Extracellular Vesicle RNA Sample Prep, Done Right

Team Dynamic Matrices | 4 August 2026

Extracellular vesicles carry RNA from their cell of origin into blood, urine, saliva and CSF, which makes them one of the most promising sources of non-invasive biomarkers. They are also among the hardest samples to prepare well. Most EV RNA studies that fail do so before the sequencer is touched: in isolation, in extraction, or in a QC step that measured the wrong thing. Here is how we approach it.

Why isolation is hard

EVs are 30 to 150 nm for small EVs (exosomes) and up to a micron for larger classes. They are outnumbered in plasma by lipoproteins of similar size and density, by protein aggregates, and by free protein at roughly a million-fold excess by mass. Every isolation method is a compromise (Théry et al., MISEV2018; Welsh et al., MISEV2023):

• Differential ultracentrifugation at 100,000 g or more pellets EVs together with protein aggregates and lipoproteins, and the shear and pressure rupture a fraction of vesicles. Recovery is often under 30 percent.
• Size exclusion chromatography separates EVs from free protein cleanly but not from lipoproteins of the same size, particularly chylomicrons and VLDL in non-fasting plasma.
• Polymer precipitation (PEG-based kits) gives high apparent yield because it co-precipitates protein complexes, lipoproteins and free RNA. Purity is low and downstream RNA is contaminated with non-vesicular material.
• Immunoaffinity capture on CD63, CD9 or CD81 beads isolates only marker-positive vesicles and misses subpopulations, and elution from beads is harsh.
• Density gradient after ultracentrifugation improves purity at the cost of time and yield.

MISEV2023 recommends reporting the method, the recovery, and the purity (particle-to-protein ratio, markers present and absent) for exactly this reason. Choose a method for the purity your question needs, and report it.

The RNA is different

EV cargo is dominated by short fragments and small RNAs: miRNA, tRNA fragments, YRNA, and fragmented mRNA. Full-length transcripts are rare. Total RNA from a millilitre of plasma is typically in the low-nanogram to sub-nanogram range. That leaves no margin for the losses that a standard column protocol accepts without noticing.

Cell-free RNA outside vesicles (in ribonucleoprotein complexes, on lipoproteins) co-isolates with most methods and dilutes the vesicle signal. RNase treatment of the EV prep before lysis removes free RNA while the membrane protects the cargo, and is a useful control for what fraction of your signal is vesicular.

Preanalytical rules that matter more than the kit

1. Anticoagulant. EDTA or citrate for RNA; heparin inhibits RT and PCR. Record it.
2. Time to processing. Under two hours at room temperature; platelets release vesicles on standing.
3. Platelet removal. Two centrifugation steps (for example 2,500 g, 15 minutes, twice) to make platelet-free plasma before freezing.
4. Freeze-thaw. One cycle. Aliquot before freezing.
5. Fasting state. Lipoprotein load in post-prandial plasma badly contaminates SEC fractions.

Most inter-lab variability in EV RNA data comes from these five points, not from the isolation kit.

Extraction and QC

Use an extraction chemistry validated for small RNA retention (higher ethanol at binding, or phenol methods with carrier). Add a carrier (glycogen or linear acrylamide) and synthetic spike-ins (cel-miR-39 or similar) before lysis to measure recovery.

Do not use RIN. It depends on 18S and 28S peaks that EVs do not contain. Use a small-RNA or pico-RNA electrophoresis assay and report the small RNA fraction and any residual rRNA. Quantify by fluorometry, not absorbance.

Where gentle phase-separation capture fits

Two steps in this workflow suffer from surfaces and force: vesicle isolation, and any targeted capture or depletion of the RNA afterwards. Phase-separation chemistry replaces a solid surface with a soluble polymer that condenses around its target. For vesicles this avoids high-g pelleting, non-specific co-precipitation and marker-restricted capture, and we cover our approach to EV isolation from culture media in a separate post. For the RNA, LASSO capture at low input avoids the surface-adsorption losses that dominate bead workflows at nanogram scale, which is where EV RNA lives.