for Physical Chemists by Physical Chemists
Exercise to create a diverse set of reliable, relevant, peer-validated, solubility data
During the PhysChem Forum in December 2005 the issue of solubility and the lack of consistency between published data was raised a number of times. There is also a great deal of concern over the validity of some of the data which is now available in the literature and on-line which further confuses the area. A number of examples were presented at the meeting showing where data had been mis-quoted and then subsequently cited a number of times until it had become reality (Antonio Llinás Martí, Pfizer Institute for Pharmaceuticals Materials Science, University of Cambridge).
While there is no clear definition for the term solubility this is always likely to be a cause for concern. This has been highlighted by a number of companies presenting how they each determine their solubility data, which is naturally measured in a method which provides them with the most relevant data for their particular compounds, disease areas and strategies. There are differences in sample concentration, co-solvent presence, co-solvent concentration, incubation times, thermodynamic methods, kinetic methods etc, all of which having been carefully selected from months of validation work, to give the answer to the most appropriate questions given the individual group’s requirements.
This is, of course, how it should be. Simply because Company X measures their compounds under a specific set of conditions does not automatically follow that this is the best method for Company Y to follow, and neither does it mean that any method that differs from this is “wrong”. However, what it does mean is that it is extremely difficult to validate the methods or to cross check the measurements with published data since so many conditions which are critical to the solubility processes are varied.
What we are proposing, is to pull together as comprehensive a set of compounds as possible which are commercially available, well characterised and structurally diverse, yet relevant to drug discovery. The aim is to put these compounds through as many of our in-house assays as possible, providing a set of data which clearly documents the variations in methods from one company to another. Over the course of time, this would grow to include more assays (permeability, lipophilicity etc) and more compounds and so allowing those wishing to validate their assays to make a more informed choice as to the data they use for comparison. This has already begun with some 100 or so compounds being studied by my colleagues at UCB, Slough and at Organon, Newhouse and the data will be released very shortly.
Our request is simple. Please take this list of compounds and put it through your assays. Share your data and your assay conditions with the PhysChem forum and send us any compounds that have not been included in this list so far that you have had experience with. All this information can then be discussed as an ongoing exercise at the next forum in December. This exercise is by no means intended to be judgemental of data and assays in any way. It is merely designed to help us create a set of values for easily accessible compounds which are validated, reliable and relevant.
Please send any contributions or comments you wish to make to firstname.lastname@example.org
Starting from 10mM DMSO stocks, a 2% DMSO solution in 0.05 M phosphate buffered saline, pH7.4 buffer (ionic strength adjusted) was incubated at room temperature (22 ± 2°C) for 24 hours with vortex mixing (1500 rpm). After mixing, samples were filtered through a filter (Millipore Multiscreen Solubility plate, 0.4 µm, modified PCTE). Sample analysis is achieved using gradient HPLC with a mobile phase of water/acetonitrile+0.1% formic acid (95/5 to 5/95). Detection was at 230 and 254 nm. A calibration curve was produced by injecting three volumes (4, 8 and 12 µl) of a 50 µM DMSO solution of the compound under test, and this curve is used to quantify how much material is in the filtrate of the sample solutions. Each experiment was performed in duplicate.
AKAS – Automated Kinetic Aqueous Solubility
Starting from 10mM DMSO stocks, 5% DMSO solution in pH7.4 buffer was incubated at room temperature for 90 minutes with gentle shaking using a Millipore 96 well Multiscreen Solubility hydrophobic filter plate (polycarbonate, 0.4mm). The filtrate was analysed by UV plate reader from 240nm to 400nm at 1nm increments. The experiment was carried out in duplicate. (Data is also available at pH3, pH5 and pH9 upon request - omitted from data set for simplicity and clarity.)
The sample concentration was determined by reference to a 4 point calibration curve with serial dilutions from 10mM DMSO stock in 20% acetonitrile / water. The wavelength for analysis was determined from the calibration and sample spectra with the use of two extra standards spiked with 10% pH3 buffer and 10% pH9 buffer for selection of isosbestic point for compounds where there was a spectral shift upon ionisation.
QSol – Pseudothermodynamic Solubility
Starting from solid material, 0.5mg sample was incubated in pH7.4 buffer at room temperature for 90 minutes with gentle shaking. The samples were transferred to Millipore 96 well Multiscreen Solubility hydrophobic filter plates (as used in the AKAS assay). The filtrate was analysed by gradient elution RP-HPLC using UV detection at 254 nm and the concentration determined by reference to a 3point calibration curve. The experiment was carried out in duplicate. (Data is also available at pH5 upon request)
For more information, please send an email to email@example.com
Send mail to
questions or comments about this web site.