In a move that was welcomed by pharmaceutical and biotech companies, the US Food and Drug Administration last week released for public comment its draft guidance on the submission of pharmacogenomic data. The guidance document is also good news for informatics firms who plan to bridge the genomics knowledge gap that exists between pharma and the FDA. Just as an entire cottage industry of contract research organizations sprang up to assist pharma with the clinical trials process, some informatics companies are positioning themselves to act as analytical go-betweens once the pharmacogenomics data submissions come pouring, or even trickling in.
The draft guidance (available at http://www.fda.gov/OHRMS/DOCKETS/98fr/2003d-0497-gdl0001.doc) lays out the fundamental conditions for submission of pharmacogenomic data derived from gene expression or SNP-genotyping experiments as part of the regulatory process. More importantly, from the perspective of pharmaceutical firms who have been reluctant to submit information that might be used against them, the guidance explicitly spells out the conditions for voluntary submission of pharmacogenomics data.
Seeking “to be prepared to appropriately evaluate the anticipated future submissions,” the FDA wants to experiment with as much experimental data as possible. To encourage submissions, the agency established a new category called Voluntary Genomic Data Submissions (VGDS), and created a cross-center Interdisciplinary Pharmacogenomic Review Group (IPRG) to review VGDS, work on ongoing policy development, and advise review decisions regarding pharmacogenomics data.
“The FDA will not use information submitted through the voluntary process for regulatory decision making on INDs or NDAs,” the guidance document clearly states. Instead, the agency “intends to gain experience and to develop an aggregate genomic knowledge database from multiple VGDSs that could be used to rationally facilitate the use of pharmacogenomics in drug development and to share what general knowledge is learned from the data repositories.”
The draft is a “great start,” according to Les Browne, COO of Iconix Pharmaceuticals. “It’s really a document they’re putting out there to stimulate discussion … and it starts the process of getting genomic data into the drug evaluation and approval process.”
Representatives from Iconix, along with other industry players, plan to attend a meeting sponsored by the Drug Information Association in Washington, DC, Nov. 13-14, where the FDA will discuss the agency’s pharmacogenomics plans. The draft guidance is likely to be a primary topic.
The comment period for the draft guidance closes Feb. 2, 2004.
Giving Birth to the Genomics CRO
Doug Dolginow, senior vice president of pharmacogenomics at Gene Logic, said the guidance should encourage pharmaceutical and biotech firms “to immediately begin to generate data in the preclinical and clinical areas using genomic technology … without having a negative regulatory or unexpected regulatory impact on that.” The upshot, he said, is “quite enabling for companies like ours.”
Gene Logic acquired CRO TherImmune earlier this year in anticipation of genomic data being fed into the regulatory process. The FDA’s blessing “makes it a little bit easier to conduct our business,” Dolginow said, especially “now, as a CRO-based company where we’re providing submissions to the FDA and working with our customer base to do that.”
Enodar BioLogic is also staking a claim in the territory between pharmaceutical firms and the FDA, and already calls itself a “GCRO,” or genomics CRO. Lue Ping Zhao, Enodar’s CEO, said he gives the guidance document “high marks” for striking a balance between “regulation and flexibility.”
Noting that the document could have been a bit “more explicit” regarding analytical methods and other informatics-related guidelines for submitting data, Zhao said that the FDA’s open-endedness in this area ultimately provides an opportunity for companies like Enodar. “Rigorous [statistical] analysis [of pharmacogenomics data] will be important in order to sustain drug companies’ success in producing new drugs,” he said, “and this is exactly the place where we would like to excel.”
No Standards Required?
Kurt Jarnagin, vice president of biological sciences and chemical genomics at Iconix Pharmaceuticals, questioned whether the FDA’s avoidance of formatting or standards issues would discourage submitters. According to the guidance document, the FDA is not offering recommendations for specific formats for the VGDS because “consensus standards do not exist for presenting and exchanging genomic data.” While Jarnagin agrees that the standards issue is a tricky one, he noted that with the decision to avoid the issue, “the agency is taking on the onus of accepting the entire variety of formats that might arise.”
If the FDA “really wants to get all of the raw data,” Jarnagin said, “it needs to promote, adopt, or establish an exact format for submissions, or else they will not get in a format that their reviewers can understand.”
Also helpful, Jarnagin said, would be a “standard input tool” — along the lines of Sequin for Genbank submissions — that would make it easier for drug companies to submit their data.
Nevertheless, he said, “We’re delighted to see it. Any progress is better than none.” In addition, he said, databases like Iconix’s DrugMatrix, which the FDA is currently using to evaluate gene expression, may eventually become de facto standards themselves.
The draft guidance distin- guishes between “known” biomarkers and “probable” markers in deciding whether submission of pharmacogenomics data should be mandatory, but Jarnagin pointed out that the line between the two is fuzzy, and even discerning whether a biomarker is “probable” based on gene expression data can be a challenge. “They don’t have standard values, or experience to tell them what the standard values might be, so they will have to rely on sponsor findings, literature findings, or appeal to a database such as the one that Iconix has provided — those are the choices that the agency will have,” he said.