Integrated multi-omics and machine learning identify Casp8 and Pycard as central regulators of pyroptosis and necroptosis in spinal cord injury

Programmed cell death pathways exacerbate secondary damage after spinal cord injury, yet their shared regulators and tractable therapeutic points remain elusive. We aimed to identify cross-pathway regulators that integrate pyroptosis and necroptosis after spinal cord injury and to evaluate a repurposable therapeutic capable of modulating these regulators in vivo. Bulk RNA-sequencing datasets (GSE47681, GSE5296, and a Figshare cohort) were integrated with single-cell RNA-sequencing data (GSE162610). Differential expression analysis, immune-infiltration deconvolution, weighted gene co-expression network analysis, and three machine-learning algorithms (least absolute shrinkage and selection operator, random forest, and support vector machine-recursive feature elimination) were combined to pinpoint key genes that modulate key types of programmed cell death. A two-gene ridge regression score was derived to quantify pathway activity across cohorts. Single-cell mapping and spatial profiling localized the candidate regulators to myeloid lineages within the injury core. Drug-gene interaction screening and molecular docking with AutoDock4 were used to prioritize a clinically available compound, which was then tested in a thoracic level 8 contusion mouse model with electrophysiological, behavioral, and molecular assessments. The integrated analysis converged on Pycard and Casp8 as central co-regulators linking pyroptosis and necroptosis. The two-gene ridge regression score robustly stratified samples in training and independent validation datasets. The score also tracked a temporal surge in the activity of programmed cell death that coincides with peak inflammatory cell infiltration. Drug-gene interaction screening prioritized decitabine, and molecular docking showed binding energies of < -5.0 kcal/mol for Pycard and Casp8. In vivo, mice received intraperitoneal decitabine (1.5 mg/kg daily for 7 days) after injury with blinded assessments. Decitabine reduced activation of NOD-like receptor family pyrin domain containing 3 and phosphorylated mixed lineage kinase domain-like protein. Macrophage polarization was shifted away from an inflammatory phenotype. Motor-evoked potential amplitudes increased, and Basso Mouse Scale locomotor scores improved, indicating mitigation of pyroptosis and necroptosis with functional benefit after spinal cord injury. This work addresses the need for unifying regulators bridging multiple programmed cell death programs after spinal cord injury and a multitarget intervention dampening convergent inflammatory pathways. A reproducible pipeline linking multi-omics integration and ensemble machine learning to therapeutic nomination and in vivo testing was established. The key contributions include the identification of a Pycard-Casp8 inflammatory hub that integrates pyroptosis and necroptosis, the development of a two-gene scoring paradigm that generalizes across datasets and time points, and the demonstration that decitabine functions as a multitarget neuroprotective strategy, resulting in measurable improvements in electrophysiological, behavioral, and molecular outcomes in a mouse model of spinal cord injury.