Low sample volume is one of the most common constraints in extracellular vesicle (EV) research on plasma, and it is often a hidden one. Biobank aliquots are small and finite, longitudinal cohorts cannot be resampled, and every freeze-thaw cycle changes the sample. Isolation-first EV workflows make the problem worse: they consume hundreds of microlitres of plasma per readout and lose a variable share of vesicles before any measurement starts. The real question for an EV study is how much information each microlitre of plasma can deliver.
Key takeaways
- EV isolation methods differ widely in yield and purity from the same starting volume, and every added purification step costs vesicles [4].
- Detection chemistry sets the volume floor: in one systematic comparison, a bead-based multiplex assay detected CD9 from 1,000-fold fewer EVs than Western blot [6].
- Splitting an aliquot across single-marker assays wastes sample; multiplexed surface protein readouts get more data from the same well.
- Pre-analytical handling (tube type, processing delay, freeze-thaw) has measurable effects on EV metrics and matters more when there is no spare sample [9, 10].
- MISEV2023 asks researchers to report input volumes and pre-analytical variables, which makes low-volume studies easier to compare and replicate [1].
How Much Plasma Does EV Analysis Actually Need?
The plasma volume an EV assay needs depends mostly on two things: how many vesicles the isolation step loses, and how many vesicles the detection method needs to produce a signal. Published isolation comparisons start from 200 µL to 3 mL of serum or plasma [4, 5], while sensitive isolation-free or capture-based surface protein assays have profiled EVs from as little as 10 µL [7].
That range spans more than two orders of magnitude, and the reason is not just assay sensitivity. Blood is a hostile matrix for EV measurement. Johnsen and colleagues compiled EV concentration estimates for healthy human blood from 2013–2018 studies. Values spanned more than seven orders of magnitude, with a geometric mean around 1010 EVs per mL [3]. The same analysis estimated total lipoprotein particles at roughly 1016 per mL, which would make EVs about one in a million particles in plasma [3]. Every workflow has to find that minority population, and most do it by discarding the rest of the sample.
For a researcher planning a study, this means the volume required is a property of the whole workflow. It cannot be read off the specification sheet of the final detection step.
Where the Volume Goes in an Isolation-First Workflow
The conventional EV workflow has three stages: separate vesicles from plasma (ultracentrifugation, size-exclusion chromatography, precipitation, density gradient, or affinity capture), characterize the isolate, and then measure the proteins of interest. Each stage takes its share of the sample.
Yield varies by method, and purity costs vesicles
Brennan et al. compared common isolation approaches from the same 200 µL of pooled human serum, which contained about 1.4 × 1011 particles [4]. Total particle recovery differed sharply between methods. The combined protocols designed to strip out lipoproteins and soluble proteins recovered the fewest particles. Adding a cushion step before ultracentrifugation cut small-EV yield by more than half, and the two-step SEC-plus-density-gradient protocol recovered roughly 6 × 107 particles [4]. The authors stated the trade-off directly: better removal of contaminants came with a significant loss of EVs.
Figure 1. Particle recovery from 200 µL human serum by isolation method.
| Isolation approach | Particles recovered (approx.) | Total protein co-isolated |
|---|---|---|
| Starting serum (200 µL) | 1.4 × 1011 | n/a |
| Polymer precipitation | 5.6 × 1010 | 911 µg |
| Size-exclusion chromatography (first fraction) | 4.2 × 1010 | 17.1 µg |
| Ultracentrifugation (small-EV yield) | 5 × 109 | ~772 µg |
| Cushion + ultracentrifugation (small-EV yield) | 2.3 × 109 | not reported here |
| SEC + density gradient | 6 × 107 | not reported here |
Data source: Brennan et al., 2020 [4]. Particle counts were measured by nanoparticle tracking and include non-EV particles; the authors note that counts in less-pure fractions overestimate EVs. Suggested rendering: horizontal bar chart, log-scale x-axis.
The table shows that the most effective clean-up removed more than 99.9% of starting particles. Some of that loss is the point, because lipoproteins and protein aggregates are not EVs. The problem for a low-volume study is that the same steps also remove vesicles, and the loss is not constant from sample to sample.
