Abstract
BACKGROUND AND AIMS: Liver fibrosis (LF) is the common pathological process of almost all liver diseases, but the pathogenesis is extremely complicated and has not been fully clarified. Therefore, we want to explore LF's complex pathogenesis and identify key genetic markers that can predict LF prognosis. METHODS: From Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/), we contained datasets GSE84044 and GSE130970. And involved GSE15654 whose survival data were for the validation of our model's prognostic value. Then, Bulk RNA sequencing was used to establish a LF-related prognostic model and to determine key genes.The key genes were later analyzed using single-cell RNA sequencing (scRNA-seq) and the role of macrophages in LF was investigated. RESULTS: We established a comprehensive LF-related prognostic model (FMSig, i.e., fibrosis macrophage-related prognostic signature) containing four genes (IER3, AKR1B10, ADCY1, and PGM1). FMSig can distinguish between survival outcomes in dataset GSE15654. IER3 showed a higher hazard ratio than the other three FMSig genes. Moreover, further scRNA-seq analysis showed that IER3 is highly expressed in myeloid cells while the other three genes were rarely expressed in six immune cell types and other cell clusters. After reclustering the myeloid cells into seven clusters, we found that IER3 was highly expressed in myeloid cell cluster 3, shown to be related to anti-inflammation and lipid metabolism functions. Based on pseudotime analysis, we suggest that myeloid cell type 3 is transformed from myeloid cell type 0 in the liver cirrhosis microenvironment. On immunohistochemical staining, LF showed significantly higher IER3 expression than healthy controls. CONCLUSION: This study established a predictive model, FMSig, to evaluate LF prognosis, and showed that IER3 and macrophages are related to LF.