Skip to main content
ENDO 2026 Main banner

Diabetes and Vascular Disease

SUN-539 - Plasma Extracellular Vesicles as Predictive Biomarkers for Developing Type 1 Diabetes

Sunday, June 14, 2026
9:00 AM - 4:00 PM CT
Location: ENDOExpo; Poster Floor
Disclosure(s):
Ernesto S. Nakayasu, PhD: No financial relationships to disclose

Plasma extracellular vesicles (EVs) are considered potential sources of disease biomarkers since they carry signatures of their cellular origin and physiopathological processes. Here, we assess the potential of plasma EV proteomics analysis to identify predictive biomarkers of developing type 1 diabetes (T1D). As plasma EV proteomics analysis is challenging due to co-isolation of contaminants, we built upon our previous concept of separating the characterization and quantification of EV proteins in different experiments (DOI: 10.1002/pmic.70036). This approach allows us to focus on EV recovery and maximizing quantification of individual proteins. Here, we used a method based on strong anion exchange beads (Mag-Net) to enrich plasma EVs followed by mass spectrometry analysis. Samples consisted of 19 donors with islet autoimmunity, determined by the presence of circulating autoantibodies against islet proteins, vs. 17 control individuals. The analysis led to the identification and quantification of 5482 proteins, a 3.2-fold increase in coverage compared to our previous plasma proteomics analysis to identify biomarker candidates for T1D. These proteins also included 1315 of 1717 proteins (77%) that we previously validated as EV proteins (DOI: 10.1038/s41597-023-02748-1). Statistical analysis identified 448 proteins to be quantitative or qualitatively differentially abundant in AAB+ vs control volunteers, including 69 proteins previously validated to be present in EVs. Moreover, a functional-enrichment analysis identified 25 pathways enriched among the differentially abundant proteins, including pathways related to autoimmune response and lipid metabolism. Lastly, machine learning analysis using a random forest model identified a multivariate panel of proteins that can predict autoantibody positivity with receiver operating characteristic-area under the curve of 0.805. In summary, proteomics analysis of plasma EVs represents an exciting approach for developing biomarkers for T1D.