Co-isolated lipoproteins compete for the signal
Less aggressive methods keep more vesicles but carry more contamination. Precipitation and SEC isolates in the Brennan comparison had the highest APOB and APOE levels [4]. Johnsen et al. noted that precipitation produced EV concentration estimates about ten-fold higher than ultracentrifugation, and SEC about 15-fold higher, partly because of co-isolated apoB-containing lipoproteins [3]. Karimi et al. showed that combining a density cushion with SEC reduced lipoprotein contamination about 100-fold, allowing them to identify 1,187 proteins in plasma EVs [12]. The extra step brought its own losses and hands-on time.
No single method fits every study
Veerman et al. ran five isolation methods on 250 µL and 3 mL plasma inputs. They found that each method yielded a different protein profile and enriched different EV subpopulations [5]. They concluded that the choice should depend on sample type, EV subtype of interest, available volume, and budget. For a biobank study with 100 µL per participant, that conclusion rules out several methods before the assay is even chosen.
There is also a selection effect. Affinity capture defines the population you measure. Spitzberg et al. reported that routine CD63 affinity purification could miss up to 80% of oncogene-positive EVs, and that tetraspanin capture cocktails missed 36–47% of KRAS-mutant EVs in their cell-line models [8]. Losing volume and losing subpopulations are two sides of the same problem. (We cover the subpopulation side in more detail in Isolation Picks Your EVs for You.)
Detection Sensitivity Sets the Volume Floor
Even with perfect recovery, the detection method decides how much plasma a measurement needs. This is where low-volume studies gain or lose the most.
Wiklander et al. systematically evaluated a multiplex bead-based flow cytometry assay that captured EVs on 37 antibody-coated bead populations plus two isotype controls [6]. Tetraspanins (CD9, CD63, CD81) were detectable from 5 × 106 particles. Less abundant markers needed about ten-fold more input. In a side-by-side test with the same EVs, Western blot needed 5 × 109 EVs to detect CD9. That is a 1,000-fold difference in input for a single marker [6]. The assay also worked on unprocessed, pre-cleared cell culture supernatant, removing the isolation step from that workflow.
Figure 2. Illustrative plasma-equivalent volume needed to detect CD9 by two detection methods.
| Detection method | EV input to detect CD9 [6] | Plasma equivalent at 100% recovery | Plasma equivalent at 25% recovery |
|---|---|---|---|
| Western blot | 5 × 109 EVs | ~500 µL | ~2,000 µL |
| Multiplex bead-based capture | 5 × 106 EVs | ~0.5 µL | ~2 µL |
Illustrative. Calculated by assuming ~1010 EVs per mL of plasma (geometric mean from Johnsen et al. [3]) and hypothetical recovery rates of 100% and 25%. Detection inputs come from cell culture EVs counted by nanoparticle tracking [6], so real plasma values will vary and low-abundance markers will need more input. Suggested rendering: grouped bar chart, log-scale y-axis (µL).
The numbers are approximate, but the order of magnitude matters. A Western blot readout can use up an entire biobank aliquot for a single protein. Sensitive capture-based formats move the problem from “Do we have enough plasma?” to “How many markers can we measure from what we have?”
Recent work has pushed the input lower still. He et al. profiled EV surface proteins from 10 µL of peripheral plasma in under four hours [7]. Their method used a rapid membrane-targeting magnetic capture step (49.6% EV recovery in one hour) followed by signal amplification. In 90 participants, plasma EV EpCAM distinguished breast cancer patients from healthy donors with an AUC of 0.885 [7]. The authors framed the low input as essential for clinical samples, where volume is limited and EV surface proteins are scarce.
Why Multiplexing Matters More When Samples Are Small
With a single-marker assay, a study that needs five readouts splits the aliquot five ways, or runs five assays on five separate isolates. Each split shrinks the input per readout and adds a new source of technical variation. Multiplexing avoids this: all markers are measured on the same vesicles, in the same well, at the same time.
