Profiling biopharma medicines

July 1, 2007
Susan Aldridge
Pharmaceutical Technology Europe
Volume 19, Issue 7

... the biotech industry could make characterization of its products easier by paying more attention to downstream processing and purification issues, creating a cleaner product that is easier to identify.

A data package of nuclear magnetic resonance, infrared, ultraviolet and mass spectrometry (MS), plus elemental analysis, is the minimum you need to characterize a small molecule. It should be enough for publication if you are an academic, but if your compound is going into patients, you will need much more than this to satisfy the regulatory authorities. If the compound is a biopharmaceutical drug, then characterization becomes even more difficult because of the size of the molecules involved — be they recombinant proteins, monoclonal antibodies or DNA vaccines. Characterization is now one of the challenges faced by companies embarking on the manufacture of biosimilar drugs, as they must prove that their product has at least similar physical and chemical properties to the original.

Unlike most small molecules, proteins can change their structure and composition depending on conditions, so characterization must take this into account. A change in pH or temperature may make a protein drug denature, or form a complex, or even a high molecular weight aggregate with other protein molecules.

Changes in manufacturing conditions can lead to new patterns of glycosylation — the addition of sugar molecules to certain amino acid residues — which can alter the molecular weight and biological properties of the protein. A range of analytical technologies is required to capture these complexities, and there is still plenty of room for innovation.

Proteins typically have molecular masses of several thousand Daltons, which used to be way beyond the scope of MS. However, this is a technology that has increased its power, while decreasing in price. The mass spectrometer can now accurately determine protein masses to within 1 Da. Given that most proteins have a molecular mass of several thousand Daltons, this is an impressive development, and MS is now the 'gold standard' in protein identification.

But the mass of a protein cannot determine whether it has the correct three-dimensional (3D) structure, the one that will have the required biological activity. In biotech production, the host cell — be it microbial or mammalian — can generally be relied upon to produce the right amino acid sequence of the desired protein, but it cannot always follow through with folding it into the right shape. This is a particular problem with microbial production; mammalian cell hosts have more 'machinery' with which to perform post-translational modifications — that is, folding and glycosylation. Techniques such as light scattering, ultraviolet and visible spectroscopy, and circular dichroism can provide useful information on whether or not the protein has been folded correctly during its manufacture. Instruments, such as the Malvern Zetasizer (Malvern Instruments Ltd, UK), use light scattering to determine not just the size of a protein, but also its oligomeric state — monomer, dimer or more complex aggregate — in solution.

Determining 3D structure, where each atom of a molecule is located in space, goes beyond knowing a molecular formula and is the only way of gaining insight into a small molecule or a biologic drug's pharmacological activity. X-ray crystallography (XRC) is another technology that has undergone amazing developments, with robotic handling allowing for routine, high throughput applications. However, the 3D structure of a protein in solution, where it is physiologically active as a drug, could differ from its structure in the solid state. And, although there are thousands of known protein structures from XRC in databases, some important proteins, such as the membrane proteins, cannot readily be crystallized. Efforts to understand and predict their structures depend upon mathematical methods and computer modelling, and these are still far more demanding than predicting small molecule structures.

Changes in glycosylation, host cell protein content or folding from batch to batch are a particular problem in biotech manufacturing. But conventional manufacturing is not immune to dramatic changes in product with slight changes in manufacturing conditions. A drug may alter its polymorphic form; that is, the way the molecules are packed together in the crystal. The tale is probably well known enough by now, but when Abbott's HIV drug ritonavir began to precipitate out of its semisolid capsule formulation, the company's scientists learned that the drug molecule had changed its polymorphic form. The new form of ritonavir, induced by some slight change in manufacturing conditions, had different and less favourable properties compared with the form that had originally been marketed. Abbott was forced to find a way back to the original form, which it did. The result is that both companies and the regulatory authorities pay far more attention to the 'supramolecular' structure of small molecule drugs, which requires extra solid state analytical work with XRC and other techniques. In other words, looking at large-scale structures is as important for small molecules as it is for biologic drugs.

One urgent requirement for characterizing biologic drugs is the development of high quality reference materials to which a product can be compared, and there are many efforts being devoted to this goal. For instance, the UK's National Institute for Biological Standards and Controls states that its mission is to safeguard public health by helping ensure the quality of biological medicines. At the heart of this work is the preparation, storage and distribution of World Health Organization International Standards and Reference Materials, to provide benchmarks for product quality. This is a vital contribution, but as biotech medicines become more complex, the analytical scientists will have to work hard to keep up with the demand for references.

Meanwhile, the biotech industry could make characterization of its products easier by paying more attention to downstream processing and purification issues, creating a cleaner product that is easier to identify. Advances in affinity chromatography, as described by Professor Chris Lowe of the Institute of Biotechnology at Cambridge University at the recent BioProcess International conference, look promising in this respect. This innovative approach involves the application of molecular modelling to design and optimize ligands that are carefully matched to a particular biopharmaceutical product. When they are attached to a column, they can capture the product with a very high specificity so that when it is eluted it should be much easier to characterize.

If the characterization issue is hard for small molecules and proteins, then it is even more difficult for more advanced therapies, for example, stem cells. Companies, such as Dynal/Invitrogen, are developing technologies for flagging up marker proteins on the surfaces of stem cells, and sorting them into purified populations. This is a world away from passing a small molecule drug down a column, but as medicines become increasingly complex, advanced analytical technologies are needed to characterize them.