Paper
- Alzforum Recommends
Atri A, Frölich L, Ballard C, Tariot PN, Molinuevo JL, Boneva N, Windfeld K, Raket LL, Cummings JL. Effect of Idalopirdine as Adjunct to Cholinesterase Inhibitors on Change in Cognition in Patients With Alzheimer Disease: Three Randomized Clinical Trials. JAMA. 2018 Jan 9;319(2):130-142. PubMed.
Recommends
Please login to recommend the paper.
Comments
Certara
Despite these setbacks, I would disagree that the 5-HT6 antagonist concept is dead. Although we don’t have the details, it is likely that the clinical trial design did not take into account the pharmacodynamic (PD) interactions of 5-HT6 antagonists with co-medications, amyloid status, and some common genotype variants. As an example, using our ADAS-Cog calibrated Quantitative Systems Pharmacology (QSP) model (Roberts et al., 2012) we showed that donepezil has a negative PD interaction with 5-HT6 antagonists, yet everybody tested the compound exclusively in combination with AChE-I.
We also have implemented the PD interactions of oligomeric Aβ in AD, which not only provides an explanation for many of the failed amyloid trials, but also allows us to quantitatively assess the impact of different amyloid loads on other symptomatic treatments. The dose response of 5-HT6 antagonism, by virtue of its indirect action on glutamate and GABA neurotransmission, is likely influenced by the amyloid status of the individual patient. All these insights could help tremendously to improve clinical trial design, especially in terms of allocating subjects to the different treatment arms or selecting the right patients.
However, as long as clinical trial design is performed without taking into account these interactions, we are counting on sheer luck to get to a positive outcome, but the odds are massively stacked against us.
In order to “validate” these predictive QSP outcomes, it would be great to have access to the individual patient characteristics in these failed clinical trials. I think the patients deserve it, as only then can we learn from our mistakes.
References:
Roberts PD, Spiros A, Geerts H. Simulations of symptomatic treatments for Alzheimer's disease: computational analysis of pathology and mechanisms of drug action. Alzheimers Res Ther. 2012 Nov 26;4(6):50. PubMed.
View all comments by Hugo GeertsCritical Path Institute
Thank you for posting this important feature, Alzforum colleagues.
I was one of many drug hunters 10 to 20 years ago, chasing lead molecules as selective and potent as we could find, and testing these leads in animal models to demonstrate their effect on cognitive performance, immediate early gene expression, neuronal survival, and neurogenesis. We all hoped that the effects we were observing in animals would hold up when tested in people suffering from disorders of cognitive impairment. In later years, promising new PET ligands were developed to aid in defining receptor occupancy and helping to choose doses for the clinic.
Yet, fast forward to 2018, it's devastating to see that it has taken 20 years, many different companies, more than 600 published studies focused on these receptors, thousands of patients, and numerous trials to generate confidence that this mechanism has been tested adequately. Each of these molecules clearly has distinct properties that make it challenging to draw the true reasons for the negative outcomes. It's critical for the field to look not just at one trial at a time and to apply learnings from these failures. CAMD's Alzheimer's disease unified clinical trial database, consisting of patient-level data from 24 different trials (>6,500 patients), serves the purpose of collective learning across a multitude of trials (Neville et al., 2015).
I would also like to congratulate Pfizer for its application of Critical Path Institute's CAMD AD drug disease trial model as a tool to render a “no-go” decision on its leading 5HT6 antagonist, SAM-531. The authors, from Pfizer, Metrum, CAMD, reported on their AAIC poster:
Methods: “A disease progression model describing the change in ADASCog over time under a variety of covariate settings had recently been developed on behalf of the CAMD initiative and was used in the analysis and interpretation of the SAM-531 interim data.”
Results: "Use of this model-based comparator suggests a positive dose-response for SAM 531, albeit not with a magnitude of effect sufficient to support further development of the compound at this time." (Rogers et al., Model-based analysis to support strategic decision making: A case study from the development of a 5HT6 antagonist for the treatment of Alzheimer’s disease. ICAD 8(4): 585; 2012)
This application of the model in 2012 preceded regulatory endorsement of the AD clinical trial simulation tool in 2013 by both FDA and EMA.
Companies sometimes convey the importance of “failing fast” as a measure of success, yet such decisions are often invisible. They should be recognized. Model-informed drug development is a powerful tool and will be a key focus of FDA's goals in the Prescription Drug User Fee Act (PDUFAVI).
Many in our field have commented that a main reason for the failures in AD is overemphasis of the amyloid hypothesis. Yet in this case, negative results from non-amyloid targets are continuing at a pace that is unsustainable. My views align with Dr. Bennett’s editorial words: “The lack of progress in the treatment and prevention of Alzheimer disease is frustrating for patients, families, physicians, researchers, industry, funders, and policy makers. Understanding the causes for these failures is essential for informing future trials.”
View all comments by Diane StephensonMake a Comment
To make a comment you must login or register.