MISEV2023 is most often misquoted on the marker question. The common shorthand — "EVs have CD9, CD63, and CD81 on them" — is not what the guideline says. The guideline says that demonstrating a preparation is EV-enriched requires a panel of markers covering several categories, and that the panel choice depends on what is being claimed about the preparation. The single-marker positive result, especially on a Western blot, is not sufficient evidence on its own.
Understanding why is the difference between a defensible characterization record and a marketing-grade one.
What the canonical markers actually are
MISEV2023 organises EV markers into three working categories.
Category 1 — Transmembrane EV-associated proteins. The tetraspanins CD9, CD63, and CD81 are the canonical examples. They are enriched in EVs across many biogenesis routes and serve as positive markers that the preparation contains material of EV origin. Different EV subpopulations express the three tetraspanins to different extents, which is also informative.
Category 2 — Cytosolic proteins recoverable in EVs. TSG101 and Alix are the canonical examples. Both are associated with the ESCRT machinery involved in multivesicular body biogenesis. Their presence in a preparation supports an exosome (MVB-derived) origin and helps distinguish exosomes from microvesicles, which bud from the plasma membrane and are typically lower in ESCRT-associated proteins.
Category 3 — Markers of non-EV contaminants. This is the category most often skipped in commercial characterization. MISEV2023 expects evidence of the absence of contamination — typically high-density lipoproteins (apolipoproteins like APOA1, APOB), serum albumin, and cellular debris markers, depending on the source biomass. A preparation that is enriched for CD9 but also enriched for apolipoproteins is co-purifying lipoproteins along with the vesicles. That is information the buyer needs.
Why no single marker is sufficient
Three reasons, all of them load-bearing.
First, no marker is exclusively expressed by EVs. CD9, CD63, and CD81 are enriched in EVs relative to whole-cell lysates, but they are also present on cellular membranes and in cellular subcompartments. A Western band for CD9 in a preparation tells you the preparation contains CD9-bearing material. It does not, on its own, tell you that material is exosomes versus, for example, plasma-membrane fragments.
Second, no marker is uniformly expressed across all EV subpopulations. Different EV-producing cells release vesicles with different marker profiles. An exosome population enriched for CD63 might be relatively depleted of CD9. A preparation negative for one tetraspanin is not necessarily not-EVs — it might be EVs of a subtype that does not strongly express that marker. This is why MISEV2023 expects a panel.
Third, contamination is invisible to single-marker positive testing. A preparation can be CD9-positive, TSG101-positive, look great on a Western, and still co-purify a substantial fraction of non-vesicular contaminating particles that drive the bulk of the measured biological activity. Without an absence-of-contaminants screen, the positive markers tell only half the story.
What plant-derived EVs require differently
The canonical tetraspanin and ESCRT markers were defined in mammalian systems. Plant-derived EVs (PDEVs) have a partially overlapping, partially distinct molecular toolkit, and the field is still working out which markers serve the same diagnostic role for plant-source preparations.
MISEV2023 explicitly acknowledges this. For PDEV preparations, the guideline expects characterization using markers appropriate to the source — typically a combination of conserved protein homologues, plant-specific membrane components, and a strict contamination panel that is especially attentive to plant cell-wall fragments and protein-pigment complexes that travel with the vesicles through some isolation workflows. Calling a plant preparation an "exosome" on the basis of mammalian tetraspanin homology alone is exactly the kind of shortcut MISEV2023 was written against.
For Aloe barbadensis-derived EVs, the working characterization panel in our process spans plant-EV-relevant protein markers identified in published proteomic surveys of PDEVs, an absence-of-contaminants screen targeting plant-cell-wall components and chloroplastic remnants, and TEM confirmation of bilayer-membrane morphology consistent with a vesicular preparation. The combination, not any one of these elements, is what supports the EV-enrichment claim.
What a defensible panel looks like on a CoA
Under MISEV2023, a marker section on a Certificate of Analysis should include, at minimum:
- At least one Category 1 marker (transmembrane EV-associated protein) reported as positive with a clear band on the appropriate detection method.
- At least one Category 2 marker (cytosolic EV-associated, e.g. TSG101 or Alix for mammalian preparations) — relevance depends on the biogenesis claim being made.
- An absence-of-contaminants screen targeting the contaminants most likely to co-purify given the source biomass and isolation method.
- Source-appropriate substitutions for PDEV preparations, using markers validated for plant EVs rather than mammalian tetraspanins alone.
- Method-of-detection disclosure — Western blot, ELISA, flow cytometry, mass spectrometry — with sufficient information that a third party can interpret the result.
A CoA that reports "CD9 positive" with no contamination screen, no second marker, and no detection method is not characterizing the preparation. It is asserting a single fact that, on its own, does not support the conclusions a marketing page draws from it.
What clinicians and partners can ask
Three questions on the marker panel, paralleling the ones on isolation and counting:
- Which markers are reported on the batch-specific CoA, and which MISEV2023 category does each one address?
- Is an absence-of-contaminants screen included on every batch, and what does it target?
- For plant-derived preparations, are the markers chosen appropriate to plant biology, or were mammalian markers used as a stand-in?
The answers, taken together with the isolation method and the particle count, are what a defensible EV product specification consists of. Anything less is a partial picture, and MISEV2023 is explicit about not wanting to settle for that.