The Food and Drug Administration has been historically cautious of biosimulation technology, but its current endorsement of the Entelos technology platform might just change the landscape for not only this biosimulation shop, but its peers in the space.
The FDA, through its Center for Drug Evaluation and Research, said this week that it has signed a two-year Cooperative Research and Development Agreement with Entelos with the goal of building a “virtual patient” to elucidate the effects of drug-induced liver injury.
Built upon the Entelos PhysioLab simulation platform, the so-called “DILI PhysioLab” will be a computer-based mechanistic model of the human liver that will guide development of preclinical assays and biomarkers to help identify drug combinations and patients with a heightened risk for liver injury.
The model will include a “cohort” of virtual patients that will represent “tolerator, adaptor, and susceptible” patient phenotypes, the company said. The model will help predict which patients are prone to hepatic injury following treatment with a specific combination of drugs.
The project is in line with the Critical Path Initiative, announced by the FDA in 2004. The FDA has identified modeling and simulation among several high-priority areas that promise to speed drug development by decade’s end [BioInform 03-24-06].
In its Critical Path “Opportunities List,” which identified 76 initial projects that it considered to be research priorities, the FDA said that in silico testing “could reduce the risk and cost of human testing by helping [drug makers] make more informed decisions on how to proceed with product testing and when to remove a product from further development."
In a written response to BioInform this week, FDA spokeswoman Susan Cruzan said that the agency chose to support Entelos on DILI modeling “after holding several scientific meetings with Entelos over the past year.”
She added, “The FDA believes that understanding and possibly being able to predict which drugs will cause drug-induced liver injury is important to public health.”
FDA officials involved in the collaboration could not be reached by press time.
Entelos CEO James Karis told BioInform that several unnamed pharmaceutical partners are funding the FDA-Entelos collaboration, but he declined to disclose the partners’ identities or the financial details of the agreement.
As for why the project has attracted the interest of both pharma and the FDA, CSO Mikhail Gishizky said, “There is a significant effort to do public-private partnerships to bring in technologies that will help [the FDA] make informed decisions … and the FDA realizes pharma participation is very important.”
Karis said this is Entelos’ first such collaboration with the FDA. “I think they have a number of initiatives under way [in this area of study] … but in terms of the history of the FDA, this is pretty novel.”
Indeed, the agency has taken several steps toward biosimulation since publishing the Critical Path document. Last June, for example, the FDA signed a CRADA with Pharsight under which the firm agreed to use its suite of pharmacokinetic/pharmacodynamic software tools to develop a data repository to support drug-disease modeling within the agency [BioInform 06-09-06].
The agreement with Entelos is a step beyond PK/PD, however, and moves the agency into the higher-risk area of mechanistic modeling and prediction.
Karis noted that the importance of hepatotoxicity testing in drug development — coupled with heightened safety concerns in the wake of recent high-profile drug withdrawals — likely convinced the agency to take the plunge.
“The FDA views DILI as a major issue, and a major reason for drug withdrawal [and subsequently] creating problems going forward,” he said.
“This is definitely part of the Critical Path Initiative,” Karis added, noting that the FDA “approached us to see if modeling could be applied to issues such as this, and how would the technology help them make better decisions?”
Colin Hill, CEO of biosimulation firm Gene Network Sciences, told BioInform that he sees the FDA-Entelos DILI collaboration as a positive step forward for the field, and “the culmination of some years of effort.”
Hill added, “Why do I think they [the FDA] put a stamp of approval on this? Ultimately, because there is a growing recognition that biosimulation methods are a necessary complement to existing data and technology for the prediction of both drug efficacy and drug safety.”
He cited the aforementioned Pharsight-FDA announcement as proof that collaborations such as those with Entelos are part of a trend.
“We see this as opening the door to even broader, more sophisticated approaches to drug evaluation that are sorely needed,” Hill said.
Even so, he has a caveat.
“Literature-based approaches at least allow for a first [step towards] evaluation of drug safety. But, so much of what drugs do is not in the literature. Approaches that are more unbiased and more data driven will have to be employed to get at the next level of predictive power and safety evaluation,” Hill said.”
Hill added that this is the role that GNS plays, coming in after the “first step.”
“We speak to Jim Karis from time to time and understand that our approaches are complementary,” Hill said.
“The FDA believes that understanding and possibly being able to predict which drugs will cause drug-induced liver injury is important to public health.”
Yet if Hill is wary, the FDA is more effusive in its support of the project.
Cruzan said that the agency hopes the Entelos partnership will lead to a “better understanding” of DILI, which she called a “growing public health problem.”
“Toxicities are typically categorized as either intrinsic — effects that are generally predictable using current tests, dose-related, and affect most individuals — or idiosyncratic — effects that are unpredictable, have no apparent dose response, and only affect a small group of susceptible people (13.9 to 24 in 10,000),” Cruzan explained via e-mail.
“It is worth noting that DILI has become the leading cause of acute liver failure in the United States, with idiosyncratic toxicity accounting for 13 percent to 17 percent of the cases, and estimated to account for 3 percent to 9 percent of all reported adverse drug reactions,” Cruzan wrote.
Virtual Patients, Real Results?
Karis, who emphasized that the DILI PhysioLab is still under development and will likely “never [be] fully finished,” is still certain of at least one thing: the model fits the problem.
As such, “the current plan is [to reap] deliverables after 12 months; significant deliverables 12 months later, but depending on partner input there could [also] be interim deliverables.”
Upon conclusion of the project, Entelos believes that the “development initiative members” will be able to use DILI PhysioLab to evaluate the effects of specific drugs on human hepatic function; identify patient phenotypes that are predisposed to potentially developing DILI when treated with a specific drug or drug combination; identify biomarkers for patients most at risk for developing DILI when treated with a specific drug or drug combination; and guide assay development for better predictors of human response.
Toward that end, the process involves “stages” in which “virtual patient cohorts” will embody five known pathophysiologic outcomes associated with DILI: hepatocellular necrosis; hepatocellular apoptosis; cholestasis; steatosis; and mitochondrial dysfunction.
Karis said that the project is focused on hepatotoxicity because it’s “one of the major reasons for drug withdrawal, [and] one of those intractable problems.” He cited Pfizer’s withdrawal of Resulin from the market in 2000 and AstraZeneca’s decision to discontinue clinical trials for Exanta in 2006 as recent examples of high-profile drug failures related to liver injury.
As such, Karis said, “the FDA has been pressured to make drugs faster, but also make them safe.”