(217ct) Rational Design of Polymeric Delivery Vehicles for Anti-HIV and Anti-Chlamydial Microbicides
The HIV/AIDS pandemic remains a major global health concern. As of the end of 2010, 34 million people were currently infected and approximately 2/3 of these cases resided in sub-Saharan Africa. Due to differences in reproductive anatomy and fluid transfer during intercourse, women have a much higher risk of infection compared to men and in some countries women account for almost 60% of the total infection rate. Current preventative measures are not adequately combating the pandemic and there is a need for novel prophylactic measures. One possibility is microbicides, which are vaginally or rectally delivered topical products that combine an active therapeutic and/or prophylactic agent that provides a preventative barrier with a delivery vehicle consisting of a polymer solution, gel, or cream. A major challenge delaying development of such products is a lack of knowledge of delivery vehicle design. Previous microbicide studies focused primarily on the antiviral activity of the active agent and failed during clinical trials. These failures create a need for next generation microbicides with rationally designed delivery vehicles. Computational molecular design (CMD) provides one such method for rational design.
CMD can be used to rationally design novel molecular structures by optimizing target physical properties of a polymeric delivery vehicle. To accomplish this quantitative structure property relations (QSPRs) are generated for a model building set of candidate delivery vehicles with known molecular structures. In this work that model building set consisted of a variety of cellulose ethers, synthetic polymers derived from cellulose, with different side group substitutions, degrees of substitution, and chain lengths. These polymers were selected based on their use in existing formulations for vaginal excipients and other pharmaceutical applications. The structural property data of these known polymers is incorporated with physical property data to generate the predictive property models. The physical properties considered include power-law consistency and shearing thinning index, which describe the flow behavior and retention of the polymer solution, and anti-chlamydial activity, which relates to the prophylactic activity of the delivery vehicle since there is a correlation between chlamydia and HIV infection. The molecular structure descriptors were then correlated with the three physical properties in non-linear property models and validated with χ2.
These QSPRs were then incorporated into an optimization problem formulation and solved to generate novel molecular structures. These structures are generated by minimizing the difference between ideal target properties and the target properties predicted by the QSPR model. The nonlinear correlations required a mixed-integer nonlinear program (MINLP) problem formulation and a stochastic solution method. One such method is Tabu search, which is a heuristic approach that uses guided local search to explore the entire solution space and solves for many local optima, resulting in the pool of possible novel structures. A range of structures are generated to ensure the selection of a novel cellulose ether candidate that can be feasibly synthesized. The ideal target property for anti-chlamydial activity is to maximize as close to one hundred percent as possible, while the ideal target viscosity properties vary across patient populations and are obtained from concurrent spreading simulations and relevant literature. To account for this variation the optimization problem will be solved several times for different ideal property values that correspond to differences in the size and mechanical properties of vaginal tissue among different groups of women.