Peripheral and central immune features in insomnia: Integrative bulk and single-cell transcriptomic analyses reveal core hub genes and candidate compounds.

Insomnia is closely associated with immune dysregulation, yet the overall pattern of peripheral-central immune disequilibrium and its underlying molecular basis remains incompletely understood. To characterize the peripheral-central immune features associated with insomnia, identify key immune cell populations and core molecular programs, and prioritize candidate therapeutic compounds with preliminary experimental validation. Peripheral blood bulk transcriptomic dataset GSE208668 was analyzed using differential expression analysis, weighted gene co-expression network analysis (WGCNA), and functional enrichment analysis to identify insomnia-associated genes. Protein-protein interaction network analysis and machine learning models were then applied in independent peripheral blood datasets to refine core genes. Immune deconvolution and peripheral blood single-cell transcriptomic dataset GSE213496 were used to determine the immune-cell context, cell-type localization, and intercellular communication features of these genes. The brain single-cell transcriptomic dataset GSE137665 was further analyzed to assess central alterations. Drug prediction, molecular docking, and molecular dynamics simulations were performed to prioritize candidate compounds, followed by in vitro validation of resveratrol in an LPS-induced THP-1 macrophage model. A total of 5,321 differentially expressed genes associated with insomnia were identified, and weighted gene co-expression network analysis highlighted the turquoise and blue modules as key insomnia-related modules. Integrative analysis yielded 390 intersecting genes enriched mainly in immune, inflammatory, and oxidative stress-related pathways. Protein interaction analysis and machine learning further identified six refined core genes: FN1, HMOX1, HSP90AA1, IL10, MYD88, and NFE2L2. Because IL10 was not stably detected in the single-cell datasets, the remaining five genes were used for downstream single-cell analyses. Immune deconvolution suggested selective peripheral immune remodeling in insomnia, characterized by increased resting CD4 memory T cells and M2 macrophages, together with reduced activated NK cells. Peripheral single-cell analysis showed that HMOX1, HSP90AA1, MYD88, and NFE2L2 were mainly enriched in neutrophils, inflammatory macrophages, conventional dendritic cells, and selected lymphocyte populations, whereas FN1 showed a more restricted distribution pattern. CellChat analysis indicated enhanced intercellular communication under the sleep deprivation-related condition. In contrast, brain single-cell analysis revealed comparatively modest but detectable central alterations, including enrichment of ependymal cells, slight increases in excitatory neurons, brain endothelial cells, and choroid plexus stromal fibroblasts, together with heterogeneous expression of the core hub genes and selective rewiring of intracerebral communication networks. Drug prediction consistently prioritized quercetin and resveratrol, and structural analyses supported stable interactions with key targets. In THP-1 macrophages, resveratrol downregulated MYD88 and HSP90AA1 while further upregulating HMOX1. Insomnia appears to be associated predominantly with a peripheral-centered immune disequilibrium pattern, characterized by selective remodeling of innate immune-related populations, enhanced inflammatory and oxidative stress programs, and increased intercellular communication. The MYD88-HMOX1-HSP90AA1-NFE2L2 axis may represent a key molecular program linking inflammatory activation and oxidative stress adaptation. Resveratrol was identified as a potential compound and has been validated through preliminary in vitro experiments.