Abstract:
The clinical research for novel biomarkers in early diagnosis and therapy surveillance of cancer diseases is a rapidly emerging field since many of the presently applied marker compounds show only unsatisfactory prediction values. Besides the analysis of the genomic profile (genomics) and the expressed protein pattern (proteomics), the metabolome, represented by the biochemical endproducts of a biological system, gains increasing interest.
Metabolites, often discussed as potential tumormarkers are the so-called modified nucleosides. In addition to the regular ribonucleosides, RNA contains a series of derived modified analogs. Modifications like methylation, hydroxylation, acetylation, isomerization and the addition of more complex side chains are posttranscriptionally implemented on the macromolecular level. After enzymatic degradation of RNA, the modified nucleosides are excreted as biochemical endproducts from the cell in the bloodstream and finally in the urine. Due to the altered cellular metabolism in tumor tissue, higher excretion rates can be observed, resulting in characteristic alterations in the metabolic profile. Considering the network characteristics in metabolism, the described phenomenon can also influence the excretion pattern of compounds, originating from pathways interconnected to the cellular RNA metabolism.
In the present work, the metabolic profile of modified nucleosides and related metabolites were analyzed with several mass spectrometric techniques.
Methods for the isolation of the mentioned compound class with affinity chromatography were developed and optimized for their application on different biological fluids and matrices (urine, cell culture supernatants, blood, tissue). For the identification of yet unknown metabolites, a selective method for the isolation of ribosylated compounds from 24h urine of tumor patients was established. The isolated metabolites were subsequently analyzed via IT-MSn and FT-ICR-MS. Combining the obtained mass spectrometric information with characteristic fragmentation patterns and molecular formula suggestion, 23 structures could be identified. Ten metabolites turned out to be novel urinary compounds, which have not been described previously in literature. Additionally, the occurrence of four modified nucleosides in human urine was verified by mass spectrometry for the first time. Besides compounds from the primarily analyzed RNA metabolism, metabolites originating from interconnected pathways like the purine biosynthesis, the histidine metabolism, the methionine / polyamine cycle as well as from the nicotinate / nicotinamide metabolism were also detected.
The excreted metabolic profiles of breast cancer cells (MCF-7) and breast epithelial cells (MCF-10A) were compared in the corresponding cell culture supernatants by means of LC-IT-MS analysis. Characteristic differences were observed in the excretion of certain metabolites, which could unambiguously be attributed to pathophysiological mutations and the resulting cell response.
The breast cancer classification potential of the excreted profile of urinary ribosylated metabolites was evaluated in another clinical study. Therefore, the metabolic profile in urine samples from 85 breast cancer patients and 85 healthy volunteers was analyzed via semiquantitative LC-IT-MS measurements. The obtained values were subsequently analyzed with bioinformatic pattern recognition via support vector machine (SVM). Instead of comparing absolute concentration values as input data, pairwise encoded metabolite ratios were used. Due to the biochemical interconnectivity of the considered metabolic pathways, dependencies are created between certain substrates. These dependencies were reflected by tumor-associated shifts of certain metabolite ratios, selected by the SVM algorithm. With the developed method, a sensitivity of 83.5% and a specifity of 90.6% were achieved with 59 metabolite ratios in the analyzed test collective. This means a considerable improvement of the classification potential, compared to currently applied breast cancer markers like CEA and CA 15-3.
The obtained results may contribute to the development of a reliable, non-invasive biomarker system for the diagnosis of breast cancer diseases.