<i>In Silico</i>Assessment of Biomolecule Reactivity with Leachables

Leachables in pharmaceutical products may react with biomolecule active pharmaceutical ingredients (APIs), for example, monoclonal antibodies (mAb), peptides, and ribonucleic acids (RNA), potentially compromising product safety and efficacy or impacting quality attributes. This investigation explored a series of <i>in silico</i> models to screen extractables and leachables to assess their possible reactivity with biomolecules. These <i>in silico</i> models were applied to collections of known leachables to identify functional and structural chemical classes likely to be flagged by these <i>in silico</i> approaches. Flagged leachable functional classes included antimicrobials, colorants, and film-forming agents, whereas specific chemical classes included epoxides, acrylates, and quinones. In addition, a dataset of 22 leachables with experimental data indicating their interaction with insulin glargine was used to evaluate whether one or more <i>in silico</i> methods are fit-for-purpose as a preliminary screen for assessing this biomolecule reactivity. Analysis of the data showed that the sensitivity of an <i>in silico</i> screen using multiple methodologies was 80%-90% and the specificity was 58%-92%. A workflow supporting the use of <i>in silico</i> methods in this field is proposed based on both the results from this assessment and best practices in the field of computational modeling and quality risk management.