How "Similar" Must Biosimilars Be?

One of the most basic issues surrounding follow-on proteins is determining what types of proteins can be considered biosimilar. One viewpoint states that only proteins that can be fully characterized, with no discernible differences in either the structure or impurity profile, can be considered biosimilar. However, this narrow view has not been adopted by the EU, whose regulations allow a degree of difference – provided it can be justified. For example, one EU-approved biosimilar product is expressed from yeast, while its reference product is expressed from E coli. Although variations in production methods create differences in the follow-on protein’s impurity profile and other characteristics, these differences may be justified based on the provision of bridging clinical and non-clinical data.

If lower-cost therapeutic proteins are to reach the market, more innovative and efficient production technologies – including transgenic production – need to be introduced. Although these technologies may create differences in impurities or post-translational modifications, it should still be scientifically possible to demonstrate that the biosimilar product provides equivalent safety and efficacy to the reference product.

In the current regulatory view, differences in primary structure or profound post-translational modifications mean that a protein must be treated as a novel compound. But such a rigid approach may not be necessary or scientifically justified. As experience with biosimilars grows, regulators may become more comfortable with allowing greater flexibility in follow-on proteins despite some differences from the reference product – provided that equivalent safety and efficacy can be proven. However, there is no regulatory precedent for this approach today.