The US Environmental Protection Agency this week presented for the first time initial data from the first phase of its ToxCast program during the Society of Toxicology's annual meeting, held this year in Baltimore, according to an official from a company who participated in the program.
In addition, the EPA has exercised the option of cell-based assay shop Acea Biosciences to participate in the ToxCast program for another year per Acea's contract with the EPA, a company official told CBA News this week.
In all, eight companies, including CBA shops Acea, Cellumen, and BioSeek, were invited to participate in ToxCast. All will participate in phase 2, "we just have not issued any task orders for phase 2 yet," Robert Kavlock, director of the EPA's National Center for Computational Toxicology, told CBA News.
"My understanding is that Acea will participate in phase 2, and my latest info from the EPA is that our participation should be authorized by the end of the summer," Yama Abassi, senior director of cell biology at Acea, said this week.
The EPA and Acea are also collaborating on a paper designed to report Acea's ToxCast data. "My understanding is that the paper has been submitted, and we should hear back fairly soon about when and where in what journal it should be published," said Abassi.
Kavlock told CBA News that the agency is actually preparing manuscripts with all its ToxCast assay sources.
He added that the FDA is also preparing a meta-analysis paper that will take the data from the biochemical, cell-based, and high-throughput assays in ToxCast, combine them with chemical properties of the compounds and "structure-activity relationship approaches," and determine what combinations of endpoints are predictive of the in vivo data that the agency already has. The paper should be ready for submission within a month, said Kavlock.
In April 2007 Acea won a contract worth $3.9 million to perform microelectronic cell monitoring under the ToxCast program (see CBA News, 4/17/07). This amount is the predicted maximum for the life of the contract, which will be for at least two years but up to five years, the agency said.
During the first phase of the program, begun in 2007, Acea used its xCELLigence system, co-developed with Roche, to observe the dose-response pattern of mammalian cells to hazardous compounds provided by the EPA (see CBA News, 11/16/07).
"We obtained profiles for different compounds, and we actually matched profiles potentially to the compounds' mechanism of action," said Abassi. So if, for example, two compounds have the same mechanisms of action, they would produce the same profile on the xCELLigence system.
Such data could be valuable to the EPA because it could enable the agency immediately to see which two compounds have similar mechanisms of action and, potentially, similar kinds of toxicity in mammalian cells and, potentially, in organisms.
"So what the EPA has done is use the profiles generated by the xCELLigence system as a way to classify the compounds based on mechanism of action," said Abassi.
The agency's goal is to build a model that could hopefully predict toxicity in animals and, ultimately, in humans.
[ pagebreak ]
"I would say that the EPA is in the best position to say just how … our [in vitro] data correlates with in vivo animal data," said Abassi. He added that the EPA is currently assembling the data, and "I do not think that they have a clear picture if the different in vitro assays from different groups do really correlate with in vivo toxicity."
Abassi said that toxicity is difficult to classify. For example, toxicity does not necessarily mean that a compound will kill cells, but rather that it could trigger a change in the morphology of the cells or inhibit a particular signaling pathway.
"With our [xCELLigence] system, we can classify these toxic compounds based on their profile," Abassi said. "Then the profile is used to infer the mechanism of action of the compounds."
During the initial proof-of-concept phase, which ran from 2007 to 2008, ToxCast examined 320 chemicals using hundreds of different high-throughput screening bioassays. Of those 320 compounds, 11 were replicate controls to which the companies were blinded, said David Manyak, executive vice president of discovery services at Caliper Life Sciences, which also had its ToxCast contract extended.
The phase 1 contracts have all been completed, the in vitro data have been compiled and have all been put into an EPA database called ToxRefDB. Approximately a dozen posters relevant to the ToxCast data were presented by the EPA at the SOT meeting in Baltimore this week, Manyak told CBA News.
On May 14-15, the EPA plans to run a "data-analysis summit" at its Research Triangle Park, NC, campus to "start releasing these more specific correlations or statistical results looking at the in vitro and in vivo correlations," Manyak said.
According to the ToxCast website, plenary talks will describe the ToxCast program and issues related to toxicity prediction, both from scientific and regulatory standpoints.
The second phase of ToxCast, for which data is expected sometime next year, will expand and validate data from the first phase, and will also generate data from more than 1,000 additional chemicals.
The second phase will also look at thousands of environmental chemicals that require prioritization by the agency because their potential toxicity is unknown.
"We are also working with pharmaceutical companies to bring failed drugs into ToxCast," that have shown evidence of hemotoxicity in clinical trials, and use the ToxCast suite of assays to try and find a correlation between in vitro toxicity and human toxicity, Kavlock said this week.
The agency recently issued four requests for proposals, which closed last week, to bring new assays into the ToxCast program and the virtual tissues projects. The RFPs included high-throughput human and rodent co-culture systems, high-throughput assays for human and rodent signaling and toxicity pathways, and high-throughput genotoxicity and mutagenicity assays.
"We are going to be making a decision in the next several months if there are any assays that we are not going to go forward with in phase 2," because the data does not look to be of high enough quality, or they are not predictive of in vivo effects, Kavlock said.