The field of metabonomics is beginning to grow rapidly due to the ability to analyse
biofluids, providing a 'snapshot' of biological processes that have happened (cf:
proteomic/transcriptomic studies, which predict what may happen), making it possible
to profile responses over time. The work described in this thesis was motivated by
the aim of profiling clinical urine samples obtained from fracture patients, with a view
to identifying potential biomarkers related to failed fracture healing. This led to the
need to develop and evaluate metabonomic approaches, specifically a orthogonal
separation approach complementary to the commonly-used reversed phase (RP)
separation methods, namely hydrophilic interaction liquid chromatography (HILI C).
Urine samples from healthy volunteers were collected and used to develop an LCMS
'metabonomic toolbox'. This development evaluated various aspects of a·
metabonomic study that are commonly poorly reported within the literature: 'study
design, sample collection storage and handling considerations, data extraction,
normalisation and scaling methods, and multivariate data analysis tools.
From the literature, the commonly-used method of normalising to creatinine was '
found to be unsuitable due to perturbations in the urinary excretion of creatinine due
to factors such as illness. Methods used to evaluate system ~tability were also
developed and added to the 'toolbox'. HILIC was successfully used as a separation
technique orthogonal to RP, producing comparable results but using different
metabolites; this highlights the fact that much potential information is P?ssibly being
lost when only RP-LC-MS methods are used for analysis.The need to use both
modes of ionisation polarity were also addressed for an increased coverage in
biofluid metabolite profiles. ,
The knowledge gained in the development of the 'metabonomic toolbox' was used for
the analysis of clinical urine samples. Despite the lack of properly time-setted
samples and none of the recruited patients suffering delayed fracture healing,
potential metabolites related to fracture healing were found. However, the samples
were very different to previously-analysed samples from healthy volunteers; they ,
showed very large amounts of protein, which had a large range of molecular weights.
These were' identified- proteomically.
Finally, ESI-Q-o-ToF MS/MS, MALDI-ToFlToF MS/MS and racemic amino acid
analysis were used for the structural determination of a pseudomonad biosurfactant,
which was identified, unexpectedly, as the cyclic Iipopeptide white line inducing
principle, WLIP.