Logo: to the web site of Uppsala University

Background

Gastric cancer (GC), the world’s fifth most common malignancy, remains a major cause of death, motivating minimally invasive protein measurements. Small extracellular vesicles (sEVs) mirror cellular state but are difficult to profile from limited material. Proximity extension assay (PEA) enables multiplex, sensitive detection. We developed a PEA-centred workflow to profile GC cell-line sEVs, benchmarked against cells and an EV-rich control, and nominate hypothesis-generating candidates.

Methodss

EVs from four GC cell lines (Kato III, AGS, MKN45, MKN7) were isolated by differential centrifugation/filtration and ultracentrifugation; prostasomes (seminal-fluid EVs; ethics-approved) were purified by ultracentrifugation, Superdex 200 size-exclusion, and sucrose gradient. EVs were characterized by TEM, NTA, and Western blot (CD9/CD63/CD81; Calnexin). Protein profiling used Olink PEA across five Target 96 panels, two commercial (Oncology, Inflammation) and three experimental (Cancer, Cellular Pathways, Neurology) with availability-based selection; single replicate per condition. NPX values were background/LOD-corrected in R, with PCA. STRING PPI used medium confidence (0.4) with FDR-adjusted enrichment. Selected proteins were validated by SP-PLA (qPCR), EV-PLA (flow cytometry), and WB.

Resultss

EVs and prostasomes were purified and characterized per MISEV2023 (TEM/NTA; CD9/CD63/CD81+, calnexin–). Using five Olink PEA panels, 460 proteins were profiled; principal component analysis separated sEVs from parental cells. We identified 71 proteins enriched in GC sEVs, including EphA2, cMet, IL-6, SCF, AXIN1, NOTCH1, and CAV1; 17 were unique to Kato III sEVs. STRING showed a highly connected network, with top pathways: Pathways in cancer, Cytokine–cytokine receptor interaction, and PI3K–Akt. Orthogonal assays (SP-PLA, EV-PLA, WB densitometry) confirmed the presence of EphA2 and cMet on GC sEVs.

Discussions

PEA-based profiling of GC cell-line sEVs nominates 71 enriched proteins (e.g., EphA2, cMet, NOTCH1, CAV1) as hypothesis-generating candidates that may reflect cell-type–specific cargo and tumor–microenvironment signaling. Key limitations include the in-vitro design, single replicates, preselected PEA panels, and the use of prostasomes rather than healthy plasma sEVs, which may affect generalizability and specificity. Translational relevance requires validation in patient biofluids with orthogonal assays and cohort-level evaluation.