Cargo profiling of extracellular vesicles from senescent glioblastoma identifies small nucleolar RNAs as candidate liquid biomarkers for therapy-induced senescence

Therapy-induced senescence (TIS) in glioblastoma (GBM) is an undesirable cell fate that contributes to GBM treatment resistance and tumor outgrowth. While emerging senescence-targeting drugs called senotherapies may be beneficial for patients who undergo TIS following standard of care, identifying candidate patients with an appreciable TIS burden is difficult. Current methods to measure TIS are tissue based and require repeated, sequential biopsies for robust analysis. The impracticality of such an effort for GBM patients therefore necessitates the development of novel, less-invasive approaches that can detect TIS clinically over time. To this end, we aimed to identify candidate extracellular vesicle (EV) liquid biomarkers for TIS by profiling senescence-associated cargo changes within GBM EVs. Using a panel of GBM patient derived cell lines, we show that radiation induces TIS across GBM models and that this is associated with a reprogramming of the cargo in the corresponding senescent-derived EVs (senEVs). In particular, senEV transcriptomes have an increased abundance of senescence-associated RNA species and enrichment of senescence-associated gene sets. Most striking, however, was the observed enrichment of small nucleolar RNAs (snoRNAs), which represented over 50% of the top 20 most upregulated species in senEVs. This signature was conserved in 4/5 GBM models of TIS and was validated by qRT-PCR. Interestingly, the increased snoRNA abundance in EVs is not a simple function of increased cellular content, as whole cells did not show consistent upregulation upon TIS. We further found that snoRNAs are likely packaged with their partner proteins, as mass spectrometry of senEVs revealed an increased abundance of snoRNA-associating proteins. Notably, a GBM model of temozolomide-induced stress that does not result in prominent senescence failed to promote increased snoRNA cargo, supporting snoRNA enrichment as a TIS-selective finding. Ongoing work is focused on identifying whether this senEV profile is conserved across tumor types and senescence-inducing stimuli, and whether nucleolar stress during TIS plays a mechanistic role in snoRNA EV packaging. Finally, to determine the feasibility of detecting these candidate biomarkers in patient biofluids, we employed a small patient cohort of matched pre-operation and post-standard of care plasma samples. Promisingly, we identified increased senescence-associated RNA such as CDKN2B and GLB1 and the snoRNA SNORA49 in post-standard of care EVs. This data suggests that senEV RNA species, specifically snoRNAs, are strong candidates for TIS biomarker development.