Publicly available, previously published and newly collected multi-omics data and literature-mined features were cross-correlated and analysed using bioinformatics tools to decipher molecular alterations related to PCa progression. This led to the generation of a knowledgebase of disease-relevant molecular alterations, molecular processes/ pathways with increased validity. Placing the data in the context of the tumour biology revealed key elements driving PCa progression.
Specifically, multi-omics analysis was performed based on seminal plasma (n=80), urinary peptidomics (n=823), tissue proteomics (n=104), and transcriptomics data (n=1,707). In brief, peptidomics analysis allowed for identification of 16 seminal plasma-derived and 91 urinary peptides that were significantly correlated with disease progression (p<0.05 Spearman's correlation). The former (among others) were fragments of most abundant seminal plasma proteins (semenogelins, lactotransferrin) and the latter were mostly collagen fragments, likely reflecting changes in the extracellular matrix remodelling, a crucial process for cancer progression. Tissue proteomics resulted in an identification of a total of 5,324 proteins, of which 2,802 were found to be significantly correlated with PCa progression; whereas the transcriptomics data (a part of Prostate Cancer Transcriptome Atlas) revealed 12,862 genes being associated with progression. 8,591 molecular features associated with PCa were also retrieved from literature and public resources. To shortlist the most credible molecular changes, the above data were cross-correlated. This gave rise to 392 proteins (a molecular signature) exhibiting the same direction of the association with disease progression across different data traits. Among them, 257 (66%) were confirmed to be associated with PCa based on literature.
Subsequently, bioinformatics analysis was conducted to decipher the biological function and pathways behind the cross-correlated data. The vast majority of the predicted pathways/ processes were linked to metabolism. Oxidative phosphorylation, branched-chain amino acids degradation, fatty acid β-oxidation I, Pyrimidine/ Purine biosynthesis, were predicted (among others) as activated, and reflected the metabolic hallmarks of PCa. Potential therapeutic agents were predicted using Connectivity Map (CMap), leading to the identification of 68 drugs/ compounds (p<0.05) that could potentially reverse the disease phenotype. Among 15 most promising findings, 7 were novel in the context of PCa. Cross-correlation of the results from the CMap and pathway analysis, revealed that 32 protein changes within the metabolic pathways are reversed by at least 3 novel therapeutic agents. Out of those, 10 proteins were involved in at least 2 pathways, and those were defined as potential drug targets and considered for validation. Three out of ten potential drug targets (i.e. mitochondrial acetyl-CoA acetyltransferase, peroxisomal multifunctional enzyme type 2, low molecular weight phosphotyrosine protein phosphatase) have been previously studied by immunohistochemistry (IHC), confirming the validity of the findings.
In parallel, dissemination to multi-target groups (e.g. scientists, pharma companies) was actively followed via multiple routes including e.g. scientific publications, presentations to meetings, project website, and direct contacts. Among other achievements, PCaProTreat was selected to be part of Innovation Radar, a Pilot study on Innovation launched by the European Commission. The IF fellow also edited a Special Issue in Proteomics Clinical Application, entitled “Clinical Proteomics on the Way towards Implementation”. The Project created a solid base for future exploitation of the findings in pre-clinical models. The initiation of new collaborative research projects is ongoing. PCaProTreat analytical pipeline may be also transformed into a service, supporting definition of molecularly-driven drugs/ therapeutic agents. In parallel, extensive training was followed to strengthen the competitiveness and the position of the IF fellow in the scientific community. These achievements were enabled by the smooth integration of the IF fellow in the Host team, and their fruitful collaboration.
