Modeling of Cellulosic Polymers for Drug Delivery in Solid Dispersions

Solid dispersion formulation is a promising method to maintain in vivo drug solubility and to improve drug efficacy. However, the exact drug stabilization and release mechanisms of the solid dispersion formulation are unclear. In this doctoral work, we present a multi-scale modeling approach to study the solvation behavior of cellulosic polymers and their interactions with the model drug phenytoin. We compare a number of atomistic force fields and find they give similar predictions for the stiffness of the cellulose chains. We then develop systematic coarse-grained (CG) force fields for two cellulosic polymers, namely methylcellulose and hydroxylpropyl methylcellulose acetate succinate (HPMCAS), based on the radial distribution functions obtained from atomistic simulations. We use the methylcellulose CG model to simulate the self-assembly of multiple 1000 monomers long polymer chains, and find that they spontaneously form ring or tubular structures with outer diameter of 14nm and void fraction of 26%. These structures appear to be precursors to the methylcellulose fibrils, whose diameter and structure are in good agreement with both theoretical and experimental results, and thus shine light on the methylcellulose gelation mechanism. We also present a simplified continuum analytical model to predict a phase map of the collapse conformations of a single self-attractive semiflexible polymer chain in solution into either folded or ring structures depending on the chains bending energy and self-interaction energy. The predicted phase map is in good qualitative agreement with simulation results for these collapsed structures. We use the HPACAS CG model to study the intermolecular interaction modes between 9 functional groups on HPMCAS and model drug phenytoin. We adopt two criteria to quantify the effectiveness of the polymeric excipients, namely 1) the ability to inhibit drug aggregation and 2) the ability to slow down drug release. We find the size of the functional group is more responsible for the former, while the intermolecular interaction strength is more responsible for the later. Therefore, hydroxypropyl acetyl group, which has both bulky size and strong interaction strength, is the most effective functional group, followed by hydroxypropyl and acetyl group, in good agreement with the results from experimental dissolution tests. In addition, we provide continuum models and predict that the drug release time from a typical solid dispersion particle with 2μm diameter ranges from several seconds to less than 10 minutes depending on the functional group. The systematic coarse-graining approach offer molecular level insights that aid the design of high performance polymeric excipients, and can be extended to cellulosic polymers with novel functional groups and additional drug candidates of interest. Thus, our multi-scale modeling approach is of great interest to the pharmaceutical and material design fields.