Clinical Trial Targets Miss the Mark for Endophenotypes Associated with Elevated Disease Risk Metrics
Gregory A. Cary, Jesse C Wiley, Anna K Greenwood, Gregory W Carter, The Emory‐Sage‐SGC TREAT‐AD Center- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
Most of the candidate disease modifying agents in clinical trials for Alzheimer’s Disease (AD) have been unsuccessful in providing significant relief to patients. While there are myriad possible reasons for the poor success rate of these trials, one potential explanation is that the targeted proteins and pathways are not the causal mechanisms of disease. The TREAT‐AD Consortium has recently developed tools to score the entire genome across several dimensions of disease risk, as well as organize genes into AD relevant biological domains of gene function. In this work we examine the historical data of AD clinical trials through the lens of AD risk and biological domains.
Method
Clinical trial data was obtained from the OpenTargets and clinicaltrials.gov platforms. For the target(s) identified for each trialed agent, we obtained the TREAT‐AD Target Risk Scores (TRS) and AD biological domain annotations. The results of gene ontology (GO) term over‐representation tests for the 253 genes targeted by agents in clinical trials were compared to the results from the top scoring genes ranked by the TRS.
Result
We find that genes targeted by agents that have entered phase IV trials tend to have a higher genetics risk score compared with genes targeted by therapeutic agents that are earlier in the clinical trial pipeline (i.e. phase I‐III). There is a mismatch between the biological domain annotations of genes targeted by clinical trials and those with the top TRS. Genes targeted by clinical trials tend to have a relatively higher fraction of annotations and are strongly enriched for terms in the Synapse, Vasculature, and APP Metabolism biological domains. Genes with elevated measures of disease risk tend to have relatively higher fraction of annotations and term enrichments in the Mitochondrial Metabolism, Proteostasis, and Endolysosome biological domains.
Conclusion
We find evidence of mismatches between the biological domains that are elevated for AD risk and those that are targeted by therapeutic agents in clinical trials. Our unbiased, genome‐scale analysis suggests that therapeutic development efforts should increase focus on the intersections of Mitochondrial Metabolism and Proteostasis, and Endolysosomal function.