(653a) Multiscale Modeling of Polymer Microsphere Drug Delivery

Proteins and pharmaceuticals can be encapsulated into biodegradable polymer microspheres of controlled size. Controlled-release drug delivery systems are being developed as an alternative to conventional medical drug therapy regimens which require frequent injections or oral dosages because some drugs become deactivated upon contact with other chemicals in the human body or are eliminated relatively quickly from the body. Also, some pharmaceuticals have poor oral bioavailability which means that the drug compounds are not readily absorbed by the body through the digestive system. Controlled-release systems have the potential to provide better control of drug concentrations and reduce the side effects of current drug therapies. These systems can be designed to provide specific drug concentrations over extended periods of time which are optimal for effectiveness of the drug and which avoid potentially harmful high and low levels of the drugs. They also can be administered locally to the part of the body where the drug is needed. Although there are clear advantages to using controlled-release systems for drug delivery, the design of controlled-release devices, such as biodegradable polymer microspheres, depends heavily on trial-and-error experiments due to incomplete understanding of the mechanisms that regulate the release processes. An accurate computational model of drug release from polymer microspheres would be useful for determining the optimal parameters for fabrication of the microspheres to yield a desired drug concentration profile. A successful mechanistic model of this type would be a useful tool for planning experiments, and once thoroughly validated, it could be used in the design of pharmaceutical manufacturing of microspheres.

This presentation describes a mechanistic model that includes the effects of autocatalytic degradation in polymer microspheres that models phenomena at length scales ranging from molecule to micropore to mesopore to hundreds of microns. Degradation effects of a large number of chemical species are considered along with bulk erosion which incorporates many highly coupled nonlinear partial differential equations for chemical reaction and diffusion. These equations are solved for small computational time steps for an extended period of time in order to capture an entire release profile. The high resolution simulation of the coupling between reaction and diffusion captures important dynamical phenomena observed in experiments that cannot be modeled with the models in the literature that have simpler numerical solution but do not take the coupling into account. By using mechanistic models for all time and length scales, the model has no empiricism or ?fitting parameters.?

Numerical techniques are described that reformulate discrete spatially-varying population balance equations that appear in the full-order mechanistic model into a continuous spatially-varying population balance equation. This extends a reformulation commonly used in modeling polymerization to spatially-varying depolymerization. Basis functions used to parameterize the spatially-varying continuous population distribution are selected to tune the resolution of the extrinsic size coordinate and an intrinsic size coordinate based on the locally appropriate scales, with the overall approach producing orders-of-magnitude speedup while reducing the memory requirements low enough for the computations to be carried out on a personal computer. A parallel implementation of the numerical algorithms is described which speeds the computation of the multiscale model enough for its use in iterative optimal design and control computations.