Often neglected steps in transforming drug solubility from single measurement in pure water to physiologically-appropriate solubility-pH.

Abstract

Background and purpose: The solubility of weakly-ionizable drugs in pure water, S_w, is commonly measured. The pH-dependent properties of the saturated solutions can be surprisingly complex in subtle ways. This commentary examines the characteristics of such measurements through case studies of 32 free acids, bases, and ampholytes (including crocetin, glibenclamide, mellitic acid, quercetin, bedaquiline, brigatinib, imatinib, celecoxib, and lysine), using published water solubility data.

Computational approach: Usually, in such saturated solutions, the ionic strength, I _w, is close to zero. When the pH is adjusted away from pH_w, the ionic strength increases, substantially in some cases (e.g. I _w > 10 M at pH 7.4 for mellitic acid and lysine). This change in ionic strength alters the activities of the species in solution. The corresponding equilibrium constants used to calculate the concentrations of these species must be adjusted accordingly. Here, the Stokes-Robinson hydration theory, slightly modified with Setschenow 'salting-out' constants to account for solvent interactions with unionized drugs, was used to estimate activity coefficients. The calculations were performed with the pDISOL-X program.

Key results: Given reliably-measured values of solubility in water (S _w) and ionization constant (pK _a) of the drugs and assuming that the Henderson-Hasselbalch equation is valid, a method is described for (i) adjusting the measured S _w values at ionic strength, I _w ~ 0 M, to values expected at reference ionic strength, I _ref = 0.15 M (or at any other reasonable reference value), (ii) determining the water pH_w in saturated solutions of added neutral-form drugs; (iii) determining the intrinsic solubility, S ₀, both at I _w and I _ref, and (iv) using analytic-continuation in the equilibrium mass action model to deduce the solubility values as a function of pH, harmonized to a selected I _ref. For highly soluble drugs, whose I _w exceeds 0.15 M, the intrinsic solubility values appear to depend on the amount of excess solid added.

Conclusion: This commentary re-emphasizes that measured S _w is not generally the same as S ₀. It is stressed that transforming measured drug solubility in pure water to an ionic strength level that is physiologically appropriate would better match the conditions found in biological media, potentially improving applications of solubility in pharmaceutical research and development.