Efficient enrichment of plasma-derived extracellular vesicles from small volumes of bovine blood

Abstract

There is a growing interest in small extracellular vesicles (sEVs). These nanoparticles, which range in diameter from 30 to 150 nm, are secreted by cells into their surrounding environment and transfer biological content to distant cells. However, the lack of consensus on sEV isolation, from bovine plasma limits their study. This work aimed to develop an optimized method to enrich sEVs from 4 mL of bovine blood plasma. To increase the yield of sEVs while reducing contamination from other particles and free proteins, sEVs were isolated from 38 bovine plasma samples of crossbred heifers using sequential centrifugation and filtration with size-exclusion chromatography. In accordance with the Minimal Information for Studies of Extracellular Vesicles (MISEV) guidelines, the sEV preparations were characterized in terms of size, particles concentration, morphology, and sEV markers. To accurately estimate particle size and distribution, we used a combination of three methods. This approach confirmed that 76% of the particles fell within the expected range of 30-150 nm for sEVs. The preparations were pure, with an average particle-to-protein ratio of 2.4 × 108 particles/µg of protein. This is comparable to or exceeds recent observations in bovine and other mammalian species when blood plasma and serum are used. Moreover, albumin, accounted for only 1.8-6.5% of the final protein abundance, indicating a 90-98% depletion relatively to raw plasma. Microscopy confirmed the presence of cup-shaped particles characteristic of sEVs. Proteomic characterization identified 417 proteins (FDR 1%, ≥ 2 peptides), corresponding to 372 unique homologous human gene names, including the cytosolic (HSPA8, SDCBP, ACT, TUB, GAPDH) and membrane (CD9, CD81) markers of sEVs. Of these proteins, 347 (93%) are referenced in Vesiclepedia, an international database of sEV proteome, suggesting a strong enrichment of sEVs during the purification process. This finding is supported by the identification of 172 significantly enriched Gene Ontology terms related to sEV annotation (P < 0.01, Fisher's one-tailed test with Benjamin-Hochberg correction) such as GO:0005615 (extracellular space) and GO:1903561 (extracellular vesicle). According to the MISEV guidelines and proteomic requirements, the proposed optimized sEV enrichment protocol is suitable for 4 mL of plasma. These results pave the way for future research into the role of sEVs in relation to animal health and performance.

Keywords: Biomarker discovery; lipoprotein removal; methodology optimization; phenotypic expressions; proteomics; vesicle-associated proteins.

Figure 1.

Workflow of plasma-derived sEV methods of preparation, purification, and analysis.

Figure 2.

A TEM analysis of sEVs isolated from bovine plasma with negative UranyLess staining. (A) Wide-field view of sEVs showing their morphology. (B) Close-up showing vesicles with the cup-shaped morphology typical of sEVs. (C) Close-up showing a cluster of aggregated particles, a common feature of plasma-derived sEVs. Scale bars are indicated on the pictures of the sEVs. Black arrows indicate particles showing sEVs structures. White arrows indicate particles showing lipoprotein structures.

Figure 3.

Characterization of bovine sEVs measured by TRPS* and protein concentration, n = 38. (A) Particle diameter distribution and relative concentration of sEVs with 95% confidence interval, based on measurements across all animals. (B) Particles concentration in PCV. (C) Protein concentration in concentrated PCV. (D) Purity of sEVs, total number of particles per total µg of protein. (*qNano, 150 nm nanopore, operated at 7 and 10 mbar pressure).

Figure 4.

Average dynamic light scattering (DLS) characterization of sEVs isolated from bovine plasma, n = 38. The intensity curve (blue line) reflects the measured relative amount of light scattered by particles (scattering intensity) of different sizes, with larger particles highlighted due to their greater scattering. The number curve (red line) shows the calculated relative number of particles, highlighting the predominance of particles within the expected sEVs range of 30–150 nm. The volume curve (grey line) shows the proportional calculated volume occupied by particles of different sizes.

Figure 5.

NanoFCM analysis of sEVs samples using FITC-labeled anti-CD9 antibody. (A) Particle size distribution on pooled fractions (B–F). Flow cytometry dot plots showing side scatter height (SS-H, X-axis) versus fluorescence intensity (FITC-A, Y-axis). (B) LI_PA-sEVs pool incubated with FITC-labeled anti-CD9 antibody. (C) Box plot of %FITC+Events for AN_PA-sEVs, SI_PA-sEVs, LI_NP-sEVs, AN_NP-sEVs, SI_NP-sEVs. (D) Isotype control*: LI_PA-sEVs was incubated with FITC-labeled isotype antibody showing low levels of non-specific binding. (E) LI_PA-sEVs without antibody incubation: out of 201 events recorded, only 8 were detected as FITC positive by the laser. (F) Negative control: PBS with FITC-labeled anti-CD9 antibody, showing minimal background fluorescence. FITC-positive (red), unstained (blue).

Figure 6.

Venn diagrams of proteins detected that match human gene name versus (A) Human protein atlas-blood plasma; (B) bovine plasma derived EVs detected by Turner et al. (2023); (C) Vesiclepedia database 5.1; (D). Vesiclepedia (Bos taurus).

Figure 7.

Network visualization illustrating relationships among the top 10 significantly enriched GO cellular component (GO:CC) terms (FDR < 0.01) generated thanks to ShinyGO. Analysis parameters included pathway size limits (min = 2, max = 5000). Nodes represent enriched GO term. The thickness of the lines reflects the percentage of overlapping genes. Node size corresponds to the number of genes associated with each term.

Figure 8.

Overview of particle types in mammalian plasma and the impact of purification steps on their populations. (A) Expected concentration ranges, sizes, and densities of plasma particles for each particle type are indicated*. (B) Expected log10(concentration) of particle populations at each step of the purification process. *These values are rough estimates and should be interpreted with caution. Lipoproteins are commonly assumed to be approximately 10³–10⁶ times more concentrated than sEVs in plasma (Simonsen, 2017; Johnsen et al., 2019; Ballantyne, 2024; Chou et al., 2024; Nieuwland and Siljander, 2024).