Exploring ABHD5 as a Lipid-Related Biomarker in Idiopathic Pulmonary Fibrosis: Integrating Machine Learning, Bioinformatics, and In Vitro Experiments

Yi Liao,Xiaying Peng,Yan Yang,Guanghong Zhou,Lijuan Chen,Yang Yang,Hongyan Li,Xianxia Chen,Shujin Guo,Qiunan Zuo,Jun Zou

DOI: https://doi.org/10.1007/s10753-024-02107-1

2024-07-24

Inflammation

Abstract:Recent studies increasingly suggest a connection between lipids and idiopathic pulmonary fibrosis (IPF). This study was aimed at exploring potential lipid-related biomarkers for IPF and uncovering the mechanisms underlying pulmonary fibrosis. IPF-related datasets were retrieved from the GEO database, and the ComBat algorithm was used to merge multiple datasets and eliminate batch effects. Weighted gene co-expression network analysis (WGCNA) was utilized to identify modules and genes associated with IPF. Potential hub genes were determined by intersecting these genes with lipid-related genes from the GeneCards database. A machine learning-based integrative approach was developed to construct diagnostic and prognostic signatures, which were validated across several datasets. Additionally, single-cell sequencing data was used to validate the expression differences of core IPF-related genes across various cell types. The effect of ABHD5 on fibroblasts was assessed using the cell counting kit-8, 5-ethynyl-2'-deoxyuridine, and cell scratch assays. The expression levels of fibrotic factors were measured using real-time quantitative polymerase chain reaction and western blot analysis. WGCNA identified a red module potentially related to IPF, and the intersection with lipid-related genes yielded 51 hub genes. These genes were used to build diagnostic and prognostic models that demonstrated robust validation capabilities across multiple datasets. Single-cell sequencing analysis revealed low expression of ABHD5 in the lung tissues of IPF patients, with a higher proportion of fibroblasts exhibiting low ABHD5 expression. Cell experiments showed that under the influence of TGF-β1, knockdown of ABHD5 slightly promoted fibroblast proliferation. Additionally, fibroblasts with low ABHD5 expression exhibited enhanced migratory capabilities and secreted more fibrotic factors. Lipid-related diagnostic and prognostic models for IPF were developed, and ABHD5 may serve as a potential biomarker. Low ABHD5 expression could potentially accelerate the progression of pulmonary fibrosis.