Validation of a Chromatographic method for LogD measurement

Introduction

This blog returns to the familiar theme of physico-chemical measurement.  Over recent weeks I have been working to establish a suite of assays using my own UHPLC instrumentation.  LogD, solubility and PAMPA permeability measurement are all now available to clients via out-sourcing, in-sourcing or consultancy to set up the methods in your laboratory which can also include bespoke software.

Here I am presenting some validation data to support the use of my LogD7.4 method.

The importance of LogD7.4

Physico-chemical properties profoundly influence ADMET parameters, compound progression and the chances of success in a drug discovery project (Leeson et al.,).  Within the range of physico-chemical parameters that can be measured to define a compound’s lipophilicity, solubility and permeability, LogD7.4 is fundamental to the understanding of the ADMET/PK behaviour of most chemotypes encountered during drug optimisation.  Data that illustrates this point is abundant in the literature and I have picked out two good examples below that show the correlation of intrinsic clearance with LogD7.4 both for a set of marketed drugs and for discovery stage compounds.

The pie charts in the figure above make it very clear how the chances of finding metabolically stable compounds (the blue sections) drop off with increasing logD - and this amounts to a lot of extra screening effort to find good compounds if you're operating in an unfavourable lipophilicity region.

Measurement of LogD7.4

The use of LogD7.4 has become so fundamental to modelling and simulation efforts that the need for rapid measurement of this ubiquitous parameter is more important than ever. It is only by measurement, in fact, that subtle shifts between slightly modified compounds can be picked up that are vital to compound optimisation but are often missed by computational methods.  The gold standard measurement method remains the shake-flask method but this is not without its problems; most notably these are associated with the practical need to analyse two layers with orders of magnitude differences in concentration but this method may also be compromised by low solubility.  In any case, greater throughput is required and so here I present an alternative chromatographic approach suitable for use as a reliable lead optimisation tool.

A liquid chromatographic method to measure logD7.4.

The method differs from some published chromatographic alternatives in the rigour with which the pH of the mobile phase is controlled using phosphate buffer and that it is an isocratic method.  Advantages over other techniques for measuring LogD7.4 such as shake-flask and titration methods include its dynamic range (six orders of magnitude or more), throughput, reproducibility, low sample requirement and that it has no solubility limitation.

Calibration

Chromatographic retention (capacity factor, k’) is calibrated using a set of neutral compounds of well-established LogD7.4 in the literature.  An example calibration chromatogram is shown below:

Test compounds are then analysed using the same system and their LogD7.4 is calculated from their retention time using the established calibration (line of best fit). The chart below (LogD7.4 versus logk’) demonstrates the calibration that was achieved in this study.

Test compound analysis

A set of known drugs, with a range of lipophilicities and acid/base properties and which were also measured using the shake-flask method, were analysed.  The table below shows a comparison of the measured HPLC LogD7.4 values with their literature and shake-flask LogD7.4 values.  The acid/base/neutral class of the molecules is also indicated together with their predicted pKa.

The chart below is a plot of HPLC LogD7.4 against shake-flask LogD7.4 values.

The above line of best fit was then used to predict shake-flask LogD7.4 from HPLC LogD7.4 with the results shown in the table below.

Discussion

There was a high correlation between shake-flask and chromatographic LogD7.4 and this was used to further refine the prediction of shake-flask LogD7.4 values.  Using this procedure, the greatest difference between HPLC predicted and actual shake flask logD was 0.64, with 7 out of the 10 compounds being within 0.33 units, typical of shake-flask inter-assay variability.

The data shows a good correlation between HPLC and shake-flask LogD7.4 as would be expected.  An exact agreement cannot be expected as the systems are chemically different.  HPLC LogD7.4 is a lipophilicity scale that has been normalised to give values consistent with octanol/water (pH 7.4) distribution values which are universally accepted by Medicinal Chemists and for which a large body of data and correlations with biological data has been built up. However, there is no octanol involved but rather distribution is between a C18 phase and a buffered aqueous phase containing a significant amount of methanol.  There is also scope for additional interactions with the underlying silica backbone which can influence the measured values.  The chemistry of the HPLC system is somewhat more complex but no less valid or biologically relevant. Technically it has advantages over shake flask in terms of speed and simplicity leading to higher throughput. It is inherently highly reproducible (thanks to the extremely high reproducibility of UHPLC systems), minimal sample is required and compound solubility is not an issue. The method here has a range of 0-6 but it could be modified by adjusting the mobile phase and with appropriate choice of calibrants.

The use of a training set of 10 known drugs with a range of properties, previously measured by the shake flask method, shows promise as a means of minimising the bias in the HPLC method and hence generating data even closer to conventional LogD7.4 values.

Acknowledgements

I would like to thank XenoGesis for providing the compounds used in this study and for carrying out the shake-flask LogD7.4 measurements.  In particular, I would like to thank Dr Richard Weaver, MD and Founder of XenoGesis, for his helpful analysis of the LogD7.4 data.

Further Information

If any of the issues discussed in these articles are of interest to you, please feel free to contact me directly on mark@discoveryanalyticalconsulting.com for further information.

Mark Portsmouth2 Comments