Bioinformatics analysis of signature genes associated with anoikis and endoplasmic reticulum stress in keratoconus.

Abstract

PURPOSE: Keratoconus (KC) is a progressive disorder of corneal thinning characterized by responses in the extracellular matrix and cellular interactions. This study used bioinformatics methods to identify key genes involved in KC development and in anoikis and endoplasmic reticulum (ER) stress. METHODS: KC and control datasets from the GEO database were analyzed to identify differentially expressed genes (DEGs). These were cross-referenced with anoikis and ER stress-related genes from Genecards. Functional enrichment, immune infiltration analysis, and machine learning techniques (LASSO, Random Forest) were used to identify candidate molecular signatures, which were then validated in an animal model. RESULTS: We identified 46 DEGs associated with anoikis and 41 DEGs related to ER stress. Functional analysis linked them to apoptosis and IL-17 signaling. Five key molecular signatures were identified: CDKN1A, MCL1, PTGS2, PTHLH, and ANGPTL4. The expression of ANGPTL4, CDKN1A, and MCL1 was consistent in the animal model. These genes are associated with inflammatory and oxidative stress responses. Twelve potential therapeutic drugs were predicted. CONCLUSION: This study identifies five candidate molecular signatures for KC related to anoikis and ER stress, offering insights into KC pathogenesis and potential targeted therapies.