Chromatography for antibody purification
The platform process for monoclonals, and where bispecifics, ADCs and fragments depart from it. Which chromatography step does what, and what to change when the standard route stops working.
Antibody purification is the most standardised process in biologics manufacturing, and that is its strength: a monoclonal antibody entering development today follows a route that has been run thousands of times. Protein A capture, low-pH viral inactivation, one or two polishing steps, viral filtration, and formulation. The template works because the molecules are similar enough for one process to fit most of them.
It also means the interesting questions are at the edges — where a molecule is not a standard IgG1, where the feed is unusually concentrated, or where a step that normally just works suddenly does not. This page covers the route and those edges.
This page is about which chemistry does what and which formats we supply. For the process itself, worked through step by step, see antibody purification: how the standard route works.
The platform process
| Step | Purpose | Typical chemistry |
|---|---|---|
| Clarification | Removes cells and debris from the harvest | Centrifugation and depth filtration |
| Capture | Binds the antibody, removes most host-cell protein and DNA in one step | Protein A affinity |
| Viral inactivation | Inactivates enveloped viruses in the acidic Protein A eluate | Low-pH hold, typically pH 3.5 for 30–60 min |
| Polishing 1 | Removes aggregate, charge variants and leached ligand | Cation exchange, bind-and-elute |
| Polishing 2 | Removes residual DNA, host-cell protein and endotoxin | Anion exchange, usually flow-through |
| Viral filtration | Removes small non-enveloped viruses by size | Parvovirus-retentive filter |
| Formulation | Concentration and buffer exchange | Ultrafiltration / diafiltration |
Three of these steps clear viruses by different mechanisms — the low-pH hold inactivates enveloped viruses, the filter removes small ones by size, and the chromatography steps contribute removal by partitioning. Regulators expect orthogonality, which is why a process rarely drops one of them even when purity would allow it.
Capture: Protein A and its alternatives
Protein A binds the Fc region, which means it works across most IgG molecules regardless of what the variable region does. That is what makes platform processing possible at all. The cost is an acidic elution — typically pH 3.0 to 3.6 — which some molecules tolerate poorly.
Where Protein A does not fit, the alternatives divide by reason. If the format lacks a suitable Fc, Protein G covers a broader subclass range, and Protein L binds kappa light chains independently of the Fc. If the problem is the acid rather than the binding, the answer is a gentler elution: the mechanisms are set out in how to elute from Protein A without acid.
Our own answer to the acid problem is Solaris® Protein A, where elution is triggered by blue light at 480 nm rather than by pH. Beta access by application.
Polishing: what each chemistry removes
| Chemistry | Removes | Typical mode |
|---|---|---|
| Cation exchange (S) | Aggregate, charge variants, leached Protein A, some host-cell protein | Bind and elute |
| Anion exchange (Q) | DNA, endotoxin, acidic host-cell protein | Flow-through |
| Hydrophobic interaction (HIC) | Aggregate, and hydrophobic variants such as unconjugated species | Bind and elute, or flow-through |
| Mixed-mode | Aggregate where a single mechanism does not resolve it | Usually bind and elute |
Aggregate is the impurity that decides most polishing strategies. It is immunogenic, it grows during processing and storage, and it resembles the product closely enough that separating it is genuinely difficult. Cation exchange and HIC attack it by different mechanisms, which is why the pair is common.
Where the platform stops fitting
Bispecifics carry two different binding arms, so the process has to remove not only aggregate but also the homodimers and half-molecules that assembly produces. Those species are close to the product in size and charge, which puts more weight on polishing selectivity than a monoclonal would.
ADCs add a conjugation step in the middle of the process. Purification afterwards has to separate species that differ only in how many drug molecules they carry, and the conjugate is usually more hydrophobic and more aggregation-prone than the unconjugated antibody.
Fragments — Fab, scFv, nanobodies — often have no Fc at all, so Protein A does not apply. Capture moves to Protein L, to an affinity tag, or to a non-affinity route built on charge.
High-titre feeds change the arithmetic rather than the chemistry. When a bioreactor delivers well over 5 g/L, the capture step becomes the bottleneck, and the usual responses are a larger column, more cycles per batch, or a continuous capture arrangement.
Where MonoCore™ fits
We make Protein A, Protein G, cation exchange and HIC capsules on one modified cellulose matrix, in a format that connects to the FPLC system you already run. Transport through the open channels is convective, so a cycle can be run at flow rates a packed bed cannot reach within its pressure limit — which is what makes running many short cycles a practical alternative to one long one. That argument is developed in rapid cycling chromatography.
What we do not claim is a capacity comparison against packed-bed resins, in either direction. Capacity depends on the molecule, the residence time and the breakthrough level you accept, and a number quoted without those conditions is not comparable to anything — the reasoning is in how to read a dynamic binding capacity figure.
- What are the steps in antibody purification?
- Clarification of the harvest, Protein A affinity capture, a low-pH hold to inactivate enveloped viruses, one or two polishing steps — usually cation exchange and either anion exchange or HIC — then viral filtration and finally ultrafiltration/diafiltration into the formulation buffer. The sequence is standardised enough that most monoclonal antibodies enter development on it unchanged.
- Why is Protein A used to capture antibodies?
- Because it binds the Fc region, which is essentially the same across most IgG molecules whatever their variable region does. That makes one capture step work for many different antibodies, which is what allows platform processing. A single Protein A step also removes the large majority of host-cell protein and DNA.
- How are antibodies purified without Protein A?
- By substituting a different ligand or a different mechanism. Protein G covers subclasses Protein A binds weakly; Protein L binds kappa light chains and so works on fragments with no Fc. Where no affinity ligand fits, capture can be built on cation exchange instead, at the cost of a less selective first step and more work downstream.
- What is the low-pH hold for?
- It inactivates enveloped viruses. The Protein A eluate is already acidic, so holding it at around pH 3.5 for 30 to 60 minutes adds a dedicated viral-clearance step almost for free. It is one of three orthogonal clearance mechanisms in the standard process, alongside size-based viral filtration and partitioning on the chromatography steps.
- How do you remove aggregates from an antibody preparation?
- Usually on cation exchange or HIC, and often on both, because they separate aggregate by different mechanisms — charge and surface hydrophobicity. Aggregate is the hardest impurity in antibody polishing: it is immunogenic, it forms during processing, and it resembles the product closely. Mixed-mode chemistry is the common fallback where neither single mechanism resolves it.
- What changes when purifying a bispecific antibody?
- The impurity profile. Besides aggregate, the process has to clear homodimers and half-molecules left over from assembly, and those are much closer to the product than the impurities a monoclonal process handles. Polishing selectivity matters more, and a gentler elution can help where the molecule aggregates under acid.