How breast cancer patients respond to treatment outcome is not easy to predict and represents a major concern in the application of therapeutics. This study aims to better define personalized treatment options and yield new applications to fighting cancer by two independent approaches. We first modelled the heterogenous genetic make-up of breast cancer patients and systematically analysed the potential impact of cancer genes on cellular fitness ( i.e. cell viability or doubling time) as response to a collection of genetic and chemical perturbations - that is sensitivity or resistance of the cells. We developed a multiplexed screening platform and investigated thousands of potential gene-gene and gene-drug interactions in an isogenic cell line model. Up to 100 genetic alterations could have been interrogated in a single well assay, increasing the number of interactions screened per well so far by up to two orders of magnitude. Considering the clinical relevance and direct implication in breast cancer treatment, we assigned the resistant interaction between PI3K inhibitors and activation of NOTCH and c-MYC as the most interesting. Thus the presented screening strategy can be considered as a powerful tool to gain deeper insight into gene function and drug action, hence to precisely guide clinical treatment decisions.
Oncogenic pathway signatures mirror the biology of cancer and clinical outcome and therefore hold the promise to guide targeted therapy. Given that much attention is being paid on how to specifically and directly attack pathways rather than individual genes, prompted us to search for a therapeutic window targeting the PI3K-PDK1-AKT signalling as this pathway is aberrantly regulated in a large proportion of breast cancer patients and associated with poor prognosis. We zoomed into the PI3K signalling cascade and obtained highly promising insight into the PDK1 regulation.To date, only few mono-ubiquitin conjugated human proteins have been reported yet in all examples this post-translational modification displays a critical regulatory function. Unexpectedly, a diverse panel of human cell lines expressed mono-ubiquitinated PDK1 at varying levels, indicating that this modification is a common and regulated process. The small molecule ubiquitin conjugates to the kinase domain of PDK1 but this attachment does not require PDK1 kinase activity. By applying a library of ubiquitin proteases, we further document the ubiquitin-specific protease 4 (USP4) as a prime enzyme that inhibits PDK1 ubiquitination in vivo and in vitro and co-localizes with PDK1 at the plasma membrane, suggesting direct deubiquitination.
Together, the regulated modification of PDK1 by a single ubiquitin moiety generates an additional, unpredictable layer of complexity in this critical signalling network. Considering the clinical attractiveness of targeting the ubiquitination machinery, our data provides potential novel therapeutic angles for drug development.