Major barriers in drug development include variable responses, efficacy, and adverse effects in humans. However, drug developers and regulators expect pharmacogenomic information, including biomarkers, to break through these hurdles and eventually guide clinical development, disease monitoring, diagnosis, and treatment.

Several approved drugs already have associated biomarkers, and Genentech's Herceptin is among the most frequently mentioned. The drug was approved in 1998 as suitable only for those 25% of metastatic breast cancer patients who overexpress the Her2 protein. The drug's labeling recommends specific Her2 gene amplification tests that can identify candidates for treatment.

Although it holds up Herceptin as a paradigm, the pharmaceutical industry still has some trepidations about using pharmacogenomic data, especially in midstage clinical studies. While the benefit is clear, for example, in avoiding adverse side effects, there's still concern that regulators will make decisions based on developing science and exploratory work.

"If we use pharmacogenomic data in enrolling patients into clinical studies, regulators will include those criteria in the label for the compound," commented Nicholas C. Dracopoli, Bristol-Myers Squibb vice president for clinical discovery technologies, at BIO 2003, the annual meeting of the Biotechnology Industry Organization. This would tie products to tests, and drug companies, most of which are not diagnostics makers, would be challenged to deliver approved tests before launching a drug, he added.

Meanwhile, the Food & Drug Administration worries that these issues are stifling communication. The agency is concerned that it may not be seeing pharmacogenomic data being collected by industry or, in the worst case, that companies are not generating the data out of fear about how regulators will use it.

"We need an approach that will enable free exchange of information in this area, help advance the science and technology, and aid in the timely development of regulatory policies," Janet Woodcock, director of FDA's Center for Drug Evaluation & Research, told the BIO meeting.

"Many of these data are not yet ready for prime time," Woodcock conceded. She promised clear policies, noting that the agency wants to integrate the new technology smoothly into the regulatory process so as not to impede it.

FDA is assessing what data are relevant and will be required to show safety or efficacy, along with considering a process for submission. Until the agency finalizes its policies--and even afterward for data it won't require--it will probably request voluntary submissions to create a knowledge base, Woodcock noted.

The agency put together an internal working group in June 2001 to begin thinking about pharmacogenomics in drug development and regulatory decision-making. It held an initial public workshop with industry in May 2002. By August, the agency hopes to have draft guidance on genomic data submission available for comment, followed by another workshop in November.

"We want to move from the current empirical process for drug development to a mechanism-based process that is hypothesis driven," Woodcock added. "Pharmacogenomics will be extremely helpful in creating a lower cost, faster process resulting in more effective and less toxic drugs."

Chemical & Engineering News
Copyright © 2003 American Chemical Society