Glioblastoma biomarkers in urinary extracellular vesicles reveal the potential for a ‘liquid gold’ biopsy

Glioblastoma (GBM), the most aggressive primary brain tumor in adults, urgently needs reliable biomarkers to monitor tumor activity and treatment response, especially for recurrent disease. Current methods, like radiological imaging, are often confounded by pseudoprogression. Liquid biopsies, which analyze tumor-derived factors in bodily fluids, offer a less invasive alternative. Extracellular vesicles (EVs), nanoparticles released by cells, have emerged as promising carriers of GBM biomarkers in various body fluids, including cerebrospinal fluid (CSF) and blood. However, collecting CSF is invasive, and analyzing blood is complicated by its molecular complexity. The urinary system is a major clearance route for circulating EVs, leading researchers to hypothesize that GBM biomarkers might also be detectable in urinary EVs (uEVs). While uEVs have been studied in kidney diseases and other cancers, their potential as GBM biomarkers remains largely unexplored. This study aimed to assess the feasibility of using uEVs as a simple, non-invasive liquid biopsy for detecting GBM biomarkers, a 'liquid gold biopsy', by performing in-depth proteomic analysis of uEVs from GBM patients and healthy controls. This would involve employing a sensitive data-independent acquisition liquid chromatography tandem mass spectrometry (DIA-LC-MS/MS) method and exploring if established GBM EV biomarkers from other biological systems are reflected in patient urine.

Previous research has demonstrated the presence of GBM biomarkers in various body fluids, including neurosurgical fluids, CSF, and peripheral blood. Extracellular vesicles (EVs) have been identified as key carriers of these biomarkers. Studies have shown the potential of EVs in other cancers, such as breast cancer, to act as diagnostic biomarkers. However, the complexities of analyzing blood and the invasive nature of CSF collection have prompted the exploration of less invasive sources, such as urine. Prior studies have also explored the use of urinary MMPs and other proteomic approaches to identify GBM biomarkers in urine. Challenges in comprehensive proteomic characterization of EVs from bodily fluids include the presence of abundant proteins like uromodulin, which can mask the detection of less-abundant potential biomarker proteins. The current study utilizes a more advanced DIA-MS approach to overcome some of these limitations.

The study involved collecting urine specimens (15–60 ml) from 24 glioblastoma patients (17 pre-operative, 7 recurrence, and 9 post-operative) and 14 age/gender-matched healthy controls. Extracellular vesicles (EVs) were isolated from urine samples using differential ultracentrifugation and characterized using nanoparticle tracking analysis (NTA) and cryo-transmission electron microscopy (cryo-EM). The EV proteomes were extracted, prepared, and analyzed using high-resolution data-independent acquisition liquid chromatography tandem mass spectrometry (DIA-LC-MS/MS). DIA-MS data were processed using Spectronaut software, and differential expression analysis was performed using Perseus software. Functional annotation of identified proteins was performed using Ingenuity Pathway Analysis and DAVID. Statistical analysis included Student’s t-tests, ROC curve analysis, and logistic regression to identify potential biomarkers. A custom GBM spectral library, generated from primary patient-derived GBM cells and GBM-EVs, was used to enhance the specificity and confidence of proteomic identifications.

A total of 6857 proteins were confidently identified in urinary-EVs, including 94 EV marker proteins. The study identified glioblastoma-specific proteomic signatures and potential urinary-EV biomarkers associated with tumor burden and recurrence (FC ≥ 2, adjusted p-value ≤ 0.05, AUC > 0.9). Analysis revealed 15 proteins differentially abundant between pre-operative GBM and healthy controls. Nine of these proteins (KRT19, RSP1, RPL8, RPL23, CSTA, ALDH1B3, RPL7A, GNAI2, FC2) showed excellent diagnostic sensitivity and specificity (AUC > 0.9). A stepwise logistic regression model identified a panel of five diagnostic uEV proteins (KRT19, RSP1, RPL8, RPL23, RPL7A) with a cumulative AUC of 0.958. Comparison of pre-operative and post-operative samples identified 72 proteins associated with tumor burden, while comparison of post-operative and recurrence samples identified 64 proteins associated with recurrence. Among these, GRN (progranulin) and PSAP (prosaposin) showed significant differential abundance across all three clinical timepoints (pre-operative, post-operative, recurrence). A logistic regression model selected ITM2B (integral membrane protein 2B) alongside GRN as the best performing uEV biomarkers for GBM recurrence. Multiple ribosomal protein subunits were also found to be significantly higher in GBM uEVs compared to healthy controls, suggesting a potential role as pan-cancer markers. Functional analysis revealed significant enrichment of pathways associated with cancer and protein folding.

The findings of this study support the feasibility of using urinary EVs as a non-invasive liquid biopsy for glioblastoma. The identification of several potential biomarkers associated with both GBM diagnosis and progression (tumor burden and recurrence) demonstrates the potential of this approach for monitoring disease and treatment response. The use of DIA-MS, along with a custom GBM spectral library, allowed for a more comprehensive and accurate analysis of the uEV proteome. The identification of ribosomal proteins and other proteins involved in pathways associated with cancer further supports the findings and highlights potential mechanistic insights. The results suggest that a panel of uEV biomarkers could potentially distinguish between different clinical stages of GBM and even differentiate between tumor recurrence and pseudoprogression.

This feasibility study demonstrates the potential of urinary extracellular vesicles as a novel source of glioblastoma biomarkers. The identified potential biomarkers warrant further investigation in larger, longitudinal studies to validate their clinical utility. Future research should focus on validating these biomarkers in larger patient cohorts, exploring the potential of uEV miRNAs, and investigating the clinical utility of these biomarkers in monitoring treatment response and predicting prognosis. This 'liquid gold biopsy' approach holds promise for minimally invasive glioblastoma surveillance.

The relatively small sample size is a limitation of this study. The short post-operative sampling time might have affected the accuracy of reflecting tumor burden reduction. The study focused on IDH-wildtype GBM and may not be generalizable to other subtypes. While efforts were made to minimize uromodulin contamination, it could still have impacted the results. Future studies should address these limitations by including larger cohorts and longitudinal sampling.