A plasma derived hypoxia signature predicts benefit from hypoxia-modifying radiotherapy and reveals dynamic humoral immune modulation in bladder cancer

Abstract

Abstract Introduction Hypoxia is a hallmark of solid tumours and drives radioresistance, yet no hypoxia biomarkers are used clinically. We explored nanoproteomics to identify a plasma protein signature for longitudinal non-invasive hypoxia monitoring. Methods Twenty-three bladder cancer patients (T2–T3b) undergoing radiotherapy (55 Gy, 4 weeks) were prospectively enroled. Plasma was collected at baseline and weekly, nanoparticle-enriched and analysed by proteomics. Hypoxia scores (HS) were derived from diagnostic biopsies using a validated 24-gene signature, and patients stratified by median HS. Results We identified 115 differentially abundant proteins (DAPs; |FC | > 1.5; p < 0.05). Seven DAPs (IGKV3, IGLV2-18, VL_4, VL_7, FN1, IGLV2-14, CAMP) correlated with HS across multiple timepoints ( p < 0.05; |r | >0.4). A two-protein signature (FN1, CAMP) was retrospectively validated, showing prognostic value in TCGA-BLCA ( n = 404; HR = 1.54; CI = 1.12–2.13; p = 0.009), BC2001 ( n = 313; HR = 1.42; CI = 1.08–1.86; p = 0.011) and in meta-analysis ( n = 150; HR = 1.48; CI = 1.23–1.78; p < 0.001). In BCON ( n = 150), it predicted benefit from hypoxia-modifying therapy (HR = 0.60; CI = 0.34–1.06; p = 0.079). The 115 DAPs formed five temporal co-expression clusters peaking at successive treatment weeks. Clusters were enriched for humoral immune pathways, with one linked to extracellular matrix remodelling. Conclusion These findings support the use of low abundance plasma biomarkers for real-time hypoxia stratification, revealing dynamic immune responses during radiotherapy associated with hypoxia.