Biomarker discovery for cervical cancer

Proteomics of human boy fluids is still in its early stage of development with major methodological challenges ahead. This implies that much attention is given to improving the methods and strategies. One major challenge is that many samples that have been acquired in the past may not fulfill the stringent requirements of storage and sample preparation to allow comparable proteomics analyses. It is therefore important to assess the factors that may affect the final proteomics result through systematic and reproducible analyses. Therefore accuracy and sensitivity of the analytical instrumentation is not the only critical factor in this research. Blood (plasma or serum) and urine can be easily sampled from patients or healthy volunteers and are therefore often the first choice when trying to discover novel biomarkers or biomarker patterns to diagnose cancer and other diseases. There are, however, drawbacks such as the masking of low-abundance by high abundance proteins and the possible effect of sampling and sample handling procedures (e.g. different times for blood clotting). A number of different approaches to deplete highly abundant proteins from human serum were studied throughout this thesis. Further, different analytical techniques were applied, such as a miniaturized, microfluidics-based LC-MS system (chip-LC-MS) to enhance overall sensitivity. It is shown that chip-LC-MS has at least twice the resolution of the previously used standard capillary LC-MS method. Since blood composition will change under the influence of external factors, the influence of clotting time on proteome of serum was studied. It was found that most proteins were not affected by clotting time except for those directly involved in this process, such as the fibrinopeptides. Next, we describe a more comprehensive approach for evaluating the influence of various pre-analytical parameters on the serum proteome. A factorial design strategy was applied to assess the importance of seven factors considered to be of relevance, including the level of hemolysis, the digestion conditions, and the storage conditions. Finally, we analyzed serum samples from cervical cancer patients at various stages of disease before and after treatment followed by data processing and statistical data analysis. While we did not discover major changes in the serum proteome using this method, subtle changes in the protein composition were observed in relation to treatment, the significance of which are being further investigated. It is thus demonstrated that the described methods are applicable to highly complex body fluids such as serum and that further studies into the relevance of the discovered changes of the serum proteome are warranted