(190ab) Predicting the Solubility of Pharmaceutical Solids by Molecular Simulation

Given the time and expense required for current rational drug design methods, molecular simulation holds a promising role as a drug design tool. We will present results of two unique approaches to predict the solubility limit of solids. The solubility limit of pharmaceutical solids is a critical thermodynamic property that is related to their bioavailability, and is also necessary for their synthesis and processing. In the first approach, free energy calculations are conducted for both the solid and solution phase of the solute to satisfy the phase co-existence criteria. In this manner, estimates of the solubility limit may be found in the complete absence of experimental data. 

The second approach is a simplified method to estimate the solubility of sparingly soluble solids. The method assumes that the activity coefficient of the solute is relatively insensitive to the solute composition, resulting in a simplified thermodynamic model that involves the calculation of the residual chemical potential of a single solute molecule in solution at a finite concentration using an appropriate free energy simulation technique. With knowledge of a single experimental solubility data point, difficult simulations of the solid phase and use of analytical reference states are avoided. 

Both methods have advantages over more empirical descriptor-based methods in that the simulations enable insight into the underlying molecular driving forces responsible for solubility trends. In this fashion, we will present simulation results that supplement solution theory to predict the existence of solubility enhancement in multi-component solvents. Results are presented for the solubility of paracetamol (acetaminophen) in pure and mixed solvents. We find that the methodology gives reasonable agreement with available experimental results, and is capable of predicting the existence of solubility enhancement in mixed solvents.