Cell-death-associated inflammatory remodeling of the SPP1/KRT5 axis in high-grade serous ovarian cancer: multi-omics, single-cell, and irradiation validation

Background High-grade serous ovarian cancer (HGSOC) is characterized by extensive cellular heterogeneity and frequent emergence of chemotherapy resistance. Therapy-induced cellular injury and regulated cell death can reshape inflammatory signaling and the tumor microenvironment, but the relationship between death-associated stress responses and secreted phosphoprotein 1 (SPP1/osteopontin)-expressing myeloid cells or keratin 5 (KRT5)-positive epithelial states remains incompletely defined. Methods The Cancer Genome Atlas Ovarian Cancer (TCGA-OV) raw RNA-seq counts and clinical annotations were analyzed in 429 tumor samples. Single-cell RNA sequencing data from GSE165897 included 22 paired treatment-naive and post-neoadjuvant chemotherapy (post-NACT) libraries from 11 patients with metastatic High-grade serous ovarian cancer (51,786 cells). GSE316118 provided RNA-seq profiles of KRT5+ and KRT5- SKOV3 populations (n = 3 per group). Differential expression, pathway enrichment, program scoring, paired sample-level comparisons, and cross-layer consistency analyses were performed. In addition, SK-OV-3 cells were exposed to 0 or 6 Gy X-ray irradiation and SPP1, HIF1A, VEGFA, and KRT5 were quantified by RT-qPCR. Regulated cell death was treated as a mechanistic context rather than directly assigned from transcriptomic data. Results In TCGA-OV, SPP1 expression was most strongly associated with the inflammatory program (Spearman r = 0.63), with weaker positive associations with hypoxia (r = 0.42) and EMT (r = 0.35); SPP1-high versus SPP1-low groups did not differ significantly in overall survival (log-rank P = 0.181). At single-cell resolution, SPP1 localized predominantly to the myeloid compartment, whereas KRT5 expression was restricted to a subset of epithelial cells. After NACT, epithelial VIM expression and the epithelial hypoxia program increased significantly, while myeloid SPP1 showed a non-significant downward trend. In the SKOV3 dataset, KRT5+ and KRT5-populations showed reciprocal exploratory KRT5-CPC and SPP1-niche signatures. The positive SPP1-inflammatory coupling observed in TCGA-OV was reproduced at sample level in the independent single-cell cohort (r = 0.71). In irradiated SK-OV-3 cells, SPP1, HIF1A, VEGFA, and KRT5 were significantly increased at 24 h (all P < 0.01). Together, these findings are consistent with therapy-associated inflammatory remodeling in a cellular injury/death context, without establishing a specific cell-death modality. Conclusion These data identify SPP1-marked myeloid inflammation and KRT5-associated epithelial states as complementary features of metastatic HGSOC and support a cell-death-associated inflammatory remodeling framework linking treatment stress with epithelial hypoxia/mesenchymal programs. The findings are associative and hypothesis-generating; specific regulated cell-death pathways require direct functional validation.