Drug Aggregation of Sparingly-Soluble Ionizable Drugs: Molecular Dynamics Simulations of Papaverine and Prostaglandin F2α.

Abstract

Aggregation in aqueous solution can have important implications on both the in vivo exposure of a drug and its pharmaceutical manufacturability. However, the drug aggregates formed can be very small and, thus, difficult to interrogate experimentally. On the other hand, at higher supersaturations where larger aggregates are supported, the chemical system is inherently metastable and therefore likewise challenging to study from an experimental standpoint. Understanding aggregation behavior is further complicated in the case of ionizable drugs where, unlike neutral compounds, there can be uncertainty in the kinds of drug molecules (i.e., charged, neutral, or both) that become incorporated into various clusters, particularly at pH values near the pKa. In this paper, we apply physics-based all-atom molecular dynamics (MD) simulations to study aggregation in the weakly basic drug papaverine and in the weakly acidic drug prostaglandin F2α. We employ in silico tools to construct simulation workflows and comprehensive cluster analysis protocols to elucidate the size distributions and dynamics of the drug aggregates formed at both an experimentally relevant concentration and at high supersaturation. We build on a previously published treatment [Solubility of sparingly soluble ionizable drugs. Adv. Drug Deliv. Rev. 2007, 59, 568-590, DOI: 10.1016/j.addr.2007.05.008] to translate the predicted aggregate distributions of each ionized drug into corresponding pH-solubility curves that can be compared directly to experiment. Our findings show that the assumption of a single predominant (charged) aggregate can be misleading in interpreting experimental pH-solubility curves, as it does not adequately reflect the rich diversity revealed in our simulations. Beyond not accounting for the distribution of ionized drug-containing clusters actually observed in solution, for both drugs we find evidence that neutral drug molecules can also participate in the aggregation phenomena. Notably, we observe that many drug molecules remain as free monomers in solution even under simulated conditions designed to mimic those where there is significant deviation of the experimental pH-solubility curve from the Henderson-Hasselbalch (H-H) equation, often taken to be a clear signpost of drug aggregation.