Isolation methods compared
Ultracentrifugation, size exclusion chromatography, tangential flow filtration and precipitation, and what each keeps and discards.
The four approaches in common use separate by different physical properties, and each produces a different mixture from the same starting fluid. Differential and density ultracentrifugation separate by sedimentation behaviour, size exclusion chromatography by hydrodynamic size, tangential flow filtration by size across a membrane, and polymer precipitation by altered solubility.
Published method comparisons consistently report that yield and purity trade against each other, and that the method has a larger influence on the composition of the final preparation than most other manufacturing choices. If a supplier will not name the method, they have withheld the single most informative fact about the product.
Four families of method
Ultracentrifugation
The historical reference method. Successive spins at increasing force sediment progressively smaller material, and a final high-speed spin pellets small vesicles. A density gradient variant adds separation by buoyant density, which improves purity because non-vesicular material of similar size often has different density.
Strengths: well established, needs no proprietary reagent, and the gradient variant is one of the better purity approaches. Weaknesses: slow, hard to scale, and the high forces involved are associated in the literature with vesicle aggregation and damage.
Size exclusion chromatography
The sample passes through a column packed with porous beads. Small molecules enter the pores and are retarded; large particles pass around them and elute earlier. Vesicles come off ahead of most free protein.
Strengths: gentle, good at separating vesicles from free protein, reproducible. Weaknesses: dilutes the sample, so a concentration step is usually needed afterwards, and it does not separate particles of similar size from each other, including lipoproteins.
Tangential flow filtration
Fluid flows across a membrane rather than through it, so material that would otherwise clog the surface is swept along. Small molecules pass through, larger material is retained and concentrated.
Strengths: scalable, which matters commercially, and gentler than dead-end filtration. Weaknesses: separates by size only, so anything of similar size is retained together.
Polymer precipitation
A water-excluding polymer is added, reducing the solubility of particles so they can be collected by a low-speed spin.
Strengths: fast, cheap, high yield, requires no specialised equipment. Weaknesses: it precipitates much else besides vesicles, including protein aggregates, and the polymer itself remains in the preparation unless removed. It is the method most associated with low purity in published comparisons.
Isolation method choice has a minor effect on the final preparation compared with cell source.
- Proposed mechanism
- All isolation methods recover the same vesicle population, differing mainly in yield.
- What has been shown
- Method comparison studies applying different isolation techniques to the same starting fluid consistently report differences in particle yield, in purity as measured by particle to protein ratio, and in the profile of co-isolated material. This is one of the better replicated practical findings in the field.
- Highest level reached
- In vitro only
- Main confounders
- Comparisons differ in starting material and downstream analysis, so exact magnitudes vary between studies even where the direction is consistent.
GradeSUPPORTED, CONSISTENT
What would change thisLittle. This is among the more settled points in the field. A supplier arguing that method does not matter is arguing against a well replicated observation and should be asked for their evidence.
The yield and purity trade
The central practical tension is simple. Methods that recover the most material also recover the most unwanted material. Methods that produce the cleanest preparation lose more of the intended product and cost more per unit.
This has a commercial consequence worth stating plainly: purity is expensive, and in a market where the purchaser cannot measure purity, there is no price signal rewarding it. That is a structural problem, not an accusation against any supplier. It is also why published characterisation matters so much, because it is the only mechanism by which a purer product can distinguish itself.
| Method | Separates by | Typical yield | Typical purity | Scales |
|---|---|---|---|---|
| Differential ultracentrifugation | Sedimentation behaviour | Moderate | Moderate | Poorly |
| Density gradient ultracentrifugation | Buoyant density | Low | High | Poorly |
| Size exclusion chromatography | Hydrodynamic size | Moderate | Good against free protein | Moderately |
| Tangential flow filtration | Size across a membrane | High | Moderate | Well |
| Polymer precipitation | Altered solubility | High | Low | Well |
The table describes general tendencies reported across method comparison studies. It is not a specification for any product, and a well executed implementation of a weaker method can outperform a poorly executed implementation of a stronger one. The value of naming the method is that it tells you which problems the manufacturer has to have solved.
Combinations, and why they are a good sign
Methods that separate by different properties can be combined, for example filtration to concentrate followed by chromatography to remove free protein, or chromatography followed by a density step. Combining orthogonal methods is the standard way to improve purity, and a supplier describing a combination is describing a process designed for purity rather than for yield alone.
Conversely, a single-step precipitation process described in marketing terms rather than by method name is the pattern most associated with a concentrated conditioned medium sold on a vesicle label, as discussed in conditioned media and the secretome.
What the method cannot do
No method in routine use separates vesicles by biogenesis, for the reasons in how extracellular vesicles are made. No method removes everything unwanted. And no method compensates for a poor starting fluid: if the culture was stressed, or the medium contributed particles, isolation concentrates that too.
This is why isolation and culture cannot be evaluated separately, and why the specification questions in what a real specification sheet shows cover both. A clean method applied to a poor starting fluid gives you a cleaner version of the wrong thing.
The reader's position
You are unlikely to be able to verify any of this independently. What you can do is notice what is disclosed. A supplier who names the method, names the order of steps, and publishes a purity measure has given you something checkable. A supplier who describes a proprietary technology has given you a noun. The difference between those two documents is the most reliable quality signal available to a purchaser in this market, and it costs nothing to look for.
Questions readers ask
Which isolation method is best?
There is no single best method. Density gradient approaches tend to give the highest purity and the lowest yield; filtration and precipitation give higher yield and lower purity. Combining methods that separate by different properties is the standard route to better purity.
Why does the method matter more than the brand?
Because method comparison studies consistently show that applying different methods to the same starting fluid produces preparations differing in yield, purity and co-isolated material. The method determines what is in the vial in a way the brand does not.
What is wrong with precipitation methods?
Nothing, if the product is described accurately. Precipitation is fast, cheap and high yielding, and it also brings down protein aggregates and other material, and leaves polymer in the preparation unless removed. It is associated with the lowest purity in published comparisons.
Can any method isolate exosomes specifically?
No method in routine use separates vesicles by how they were formed, because all of them separate by physical properties and biogenesis is not a physical property. This is a limit of the technology, not of any particular supplier.
What should a supplier disclose about isolation?
The methods used, in order, and what was measured after the process, including at least one purity indicator. A trade name for a proprietary process names the product rather than describing the separation.