For injectable biopharmaceuticals, the subcutaneous route of administration is increasingly preferred over intravenous administration. However, one of the challenges in the development of subcutaneously administered biopharmaceuticals is a reduced bioavailability, which is difficult to predict. Since animal models do not reliably reflect bioavailability in patients, in vitro models could help to develop drug candidates. The purpose of this study was to evaluate a versatile set of in vitro tools for their suitability to predict bioavailability of biopharmaceuticals after subcutaneous administration.
We examined seven commercially available biopharmaceuticals using three instruments, i.e., the Subcutaneous Injection Site Simulator (Scissor), the Osmomat 050, and a dialysis system using three artificial extracellular matrices, two dissolution apparatuses, i.e., the USP4 and the USP7, and two evaluation tools, i.e., the affinity-capture self-interaction nanoparticle spectroscopy (AC-SINS) and the Developability Index (DI). Results were evaluated for their usefulness to predict the bioavailability and other pharmacokinetic parameters in humans using the Pearson correlation. Results: None of the tested instruments and methods could reliably approximate bioavailability. Only pressure values derived with the Osmomat 050 instrument correlated with Cmax with a Pearson correlation coefficient greater than 0.8.
No single in vitro method confidently predicted the bioavailability in humans. We only found a correlation to maximum plasma concentration values for one of the tested approaches. However, a more focused evaluation would be necessary to confirm our findings and test combinations of orthogonal methods that may improve the confidence of such a prediction.
Dhananjay Jere and Lara Nonis