Discussions at the US FDA about how to consider microarray data in new drug applications have been gaining momentum during the past year, and over the summer a plan began coming into focus. A lack of standards or controls among data generated by various microarray technology platforms has long hindered FDA’s ability to consider them. And the agency’s uncertainty about the technology has made drug developers reluctant to submit the data, for fear that doing so would result in an application being hung up even longer in the approval process.
The agency’s Office of Testing and Research launched two separate gene expression database initiatives earlier in the summer (see “Guidelines on the Way? FDA Delves into Microarray Data,” in GT’s August 2003 edition), and when FDA Commissioner Mark McClellan spoke at the Drug Discovery and Technology meeting in Boston in August, he hinted that guidelines for using microarrays would be forthcoming. Now another prong has been added to the multi-part initiative.
In late August, the agency’s Center for Drug Evaluation and Research recruited Gene Logic to conduct a bioinformatics analysis to identify genes that are consistently expressed among Affymetrix, Agilent, and Amersham Biosciences’ microarray platforms.
The data will be used to provide drug developers with a method for calibrating their microarray analyses. “Blood tests are done all the time on all types of machines,” explains Doug Dolginow, Gene Logic’s senior VP for pharmacogenomics. “How do you know if a result such as your cholesterol level is correct on one machine versus another? You have control material that you run every time you run the assay that shows you’re in the range.”
By looking across the thousands of genes that are common among the three microarray platforms, Gene Logic will determine which are consistently found at a rank order level that is the same from tissue type to tissue type. That way, Dolginow says, whenever pharmacogenomic or toxicology data are submitted to FDA, “you would expect to see those same genes present in the rank order that’s expected, and you would expect to have some assurance that the technical quality of the data is good.”
Karol Thompson, a principal investigator with the Center for Drug Evaluation and Research who has been working separately with the DC-based nonprofit International Life Sciences Institute and FDA’s Genomics and Proteomics Working Group to incorporate gene expression biomarkers into drug safety evaluation, will manage the Gene Logic/FDA collaboration.
CDER says it is “poised to contribute substantially to the evolution of guidance to review staff and to the regulated industry dealing with microarray gene expression data generation and interpretation.” The agency’s broader effort focuses on creating a “central multi-platform-compatible database capable of accepting public domain as well as user-generated and sponsor-submitted gene expression data.”
When Gene Logic has completed the platform comparisons — by year end, according to Dolginow — FDA will make the results freely available on the Web. How control material will be distributed remains a question. It looks like a business opportunity for someone, Dolginow says.