Abstract

Exploiting the Unique Protein Distribution in Different Plasma EV Types For Biomarker Discovery

Sai Kiang Lim, Research Director, A*STAR Institute of Medical Biology

Intercellular communication via secreted lipid membrane vesicles or extracellular vesicles (EVs) is an emerging research area.  Many cell types are known to secrete EVs and these EVs are heterogenous in size, density, composition and biogenesis pathway.  EVs in the size range of 100-200 nm are among the most intensely studied EVs, and are frequently referred to as exosomes which are EVs derived from endosomes.  Recently, we demonstrated by pulse chase experiments that Mesenchymal Stem Cell (MSC) exosomes can be identified by their high affinity for cholera toxin B chain (CTB) which binds GM1 gangliosides.  MSC was also observed to secrete at least two other EV types that can be distinguished by their affinity for lipid-binding proteins namely, AnnexinV (AV) that binds phosphatidylserine and Shiga toxin B chain (STB) that binds globotriasosylceramide.  These three EV types have different protein and RNA cargo.  Several other cell types were also observed to produce CTB-, AV- and ST-binding EVs (CTB-EVs, AV-EVs and ST-EVs).    Evaluation to date suggest that these different EV types are unique, and carry different protein and RNA cargos.  We have since demonstrated that plasma also have CTB-EVs, AV-EVs and ST-EVs.  Hence plasma biomarkers can now be analyzed in four permutations: plasma, plasma CTB-EVs plasma AV-EVs or plasma ST-EVs.  The distribution of some plasma proteins in these four permutations is different among individuals, and could be associated with diseases.   As such, plasma CTB-EV and AV-EV could be sources of novel biomarkers or novel permutations of known biomarkers.  Here I will describe using this approach to discover three biomarkers in plasma, plasma CTB-EV and plasma AV-EV of 28-32 weeks pregnant women that could predict development of PE at an average of seven weeks later.  This study was done using a prospective biobank of 843 low risk patients with a 2.1% PE risk.  At 100% sensitivity, the three biomarkers have a combined AUC of 0.96, 78.6% specificity and a PPV of 9.9%.  I will also present the discovery of ovarian cancer biomarker candidates using similar approaches.