Heterogeneous vesicles need more than one marker
The case for multiplexing is not only about saving volume. EV populations are heterogeneous at the single-vesicle level. Mizenko et al. showed that tetraspanins are unevenly distributed across individual EVs, that tetraspanin profiles depend on the EV source, and that capturing EVs by a single tetraspanin can bias the apparent marker profile and reduce sensitivity to disease-associated surface markers [11]. Across the three EV sources they tested by flow cytometry, including human serum, only 1.3–13.6% of light-scattering events were also above background for CD9 [11].
The practical point is that a single capture marker can make a population look less abundant or less varied than it is. A panel that reads several surface proteins, including more than one tetraspanin, gives an internal check on capture bias. That check would otherwise need extra sample.
Same well, same vesicles, cleaner ratios
Measuring markers together also improves interpretation. When markers come from different aliquots or separate isolations, the ratio between them carries the variance of each preparation. When they come from one well, the ratio reflects the biology of that sample. For longitudinal designs, where the question is how a participant’s EV profile changes over time, this reduces the noise that would otherwise hide small effects.
This is the gap the LuminEV™ Research Kit was built to address: multiplex EV surface protein profiling directly in plasma, cell culture media, or purified EVs, without a separate isolation step. Because it runs in a 96-well format, a whole cohort can be profiled on the same plate layout, with tetraspanin markers measured alongside the proteins of interest.
Protecting Low-Volume Samples Before They Reach the Assay
When a study has no spare plasma, pre-analytical handling becomes part of the assay. A sample that was mishandled at collection cannot be replaced by thawing another aliquot.
Dhondt et al. benchmarked 11 blood collection tubes and three processing intervals (1, 8, and 72 hours) across nine EV performance metrics. They found that tube type and processing delay significantly affected haemolysis, platelet activation, ex vivo release of blood-cell EVs, EV recovery, and EV molecular profiles [9]. Platelet activation matters for EV studies because platelets are a major source of plasma EVs, and vesicles released after the blood draw dilute the signal from the tissue of interest. Clayton et al., writing for the ISEV blood EV task force, laid out a roadmap for standardizing collection, handling, and storage, and called for quality control procedures and reference materials to make blood EV studies comparable [10].
For low-volume work, a few habits protect the data:
- Aliquot at first processing. Store plasma in volumes that match the planned assay input, so each analysis thaws only what it uses and freeze-thaw cycles are not repeated.
- Record everything MISEV2023 asks for. The updated guidelines ask researchers to report pre-analytical variables, input volumes, and the methods used to separate and characterize EVs [1]. The earlier MISEV2018 framework set the baseline for these reporting fields [2].
- Keep isolation steps consistent across the cohort. If an isolation step is necessary, use the same method, operator, and batch protocol throughout. Brennan et al. and Veerman et al. both show that switching methods changes what is measured [4, 5].
- Plan the panel before thawing. Decide which markers matter, including tetraspanin controls, so one thaw delivers every readout the study needs.
Choosing a Workflow That Fits the Sample You Have
Low sample volume does not have one fix, but it does set the order of decisions. Start with the volume available per participant and work backwards: how many readouts are needed, which detection method can produce them from that volume, and whether an isolation step is required for the biological question.
Figure 3. Two ways to spend a 100 µL plasma aliquot.
| Step | Isolation-first, single-marker | Isolation-free, multiplexed |
|---|---|---|
| 1 | Thaw aliquot | Thaw aliquot |
| 2 | Isolate EVs (UC, SEC, or precipitation); variable loss | Add plasma directly to capture well |
| 3 | Characterize isolate (particle count, purity markers) | Capture and detect several surface proteins in parallel, tetraspanins included |
| 4 | Split isolate across single-marker assays | Read all markers from the same vesicles |
| 5 | Each marker gets a fraction of the input | Leftover plasma stays available for orthogonal assays |
| Outcome | Few markers per aliquot; ratios carry the variance of each preparation | More markers per aliquot; ratios come from one well |
Conceptual comparison (illustrative). Suggested rendering: two-lane flowchart with sample volume shown as a shrinking bar at each step.
