Wyeth in June launched an internal R&D protocol that it said will rely more on biomarker and other pharmacogenomic data to improve its drug-discovery and -development programs.
The new system, which Wyeth calls Learn and Confirm, will result in broad management changes, alter the way the firm populates clinical trials, and modify how its R&D teams interact. In the end the protocol may result in more pharmacogenomic-based therapies.
The new format will also engender changes in data collection, data access, clinical trial decision making, patient recruiting, and how researchers ask scientific questions. It will also take advantage of the US Food and Drug Administration’s recent willingness to allow drug makers to employ adaptive clinical trial designs.
“As things come through [the pipeline], the question is, ‘Are you considering pharmacogenomics and translational medicine, and if not, why not?’” Matt Bell, head of Wyeth’s Learn and Confirm implementation office, told Pharmacogenomics Reporter this week.
Specifically, Wyeth expects that in two years the strategy will enable 75 percent of its drug program to have some kind of pharmacogenomic component, such as biomarker data stratifying trial populations, according to a company spokesperson. The company does not have a record of how many compounds currently use these strategies, he added.
The first component of the model — the Learn phase — is where Wyeth will rely most heavily on biomarkers and pharmacogenomic approaches. Through the FDA’s recognition of adaptive clinical trial designs, the Learn phase will replace the traditional clinical trial phases 0, 1, and 2 and focus on identifying patients who are best suited for clinical trials, understanding a drug’s medical value, and identifying possible safety risks.
The second component — Confirm — closely resembles phase 3 studies and relies less on biomarkers or other pharmacogenomics components.
Wyeth said it also hopes it will help it push “more compounds through [the pipeline], inform better decisions, [and] kill compounds early [when necessary],” said Bell.
“Where compounds have any glint of activity, we’re doing everything we can to identify that glint of activity, and applying our resources to it, so that we don’t kill things that we shouldn’t be killing,” he said.
That means a lot of reliance on biomarkers. As part of Learn and Confirm, Wyeth drug-development teams are encouraged to examine each compound coming through the pipeline for ways that the company can employ a biomarker-based or pharmacogenomic approach that can help move it downstream.
Compounds already in the pipeline will have some changes made to their development procedures, but it might take five to seven years for drugs to reach the market that have benefited from the full Learn and Confirm process, said Bell. According to industry data that timeframe on average currently stands at between eight and 10 years.
The most biomarker-dependent part of the Learn phase is Wyeth’s plan to use smaller or more informative patient populations in clinical trials. “It’s really understanding in which population the drug works, and using that as a tool to drive drug development, as opposed to doing these broad-based indication studies that we’ve done historically,” said Bell.
To help it define patient populations Wyeth will use prognostic indicators such as genomic as well as clinical factors, said Bell. “There’s been a slow move in industry towards that, but we’re really embracing it in Learn to really make that part of how we make decisions,” he said.
The Learn component also involves several other changes meant to increase drug-discovery efficiency, such as grouping compounds under development teams that specialize in a particular disease or indication.
“As things come through [the pipeline], the question is, ‘Are you considering pharmacogenomics and translational medicine, and if not, why not?’”
Wyeth currently focuses on six main therapeutic areas: women's health and musculoskeletal disorders; cardiovascular disease and metabolic disorders; oncology; neuroscience; inflammation; and vaccines.
Implementing Learn and Confirm also “involve[s] an awful amount of changes in management and changes in processes in how we do business,” said Bell.
On the clinical operation level, Wyeth will be making changes in data collection, data access, clinical trial decision making, patient recruiting, and how researchers ask scientific questions, he added.
Learn and Confirm also seeks to take advantage of the FDA’s increasing willingness to allow drug developers to employ adaptive clinical trial designs. For example, the agency further encouraged adaptive designs in May with a guidance document outlining Bayesian statistical analysis for trials.
Asked whether the Learn and Confirm strategy could lead to more pharmacogenomic therapies from Wyeth, Bell said, “The model will potentially lead you to much more specific therapies, and the logical extension of that is that you’re using diagnostic products to help you with that. I don’t think that Wyeth is about to become a diagnostics company, but I do think that diagnostic companions will become an increasingly important part of what we do.”
That task will fall largely on a group of Scottish researchers. In April Wyeth announced a five-year, $57 million collaboration with four Scottish Universities and business incubator Scottish Enterprise that figures heavily in Wyeth’s diagnostics plans.
The partnership created a 50-employee central research lab at the University of Dundee to coordinate drug and diagnostic research with labs at the Universities of Aberdeen, Dundee, Edinburgh, and Glasgow, as well as National Health Service Scotland.
“The therapeutic intellectual property is really something that Wyeth can drive, commercialize, and make the best of,” Bell said. “The Scottish parties really see an opportunity in taking the diagnostic IP and driving that. So we already have a mechanism to bring diagnostics in earlier in the process than we used to,” although Wyeth probably will not be entering the diagnostics field itself, he added.
The research in Scotland, which is centered on translational medicine and biomarkers, as well as figuring out how to use “genomics and other tools to understand how therapeutics are working, which patients to apply those compounds to, and how to move compounds forward,” said Bell.
The company began putting the Learn and Confirm model in place in February as a phase of its Project Springboard restructuring plan begun in 2005 by CEO Robert Essner.
The model is based on concepts introduced by L.B. Schreiner in an article entitled “Learning Versus Confirming in Clinical Drug Development,” which was published in the March, 1997 issue of Clinical Pharmacology and Therapeutics.