Unraveling the regulatory nexus of aggrephagy in ALS: identification of novel candidate biomarkers and molecular triggers.

backgroundAmyotrophic lateral sclerosis (ALS) is a progressive and ultimately fatal neurodegenerative disorder involving multiple systems, with motor neuron degeneration as its primary feature. This disease can be classified into sporadic and familial types. Genes such as SOD1, C9orf72, FUS, and TDP-43 have been identified as the main causative genes for familial ALS. Multiple bioinformatics tools combined with an experimental verification strategy have helped in understanding the association of a selective autophagy pathway called aggrephagy with the disease.

resultsThe transcriptome data of spinal cord tissue from SOD1-G93A mice was obtained from the Gene Expression Omnibus (GEO) database. Based on the GSE281064 dataset, we investigated aggrephagy-related transcriptional alterations in the SOD1-G93A mouse model of ALS. After comparison with the aggrephagy-related genes (AGGRGs) set included in the GeneCards database, 49 candidate genes closely related to the autophagy process were obtained. Functional enrichment analysis showed these genes participate in extracellular matrix remodeling, hyaluronic acid and glycosaminoglycan metabolism, tumor necrosis factor regulation, and lysosomal function, indicating central roles in inflammation, apoptosis, and metabolic disorders. Based on feature selection algorithms, this study employed machine learning methods such as random forest (RF), extreme gradient boosting (XGBoost), and Boruta to conduct multi-angle screening of candidate genes. The intersection of the results ultimately identified three key genes: Ctsb, Kif11, and S100a6.

conclusionsIn both the training data and external validation data, Ctsb and S100a6 showed significant upregulation and demonstrated excellent discriminatory capabilities. The nomogram constructed based on Ctsb and S100a6 expression showed potential for distinguishing SOD1-G93A model samples from nontransgenic controls. The predicted probability demonstrated the potential of these two as candidate biomarkers. Meanwhile, further validation is needed in larger independent population cohorts in the future. The SOD1-G93A mouse model and SOD1-G93A-expressing NSC34 cell model showed expression patterns of Ctsb and S100a6 consistent with the bioinformatics findings. Through S100a6 overexpression and knockdown experiments in an NSC34 motor neuron-like ALS model, we found that S100a6 impaired autophagy and promoted SOD1 aggregation. These findings further validate its potential as a biomarker and provide new insights into the pathogenesis of ALS.