Antimetabolites during the sustained phase accounts for differences in accuracy between these two models. Although the model from Balakrishnan et al.36 predicts a smooth continuation of the curve, characteristic of drug release from a single layer model, the duallayer model transitions into the sustained phase of the release profile more accurately, representing the experimental data. Assuming that for a certain drug, release was exclusively dependent on diffusion, normalized drug release profiles should be independent of initial drug loading within the coating. From the experimental data presented in Figs. 7a and 7b, it is noteworthy that RQ1 and RQ2 drug loadings result in different normalized release profiles, indicating that drug release from the polymer is not solely diffusion driven. However, by using two diffusion coefficients to account for the complex molecular interactions that drive biphasic release profiles, we can use the proposed model to assume diffusion mediated kinetics. A family of predicted drug release profiles for RESV and QUER survivin revealed that the majority of RESV and QUER within the RQ1 coatings resides in the sustained release domain of the polymer compartment.
In the RQ2 coating, a greater proportion of the initial TGF-beta QUER load redistributes to the burst compartment compared with RESV, even though the loading of RESV is twice that of QUER. Thus, the equilibrium state of QUER, a less lipophilic drug compared with RESV, appears to favor burst release at relatively low initial loading. Given that arbIBS is a hydrophobic polymer,47 it may have greater capacity for lipophilic compounds such as RESV, lacking the ability to sequester and provide controlled release of less lipophilic compounds such as QUER. Thus, drug polarity appears to influence the apparent saturation point of drug within the polymer system, after which increased drug loading results only in exacerbating the initial burst drug release phase. As shown in Figs. 7c and 7d, a robust burst phase from the polymer rapidly saturates the arterial compartment, followed byminimal delivery at later time points. Because of potential arterial toxicity, this type of capecitabine profile may be deleterious for a therapeutic associated with a narrow TI. Conversely, drug distribution primarily within the sustained domain features a limited burst phase, followed by greater drug elution later in the release profile.
Note that the predicted arterial profiles are actually triphasic. The first phase is characterized by an initial saturation of the tissue within the first 24 h. This is immediately followed by a period of rapid arterial drug loss driven by large concentration gradients within arterial tissue. The third phase is revealed as the arterial concentration transport gradients resolve over time. After about 2 weeks, a balance is achieved between a slow influx of drug from the polymer coating and the loss of drug from the artery into either the adjacent tissue or blood compartment. The simulations reveal that, relative to QUER, lower transmural and planar arterial diffusion of RESV may promote the residence time of RESV in arterial tissue, resulting in normalized profiles that are potentially two orders of magnitude greater than QUER. This disparity in arterial drug level is independent of polymer release profile.