Scientists created a new computational method that could make the discovery of new drugs for cancer and other diseases more efficient. By computational analysis alone, the researchers successfully identified four drugs that could be candidates for the treatment of hepatocellular carcinoma, a type of liver cancer for which no effective therapy exists. Of the four, pyrvinium pamoate, a drug traditionally used to treat pinworms, was shown in to be the most effective against hepatocellular carcinoma in mouse models, warranting further clinical development. Study results were published in Nature Communications on July 12, 2017.
The computational method involved mining immense amounts of data from four open-access databases. The goal was to identify drug candidates that could reverse cancer gene expression.
First, to identify cancer gene signatures, the research team probed The Cancer Genome Atlas for data and compared available gene signatures for tumors with adjacent healthy tissue (from human tissue samples). They then searched the Library of Integrated Network-based Cellular Signatures L1000 data set to determine the effect compounds had on the gene expression of those signatures. Next, they mined the database ChEMBL for drug-efficacy data. Lastly, the team probed the Cancer Cell Line Encyclopedia to map cell lines in the other databases.
The team then quantified the likelihood that each identified drug candidate could reverse cancer gene expression; this measurement was called the Reverse Gene Expression Score. The drugs that were determined to be the most likely to reverse cancer gene expression were then tested in mouse models, and pyrvinium pamoate was deemed to be most effective to shrink tumors.
Screening for drug candidates using this systematic computational approach is faster and less expensive than the traditional method of running biological experiments, according to the study authors.
The current application of this computational model is to screen compounds in preclinical studies, explained lead study author Bin Chen, Ph.D., assistant professor with the Institute for Computational Health Sciences and the Department of Pediatrics at the University of California, San Francisco. In the future, he predicts this method could be used in clinical trials to match patients to a targeted therapy. His hope is that this approach could also be applied to other molecular data, such as proteomics.
Now the goal is to move pyrvinium pamoate and drug candidates for other diseases to clinical trials. However, doing so has proven difficult because of lack of funding.
Dr. Chen said the National Institutes of Health usually have “very limited” funds for early stage clinical trials. Also, pharmaceutical companies usually have “very limited” interest in developing repurposed drugs, particularly in the case of those that are no longer protected by a patent.
“Most of the repurposed drugs we’ve been looking at happen to also be off patent,” said senior coauthor of the study Atul Butte, M.D., Ph.D., director of the Institute for Computational Health Sciences at the University of California, San Francisco. “So these are extra challenging [in terms of finding] commercial interest.”
“I definitely do believe high rewards are possible for some of these molecules, but it’s going to take commercial creativity to figure out how to regain value on some of these compounds, at least enough to justify the costs of clinical trials.”