Multi-omics molecular profiling of prostate cancer
MetadataVis full innførsel
Prostate cancer is the most common form of malignancy inflicting men in the west and manifests itself as a heterogeneous disease with a range of different clinical outcomes. Whereas some prostate tumors grow slowly and may never become clinically significant, other tumors are aggressive and require early and radical treatment. The different radical treatment options for prostate cancer have several substantial side-effects, and it is therefore of interest to both the patients and the healthcare system to avoid unnecessary treatment. Hence, it is important to identify which tumors are aggressive and should be treated. Molecular characterization of tumors is important for finding and developing novel diagnostic and prognostic biomarkers that can accurately identify which patients would benefit from radical treatment. The overall goal of this thesis was to gain increased biological knowledge of prostate cancer tissue through a multi-omics approach. The aim was to identify metabolic and transcriptomic biomarkers that are descriptive of prostate cancer development and can predict tumor aggressiveness and clinical outcome. Transcriptomics is the study of the complete set of RNA transcripts, or gene expression, made from DNA in a biological sample, e.g. a tissue sample, whereas metabolomics investigates the chemical processes involving metabolites. The methods used in this work included; gene expression analysis through DNA microarrays, histopathology, immunohistochemistry and metabolic profiling through high resolution magic angle spinning magnetic resonance spectroscopy (HR-MAS MRS) and matrix assisted laser desorption/ionization time-of-flight mass spectrometry imaging (MALDI-TOF MSI). The first paper investigated the gene expression of secreted frizzled-related protein 4 (SFRP4) across nine independent patient cohorts. SFRP4 was shown to be significantly associated with prostate cancer and aggressiveness, and was significantly inversely correlated with the metabolites citrate and spermine, two metabolites whose reduced levels previously are identified as prognostic markers for prostate cancer. Further, SFRP4 gene expression was a predictor of recurrence and metastatic disease, making SFRP4 a potential biomarker for prostate cancer stratification. Reactive stroma is a tissue feature commonly present in prostate tumors and is reported to predict a worse clinical outcome. In the second paper, metabolic and transcriptomic profiling were performed on prostate cancer tissue samples with low and high reactive stroma content. Genes and metabolites related to immune processes were particularly significant and enriched in samples with high reactive stroma content compared to low reactive stroma. The expression of genes related to remodeling of the extracellular matrix (ECM) were also increased. Finally, in concordance with previous publications, high reactive stroma content was a significant predictor of recurrence even when accounting for Grade Group, which is the standard histopathology grading system for prostate cancer aggressiveness. This finding supports integrating reactive stroma assessment with the Grade Group system in the clinic, although standardization of this assessment is needed. MALDI-TOF MSI is an emerging method which allows measurement of the spatial distribution of biological molecules, including metabolites and lipids. This is of special value for analyzing malignant tissues such as prostate cancer which commonly have a heterogeneous tissue composition. In the third paper, we tested the applicability of MALDI-TOF MSI for metabolic profiling of prostate cancer tissue sections, using both ion modes. By pairwise comparing benign epithelium, stroma and cancer tissue regions, we identified alterations in key metabolic processes. Notably, reduction of citrate and spermine in cancer tissue compared to benign epithelium were validated, higher levels of the antioxidant taurine were found in stroma compared to both benign and cancer epithelium and higher levels of phospholipids indicated altered lipid metabolism. Interestingly, we identified elevated levels of carnitine and acetylcarnitine, key metabolites of the carnitine shuttle which facilitates oxidation of fatty acids, in cancer tissue compared to both stroma and benign epithelium. This is a novel finding in prostate cancer tissue and further investigation of these metabolites as biomarkers is warranted. Together with generally increased levels of phospholipids in cancer tissue, this finding supports lipid metabolism as a key player in prostate cancer development. In conclusion, by using a range of different metabolomics, transcriptomics and histology methods for analyzing prostate cancer tissue samples, several molecular alterations were identified in this thesis. SFRP4 gene expression is a potential biomarker for prostate cancer stratification, although further assessment of its applicability as a biomarker is needed. Molecular characterization of reactive stroma showed an increase of immune processes and ECM remodeling, and added extra support for using reactive stroma assessment in clinical histopathology. Through MALDI-TOF MSI analyses of prostate tissue, several biological molecules were detected as altered between different tissue types, such as metabolites involved in lipid synthesis, �����-oxidation, prostatic secretory function and inflammation. The differences in metabolite distributions between the defined tissue types, showcase MALDI-TOF MSI as valuable metabolomics tool for analyzing heterogeneous tissue samples and for biomarker discovery.