Drug development in Alzheimer's disease (AD) requires a considerable investment of time and re- sources, often with little reward as the vast majority of medications ultimately prove unsuccessful. Drug repur- posing, in which medications that already have been approved for treatment are evaluated for therapeutic effects in other disorders, has the potential to markedly increase the number of agents in the drug development pipeline but requires methods for effective screening of candidate medications for activity. In silico or computational ap- proaches to medication screening are rapidly growing, and have been successful in illnesses such as cancer, but their application to AD remains understudied. There is also intense interest in drug repurposing approaches that will utilize the vast amounts of clinical data that are being collected from epidemiological studies and clinical encounters documented through electronic health records (EHRs). In this proposal, we present a novel approach to drug repurposing that uses large-scale data mining (i.e., pattern recognition) algorithms applied to concurrent medication taken by participants in AD clinical trials and in Medicare administrative data to determine which of these medications show potential therapeutic bene?ts. With over 30 years of AD clinical trial data available to us through a recently developed meta-database and 10 years of prescription data available through Medicare Part D, the administration of concurrent medications to patients as part of their routine clinical care constitutes a large- scale natural experiment. This information can be harnessed for AD treatment discovery if appropriate methods can be developed to detect effects on disease progression within this high-dimensional data. Data mining al- gorithms that discover patterns of associations in data, rather than testing predetermined hypotheses, are well suited to application in large-scale screening for drug repurposing. Using our meta-database and Medicare data, we will be able to evaluate most of the more than 6,000 currently available prescription medications for ef?cacy in AD using well-accepted endpoints for measuring disease progression. The discovery phase will be followed by a validation phase of promising candidate medications in independent data sets, as well as identi?cation of plausible gene targets for each medication from the biomedical literature. This study will set the groundwork for a series of follow-up in vivo studies to conclusively demonstrate effects of selected medications for AD, expanding the current armamentarium for treating this common and debilitating disorder.
In this proposal, we will explore drug repurposing, in which medications approved to treat other illnesses are evaluated for potential bene?ts in Alzheimer's disease (AD), to improve the ef?ciency of the drug development process, which currently ends in failure for most candidate medications despite considerable investment of both time and resources. We will apply data mining, or pattern recognition, algorithms to the concomitant medications taken by subjects in previous AD clinical trials that are part of our meta-database of more than 6,500 individuals and by participants in Medicare Part D with more than 787,000 individuals. This approach will allow us to screen hundreds of drugs for effects in slowing the progression of AD, which will guide future clinical trials of novel therapeutic agents.
| Babulal, Ganesh M; Quiroz, Yakeel T; Albensi, Benedict C et al. (2018) Perspectives on ethnic and racial disparities in Alzheimer's disease and related dementias: Update and areas of immediate need. Alzheimers Dement : |
