An experimental and bioinformatic tissue proteomics-strategy applied to human hepatocellular carcinoma focussing on functional data interpretation

The main challenge of tissue proteomics arises from the intrinsic complexity of samples which impedes reliable data interpretation. Tissues are composed of different cell types and a specific extracellular matrix. Although tumors are supposed to arise from a single cell type (e.g. HCC from hepatocytes or liver stem cells), their composition is additionally influenced by e.g. the infiltration of immune cells, by changes in the abundance of cells constituting also non-tumorous tissues and by epithelial-mesenchymal transition of cells due to an increase of developmental plasticity and loss of differentiation state usually accompanying tumor progression as well as the extracellular matrix as crucial constituent of the tumor microenvironment. The aim of this work was to determine functional differences of non-tumorous liver and HCC tissue via proteomics methods (shotgun, mass spectrometry, 2D-PAGE). For data interpretation, the focus was laid on basic cell biological and metabolic alterations resulting in the loss or gain of functions rather than those occuring at the single gene (protein) level. Proteins were functionally clustered based on several databases (e.g. KEGG, PubMed) as well as biochemical and toxicological literature. Moreover, functional proteome alterations were correlated to physiologically, histologically, and cytologically observable alterations. In order to enable the comparison of immunohistochemically (from www.proteinatlas.org) and tissue proteomics-derived protein expression, a method for the elaboration of histologic data was introduced. In addition, proteome data were compared to those of 'Encyclopedia of Hepatocellular Carcinoma genes Online' (EHCO), a database compiling eight gene set collections including PubMed, SAGE, microarray, and proteomics data. HCC tissues generally exhibit reduced secretion performance and concomitantly enhanced cytoplasmic protein synthesis (see paper 'A novel technique to specifically analyze the secretome of cells and tissue') although liver-specifically synthesized plasma proteins were found to be enriched in HCC compared to non-tumorous liver at the tissue proteome level. Hence, the quantitative and qualitative aspects of the tumor secretome are discussed as well as functional groups with influence on this phenotype including protein synthesis, import of proteins into the ER as prerequisite for secretion as well as chaperones and protein degradation. Furthermore, alterations of enzyme expression involved in glycolysis (Warburg effect), glycogen metabolism and fatty acid metabolism (altered composition of lipid composition of tumors) were determined. In additon, some liver-specific functions concerning detoxification in a broader sense are covered (urea cycle, phase I- and II-system). Moreover, the involvement of e.g. c-Myc and Ras in tumor development or maintenance is indicated, the first based on the coordinated expression levels of five independent proteins, the second by being highly upregulated in HCC tissue. In addition, some other tumor-relevant proteins such as anti-apoptotic and drug resistance-mediating sorcin, the supposedly locally enriched cardiotrophin-1, and HRPAP20, which enhances the growth and survival of hormone-responsive tumor cells but has not been associated with primary liver cancer thus far, were found to be pronouncedly more abundant in HCC. (author)