Isolation still has a place. Proteomics, RNA sequencing, and functional studies usually need purified vesicles, and MISEV2023 describes when separation is necessary to support a claim [1]. But for surface protein profiling across large or longitudinal cohorts, where the aim is to compare samples with each other, removing the isolation step removes the largest source of sample loss and much of the variability between preparations. Platforms such as LuminEV were designed for that use case.
What comes next
Plasma EV research is moving from small case-control comparisons to biobank-scale cohorts, and blood-based biomarkers are increasingly expected to track change within one person over time. In that setting, the sample is the limiting resource, and methods will be judged by how much reliable information they get per microlitre. Single-vesicle methods, more sensitive capture chemistry, and standardized pre-analytics under MISEV2023 are all moving in that direction. The studies that benefit most will be the ones that plan their EV panels around the volume they actually have.
References
- Welsh JA, Goberdhan DCI, O’Driscoll L, et al. Minimal information for studies of extracellular vesicles (MISEV2023): From basic to advanced approaches. J Extracell Vesicles. 2024;13(2):e12404. https://doi.org/10.1002/jev2.12404
- Théry C, Witwer KW, Aikawa E, et al. Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines. J Extracell Vesicles. 2018;7(1):1535750. https://doi.org/10.1080/20013078.2018.1535750
- Johnsen KB, Gudbergsson JM, Andresen TL, Simonsen JB. What is the blood concentration of extracellular vesicles? Implications for the use of extracellular vesicles as blood-borne biomarkers of cancer. Biochim Biophys Acta Rev Cancer. 2019;1871(1):109-116. https://doi.org/10.1016/j.bbcan.2018.11.006
- Brennan K, Martin K, FitzGerald SP, et al. A comparison of methods for the isolation and separation of extracellular vesicles from protein and lipid particles in human serum. Sci Rep. 2020;10:1039. https://doi.org/10.1038/s41598-020-57497-7
- Veerman RE, Teeuwen L, Czarnewski P, et al. Molecular evaluation of five different isolation methods for extracellular vesicles reveals different clinical applicability and subcellular origin. J Extracell Vesicles. 2021;10(9):e12128. https://doi.org/10.1002/jev2.12128
- Wiklander OPB, Bostancioglu RB, Welsh JA, et al. Systematic methodological evaluation of a multiplex bead-based flow cytometry assay for detection of extracellular vesicle surface signatures. Front Immunol. 2018;9:1326. https://doi.org/10.3389/fimmu.2018.01326
- He J, Li H, Mai J, et al. Profiling extracellular vesicle surface proteins with 10 µL peripheral plasma within 4 h. J Extracell Vesicles. 2023;12:e12364. https://doi.org/10.1002/jev2.12364
- Spitzberg JD, Ferguson S, Yang KS, et al. Multiplexed analysis of EV reveals specific biomarker composition with diagnostic impact. Nat Commun. 2023;14:1239. https://doi.org/10.1038/s41467-023-36932-z
- Dhondt B, Pinheiro C, Geeurickx E, et al. Benchmarking blood collection tubes and processing intervals for extracellular vesicle performance metrics. J Extracell Vesicles. 2023;12(5):e12315. https://doi.org/10.1002/jev2.12315
- Clayton A, Boilard E, Buzas EI, et al. Considerations towards a roadmap for collection, handling and storage of blood extracellular vesicles. J Extracell Vesicles. 2019;8(1):1647027. https://doi.org/10.1080/20013078.2019.1647027
- Mizenko RR, Brostoff T, Rojalin T, et al. Tetraspanins are unevenly distributed across single extracellular vesicles and bias sensitivity to multiplexed cancer biomarkers. J Nanobiotechnology. 2021;19:250. https://doi.org/10.1186/s12951-021-00987-1
- Karimi N, Cvjetkovic A, Jang SC, et al. Detailed analysis of the plasma extracellular vesicle proteome after separation from lipoproteins. Cell Mol Life Sci. 2018;75(15):2873-2886. https://doi.org/10.1007/s00018-018-2773-4
