One of the most promising areas of human cancer research is the study of signal transduction, or cell signaling. Communication between and within cells in the body is accomplished through signal transduction, as stimuli generated in one part of an organism travels through the circulatory system and initiates a response in specific target cells.
As cells mutate, they can affect these signaling networks, causing the formation and growth of cancer. Cancer treatments such as chemotherapy and radiation therapy attack any actively growing cells, not just cancer cells.
However, by better understanding the signaling pathways that are hyperactive in cancer cells, scientists can develop treatments that target just the cancer cells, sparing normal cells. This type of targeted treatment, which blocks signals that promote cancer cell growth, has the potential to kill cancer cells more effectively while avoiding harmful side effects.
Because cancer is such a complicated disease, the mathematical possibilities of protein interactions and combinations of drugs that could block signals in the human body that promote the uncontrolled growth of cancer cells can be overwhelming, and the behavior of these networks is difficult to predict.
Researchers use mathematical models to more accurately predict how certain drugs will inhibit the growth of cancer cells and use these predictions to develop the most effective combinations of drugs to treat specific cancers.
Computational systems biologist Richard Posner, a professor in Northern Arizona University's Department of Biological Sciences, has received a $1.4 million renewal grant from the National Institutes of Health (NIH) to continue building software tools for modeling cancer pathways responsible for aberrant signaling of growth and proliferation.
Modeling a cellular regulatory network is challenging because the number of molecular species in these networks is too large to use traditional modeling approaches. Normally, researchers need an equation for each molecular component they are tracking. But in a cellular network, the number of components is too large to be written down by hand. Our software enables a modeler to describe molecular interactions at a high level in terms of rules and this rule-based description is then automatically turned into equations by the software."
Richard Posner, Computational Systems Biologist and Professor, Department of Biological Sciences, Northern Arizona University
Posner added, "In future work, we will focus on new tools for learning model parameters from data, which is challenging because practical problems require scalable algorithms that aren't available in existing toolboxes. We're trying to provide solutions."
Posner and computational systems biologist William "Bill" Hlavacek in the Theoretical Division at Los Alamos National Laboratory are the principal investigators of the research project, "Hardening Software for Rule-based Modeling," which began in May 2014 with a $1.34 million grant from the NIH National Institute of General Medical Sciences.
The project was recently renewed through April 2024 with the additional funding, which will support enhancements of the PyBioNetFit software package and related tools that Posner and Hlavacek developed for rule-based modeling.
In rule-based modeling, the mechanistic facts about protein-protein interactions, post-translational modifications and other processes are captured as rules encoded in a formal language akin to a programming language. "Rules stipulate the context necessary for specific interactions to occur and provide a computer-readable representation of our knowledge of a system," Posner said.
The modeling tools developed through the ongoing project have been used by a number of groups in the U.S. and abroad. In recent work, Posner and Hlavacek collaborated with Oleksii Ruhklenko and Boris Kholodenko at University College Dublin to apply these tools to design synergistic combinations of kinase inhibitors that suppress signaling by mutant RAS and BRAF oncogene products.
The modeling predicted that a pair of kinase inhibitors both targeting mutant BRAF can potently suppress growth signals so long as the two drugs recognize distinct conformations of the protein, even if each drug alone is ineffective.
To advance this work, NAU assistant professor and developmental biologist Matthew Salanga is leveraging a zebrafish model for melanoma to evaluate novel kinase inhibitor combinations suggested by the modeling results.
Study shows AI can predict prognosis in triple-negative breast cancer