Which chromatography media are suitable for AAV polishing?

Because an AAV capsid is roughly 25 nm across, it is largely excluded from the pore network of a conventional porous resin, whose pores sit in the tens of nanometres. There are exceptions in routine use — resins built with wide through-pores, and core-bead media designed for flow-through polishing — but for the rest, convective media are the usual choice.

Why particle size decides the format

The argument is geometric, and it is the reason a polishing step for viral vectors looks different from a protein process.

• A monoclonal antibody is about 10 nm across. It diffuses into the pores of a chromatography bead, which is where most of the surface area sits.
• An AAV capsid is about 25 nm, a lentiviral vector about 100 nm. Both are largely excluded from that internal pore volume. What remains is the outer surface of the bead — a small fraction of the available area.

What follows is geometric rather than numerical. For a particle that cannot enter the pores, only the outer surface of the bead is in reach, and reaching even that still costs residence time. What that means for your step is a number to measure on your own feed rather than to read from a comparison — the datasheet states the conditions we measure under, and how to read a dynamic binding capacity figure sets out which of them have to match before a number transfers to your feed.

Convective media — monoliths and membrane adsorbers — carry the vector past the ligand in the flowing liquid instead. The ligand sits on the channel wall, in reach. That is why literature and suppliers converge on this class for viral vectors.

The polishing step in practice

• Host cell protein, host cell DNA and process impurities are usually cleared by anion exchange (Q). Depending on the mode chosen, either the vector binds and impurities separate by charge, or the reverse.
• Full versus empty capsid enrichment runs on anion exchange with a shallow salt or pH gradient. The separation is small, depends strongly on serotype, and requires method development.
• An orthogonal polishing step can be added with cation exchange (S) where the anion exchange step alone does not clear a specific impurity. Size exclusion as a group separation is the other route: nothing binds, so there is no capacity limit and no gradient to develop, at the cost of throughput and dilution — which makes it the safe option when the vector does not tolerate a binding step. Anion exchange membrane adsorbers in flow-through are the third option at this step and are widely used.
• Hydrophobic interaction is used less often than anion exchange at this step, but published methods exist both for aggregate removal and for serotype-dependent full-and-empty enrichment.

Mixed-mode media are the fourth option, and they are in real use at this step: a ligand that combines charge with a hydrophobic or hydrogen-bonding element can clear impurities that a single-mode anion exchanger leaves behind, and some groups use them for full-and-empty enrichment where a salt gradient alone does not resolve. We do not offer a mixed-mode chemistry, so the honest statement is that it belongs on your shortlist and not on ours.

Serotype matters. AAV8 behaves differently from AAV2 or AAV9 in surface charge and in how tightly it holds onto an anion exchanger. A gradient that separates full from empty capsids for one serotype rarely transfers unchanged to the next.

Being honest about the hard part: full-and-empty separation is where published methods vary most. We have no data of our own on it — so we do not claim any. Density-gradient ultracentrifugation in caesium chloride or iodixanol remains the reference both for the separation itself and for the analytics you measure a gradient against; it does not scale, which is why the chromatographic route exists at all, but it is the yardstick. If that separation is the core of your process, ask us what we can and cannot say before you rely on a general claim, ours or anyone else's.

MonoCore™ for AAV polishing

• MonoCore™ Q — strong and weak anion exchange; polishing, impurity removal, full/empty enrichment.
• MonoCore™ S — strong and weak cation exchange; orthogonal polishing.
• Both on a modified cellulose monolith, 15 µm standard channel size, 1.8 mL and 5.4 mL, larger on request, UNF 10-32 connectors.

Three properties matter in this step.

• A hydrophilic cellulose backbone. Low non-specific adsorption keeps recovery up when the product is expensive and the load is small.
• An open channel structure. Polishing feeds are usually clean, so blocking is rarely the limiting factor at this step — the open channel matters most on a less-clarified intermediate or in a long campaign. It is not a capacity argument either — channel size is chosen against the particle that has to move through it, not maximised.
• Predictable scale-up. Pressure and flow profile stay consistent from the lab format to production, which means less method re-optimisation when volumes grow — less, not none.

Against membrane adsorbers, a monolith gives a more homogeneous flow path and a lower void volume, so the elution peak stays narrow and the product is less diluted. Against packed-bed resins, the gain is speed and robustness, not resolution: a monolith is resin-class in resolution, and we do not claim more.

Frequently asked questions

Is a monolith better than a membrane adsorber for AAV?

It is different. Membrane adsorbers are fast, disposable and cheap per run. A monolith has one continuous flow path rather than a stack of pleated layers, so flow distribution is more uniform and void volume smaller — which shows up as a narrower elution peak and less product dilution. Where a fully disposable flow-through step is the goal, a membrane is often the simpler answer.

Which media size handles a 5 L harvest?

After capture, the polishing load is small compared with the harvest volume. A 1.8 mL or 5.4 mL capsule usually covers this scale; the deciding numbers are the amount of vector to be bound and the cycle time you want, not the harvest volume.

Can anion exchange separate full from empty capsids?

Partially, and it is the most common approach. The two populations differ in surface charge because of the packaged genome. The separation is narrow, serotype dependent and requires careful gradient development.

Do I need to clarify the feed before polishing?

Polishing follows capture, so the feed is usually already clarified. The open channel structure tolerates residual turbidity better than fine-pored media, but it does not replace a clarification step.

How many cycles does a capsule last?

For ion exchange media the limit is set by cleaning, not by the number of bind-and-elute cycles. See the datasheet for validated cleaning conditions.