Dynamic changes in pyroptosis following spinal cord injury and the identification of crucial molecular signatures through machine learning and single-cell sequencing

Abstract

The pathological cascade of spinal cord injury (SCI) is highly intricate. The onset of neuroinflammation can exacerbate the extent of damage. Pyroptosis is a form of inflammation-linked programmed cell death (PCD), the inhibition of pyroptosis can partially mitigate neuroinflammation. It is imperative to delineate the principal cell types susceptible to pyroptosis and concomitantly identify key genes associated with this process. We initially defined the pyroptosis-related genes (PRGs) and analyzed their expression at different time points post SCI. The results demonstrate a substantial upregulation of differentially expressed genes (DEGs) related to pyroptosis on the 7 days post-injury (dpi), these DEGs in the 7 dpi are closely related to the inflammatory response. Subsequently, immune infiltration analysis revealed a predominant presence of inflammatory microglia. Through correlation analysis, we postulated that pyroptosis primarily manifested within the inflammatory microglia. Employing machine learning algorithms, we identified four pyroptosis-related molecular signatures, which were experimentally validated using BV2 cells and spinal cord tissue samples. The robustness of the identified molecular signatures was further confirmed through single-cell sequencing data analysis. Overall, our study elucidates the temporal dynamics of pyroptosis and identifies key molecular signatures following SCI. These findings can provide novel evidence for therapeutic interventions in SCI.