Chemotherapy resistance: from proteomics to drug targets
Research Team: Dr. Yu Yu, Mrs. Mudra Binju
CIC Specialists: Dr. Kevin Chai, Dr. Nancy Tippaya
Ovarian cancer remains the deadliest of all gynecologic malignancies because of its non-specific symptoms and propensity to metastasise early coupled with the lack of an effective screening test. Despite standard therapeutic strategies, which include cytoreductive surgery and platinum/taxane-based chemotherapy, most women with advanced ovarian cancer will experience a relapse. It is common for the tumour to develop chemoresistance and recur. Thus, it is necessary to develop new therapies.
Protein kinases are involved in many essential cellular processes by downstream substrate phosphorylation on tyrosine, serine or threonine amino acid residues. In malignant tumours, kinases are often hyperactive, overexpressed and dysfunctional. Therefore, kinase inhibitors are widely used for treating solid cancers. However, phosphorylation signalling is a highly robust and evolvable system that is non-linear and has a network-type architecture. Thus, targeting single kinase is unlikely to produce long-lasting tumour regression in patients, and it is common for a tumour to develop resistance. Dr Yu Yu and Mrs Mudra Binju from the School of Pharmacy and Biomedical Sciences and Curtin Health Innovation Research Institute (CHIRI) are teaming up with Dr Kevin Chai and Dr Nancy Tippaya from the CIC to develop a new combination targeted therapy for resistant ovarian cancer.
The team focused on identifying protein kinase-substrate relationships using raw phosphorylation peptide mass spectrometry data in ovarian tumours which recurred within 6 months of chemotherapy, thus classified as resistant tumours. We predicted kinase-substrate associations using a machine learning-based tool, NetworKIN, to explore potential drug targetable cellular phosphorylation networks. After which, the top combination kinase-substrate targets were prioritised using Cancer Hallmarks Analytical Tool, a text mining and natural language processing tool that considers cancer-related functions of a protein. The combination with the highest cancer hallmark coverage was hypothetically considered the most crucial.
This analytical pipeline allows researchers to have a logical framework to translate proteomics data into drug combination targets that can be tested back in the laboratory. As a result, the team is currently testing new drug combinations in patient-derived ovarian cancer organoid model. We can also apply this knowledge to phosphorylation proteomics studies in different cancers. The data would allow us to re-evaluate the system and hopefully generate useful treatment for resistant tumours.