Oral delivery of poorly water-soluble drugs (PWSDs), which predominate the development pipeline, poses significant challenges. Weakly basic compounds, such as atazanavir, represent a critical class of PWSDs as even the administration of the crystalline solid may invoke supersaturation during gastric-intestinal transfer. The absorption advantage afforded by this supersaturated state can be offset by inherent metastability and a tendency to revert to the lower energy crystalline state. Therefore, it is important to understand the physiological factors that can affect crystallization to improve in vitro-in vivo predictiveness and to regulate inter-individual responses. The first aim of this study was to elucidate the influence of lyso-phosphatidylcholine (lyso-PC) and sodium oleate on crystallization, as the products of phosphatidylcholine (PC) hydrolysis and the major lipid components of human intestinal fluid (HIF) and updated fasted state simulated intestinal fluid version 3 (FaSSIF-V3). Secondly, as an individual's bile acid pool is unique, dynamic and related to gut microbiome community structure, it was of interest to investigate the impact of bile acid pool variations on crystallization from supersaturated solutions. To study the effect of PC hydrolysis, media with 2.8 mM sodium glycocholate (GCA) and sodium taurocholate (TCA) (1:1) but varying concentrations of PC, lyso-PC or sodium oleate were prepared. To investigate the influence of inter-individual variations in intestinal bile acid pool size and composition, media simulating the profiles of six healthy Western volunteers were prepared based on previously published data. The crystalline and amorphous solubility of atazanavir, a weakly basic antiretroviral drug, was firstly determined in these media. Nucleation-induction time experiments were then performed at an equivalent extent of supersaturation in each medium (corresponding to the amorphous solubility). At a constant level of GCA/TCA, increasing concentrations of both PC and lyso-PC accelerated crystallization onset; however, this was at least 2-fold more pronounced with lyso-PC at a given molar concentration. The addition of sodium oleate was also observed to induce crystallization. Interestingly, substituting GCA/TCA with the bile salt fraction of other biorelevant media partially circumvented the crystallization-inducing effect of phospholipids and their digests. The presence of dihydroxy bile salts was found to be particularly significant in decelerating the crystallization process. Nucleation-induction times in simulated volunteer pools were found to be dependent upon bile salt concentration, with higher bile salt levels generally prolonging supersaturation. Differences of up to 6-fold were observed. This study demonstrates that the choice of biorelevant medium used to evaluate supersaturating formulations can influence the observed crystallization kinetics. While the presence of lyso-PC and sodium oleate in FaSSIF-V3 medium is more physiologically relevant, further attention should be paid to the bile salt fraction when designing a biorelevant medium for supersaturating formulations. In vivo, inter-individual differences in the amount and types of bile acids and phospholipids present may influence the behaviour of supersaturating formulations.
Keywords: Bile acid; Biorelevant media; Crystallization; Microbial bile acid metabolism; Phospholipid; Simulated intestinal fluid; Supersaturation.
Copyright © 2020 Elsevier B.V. All rights reserved.