The technical innovation shows, using Physiomics' SystemCell software, complex virtual cells "growing" in the computer to mimic the experimental behaviour of many cancer cells and their responses to drugs. This important development paves the way for realistic and predictive models of tumours' drug responses. It also meets the criteria laid down by Physiomics' Remuneration Committee on key technical criteria such that 25 percent of the unapproved options granted on 26 June 2005 are now available for exercise.
A better predictive system for testing cancer therapeutics is expected to be enormously valuable to pharmaceutical companies and could confer significant patient benefit through better treatment. Currently, only 5 percent of all cancer drugs that are tested in clinical trials are approved for patient use, half the industry norm ("Can the pharmaceutical industry reduce attrition rates?" - Kola, I. and Landis, J. Nature Reviews DrugDiscovery, 3 (2004) 711-715).
The new work on cell growth examines why small genetic differences between cancer cells mean that some die when treated whereas others survive and grow. This is not obvious from standard cell biology work, and it opens up the possibility of optimising treatment using appropriate drug combinations depending on patient profiles.
The data used to develop this model is from the scientific literature and focuses specifically on an aurora kinase inhibitor drug candidate, VX-680, developed by Vertex Pharmaceuticals Inc. and licensed to Merck & Co. Inc. Published data on an aurora kinase inhibitor being developed by AstraZeneca has also been used. Aurora kinase inhibitors are a new type of therapeutic that has shown excellent cancer killing ability in preclinical studies. Some 16 aurora kinase inhibitors are under development by companies such as Astex Therapeutics Ltd., Chroma Therapeutics, Cyclacel Ltd., Millennium Pharmaceuticals Inc., Rigel Pharmaceuticals Inc. and Serenex, in addition to Merck & Co and Vertex.
Aurora kinases are key enzymes involved in the later stages of cell division, particularly the formation and control of the machinery that enables cells to divide. When applied to cancer cells, inhibitors of these enzymes prevent the cells from dividing properly. This can lead in some tumour types to cell growth being halted quickly. In other types of cancer cells, growth continues for more cycles before the cells start to die. Understanding the differences between these aspects has been problematic.
A key part of Physiomics' work in this area has been reproducing two biomarkers of drug action. These are the extent of aurora kinase modification of a protein used to package DNA (histone H3), and a measure of the number of gene copies that a cell has (polyploidy). In addition, the model starts to explain why some cell types respond well to the drug and others do not. In the next stage of the development, different drug doses and schedules plus combinations with existing therapeutics could be explored.
Physiomics believes that computer simulations could help to optimise the transition from preclinical to clinical development and help in the optimisation of dose scheduling, one of the major problems in clinical trials. Studies (Tufts Center Impact Report 2002) have shown that improving the drug development success rate by 10 percent overall could save $242 million per drug. Physiomics and Bayer Technology Services collaborate on Clinical Response Prediction in the area of cancer.
"For the pharmaceutical industry the virtual approach could represent huge savings in time and money by predicting the effective drug dose and schedule for different cell types. This currently is done step by step in the laboratory, or in inflexible clinical trials", stated Dr. John Savin, CEO of Physiomics. "Our technology is also very relevant for working out how best to use the new generation of targeted therapeutics and for reducing their clinical development risks."
Using the Physiomics advanced cell cycle model and SystemCell software, populations of virtual cells have been "grown" in a computer and a virtual drug that targets aurora kinase administered. The virtual cell population quantitatively reproduced the observed experimental behaviour published in the literature.
One type of responsive cell stopped growing after drug administration due to the action of the drug on the individual cells in the population model. Another cell type with a common cancer-causing alteration (reduction of p53 levels) kept on growing with two, four and eight times the normal amount of DNA. Cells with higher levels of DNA then started to experience cell death.
In both cases, the virtual experiment reproduced results acquired over five days of actual experimental time in under an hour of computational time. This indicates that the model should have potential in predicting the effect of human drug doses, as well as the activity of new drugs and drug combinations.
Physiomics plc is a computational systems biology services company applying simulations of cell behaviour to drug development to reduce the high attrition rates of clinical trials. As 80-90 percent of all clinical drug candidates fail to reach the market, estimates show that an overall ten percent improvement in success rates could reduce the cost of one drug's development by as much as $242 million, from the current estimate of around $800 million.
Physiomics is currently focused on optimising the design of cancer clinical trials as a service to major pharmaceutical companies. In addition, it has secured an option to in-license two innovative molecules in the cancer area from Cronos Therapeutics. Physiomics has a major collaboration with Bayer Technology Services GmbH to develop globally clinical response prediction services using both companies' proprietary technologies and mathematical cell models, including the SystemCell technology developed by Physiomics. In June 2005, Physiomics became a distributor of Bayer's PKSim product for understanding the way pharmaceuticals are distributed round the body. Physiomics, based in Oxford, United Kingdom, was founded in 2001.