BioRationality: New FDA Guideline Invites Biosimilar Developers to Reduce the Cost of Testing

Sarfaraz K. Niazi, PhD, examined the new FDA guideline that acts as an invitation to biosimilar developers to decrease the cost of testing, potentially opening the door for companies to reduce nonclinical testing of biosimilars entirely.

I gasped in May 2023 when I read the FDA guidance “Generally Accepted Scientific Knowledge (GASK) in Applications for Drug and Biological Products: Nonclinical Information” that leaves the doors open to reduce nonclinical testing of biosimilars. The guidance describes the 2 circumstances in which sponsors can rely on GASK in their product development programs to meet relevant approval requirements:

  • When a product contains a substance (either naturally derived or synthesized) that occurs naturally in the body, one can rely on GASK regarding that substance and its known effect on biological processes instead of conducting certain nonclinical studies
  • When a sponsor has demonstrated a drug’s impact on a particular biological pathway, and then relying on GASK to conclude that certain nonclinical studies are not necessary to support approval and labeling of the drug are not required.

For example, in some cases, a drug has either on- or off target impacts on a biological pathway or molecular mechanism of action that is known to result in adverse effects at clinically relevant exposures based on the operation of the biologicalpathway.

Thus,it may be appropriate to rely on GASK regarding the impact of the pathway rather than to conduct specific pharmacology and toxicology studies intended to measure the impact of the pathway. Therapeutic proteins act by receptor binding, which leads to a pharmacodynamic response that brings the pharmacological response that forms the basis of the clinical response.

Unlike chemical drugs, biological drugs' toxicity depends entirely on pharmacology; this established scientific knowledge (GASK) should support claims of the same toxicology profile if the receptor-binding properties are highly similar. This will reduce in situ testing in the analytical assessment of biosimilar candidates. There is no need to include dozens of tests, even in an orthogonal mode, to prove what is already proven. The same conclusion will be drawn in functional testing, leaving the testing to only those process-related attributes such as post-translational modifications, aggregates, impurities, and other readily analyzed attributes.

One limit of the GASK allowance may extend to structure testing since the 3D structure depends entirely on the primary amino acid chain sequence. The 3D structure prediction algorithms AlfphaFold2 and ESMFold will always yield the same 3D if the identical sequence is readily analyzed with the highest confidence.

The FDA guidance is critical since it acknowledges that the developers have been questioning this concession, but the terminology “generally accepted” is very important. How does one establish that a fact is generally established? It is based on “scientific knowledge,” but what constitutes scientific knowledge? Is it a peer-reviewed publication or a general belief in the scientific community? The burden of establishing these 2 elements will lie on the developer.

Still, since the doors are open, I anticipate competent developers taking the most advantage of these newly found concessions to reduce their cost of development and avoid the possibility of redundant testing leading to contradictory findings. This forward-thinking of the FDA teaches us to bring rational approaches to developing biosimilars and not follow the guidelines that had considered them new biologics when biosimilars were introduced.

Biosimilarity should be established based on comparison with the reference product for the product-related attributes and compliance with scientifically justified process-related attributes, an idea that has long been debated; what is a critical quality attribute? Initially, the FDA and all other agencies included just about every process-related attribute as critical attributes, despite realizing that the release specification attributes are process-related and complied with every batch. Essentially, this meant that what the FDA has accepted as good enough for patients may not be good enough for the FDA.

An excellent example to consider is the protein content that is generally allowed a 3% variability, but if this is also declared as a critical attributed and compared with the reference product, the test may fail even if the variation is less than 3%. This was the subject of the first FDA guidance, that was withdrawn to overcome this problem. The GASK also opens the doors for developers to challenge the FDA on what is a critical quality attribute. The GASK application will remove all testing that is part of the release specification from the analytical assessment. However, this may still require testing reference products to establish release specifications.

Overall, it is a tremendous scientific step taken by the FDA. It is now up to the developers to use these concessions best.

Reference

Jumper J, Evans R, Pritzel A, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596(7873): 583-589. doi: 10.1038/s41586-021-03819-2