Presently, the formulation of stable crystalline nanosuspensions is largely an empirical process. Each new drug compound is screened with various surfactant systems to find the most stable formulation. This process can be time-consuming, and can require a prohibitive amount of drug in early development phases. This work was undertaken to develop a molecular modeling approach to surfactant screening, which could identify stabilizing surfactant systems for a particular drug, based on its crystal structure.

We have screened several model pharmaceutical drug crystals with surfactants, including the nonpolar surfactant, Polysorbate 80. The nanosuspensions were prepared by high-pressure piston gap homogenization of the drug in an aqueous surfactant mixture. Aliquots of each suspension were stress-tested to rapidly assess the physical stability of the formulation. The various combinations of model drugs and surfactants provided suspensions with a wide range of stabilities to model.

The molecular modeling was performed in Materials Studio (Accelrys). For each model drug, the adsorption of both the surfactant and a molecule of the drug were modeled on a face of the drug crystal. We calculated the binding energy of each drug with these additives, and the ratio of the surfactant binding energy to that of the drug correlated well with the rank order stabilization of the crystalline suspensions. A ratio of approximately 3 correlated to moderate stability; a ratio larger than 3, correlated to excellent stabilization. For the drugs and surfactants with poor stabilization, the ratio was near 1.

The use of molecular modeling for preliminary formulation of suspensions offers a way to visualize the interactions of surfactants on the surfaces of crystal particles. It can be less time-consuming than empirical formulation screening, and does not use valuable and limited drug material. The technique could help to prioritize the formulations tested in empirical laboratory experiments, so that only the most promising surfactant systems would need to be screened to find a stable formulation.