Blood-derived small extracellular vesicle-based liquid biopsy for repeatable genomic and transcriptomic monitoring in adult-type diffuse glioma.

Abstract

e14081 Background: Adult-type diffuse glioma is an aggressive primary brain tumor characterized by profound spatial and temporal heterogeneity. Although, clinical decision-making relies primarily on tissue biopsy, pathologic examination of resected tumor typically involves only a small portion of the heterogeneous tumor mass, limiting molecular interpretation. Moreover, repetitive sampling of intracranial tissue is not feasible. Small extracellular vesicles (sEV), released by all cell types and detectable in circulation, can carry nucleic acids originating from intracranial tumors, capable of crossing the blood–brain barrier. Since sEV can be isolated from blood repeatedly and non-invasively, they represent a promising platform for longitudinal genomic and transcriptomic profiling. Methods: DNA isolated from plasma sEV of glioma patients was sequenced and processed through a standardized variant calling and hard-filtering workflow (depth, mapping quality, genotype quality, strand-bias metrics). Variant recurrence across plasma-derived sEV (n=5) and concordance with matched solid tumor DNA (n=3) were assessed. RNA was also isolated from plasma sEV (n=5 healthy; n=3 glioma) and analyzed by capture-based whole-transcriptome sequencing. Results: Across sEV-DNA samples, we detected ~30–70 high-confidence variants per patient after filtering, with multiple loci recurrent across individuals. A subset of these variants was shared across all samples, supporting technical reproducibility. In matched sEV–solid biopsy datasets, more than 200 shared variants were observed in a representative paired analysis, demonstrating measurable overlap between sEV and tissue DNA while also indicating that sEV profiles contain additional variants not captured by single-site biopsy sampling. Importantly, sEV RNA whole-transcriptome sequencing detected expression of >9,000 genes and identified ~500 dysregulated genes in glioma sEV relative to healthy controls. sEV from glioma patients broadly showed transcriptional suppression compared to healthy controls, with 13 transcripts uniquely detected in glioma sEV and 43 were exclusive to controls. Notably, several recurrent sEV-DNA variants, including those annotated to WBP1L, ATXN1, and MLLT3, were consistently detected across all five glioma sEV-DNA samples, and their corresponding transcripts also showed altered expression in sEV-RNA profiling. This concordance between DNA-level variation and transcriptomic signal strengthens confidence in the sEV-derived mutation calls and highlights their potential for uncovering previously unrecognized molecular alterations relevant to glioma biology. Conclusions: Combined sEV- DNA and -RNA profiling provides a feasible, non-invasive, and repeatable approach to capture complementary genomic and transcriptomic features of glioma.