Protein profile in urinary extracellular vesicles is a marker of malignancy and correlates with muscle invasiveness in urinary bladder cancer

Urinary Bladder Cancer (UBC) ranks among the most prevalent cancers worldwide, has a high recurrence rate and unpredictable treatment responses. Thus, biomarkers are urgently needed. Extracellular vesicles (EVs) are released from both cancer- and immune cells and provide a snapshot of the originating cell. They are abundant in urine and are therefore candidate biomarkers for UBC. Isolated urinary EVs from 39 UBC patients were compared with EVs from healthy controls, prostate cancer patients and whole urine. Samples were from bladder urine at time of both transurethral resection of the bladder tumour (TURB) and cystectomy, as well as urine taken from the ureter at cystectomy. EVs were isolated by tangential flow filtration and differential ultracentrifugation and their protein composition was detected by Proximity Extension Assay (PEA; Olink, immuno-oncology panel). In UBC patients, the proteomic signature of bladder urine EVs differed from ureter urine EVs from the same individuals, and from bladder urine derived EVs of both healthy and prostate cancer controls. Pairwise comparison was performed with matched whole urine revealing proteins solely detected in isolated vesicles. Additionally, a distinct signature was identified in bladder urine EVs correlating with muscle invasiveness, and a trained classifier could predict UBC with 92 % accuracy. Some differentially expressed proteins, HO-1 and MMP7, were analysed by bead-based flow cytometry, where HO-1 was detected on the EV surface. Taken together, these results strengthen the rationale of using EVs as non-invasive biomarkers and prognostic tools for UBC. •-Urinary EVs from urinary bladder cancer patients have a specific protein signature.•-More immune-oncology proteins are found in isolated EVs than in whole urine.•-EVs from muscle invasive cancer patients upregulate a specific set of proteins.•-A machine learning model classified muscle invasiveness based on EV profile.•-EVs are promising biomarkers for detection, prognostics and diagnostic selection.