Drug-Target Database and Algorithm Help Match Old Drugs to New Uses

There are a large number of drugs that silence many thousands of cancer-causing genetic abnormalities. Some of these drugs are in use now, but many of these drugs are not being used or could be used beyond the disease for which they were first approved. Repurposing these drugs depends on matching drugs to targets. A study published May 7, 2014, in the journal Bioinformatics reported on a new database and pattern-matching algorithm that allows researchers to assess reasonable drugs and drug combinations, and also recommends a new drug combination to treat drug-resistant non-small-cell lung cancer.

“Most cancers have more than one genetic alternation. And even genetically targeted drugs tend to affect more than only their stated target. And so the challenge is matching drugs with many effects to cancers with many causes in a way that best maps the drugs’ effects onto the intended targets,” stated Aik Choon Tan, PhD, an investigator at the University of Colorado (CU) Cancer Center (Denver, USA) and associate professor of bioinformatics at the CU School of Medicine.

There are approximately 500 kinases in the human genome, each of which represents a potentially important drug target. Dr. Tan portrayed the database as a spreadsheet with 500 columns, each column representing a kinase. Heading each row is a drug and then in each column cell is that drug’s activity against the kinase. “Imagine you know a cancer is caused by five kinases acting in unison,” Dr. Tan noted. “Our approach would allow you to query the database for this pattern and discover the drug or combination of drugs that best match the genetic needs.”

Because many of these drugs have already earned US Food and Drug Administration (FDA) approval for use in other diseases, the processes of repositioning these drugs for new diseases is much less involved and costly than if drug developers had started fresh. Dr. Tan and colleagues put the technique to use to recommend drugs that could turn off the kinases that non-small-cell lung cancer uses to create resistance to existing treatments. It has been a key conundrum—many lung tumors depend on over-activation of the gene EGFR (epidermal growth factor receptor), but then when EGFR inhibitors such as gefitinib or erlotinib are used, the cancers tend to trigger other “kinases” that allow the cancer to bypass around this dependence. Dr. Tan and colleagues explored what exactly are these kinases that allow lung cancer to evade gefitinib, and what other drug might inactivate them.

The solution may be in the drug bosutinib, developed by Pfizer (New York, NY, USA), which earned FDA approval in 2013 for the treatment of chronic myeloid leukemia. The drug out-competes the body’s energy source, adenotriphosphate (ATP), for space in kinases and so keeps them from being activated. Furthermore, bosutinib may suppress the activity of exactly the kinases that EGFR-dependent lung cancers require to mutate around the challenge of EGFR inhibitors.

In research on EGFR-dependent lung cancer cell lines, Dr. Tan and colleagues show that the drugs gefitinib and bosutinib “showed additive and synergistic effects.”
In a mechanism that Dr. Tan hopes will become common, his group will now provide data about this rational combination to other researchers at the CU Cancer Center and elsewhere who will help move the drugs toward a human clinical trial.

The K-Map database free for use and it is availble online (please see Related Links below).

Related Links:
University of Colorado Cancer Center
K-Map database