June 20, 2005

Researchers Create Method for Predicting Most Effective Use of Cancer-fighting Drugs

UC Irvine researchers have developed a method that can help doctors choose the best combination of drugs for fighting cancer "“ a development that may lead to more effective treatment strategies.

The method is based on a mathematical model that determines when a cancer becomes drug-resistant during therapy. It does this by estimating the probability of finding drug-resistant mutant cells at different stages of cancer progression and examining the fate of individual mutant colonies during therapy. It then pinpoints under what circumstances these mutant cells become a problem and the therapy stops being effective.

The finding will help physicians determine when a combination of drugs will be more effective in fighting a cancer.

Researchers Natalia Komarova and Dominik Wodarz present their findings in this week's online edition of the Proceedings of the National Academy of Sciences.

The researchers applied their model to chronic myeloid leukemia, a cancer of the blood and bone marrow that is treated today with a bone marrow transplant, drug therapy, or both. The drug used for treating CML works by attacking a specific oncogene "“ the gene responsible for the initiation and progression of CML. The treatment is successful during the early stages of the progression of the disease. In later stages, however, the cancer develops a resistance to the drug because the oncogene mutates. Once the drug can no longer recognize the cancer cells, the treatment fails.

Using their model, the researchers concluded that mutant cells existed before treatment began and that the drug was not effectively eliminating these cells. "This is important because a good knowledge about the origin of cancer cells' resistance to drugs is the first step towards more effectively breaking this resistance," said Komarova, assistant professor of mathematics with a joint appointment in ecology and evolutionary biology, and the first author of the paper. "It means, too, that simply changing the intensity of a treatment will not prevent resistance to a drug. It also suggests that early prognostic screenings should be developed to determine the "Ëœresistance profile' of each patient, and it emphasizes the crucial importance of early start of therapy."

The researchers' mathematical model suggests that for CML, a cocktail of three drugs is needed to reduce chances of treatment failure.

Currently, the model is specific to cancer, but the general strategy also can be applied to other drug therapies, for example, therapies for infectious diseases. "Now we are working on expanding our model to make it applicable to different cancers," said Wodarz, associate professor of ecology and evolutionary biology, and the paper's co-author.

In collaboration with researchers in the United Kingdom, Komarova and Wodarz will work in the near future on more detailed models of leukemia. They also will focus on the role stem cells play in leukemia, as well as on the different mechanisms of drug resistance that are associated with the disease.


In their model, the researchers consider a population of cells. Each time a cell divides in this group, there is a small probability that a mutation occurs, generating resistant mutant genes.  Approaching the issue mathematically, the researchers pose the following question: Given a certain size of a colony of cells, what is the probability that there are resistant mutants inside this colony?

The model addresses the question in terms of probability, not certainty. While the model cannot predict at what moment a mutant is generated, it predicts the likelihood that such a mutant will be generated "“ a feature that is not a weakness of the model but captures instead a fundamental principle underlying the biological process of cancer growth and resistance generation.

On the Web:

University of California - Irvine