It’s Complicated

Generic medicines are now an established part of the pharmaceutical supply chain and offer significant savings to health services, insurers and patients alike. More recently, biosimilars have entered the fray, and while the savings are not as great as with small molecule generics, they still help cut the costs of medicine. However, there is also a third category of product that falls between the two – the complex generic. These are products that may include complex: active ingredients, formulation, route of delivery, or even a mixture of ingredients.

The key to creating a new generic or biosimilar medicine and gaining regulatory approval is proving that it is safe and comparable to the originator product. For a small molecule generic, proving “sameness” between the two is relatively straightforward, relying heavily on a sub-set of well-defined analytical methods. Biologics are very different because the exact nature of the product depends on how it is manufactured, leading regulators to demand clinical studies that prove the biosimilar is functionally comparable to the reference product.

Many of the complex generics currently being developed are peptides – albeit less than 40 amino acids long – and proving sameness for the active ingredients can be tricky. The same often applies to other molecule types that can be considered “complex” such as polyamino acids (Copaxone (glatiramer acetate), which is a random combination of four amino acids) carbohydrates (which can also be sulfated as Enoxaparin or pentosan polysulfate), and naturally derived mixtures, such as oestrogens. Unfortunately, it is possible for an experienced generics company to approach complex generic submissions as generic ones, only for the application to be questioned by regulators.

European regulators tend to consider some of these complex generics products to be more like biologics (e.g., Enoxaparin), thus requiring clinical work. But the story is different in the US where regulators are looking for proof that the molecules are the same, similar to a small molecule generic. While draft guidances were recently published for Enoxaparin and glatiramer acetate, they only provide the general areas where sameness needs to be demonstrated – and no details on how to actually demonstrate it. There’s also limited technical direction – certainly not to the same level as a general chapter in the US Pharmacopeia – the guidance simply says that equivalence must be proved. In some respect, this is in agreement with a lot of guidances from regulatory authorities. However, for biological products, other documents such as the ICH Q6B guidelines do offer a list of critical parameters and possible techniques to be applied when characterizing a protein. In the case of complex generics, there is very little documentation to be used and, when it does exist, caution needs to be exercised on how to put the information provided into use. For example, the guidance for Enoxaparin refers to complex documents such as a petition that spans over almost a decade, which discusses what might be required and refers to about 133 publications that readers will want to check. Much of this might be obsolete, having been superseded by more recent and applicable research.

For complex generic peptide APIs, the FDA specifies that physicochemical properties, primary sequence, secondary structure, oligomer structure, and biological activities must all be assessed. While many complex APIs may be comprised of chains of amino acids, they aren’t proteins, so the typical protein toolbox isn’t readily applicable. In fact, some are heterogeneous mixtures that may or may not have specific signatures or modifications, such as glatiramer acetate – for which there are no off-the-shelf tools at all.

For primary sequence or impurity characterization, there is a widespread, and mistaken, belief that mass spectrometry analysis will suffice, but this is rarely the case. If a peptide includes an unnatural amino acid that is an enantiomer of the naturally occurring version, mass spectrometry cannot unequivocally identify this because their mass will be the same. A technique such as chiral chromatography will be required in conjunction with sequence analysis if sameness is to be proved.

For these peptides, determining secondary and higher-order structures is quite complex. The techniques applicable to proteins simply aren’t appropriate for smaller peptides, and the list of techniques suitable for this class of compounds is decidedly limited. How can these be used to create a comprehensive analytical strategy to prove sameness? To complicate matters even more, the FDA is clear that orthogonality in the definition of each quality attribute is recommended.

My take on this would be that the solution must use a lot of experimentally-driven evidences and an appropriate analytical strategy. The costs and timelines associated with this work are significant – and it would be easy for generics companies to embark on developing a complex generic, without fully realizing how much more challenging the process is, compared with a traditional small molecule. Even with a good analytical strategy at hand, there is the challenge of comparing it to the reference listed drug. Some of these peptides are formulated at extremely low concentrations – often less than a milligram per millilitre, and even down to the micrograms level. Vasopressin, for example, is typically formulated at approximately 37µg/ml, and calcitonin at 33µg/ml. Biophysical techniques to determine secondary structures are not applicable at such low concentrations and for such short chains. The formulation of the reference product also poses problems. Not only are they usually of low concentration, they are formulated with the inclusion of bacteriostatic ingredients, which are ultraviolet (UV) absorbents. Most secondary structure analysis techniques are based on UV methods, meaning these cannot be used on the formulated product.

New methods will have to be brought to the FDA that will work. But for the analytical scientist, this isn’t as simple as finding the best method and running with it; it must also be demonstrated that the other methods won’t work.

In my view, the key for all analytical sameness studies is in the preparation, planning and understanding of the technical and scientific challenges each complex generic API presents. Only if these are properly evaluated and defined in advance can any analytical package have a chance of being favorably looked upon. With the right planning, companies will be able to purchase enough reference listed drug (RLD) material for all phases of the study, design the fit-for-purpose studies for each of the quality attributes to be followed, and perform the experimentally-defined selection demonstrated-to-be-fit methods. Only then will this ultimately lead to straightforward analytical comparability studies.