Identification of glycolysis-related gene signatures for prognosis and therapeutic targeting in idiopathic pulmonary fibrosis

Background: While it’s known that glycolysis, the process of glucose breakdown for energy, plays a significant role in the development of fibrosis (scarring of tissue), the specific genes involved in glycolysis in the context of idiopathic pulmonary fibrosis, a progressive and fatal lung disease, are not well understood.

Methods: To investigate this, researchers analyzed three gene expression datasets related to IPF that were publicly available in the Gene Expression Omnibus database. They also compiled a list of glycolysis-related genes from the Molecular Signatures Database. Using a statistical method called “limma” in the R programming language, they identified glycolysis-related genes that were expressed differently in IPF patients compared to healthy individuals. To pinpoint the most important glycolysis-related genes for diagnosing IPF, they used two machine learning techniques: least absolute shrinkage and selection operator (LASSO) regression and support vector machine-recursive feature elimination (SVM-RFE). A model to predict prognosis (disease progression) was then built using LASSO regression, and its accuracy was evaluated using time-dependent receiver operating characteristic curves. To see which cell types within the lung express these genes, they analyzed single-cell RNA sequencing data. Further analyses, including immune infiltration analysis, Gene Set Enrichment Analysis, and Gene Set Variation Analysis, were performed to understand the potential biological mechanisms involved. Finally, they used a mouse model of bleomycin-induced pulmonary fibrosis to experimentally validate their findings using reverse transcription-quantitative polymerase chain reaction.

Results: The analysis identified 14 glycolysis-related genes (VCAN, MERTK, FBP2, TPBG, SDC1, AURKA, ARTN, PGP, PLOD2, PKLR, PFKM, DEPDC1, AGRN, CXCR4) as potential diagnostic markers for IPF. Among these, seven genes (ARTN, AURKA, DEPDC1, FBP2, MERTK, PFKM, SDC1) were used to create a prognostic model that showed good predictive ability (AUC ranging from 0.831 to 0.793). The single-cell RNA sequencing data revealed that the expression of these glycolysis-related genes varied across different cell types in the lung, with particularly high expression in macrophages and fibroblasts, key cell types involved in fibrosis. Immune infiltration analysis suggested a link between these genes and imbalances in the immune response in IPF. The experimental validation in the bleomycin-induced fibrosis mouse model confirmed that several of these glycolysis-related genes, including AURKA and CXCR4, were upregulated. Finally, drug prediction analysis identified potential therapeutic inhibitors, such as Tozasertib for AURKA and Plerixafor for CXCR4.

Conclusion: This study identifies specific glycolysis-related genes as potential biomarkers for predicting the prognosis of IPF and highlights their role in influencing immune responses within the fibrotic lung environment. Notably, AURKA, MERTK, and CXCR4 were found to be associated with biological pathways linked to the progression of fibrosis and represent potential targets for new therapies. The findings of this research provide valuable insights into the metabolic changes occurring in IPF and suggest that targeting glycolysis-related pathways could be a novel strategy for developing antifibrotic treatments.