Interest in metabolite profiling has grown considerably over the last several years, largely due to its applications to biomarker discovery. Mass spectrometry is an important tool in metabolite profiling, offering high sensitivity, accuracy and a large dynamic range. Furthermore, the inherent advantage of monitoring these small molecules over proteins is the relative ease of quantitation using mass spectrometry. Multivariate analysis methods can facilitate the discovery of hidden structure in the resulting spectral data and lead to the discovery of new biomarkers. This approach allows for a greater understanding of dynamic metabolite profiles of disease as well as in determining drug toxicity.

Here, reverse phase liquid chromatography/mass spectrometry with electrospray ionization (LC/MS) and differential profiling were used to investigate serum metabolite profiles as a possible diagnostic for prostate cancer. In order to focus on metabolite detection, proteins were first precipitated from the serum samples. Reverse-phase chromatography and mass detection were performed using an Agilent 1100 LC/MSD and a Micromass Q-TOF Micro. The C₁₈ column was eluted under gradient conditions of 5 to 90% acetonitrile (0.1% formic acid), with chlorpromazine as an internal standard. Mass spectral data were acquired from 100-1000 m/z in positive mode. The simultaneous measurement of m/z, retention time, and peak intensities of ions in plasma from both groups created complex data sets, which were analyzed using an in-house Differential Profiling program. Normalization of the data by retention time relative to the internal standard enabled accurate comparisons of all data sets. Ratios of ions were compared between the patient groups and, of over 2000 potential metabolites detected, approximately 10 are being investigated as possible biomarkers for the detection and diagnosis of this disease.

The ultimate goal of metabolite profiling studies is similar to proteomics; a complete understanding of the structure and function of each individual metabolite. To this end, we have undertaken the development of a web-based database, which allows for the comparison of metabolite profiles by biofluid type, disease, gender and age (http://metlin.scripps.edu). Ultimately, the inclusion of MS/MS fragmentation data for these small molecules would enable confirmation of identity.