Abstract

Increasingly major pharmaceutical companies are using molecular modelling approaches to develop liquid chromatography methods for the many compounds taken through pre-clinical drug development. This involves mapping out many compound structures and the properties of many LC stationary phases, and is only likely to be effective for large, well-resourced pharmaceutical companies. An alternative approach is to use LC stationary phases with very specific selectivity based on one single retention mechanism and prepare mixtures of such phases as required for method development. Two techniques that have employed this approach used coupled cartridges (phase optimised liquid chromatography or coupled columns (stationary phase optimised selectivity liquid chromatography) to achieve the optimal mixture of retention mechanisms, but interest in these techniques have declined partly because of their incompatibility with the very short columns used in modern high performance chromatography when utilising modern particle technologies such as sub 2 µm sizes or superficially porous particles.

The intention of this research project was to develop improved high performance liquid chromatography related substances assay methods by exploiting selectivity focused approaches combined with modern stationary phase particle technologies. Use of sub-2 µm silica and superficially porous particles was preferred to enable a smoother method transfer from the drug discovery and development stage through to the post market authorisation of an investigational new drug.

The first selectivity focused approach addressed was ion-pairing liquid chromatography, updated with volatile additives to overcome the mass spectrometry compatibility issues of traditional salt-based additives. A selection of low molecular weight perfluorinated carboxylic acids were characterised for their performance as ion-pairing additives against octyl sulfonic acid in the presence or absence of a counter-ion agent, N,N-dimethyl octylamine. Heptafluorobutyric acid provided the optimal balance of retention strength and selectivity to concentration ratio, baseline resolving the critical impurity separation of paroxetine from cis-paroxetine and desfluoro-paroxetine. Furthermore, the elimination or substitution of N,N-dimethyl octylamine with triethylamine resulted in the loss of resolution between the cis-paroxetine and desfluoro-paroxetine impurities. This suggested that ion-pairing acted as a binary mixed mode separation mechanism of dynamic ion exchange and hydrophobic partitioning of ion-pairing agent-analyte complexes with the addition of a counter-ion agent enhancing selectivity further by adding an ionic repulsion mechanism, forming a trinary mixed mode system that overrides the underlying stationary phase. This ion-pairing: counter-ion approach was used to develop related substances assay methods for active pharmaceutical ingredient and oral solution formulations of naltrexone, chlorphenamine, baclofen, and loperamide alongside separations of E- and Z- isomers of doxepin and dosulepin.

The second selectivity focused approach addressed was mixed mode liquid chromatography. By conducting a series of mobile phase scouting experiments in parallel on two stationary phases of orthogonal selectivity to one another with equivalent particle sizes, a facile two-dimensional plot was generated that could predict the optimum ratio of the two orthogonal stationary phases and the associated retention factors and selectivity of a drug and its related substances. Initial attempts at method development for a chlorphenamine and flurbiprofen related substances separation were frustrated by a lack of stationary phases orthogonal to C18 with modern particle sizes. Further progress was made with the baclofen related substances assay, with the packing of a ZWIX (+)-NH2 (72:28) stationary phase. Significant differences were observed in the retention order of baclofen oral solution related substances with this mixed mode column, attributed to a heterogenous mixture of the NH2 and ZWIX (+) particles across the length of the column.

The chiral selectivity towards baclofen hindered development in this case but could lead to mixed mode stationary phases to simultaneously resolve chiral and achiral related substances for racemic drugs. Overall, the two studies of these selectivity focused approaches to liquid chromatography method development provided a suitable platform for further exploitation of the stationary phase as a programmable variable through future application to a wider range of compounds.