It is clear that the repurposing or repositioning of existing approved drugs to treat diseases of unmet medical need can be a rapid, cost effective and successful strategy. This figure, published in 2004, shows the combination of steps used in traditional de novo drug discovery compared to a repositioning approach.
In 2004, novel insights and serendipity played a key role in compound identification, suggesting a haphazard process which often relied on random clinical observations and prior knowledge. Recent technological and collaborative advances reduce the reliance on siloed information or chance and, most importantly, generate novel associations between drugs and disease.
The NCGC pharmaceutical collection, a complete non-redundant compendium of potential drugs that includes those approved for human and veterinary by regulatory agencies worldwide was made publicly available in 2011. This resource has already been used effectively to identify drugs to treat fungal meningitis, Chordoma and Chronic Lymphocytic Leukemia. Compounds from this library are screened by traditional drug discovery approaches but since they have largely been approved for human use or at least undergone some safety testing in humans, they can enter the drug development phase rapidly. Data mining using open source databases such as DrugBank, The Potential Drug Target Database, the Therapeutic Target Database, and both SuperTarget and Promiscuous can provide useful information by using different mathematical models to link the known chemical, structural, biochemical effects of drugs with diseases.
The DrDKB database developed at Stanford by Dr. Atul Butte and colleagues, is a similar database but with demonstrated effectiveness. The researchers identified 57,542 novel drug use suggestions using this database and validate their findings by checking which are already in clinical trials. They found that their drug use suggestions were 12 times more likely to be in a clinical trial than those not suggested by their database mining approach. This database has the potential to identify truly novel drug disease interactions. Dr Butte's team also mines open source gene expression data to match drugs with diseases. The ability to determine an almost complete signature of gene expression in any cell under different treatment conditions using microarrays is one of the major technological breakthroughs of this century. Analysis of this data can provide information on how gene expression levels differ under certain conditions for example in health and disease.
Analysis tools, such as Profilechaser, can be used to mine this data and to match gene expression changes that occur in disease and with gene expression changes that occur with drug treatments. Using this approach the group have demonstrated efficacy of tricyclic antidepressants in cell culture and mouse models small cell lung cancer. They have also generated data to suggest topiramate, an anti-convulsant, might be useful in the treatment of inflammatory bowel disease and validate efficacy in a chemically induced rat model of the disease. To confirm the effect in humans, another newly emerging technology, the mining of healthcare records, will be used by Dr. Seth Crocket at the University of North Carolina, to confirm whether epileptic patients receiving topiramate had fewer flare ups of inflammatory bowel disease than patients with no exposure to the drug. This approach can also identify completely novel associations between drugs and diseases.
Biovista is pioneer of systematic drug repurposing and data analysis. Their proprietary Clinical Outcomes Search Space Platform (COSS TM) analyzes data-sets of over 20 million records, comprised of scientific literature, patents, adverse event databases, chemical docking, cheminformatics and other sources to identify non-obvious correlations between drugs, targets, pathways and disease mechanisms. The company has established collaborations with major pharmaceutical companies, the FDA and several private research organizations who use the technology to support their repurposing efforts. Biovista has also established their own clinical pipeline of repositioned drugs for Neuroregenerative disease, Oncology and CNS. Another recently announced collaboration between Thomson Reuters, a world leader in pharmaceutical intelligence and Numedii (16), a company founded by Dr Butte exploit gene expression data for repurposing, will establish data mining technology to translate novel disease-drug connections into new indications. As organizations try to merge increasingly large databases of drugs, drug targets, protein interactions, off target effects, safety, gene expression patterns, patient healthcare records etc., we move into the realm of “Big Data”, the informatics buzzword of 2013.
The collaborative mining of big data to identify non-obvious drug disease associations is the future of drug repurposing.