Ask the Expert: Purification of Antibody Fragments Using Amsphere A3 Protein A Resin

In an “Ask the Expert” webinar on 13 September 2017, Gerald Platteau of JSR Life Sciences described the use of Amsphere A3 resin to purify antibody fragments. He explained the molecular binding mechanism for VHH singledomain antibodies and compared dynamic binding capacity (DBC) data with those of other affinity resins.

Platteau’s Presentation
For full-size monoclonal antibodies (MAbs), the standard capture step is based on protein A. Its binding to the Fc region has been well described as taking place at the CH₂–CH₃ interface primarily through hydrophobic interactions. Because antibodies in the same subclass have >95% homologous Fc regions, protein A affinity is a suitable capture platform for a wide antibody range.

Antibody fragments are becoming an important therapeutic class comprising a variety of molecules such as VHH single domain antibodies, Fabs (antigen-binding fragments), single VH domains, and scFvs (single chain variable fragments). Because they lack Fc regions, these antibody fragments cannot be captured by most engineered protein A ligands. Amsphere A3 ligands, however, show a high affinity to VHH single-domain antibodies. They have a higher binding capacity than that of other commercially available protein A resins, the closest of which has a 40% lower DBC.

To identify the interaction sites between Amsphere A3 ligand and a VHH domain, we obtained crystals of a VHH–PrA complex. By X-ray diffraction analysis, the complex was modeled, showing binding sites (and involved amino acids) between the VHH molecule and Amsphere A3 Protein A ligand.

DBC (in grams of product per liter of resin) values for VHHs are about 2–3× lower for VHHs than for full-size MAbs. However, the antibody fragments are ~10–11 times smaller than MAbs. So when comparing DBC values for different types of molecules, it’s better to take into account their molecular weight. Molar DBC data give a better comparison. We have observed that about 3–4× more monovalent VHHs are bound than full-size MAbs. Affinity plays a role, but spatial limitations also determine how many molecules can be bound.

We compared this resin with other types used for capturing antibody fragments. Ion-exchange resins and mixed-mode resins can have higher capacities. But using a protein A resin can shorten process development times with very high product purity and recoveries >95% after the capture step. An affinity chromatography step can be a true platform for fragment capture. Protein A also has better caustic stability than other affinity ligands produced specifically to bind certain antibody domains. That allows it to be used for many runs, lowering the overall cost per gram of purified product.

Questions and Answers
Can you explain DBC and its importance as a measurement? DBC shows how much of a product you can bind to a resin for a certain residence time. For a given flow rate and residence time, you can measure the capacity of a resin for a given molecule. The value typically is expressed as an amount of target protein that binds to a resin before a breakthrough of unbound protein occurs. Usually you identify DBC at a certain breakthrough point (e.g., 10%). You need to know a resin’s capacity to calculate the amount you need to purify your drug substance. The higher the capacity, the less resin you need, and the less expensive the process will be.

How do you detect Amsphere A3 resin as an impurity? If you are referring to protein A leaching, you can determine that with a validated enzyme-linked immunosorbent assay (ELISA) kit that can be purchased through Cygnus Technologies.

Do you have data showing the difference between your protein A ligand and mixed-mode or ion-exchange chromatography? We have not generated such data internally, but we know that it depends on which antibody fragments you have. We have performed a large screening of 110 VHH molecules, and our resin’s highest DBC was 30–35 g/L. I don’t know what capacities would be reached for these VHH single-domain antibodies with either ion-exchange or mixed-mode resins.

Will it bind to other species, such as murine IgG1 and rat IgG2a? It is very unlikely that it would bind those. For murine IgG2a, binding could occur through the Fc region. Literature and our own data show that the variable-domain binding is mainly restricted to VHH single-domain antibody or VH3 domains of the Fab region of a full-size MAb.

Does that binding work with all of VH subtypes? No, it is restricted to VH3 domains. VHHs are a different story because they come from the heavychain antibody domains and then are further modified to be single-domain antibodies. But when you look at Fabs or your single-chain variable fragments, it would be restricted to members of the VH3 gene family.

More Online
The full presentation of this webcast can be found on the BioProcess International website at the link below. Listen online at www.bioprocessintl. com/ask-the-experts/amsphere-a3-protein-a-ligand-purification-of-antibody-fragments.