Cheng Luo ¹ Baoping Tan ^1* Luoxiang Chu ^1* Liqiang Chen ^2* Xinglong Zhong ^1* Yangyang Jiang ^3* Yuluan Yan ^1* Fanrui Mo ^1* Hong Wang ^1* Shuxiong Nong ^4* Fan Yang ^1*

• ¹ Department of Cardiology, Liuzhou Workers' Hospital,The Fourth Affiliated Hospital of Guangxi Medical University, Liuzhou, China
• ² Department of Oncology, Liuzhou Workers' Hospital,The Fourth Affiliated Hospital of Guangxi Medical University, Liuzhou, China, Liuzhou, China
• ³ Rehabilitation department, Liuzhou Workers' Hospital,The Fourth Affiliated Hospital of Guangxi Medical University, Liuzhou, China, Liuzhou, China
• ⁴ Baise People's Hospital, Baise, China

Background: Fibroblasts in the fibrotic heart exhibit a heterogeneous biological behavior. The specific subsets of fibroblasts that contribute to progressive cardiac fibrosis remain unrevealed. Our aim is to identify the heart fibroblast(FB) subsets that most significantly promote fibrosis and the related critical genes as biomarkers for ischemic heart disease.Methods: The single nuclei RNA sequencing (snRNA-seq) and bulk RNA sequencing datasets used in this study were obtained from the Gene Expression Omnibus (GEO).The activity of gene sets related to progressive fibrosis was quantified for each FB cluster using the AddmoleculeScore function. Differentially expressed genes (DEGs) for the specific cell cluster with the highest fibrotic transcription dynamics were identified and integrated with bulk RNA sequencing data for analysis. Multiple machine learning models were employed to identify the optimal gene panel for diagnosing ischemic heart disease (IHD) based on the intersected DEGs. The effectiveness and robustness of the gene-derived diagnostic tool were validated using two independent IHD cohorts.Subsequently, we validated the signature genes using a rat post-myocardial infarction heart failure model. Results: We conducted an analysis on high-quality snRNA-seq data obtained from 3 IHD and 4 cardiac sarcoidosis heart samples, resulting in the identification of 16 FB clusters. Cluster2 exhibited the highest gene activity in terms of fibrosis-related transcriptome dynamics. The characteristic gene expression profile of this FB subset indicated a specific upregulation of COL1A1 and several pro-fibrotic factors, including CCDC102B, GUCY1A3, TEX41, NREP, TCAP, and WISP, while showing a downregulation of NR4A1, an endogenous inhibitor of the TGF-β pathway. Consequently, we designated this subgroup as COL1A1hiNR4A1low FB.Gene set enrichment analysis(GSEA) shows that the gene expression pattern of COL1A1hiNR4A1low FB was closer to pathways associated with cardiac fibrosis. Through machine learning, ten feature genes from COL1A1hiNR4A1low FB were selected to construct a diagnostic tool for IHD. The robustness of this new tool was validated using an independent cohort and heart failure rats. Conclusion: COL1A1hiNR4A1low FB possess heightened capability in promoting cardiac fibrosis. Additionally, it offers molecular insights into the mechanisms underlying the regulation of the TGF-β pathway. Furthermore, the characteristic genes of COL1A1hiNR4A1 FB could serve as valuable tools for diagnosing of IHD