EVENTS | VIEW CALENDAR
Modeling drug toxicity
FOSTER CITY, Calif.—Late-stage drug candidate failures and post-market recalls represent perhaps the largest costs associated with the development of new drugs, and most often, these events are triggered by unforeseen toxicity issues. Thus, the ability to identify potential toxic impacts of new molecular entities early in the discovery process will save not only dollars, but also lives.
With this in mind, biological simulations specialist Entelos announced it entered into a CRADA with the U.S. Center for Drug Evaluation and Research (CDER) to develop models of drug-induced liver injury. Working with CDER researchers, and with input from pharmaceutical industry partners, Entelos will develop a platform that will help organizations identify clinical biomarkers and develop preclinical assays for patient populations and drug combinations with increased risk for liver toxicity.
"Most of the work we've done in the past is around the question of efficacy," says Entelos CEO James Karis. "And for a long time, we've had some ideas about how do we use our fundamental modeling capability to apply it to some of the questions around toxicity. A few years ago, we started a conversation with the FDA, which was precipitated by their Critical Path Initiative, about how we could apply this technology to the really tough problem they and the industry have of drug-induced liver injury."
"One of the key things that we discussed with the FDA is how to integrate and understand all the information and knowledge that already exists in a way that allows better prediction and better understanding of how future drugs will act," adds Entelos CSO Dr. Mikhail Gizhisky.
The goal, Gizhisky explains, is to develop a platform of a human liver that represents the specific cells and processes, and then focus on drugs shown to have liver injury issues to calibrate the model and then validate it. This platform would then allow researchers to look at novel drugs and activities in a manner that is more predictive of human response than what is currently available.
As with Entelos's previous biosimulation efforts, such as its diabetes model, experimental data—and lots of it—will be key.
"The best in-silico models are based on experimental results that are closely connected to the compounds of interest," says Jack Gardner, pharmaceutical analyst with Kalorama Information, who recently authored the report Early Toxicology: Markets and Approaches. "New software allows researchers to input assay data to modify the model."
Gizhisky is quick to point out, however, that neither the FDA nor Entelos's pharmaceutical collaborators will be providing data for the project. Rather, the FDA will provide expertise in understanding of toxicity issues and will actively participate in the development of the platform.
"The FDA can certainly facilitate us getting access to data, but the data itself won't be coming from the FDA," he says.
"There is a real reluctance for [pharmaceutical companies] to share data with competitors," Gardner says. "But this is being overcome by the need to pool information to secure better results. Expect sharing to expand." The Entelos-FDA CRADA may be just such a project.
Looking farther downstream, Karis suggests his company is hoping there will be sufficient interest in the platform that companies will either license it for internal use or establish a service agreement with Entelos to push their own R&D priorities.
"From our pharmaceutical partners' perspective, the fact that drugs that are known to have benefit for a clear number of patients are being withdrawn from use because some small subset is affected adversely really hinders our ability to impact human health," Gizhisky says. "If there were a more definitive way of excluding those patients at risk, clearly there could be medicines provided for an unmet medical need."