The aqueous solubility of a compound is a critical determinant in its success or failure as a drug candidate. Given this, one might anticipate that there is a high throughput, standardised method to measure solubility in the drug discovery and development setting.
Unfortunately, traditional methods used to measure solubility are neither rapid nor cost effective for higher throughput screening. As a result, solubility measurements are often pushed down in development process where the numbers of active compounds are significantly reduced. As a result, direct experimental solubility data is not available to assess of the scope of the overall solubility profile of the drug candidates being produced. Although there is no general consensus among drug researchers that solubility measurements (at early screening phases) are necessary to improve the overall adrug space' that compound libraries occupy, there is a common appreciation of the need for faster, more affordable solubility measurements to improve the drug discovery process as a whole.
The number of poorly soluble compounds in development puts a significant burden on researchers not only in terms of identifying adifficult' compounds early, but also to increase the accuracy of other measurements that are directly effected by solubility. This burden can only be effectively addressed by adopting a solubility screening strategy based on currently available methods.
Solubility is theoretically defined as an equilibrium state. If the fundamental factors that govern solubility are understood, why are solubility measurements themselves difficult to make?
Part of the answer is obvious. There is no direct relationship between the theoretical understanding of a given phenomenon and the ease of its measurement. The rest of the answer comes from the practical difficulties associated with the particular solubility assay conditions. By their nature, traditional solubility measurements take time.
Equilibrium solubility measurements start with the solid form of the compound (crystalline form preferred). The powder is physically mixed with an aqueous buffer (defined pH) until equilibrium is reached. The length of the incubation time directly influences the accuracy of the solubility measurement because of compound sticking. To minimise this effect, researchers have looked at a number of different assay tubes. In addition to the time of the assay, the process of physically separating the solid material from the ain solution' material directly influences the accuracy of the traditional solubility measurement.
Centrifugation and filtration
Separation methods most used are centrifugation and filtration. Filtration increases the potential for non-specific sticking by introducing a new material into the separation process (membrane filter).
Speciality membranes address this issue, but no one material has alow binding' properties for all possible compound types. In addition, the filtration process plays a roll in defining the solubility measurement by introducing an arbitrary particle size cutoff for solids. Compound particles that are smaller than the pore size can pass through the membrane (usually resulting in an over-estimate of the compound's solubility). Alternatively, centrifugation does not expose the compound to anew materials'. Non-specific sticking, as a result, is less of an issue.
The particle size distribution of the solid material directly affects the rate of the separation. Longer centrifugation times can help address this issue, but complete separations for all compounds are not guaranteed.
Higher throughput solubility measurements (using DMSO stocks, shorter assay times, more rapid detection methods) can suffer from the same separation artifacts if they require a physical separation of the solids before analysis. Some high throughput methods however do not require any removal of solids before analysis.
Particle detection methods (primarily using light scatter) detect the presence of the solid form to estimate solubility and the measurements are not affected by the presence of dissolved compound. The absence of a separation step is a big plus for this approach and can explain some of the differences observed between in-solution and solid form detection techniques. u
Joseph Goodwin is Product Development Leader, BD Biosciences, Bedford, MA, USA. www.bd.com
