P1‐307: Failure analysis of dimebon using mechanistic disease modeling: Lessons for clinical development of new Alzheimer's disease therapies
Hugo Geerts, Äthan Spiros, Patrick Roberts · Alzheimer s & Dementia · 2012
Dimebon (latrepirdine) is a putative mitochondrial stabilizing compound that showed great promise in an early clinical Alzheimer's diseases (AD) trial, but failed to show conclusive benefit in subsequent clinical trials. Similar to other industries, we applied failure analysis using a sophisticated quantitative systems pharmacology approach to identify possible reasons for the clinical failure and provide suggestions to improve clinical trial design. We developed a complex computer-based cortical network model with implementation of the physiology of 12 different membrane CNS targets based upon a biophysically realistic multi-compartment model of 80 pyramidal cells and 40 interneurons. The model was calibrated using data reported for working memory tasks in healthy humans and schizophrenia patients. Alzheimer's disease (AD) pathology was introduced using realistic values for synapse and neuronal cell loss and cholinergic deficits that are calibrated to align with the natural clinical disease progression. The neurophysiology of 5-HT6R effect was calibrated using the reported clinical data on SB742457. Dimebon's pharmacology on various neuromodulatory receptors was implemented using published data. The clinical trials were simulated and sensitivity analysis was performed to identify possible drivers of the observed outcome. We identified dimebon's dopamine D1R effect as a potential liability for reduction of the functional cognitive signal in ADAS-Cog. The size of this liability is mediated by the COMT Val108Met genotype. By subtracting the ‘liability’ pharmacology, a ‘hypothetical’ compound profile was found that showed a robust beneficial effect on cognitive performance. This case-study shows the capability of a quantitative systems pharmacology approach to extract knowledge about the underlying neurobiology from a retrospective analysis of clinical trial results. The results suggest that independent of the primary disease-modifying pharmacology, great care need to be taken about off-target effects at neuromodulatory receptors that might affect the functional outcomes probed by the ADAS-Cog clinical scale. It is suggested that by applying failure analysis, common in other industries, the pharmaceutical industry would be able to identify possible issues in clinical trial design or drug selection that might improve the success of subsequent clinical trials.