April 15, 2008 (Vol. 28, No. 8)
Anthony R. Newcombe
Guidelines for Dealing with Complications that Arise with Workhorse Purification Procedure
Cation- and anion-exchange chromatography columns are used routinely for the downstream processing of therapeutic antibodies and proteins. Despite their frequent use, they are considered less glamorous than the high-profile affinity adsorbents such as protein A sepharose that enjoy celebrity status in the world of industrial antibody purification.
While the protein A adsorbents may be compared to formula one racing cars—continually nurtured, tweaked, redesigned, or redeveloped to improve performance with due care and attention—ion exchangers are often considered the workhorses of the purification process—well-characterized adsorbents with well-defined, charged chemical ligands that have been used for decades and typically perform in a predictable manner.
Ion exchangers are often used as a second chromatography step, following primary capture of an antibody or therapeutic protein to remove potential impurities such as DNA, host cell proteins, endotoxins, and viruses. Therefore, the performance and robustness of the process step may be critical.
As the name suggests, ion exchangers were initially developed for the purification of ions and small molecules including ammonium, radioactive isotopes, and amines. As such, the results were often consistent and predictable. The seemingly simple application of ion exchangers for what perhaps should be described as protein-exchange or ion:protein-exchange, however, often presents a multitude of unexpected technical challenges that have no doubt been the focus of many purification scientists’ woes.
The apparent simplicity of the ion-exchange (IEX) process for the purification of proteins may hoodwink even the most experienced of scientists. The pI, or isoelectric point (the pH at which the molecule has a neutral net charge), may be accurately predicted or determined experimentally. Therefore, a theoretical ion-exchange step for the purification of a protein of interest may be developed with the minimum of effort.
The use of buffers one pH unit above or below the pI will ensure that the protein of interest binds or flows through the column. If the buffer pH is maintained, then the purification step will perform consistently and reproducibly.
The timeline for purification development is often expected to be minimal, and subsequently, researchers head to the laboratory convinced that a proposed experimental design cannot fail. To the despair and frustration of development scientists, the complex exchange of ions, amphoteric buffer molecules, impurities, and the solubility and stability of the protein of interest may result in experimental data that is often far from predictable.
Many proteins are soluble, stable, and will of course interact predictably with ion-exchange adsorbents. Other proteins, even closely related molecules such as mAbs, however, may demonstrate specific regions of charge, hydrophobicity, or show minor charge heterogeneity. Therefore, an apparently simple IEX-platform purification step becomes complicated.
At least three critical features may need to be considered when designing an ion-exchange step for a therapeutic protein: the stability and solubility of the target protein (and impurities) under the proposed conditions for the purification step; the buffer concentration, composition, and potential interaction of the buffer molecules with the adsorbent; and the physical properties of the base matrix, in particular the effect on nonspecific binding of sample components during the loading, washing, and elution steps. In addition, temperature may also affect adsorption to weak ion exchangers and may need to be evaluated as part of the process characterization.
The surface chemistry of a desired protein may have a dramatic effect on the performance and predictability of an ion-exchange step. Although the pI of a protein defines the pH at which the molecule has no net charge, the protein will undoubtedly have regions of charge on the surface of the protein dependent on the primary sequence, the pKa of the individual amino acid side chains in question, and the complexity of the molecule.
Although some impurities such as DNA and endotoxin demonstrate a strong and predictable charge, target proteins may bind (or not bind) unexpectedly to an IEX adsorbent. In addition, the stability and solubility of the protein of interest may need to be considered.
The protein of interest may show a tendency to aggregate or precipitate during the purification process. This may be dependent on a number of factors including the protein concentration, buffer composition, and the stability of the protein following subsequent upstream processing steps such as pH neutralization following elution from an affinity column.
Elution of bound proteins from an ion-exchange column is typically undertaken using increasing conductivity or a pH shift; changes in either condition may impact the quality of the eluted protein and should be assessed carefully when using ion-exchange adsorbents in a bind-and-elute mode.
While data from characterization, stability, and forced degradation studies are often not available when developing a downstream process for a new biotherapeutic, the biophysical characteristics of the product and the potential pitfalls should be considered when developing IEX-purification processes at laboratory scale.
Buffer Composition & Concentration
Although buffer pH and conductivity are usually assessed and monitored when developing an IEX step, the buffer type and concentration is often initially chosen from a combination of historical or published data and large-scale commercial considerations such as cost and availability. Buffer composition should ideally be selected to ensure that the buffer molecules do not have the opposite charge to the bound ligand on the ion-exchange adsorbent.
Although this is sometimes unavoidable or undesirable for commercial purification processes, the buffer molecules may act as counter ions and participate in an ion-exchange process during column equilibration and even during column loading. This consideration is further complicated as many buffer species by their very nature are amphoteric and exhibit a number of charged species in equilibrium.
It is therefore important to ensure that a sufficient buffer concentration is used and that the IEX column is fully equilibrated to ensure that any buffer-ion exchange is complete prior to loading valuable product onto the column. Although many of these issues may also be applicable to affinity-purification steps, the specific and high-affinity interaction of the desired protein with the immobilized affinity ligand as the primary method of capture typically minimizes buffer composition effects.
Although the detailed characterization of a problematic ion-exchange step is desirable, industrialists are frequently constrained by timelines and commercial pressures early in the development process, and often, limited product-stability data is available. The technical challenge in this case is untangling the process- and product-related characteristics that are influencing the performance of the IEX chromatography column.
Potential issues and pitfalls may be minimized by ensuring that low buffer concentrations are avoided and ion-exchange columns are fully equilibrated, typically with >5 column volumes, to minimize potential pH and conductivity drifts that may impact the performance and reproducibility of the process step.
When developing a bind-and-elute ion-exchange step, evaluating the appropriate conditions to elute the target protein by avoiding transitions through the isoelectric point during elution in the least harsh conditions may also prove beneficial.
The development of an IEX step for the purification of a biopharmaceutical is often considered simple, predictable, and straightforward, with the column performance defined almost exclusively by the isoelectric point of the target protein and the buffer pH. Although this is the case for many proteins, the characteristics of the protein of interest, impurities, buffer composition, and adsorbent properties may need to be considered if unexpected experimental results are observed.
Anthony R. Newcombe, Ph.D., is process science manager at Protherics, Claire Newcombe is CEO at Applied Biopharm Consultancy, and Richard Francis is director of process science, also at Protherics. Web: www.protherics.com. E-mail: firstname.lastname@example.org.