total of 574 DEGs were identified between CAD and normal samples. From this intersection, 29 DIRGs were identified, of which 14 DIRGs (PTGER1, IL17RC, KLKB1, GPR32, ADM, NUPR1, SCN9A, IL17B, CX3CL1, FFAR3, PYDC2, SYT11, RORA, and GPR31) exhibited high diagnostic efficacy. Four biomarkers (ADM, NUPR1, PTGER1, and PYDC2) were identified using Support Vector Machine (SVM). Ten types of immune cells, including CD8+ T cells, regulatory T cells (Tregs), and resting NK cells, showed significant differences between the CAD and normal groups. Furthermore, increased levels of ADM, NUPR1, PTGER1, and PYDC2 were validated in PBMCs isolated from CAD patients. In addition, ADM, NUPR1, and PTGER1 were upregulated in the mouse atherosclerotic aorta.
Conclusion: Our findings revealed novel inflammatory gene signatures of CAD that could be potential biomarkers for the early diagnosis of CAD.

Keywords: inflammation, coronary artery disease, machine learning, immune infiltration