MISEV2023 is often summarised as "the EV characterization standard." That undersells it. The guideline is also, less famously, an isolation-method-honesty document. Section after section returns to the same working point: tell the reader exactly how the EVs were separated from their source material, characterize what came out of that workflow, and report it. The method is not background detail. It is part of the product specification.
This is the framing most commercial EV product pages omit. "Isolated from species X by proprietary methods" appears on a non-trivial fraction of exosome skincare labels. Under MISEV2023, that sentence is incomplete by design.
What MISEV2023 actually requires about isolation
The 2023 update sharpens three asks from earlier MISEV iterations. First, the method has to be named. Ultracentrifugation, density gradient, ultrafiltration, size-exclusion chromatography (SEC), tangential flow filtration (TFF), polymer-based precipitation, and immunoaffinity capture are not interchangeable; the choice has downstream consequences and the consumer of the preparation is owed the information.
Second, the method has to be characterized. It is not enough to assert that ultracentrifugation produces a "clean" preparation. The guideline expects the manufacturer to demonstrate, with measurements, what the chosen workflow yields in terms of particle concentration, marker enrichment, contamination, and where applicable, the integrity of the biomolecular corona and the absence of non-vesicular contaminants. These measurements should be batch-specific.
Third, where the method strips, modifies, or co-purifies something that affects the active fraction, that has to be disclosed in the characterization record. "We used ultracentrifugation and confirmed the resulting preparation by NTA, Western for CD9 and CD63, and SDS-PAGE for soluble protein contamination" is a defensible disclosure. "Proprietary isolation" is not.
Why method changes the product
Two preparations from identical biomass, separated by different methods, are not the same product. The mechanism is straightforward and has been demonstrated experimentally in our own published work.
Ultracentrifugation applies high gravitational force over extended periods. It pellets EVs efficiently but also shears more labile components of the biomolecular corona, fragments fragile vesicle subpopulations, and co-pellets contaminating non-vesicular protein aggregates that travel with the EV fraction into downstream characterization. The resulting product is often particle-rich but corona-altered.
Ultrafiltration uses a size-cutoff membrane and depends on pressure-driven flow. It preserves more of the soft corona because the EVs are not subjected to high g-forces, but it can introduce adsorption losses to the membrane material and concentration-dependent aggregation effects. SEC preserves the corona more gently than either, with its own trade-offs in dilution, throughput, and yield. TFF combines size selection with continuous flow and is the workflow most amenable to scale, with characterization implications of its own.
None of these methods is wrong. They produce different preparations from the same starting material. MISEV2023's point is that the manufacturer has to know which one they used and characterize what came out of it. A "proprietary" answer leaves the buyer guessing about a load-bearing variable.
Three implications for evaluating an EV product
First, source biomass is not a complete specification. "Aloe-derived EVs" or "platelet-derived exosomes" tells the buyer the species and tissue. It does not tell the buyer what the surface chemistry, particle integrity, or corona profile of the final product looks like. Source plus isolation method together define the active particle.
Second, batch consistency requires control over both source variability and processing variability. A manufacturer that has not characterized how its isolation method shapes the corona is not controlling one of the two main drivers of lot-to-lot drift. Asking which method is used, and how it was validated, is a legitimate clinician question.
Third, "proprietary" should not mean "undisclosed." Method-level intellectual property is real and worth protecting. MISEV2023 does not ask a manufacturer to publish a step-by-step protocol on a marketing page. It asks for a defensible disclosure to qualified clinical and manufacturing partners under appropriate agreement. The default is "characterized and disclosed under terms," not "characterized in private."
What BioThera does and why
mPDEV is produced through a gentle, scalable isolation workflow designed to avoid the mechanical and chemical damage that more aggressive methods can inflict on plant-derived EVs. The detailed protocol is proprietary and IP-protected. Process transparency documentation, including characterization of how the workflow preserves the biomolecular corona, is available to qualified manufacturing partners under NDA. Every production batch ships with a CoA reporting NTA-measured particle concentration, marker panel data, and the absence of relevant contaminants.
The reason the workflow gets discussed alongside the product, rather than buried under "proprietary," is that MISEV2023 makes it a load-bearing variable. The guideline's authors, this one included, designed it that way.
What this looks like in practice
A clinician evaluating a candidate EV product can ask four questions that operationalise MISEV2023's isolation-disclosure requirement:
- Which isolation method was used to produce this batch?
- Has the method been validated for impact on particle integrity, marker enrichment, and biomolecular corona retention?
- Are batch-specific characterization data (NTA, marker panel, contamination screen) available?
- Is the level of process transparency on offer appropriate for the product's intended use — over-the-counter cosmetic versus professional-use clinical product versus investigational therapeutic?
A manufacturer that can answer these four questions cleanly is operating inside MISEV2023's intent. A manufacturer that cannot is, regardless of marketing claims, producing a preparation whose composition is partly accidental.