The Food and Drug Administration is taking a next step in evaluating systems for integrating microarray data into its regulatory drug-approval process.
Iconix Pharmaceuticals, a five-year-old spinoff from Microcide Pharmaceuticals of Mountain View, Calif., last week announced that it would provide the Center for Drug Evaluation and Research of the FDA with research access to its flagship DrugMatrix system.
The DrugMatrix database, a collection of genomic and chemical data, will help the agency take another step in evaluating processes for the inclusion of data from microarrays. This data has the potential to provide critical insights into how drugs interact with the human body, and might be able to improve and shorten the regulatory review process as well.
Last week, BioArray News reported that in 60 days the FDA will release a draft policy document on the concept of providing a “safe harbor” for drug sponsors to introduce microarray experiment data in the drug approval process.
This database agreement speaks to the agency’s desire to develop skills and identify tools that can help it collect and critically analyze genomic data.
“This relates to what we need to do internally to get us to the point where we can work with the data submissions that will be coming to us,” said Frank Sistare, CDER’s director of the division of applied pharmacology research. “To turn data submissions into information, and information into knowledge, you have to decide how you want to see the data, and develop the skills, and have the tools to manage, help interpret, and compare such large datasets.”
The DrugMatrix System
Created in collaboration with Incyte and MDS Pharmaceuticals of Montreal, the DrugMatrix system was introduced in February 2002. Its 40-gigabyte Oracle database platform holds compound, gene, assay, and expression experiment information derived from microarray analysis.
The company takes a chemogenomics approach, using the genome as an assay (via microarray analysis) and applying the information to the drug discovery process.
The database will be installed on a dedicated mainframe computer housed in CDER, where select researchers can log onto an intranet and use a web browser to gain access to an electronic set of information on the genomic effects of drugs, chemicals and toxicants. The dataset holds microarray information, bioassay information, reference sets of public and private information; as well as histopathology and toxicology information.
The DrugMatrix database, 18 months in development, contains 550 known compounds that have been evaluated in expression experiments, with corresponding blood chemistry and histopathology; as well as another 850 compounds profiled using MDS molecular pharmacology assays, and aug-mented with full literature curation.
The data on the compounds included in the database was developed through research conducted by in vivo dosing on rats, and expression array profiling of tissues. Samples were collected at different time points, and were then assayed on microarrays, using mainly CodeLink microarrays, under an arrangement initiated in 2000. Iconix selected CodeLink as its research platform through a “bake-off” competition where it pitted arrays from Motorola, Affymetrix, and Agilent against one another, with finanancial considerations affecting the final deal. In the original agreement with Motorola’s BioChip Systems unit, Iconix was selected to validate the beta release of the CodeLink system, and then to become a high-volume user of the system. Motorola sold the CodeLink line to Amersham for $20 million in 2002.
Can data derived from one platform — in this case, the CodeLink microarrays — help yield information that is pertinent to all platforms such as Affymetrix, Agilent, and the home-brew types?
“We have some healthy skepticism on our part that the information output will be exactly the same; we will need to query that and ask, most importantly, whether the biologically meaningful conclusions will be the same,” Sistare said.
If DrugMatrix does indeed hold up as a platform-agnostic tool, so much the better to take 10,000 to 20,000 data points from any microarray experiment and match it up to the information in DataMatrix for analysis.
“The agency is not going to tell sponsors ‘you have to use Affymetrix, Amersham, or others,’” Sistare said. Data submitted, regardless of platform, will have to provide essent-ially the same answers, he said, unless the biology is different.
DataMatrix can do that, James Neal, Iconix’ CEO, told BioArray News.
“This has good cross-platform capacity,” he said. Why? Because, despite the differences in technology in the commercially built microarrays, they are still alike in being oligo-based platforms, Neal said.
“The key difference is what genes you put on the glass,” he said. “You look at the genes that are in common across two platforms and compare the biological conclusions as well.”
To test its capabilities, the company has taken electronic datasets sent to it by third parties — without knowing the source of the data — and identified the compound or biological event occurring, Neal said.
He did not have any kind of number to index the system’s accuracy.
Iconix has signed one customer, Schering Plough, for its system and is in negotiations with other pharmaceutical companies, said Neal who was executive VP, sales, marketing, and business development for Incyte Genomics before being named CEO of Iconix in August.
While pricing would vary case by case, subscriptions to the database can cost from $1-3 million per year, or $100,000 to $200,000 per compound, he said.
The size of this collection of data puts it in a unique category, but there are others, including Gene Logic and in-house pharma researchers, who are collecting similar databases of information, on a smaller scale.
Iconix, which is privately held, was founded in 1998 and has been financed by the venture capital firms Abingworth Management, Institutional Venture Partners, and Kleiner Perkins Caufield and Byers, as well as receiving equity financing from Motorola, Incyte, and MDS Pharma Services of Montreal.
For the FDA, Sistare said the database is an exploratory step that will help its staff gain experience to do this type of analytical work independently.
The database may not be the final answer in investigating drug candidates, and it may not be the ultimate gatekeeper.
Like microarray technology, it will be a work in progress.