PHILADELPHIA — Without a strong efficacy biomarker in hand Novartis will not advance an investigational drug candidate into clinical development.
“One can say that, except for the very rare instance, Novartis’ approach is that [unless] there is a drug with a really solid biomarker attached to it, we don’t develop it,” Robert Schmouder, executive director of translational medicine at Novartis, said at a Cambridge Healthtech Institute conference on translational medicine, held here last week.
For the past two years, the Swiss pharmaceutical giant has been moving away from the traditional three-phase drug development paradigm and toward what it calls a “learn-and-confirm” model.
Schmouder said pharma must do a better job of establishing the safety of drug candidates in the pre-clinical phase. At Novartis, this strategy employs modest-sized proof-of-concept studies in humans before Phase I research starts to give the company a better sense of the investigational agent’s efficacy and safety performance in large-scale clinical development.
Novartis hopes the strategy will decrease attrition rates through smaller clinical trials in targeted populations, and enable it to exploit biomarkers with broad utility across multiple indications.
To support the model, the company has already made broad changes in management and in the way its R&D teams interact. [see PGx Reporter 03-07-07].
Novartis’ strategy comes at a time when big drug makers, faced with mounting drug-development costs and high-profile drug failures, streamline their R&D operations and embrace molecular tools and adaptive clinical trial designs in early drug development.
Novartis peers Bristol-Myers Squibb, Wyeth, and Roche, among others, have similarly changed their R&D strategies and pharmacogenomic outlook.
In May 2005, Novartis announced that it was “redefining research-to-development transition through fast and rigorous proof-of-concept trials.” While the new model doesn’t guarantee approval from the US Food and Drug Administration, it is meant to give Novartis more confidence in the candidates it chooses to submit to the agency, Schmouder said.
Novartis’ new research strategy involves integrating R&D in the exploratory period, conducting small studies in a targeted population early on in development to characterize a drug candidate’s therapeutic potential, and doing studies targeted at mechanisms that test other hypotheses.
“We have this learning, or exploratory phase, and we see these proof-of-concept studies almost as a gate to move out into subsets of patients where we can better describe the actual drug activity in the patients that have the disease,” Schmouder said. “Once that concept has been proven … we can move into a confirmatory phase where large-scale efficacy trials can be conducted to prove the drug is reasonably safe and effective.”
Novartis recently conducted a successful proof-of-concept study for ACZ-885, an antibody that targets the IL-1 beta signal to treat a rare inflammatory condition called Muckle-Wells syndrome.
The company plans to use the learn-and-confirm strategy to identify biological targets with broad applicability in multiple disease markets. According to Schmouder, the IL-1 beta signal may be meaningful for certain populations in rheumatoid arthritis, asthma, type-1 diabetes, and psoriasis.
Schmouder emphasized that the use of efficacy biomarkers is the modus operandi for drug development at the company.
“On the activity side … the research people will have to provide, almost by decree … biomarkers that we know are robust, that can go into the clinic and have very good read-outs,” Schmouder said.
“We’re just trying to apply a reasonable scientific method to drug discovery, [so we can] hopefully reap some of the rewards of that and move into full development with some reasonable assuredness of success,” he added.
Novartis is not alone in incorporating biomarkers into drug development and embracing the learn-and-confirm model. Encouraged by a willingness at the FDA to accept adaptive clinical trial designs, Novartis is following a larger shift within pharma toward more predictive drug-development strategies.
For instance, Bristol-Myers Squibb uses biomarkers and pharmacogenomics to expand the indications for existing oncologics [see PGx Reporter 01-10-07].
Also, Wyeth instated a learn-and-confirm model of its own last year, hoping that in two years the strategy will enable 75 percent of its drug program to have some kind of pharmacogenomic component.
The company has said it hopes learn-and-confirm will help push “more compounds through [the pipeline], inform better decisions, [and] kill compounds early [when necessary]” [see PGx Reporter 08-09-06].
Similarly, Roche announced in January it was introducing a new model for its global R&D activities organized around disease biology areas that will allow the company to be more “efficient and effective in translating research activity in each therapeutic area into clinically differentiated medicines.”
Under this new model, “cooperation between the pharmaceuticals and diagnostics divisions in the area of personalized medicine will be further strengthened,” the company said in a statement.
Novartis’ learn-and-confirm strategy has brought about personnel and organizational changes that the company is still getting used to.
At the company, exploratory and confirmatory studies are managed by separate divisions. The head of exploratory clinical development is Trevor Mundel and confirmatory studies are led by James Shannon, head of pharma development.
“I think we are getting better and better at being able to predict the drug activity or potential efficacy of the drug. But how do we know what the safety is of the drug as we move into the confirmatory phase?”
“We see these two [functions] as so different that it would be difficult to have them as one department and still get the maximum bang for the buck,” Schmouder said. However, he added that “this [separation] does create some tensions between those two groups.”
The company is attempting to ameliorate turf battles by incorporating its clinical-development colleagues early in the exploratory research phase.
“Even in the first assessments of any kind of proof-of-concept studies we will do, the clinical-development people will be there in the room doing a kind of reality test,” said Schmouder. “They are saying: ‘This proof-of-concept study may be fine to show this drug is active, but it doesn’t help us move this drug into clinical development as quickly as we like. Can we do something else, like add an extra arm or two that will allow us to understand safety at a higher dose?’”
Additionally, Novartis has identified a breed of “physician-scientists” who “can speak the language of the basic scientist but they can also speak the language of the clinician.”
These physician-scientists can act as a bridge between research and early development, and “hopefully draw out the maximum benefit from that interface,” he said.
Safety Biomarkers Need Improvement
“I think we are getting better and better at being able to predict the drug activity or potential efficacy of the drug,” Schmouder said at the conference. “But how do we know what the safety is of the drug as we move into the confirmatory phase?
“Even if the drug does work, a lot of drugs fail during Phase III or in the confirmatory phase due to safety issues.”
Schmouder said that the use of safety biomarkers is particularly necessary in the development of small-molecule drugs. “With monoclonal antibodies it’s not as difficult, but with small molecules I think it’s tricky,” he said.
Under Novartis’ learn-and-confirm system, all first-in-human and proof-of-concept studies are reviewed by an integrated safety assessment board of several preclinical members and translational medicine physicians. “We go over the protocol and approach with a fine-tooth comb for safety,” Schmouder said.
Novartis is “working hard to move as much of that potential safety signal into preclinical studies” and gauge “what type of biomarker studies we may be able [to do] that would give us an insight into safety,” he said.
“But this is an area that’s still a work in progress,” he said. “There’s a long way to go to move forward with a reasonable assuredness of safety.”