Experimental-Computational Approaches to Identify Synergistic Effects of Anti-Cancer Drugs
MetadataShow full item record
Drug combinations are hoped to improve treatment response to anti-cancer drugs by targeting the cancers vulnerabilities at diverse trajectories. With a large number of approved agents of chemotherapeutic and targeted therapies and several anti-cancer drugs in development, the number of possible combinations extensively escalates. As drug combination effects are highly influenced by individual traits of the cancer or cancer model system, identification of effective compound combinations is challenging. High-throughput screening is used to uncover effective combination treatments by testing them in experimental systems relevant for a large range of cancer subtypes. However, efficient filtering of interesting combinations for testing, such as can be provided by predictive computational modelling, is needed to further economise screening efforts. In the scope of this thesis, we advanced the use of logical modelling to identify putative effective drug combinations for pre-clinical screening by demonstrating the successful calibration of cancer cell line specific models to predict drug combination effects for four different cell lines. Testing predictions against experimental observations for 153 drug combinations in our large drug combination screen, indicated that we could have reduced the screening load considerably and increased synergy detection rate among the proposed combinations by 2.6-fold. Improvements in modelling strategies contributed through the work in this thesis pertain to optimisation of model network topology, selection of baseline protein activity data for model calibration as well as to approaches that can be used to identify subsets of nodes in the model for which accurate acquisition of baseline activity data is most important. Drug combination hits identified in vitro are met by low bench-to-bed translational efficiency, questioning the reliability of traditional planar (2D) cancer cell line cultures as model systems for drug screening. As part of this thesis, we have performed a high-throughput screen to systematically compare drug combination effects observed in planar (2D) and spherical (3D) cell line cultures. We identify combinations that act synergistically in only one but not the other culture mode. In spheroid cultures a lower number of synergistic drug combinations was identified with stronger dependency on MEK-signalling compared to 2D cultures. This indicates that future screening platforms should encompass more complex cancer models to capture a broader range of therapeutic synergy landscape as well as highlights the relevance of signalling dynamics and activities as markers for drug response. This thesis offers strategies to support cell line specific logical modelling as a possible tool to economise pre-clinical screening efforts, and to enhance insights into design of high-quality high throughput combination screening for reliable detection of synergistic drug combination effects.
Has partsPaper 1: Flobak, Åsmund; Niederdorfer, Barbara; To, Vu; Thommesen, Liv; Klinkenberg, Geir; Lægreid, Astrid. A high-throughput drug combination screen of targeted small molecule inhibitors in cancer cell lines. Scientific Data 2019 ;Volum 6. s. 1-10 https://doi.org/10.1038/s41597-019-0255-7 Attribution 4.0 International (CC BY 4.0)
Paper 2: Barbara Niederdorfer, Vasundra Touré, Miguel Vazquez, Liv Thommesen, Martin Kuiper, Astrid Lægreid, Åsmund Flobak. Identification of strategies to enhance cell line specific drug synergy prediction using logic modeling.
Paper 3: Evelina Folkesson*, Barbara Niederdorfer*, Vu To Nakstad, Liv Thommesen, Geir Klinkenberg, Astrid Lægreid, Åsmund Flobak. High-throughput screening identifies selective synergistic drug combinations in 2D and 3D colorectal cancer cell cultures