Multi‐Omics Integration Identifies Immune–ECM Dysregulation and Candidate Biomarkers of Infected High‐Stage Pressure Injury in a Novel Aged‐Rat Model

Infected high-stage pressure injuries (IHPIs) in older, immobile patients lack validated biomarkers to distinguish infection-driven chronicity and guide therapy. We leveraged multi-omics integration to nominate biomarker candidates and mechanistic signatures of IHPI. An aged-rat IHPI model was established by combining deep-tissue magnetic compression with Staphylococcus aureus inoculation. Acute wounds (AW), non-infected high-stage pressure injuries (HPI), and IHPI were compared through laser-speckle perfusion imaging, in vivo bioluminescence for bacterial burden, histology, and immunohistochemistry. Serum IL-1β was measured by ELISA. Bulk RNA sequencing (AW n = 3, HPI n = 3, IHPI n = 5) and quantitative proteomics (AW n = 3, IHPI n = 5) were cross-layer integrated using network analysis. IHPI wounds exhibited delayed closure, purulent exudate, reduced angiogenesis, diminished collagen deposition, and elevated systemic inflammation. Transcriptomic data revealed activation of innate-immune, chemokine, and keratinization programs, while proteomics highlighted NF-κB, lysosomal, and extracellular-matrix (ECM) remodeling processes. Integration defined a translational immune biomarker panel-SELL (L-selectin)↑, ARG1↑, CD163↓, IL33↓-implicating sustained leukocyte recruitment and impaired resolution. Concurrently, MMP9-driven ECM degradation and enrichment of sarcomeric/calcium-handling proteins (Tnnt2, Casq2, Tnnc1, Myh8, Myl4) indicated deep-tissue and muscle-layer involvement. Multi-omics integration in an aged-rat IHPI model yields a translational immune biomarker panel (SELL↑/ARG1↑/CD163↓/IL33↓) for detection/stratification and a mechanistically coherent ECM-remodeling signature (MMP9↑) that explains tissue-level non-healing. Using accessible assays (IHC/ELISA), these candidates prioritize testable biomarkers for IHPI detection, stratification, and therapeutic targeting, while the model provides a translational platform for future validation.