Molecular profiling of prostate cancer offers hope for more effective drugs
Prostate cancer is diagnosed in around one in nine men. In 2018, over 1.2 million men worldwide were newly diagnosed with prostate cancer, including around 500 000 in Europe, making it the second most common form of cancer diagnosed in men. Although new drugs have been developed to treat prostate cancer, the response rate remains low, with frequent drug resistance and side effects. To develop more effective drugs, the EU-supported PCaProTreat project set out to improve current molecular-level knowledge about prostate cancer. PCaProTreat modelled Big Data gleaned from multi-omics technologies involving urinary and seminal plasma, peptidomics, tissue proteomics and transcriptomics, along with literature that defined molecular signatures for disease progression. Bioinformatics tools then determined novel drug targets and potential therapeutic drugs. “Currently, there are no drugs based on multi-omics profiles under clinical development, so this is a true innovation,” says Marie Skłodowska-Curie fellow, Agnieszka Latosinska from Mosaiques Diagnostics, the project host.
The molecular signature
Prostate cancer is a highly complex disease. Even though symptoms might be similar among patients, the set of molecules known as the ‘molecular signature’ responsible for the disease’s progression can be different between patients. As drugs act on the molecular level, mostly on proteins, this molecular difference between patients means that while a drug can benefit some, it might not be effective for all. Moreover, as the molecular environment is dynamic, tumours may develop resistance rendering a drug ineffective. Valuable samples rich in molecular information can be collected from prostatic tissue, urine and seminal plasma. Prostatic tissue is the site of the tumour growth. Seminal plasma contains proteins secreted by prostatic tissue, while urine contains proteins that were released as a result of tumour growth. To investigate changes in these samples, peptidomic, proteomic and transcriptomic data sets can be used to evaluate thousands of molecules simultaneously, and at different molecular levels, such as mRNA or protein. “For tumours to grow, molecules work together. By reverse engineering this disease-related molecular organisation, we can pinpoint the best targets for drugs to improve or even revolutionise disease management,” explains Latosinska.
The significance of metabolic reprogramming
Molecular profiles from the seminal plasma of 80 men, urinary peptidomics of 823 men, tissue proteomics of 104 men and transcriptomics data from 1 707 men, were analysed and cross-correlated. Integrating these different types of Big Data proved quite a challenge. But it was necessary to overcome the limitations inherent in describing prostate cancer based on clinical characteristics alone, such as tumour stage or prostate-specific antigen levels. This resulted in the discovery of a molecular signature for prostate cancer progression comprised of 392 proteins. The team showed that changes associated with prostate cancer progression are dominated by the increase in proteins involved in metabolic processes, suggesting that metabolic reprogramming is key to the cancer’s progression. Computer modelling predicted 68 drugs/compounds with potential to reverse the molecular signature and so fight the cancer’s progression. Among the most promising findings, seven novel drug candidates, previously untested for prostate cancer, were discovered. The team is now working towards testing these drug candidates in appropriate prostate cancer cell lines and animal models, in collaboration with pharmaceutical companies. If successful, these compounds will then be tested in human clinical trials. “If we are right that drug efficacy is determined by its ability to reverse the molecular changes underlying prostate cancer, identification of new molecularly driven drug candidates could help treatments to leapfrog a number of research steps,” adds Latosinska.
Keywords
PCaProTreat, prostate cancer, molecular, multi-omics, disease, drugs, proteins, seminal plasma, urinary peptidomics, tissue proteomics, transcriptomics, metabolic
