Molecular Dynamics Simulations on Interindividual Variability of Intestinal Fluids: Impact on Drug Solubilization

Abstract

Efficient delivery of oral drugs is dependent on their solubility in human intestinal fluid, a complex and dynamic fluid that contains colloidal structures composed of small molecules. These structures solubilize poorly water-soluble compounds, increasing their apparent solubility, and possibly their bioavailability. In this study, we conducted coarse-grained molecular dynamics simulations with data from duodenal fluid samples previously acquired from five healthy volunteers. In these simulations, we observed the self-assembly of mixed micelles of bile salts, phospholipids, and free fatty acids. The micelles were ellipsoids with a size range of 4-7 nm. Next, we investigated micelle affinities of three model drugs. The affinities in our simulation showed the same trend as literature values for the solubility enhancement of drugs in human intestinal fluids. This type of simulations is useful for studies of events and interactions taking place in the small intestinal fluid.

Keywords: MD simulations; coarse-grained simulations; drug solubility; intestinal fluids; micelles.

Figure 1

Components used in the MD simulations of aspirated human intestinal fluid (HIF): two-dimensional (2D) structure and the corresponding CG representation. Bile salt is colored gray; cholesterol, pink; free fatty acids, green; phospholipid head groups, red and yellow; and their tails, green.

Figure 2

Molecular structure and CG representation of (a) prednisolone (published by Estrada-López et al.), (b) fenofibrate, and (c) probucol. To inhibit strong self-aggregation, the Lennard-Jones potentials were reduced between beads of the same drug.

Figure 3

(a) Snapshots from the final frame of MD simulations of human intestinal fluid representing concentration from five healthy volunteers. Colloidal structures in terms of micelles can be seen. (b) The number of micelles (N) changes during the simulations. At each time point, (N) has been normalized to the number of micelles in the last frame for each system. (c) Shape factors for each micelle show their tendency to form ellipsoids rather than spheres. (d) Micelle size, described as the maximal diameter.

Figure 4

(a) Largest micelles for each human intestinal fluid simulation. Molecular compositions are displayed as pie charts. (b) A single micelle displaying how the phospholipids (green tails, red and yellow head groups) and free fatty acids (green) form a denser core than the bile salts (gray). It is also clear that cholesterol, if present, resides in the core rather than the shell.

Figure 5

Micelle affinities from simulations calculated as ratios of contacts between drug and micelle and between drug and water. (a) Micelle affinity from the largest micelle in each simulation, extracted to, and simulated in, 20 nm boxes. (b) Micelle affinities adjusted for the volume fraction of micelle and water in the larger, previously simulated, human intestinal fluid systems for five HVs (Methodsection, eq 3). Values for each HV were sampled from six simulations with drug concentrations ranging from 0.8 to 4 mM, displayed here as average values with standard deviations (bar and whiskers).