SEOUL—Protein biomarkers are a dime a dozen, so why have so few of them proven to be of particular clinical use?
According to speakers at the Human Proteome Organization’s annual conference here last week, the answers are myriad, and the solutions complicated.
There is certainly no shortage of biomarker discovery research in the proteomics field. A cursory search on PubMed with the keywords “protein” and “biomarker” brings up more than 4,500 results. And most, if not all, proteomics work being done by the pharmaceutical industry is directed at biomarker discovery.
But to date, the number of protein biomarkers that has been approved by the US Food and Drug Administration has been miniscule, and those that have been further developed into diagnostic or clinical tools are even fewer.
At a HUPO panel discussion specifically targeting biomarker research in proteomics, the overriding sense was that the failure is due to a poorly executed plan of moving these biomarkers from discovery to clinical applications.
Denis Hochstrasser, chairman of the department of clinical pathology at the University Hospital of Geneva, began by asking who in the audience of a few hundred worked in the diagnostics field and was actively involved in bringing biomarkers to market. Only a very small handful raised their hands. His point: While there is ample research into biomarker discovery, there is little work being done to move them to the clinical side.
Finding biomarkers is not difficult, he told the audience. Validating them is a challenge, but not impossible.
“Getting the diagnostic industry to move [biomarkers] to the clinical side is the biggest challenge,” he said.
It was a theme repeated by several members of the panel who said that one shortcoming in the proteomics field has been the ability of researchers to engage the diagnostic industry in their work.
“I think that’s something we have to do in proteomics, engage people on the other side of the fence,” said Mike Dunn, a professor of biomedical proteomics at the University College Dublin in Ireland.
Indeed, that sentiment has been pervasive throughout proteomics. Last fall, at a conference held by the American Association for Clinical Chemistry, several speakers made the point that the diagnostic industry must be included in the fold as proteomics tries to advance into the clinical field.
Omar Laterza at Merck Research Laboratories said at the time that even beyond the proteomics research community, drug companies need to work with diagnostic firms to validate biomarkers. Traditionally, he said, drug manufacturers have been interested in biomarkers purely from a pharmacodynamic perspective. That interest is turning toward disease biomarkers, however, and here diagnostic firms can bring expertise in assay development and validation and access to assay platforms [See PM 10/19/06].
Garbage In, Garbage Out
At last week’s HUPO meeting, however, some panelists said that the bottleneck is not strictly in the discovery-to-diagnostic continuum, but at a more fundamental level. In particular, these observers said that the proteomics field needs to see improvements in experimental design and data-handling capabilities before it can advance toward the clinic.
For example, during the conference, HUPO released details about its protein standard mixture initiative in which only six of the 24 labs that participated were able to identify all 20 proteins on their own, a 25 percent success rate [See PM 10/11/07 and 09/06/07]. At a panel discussion, John Bergeron, a professor of anatomy and cell biology at McGill University in Montreal and immediate past president of HUPO, called the ability to match mass spectrometry spectra with the proper peptide and protein the major bottleneck in the initiative.
“Getting the diagnostic industry to move [biomarkers] to the clinical side is the biggest challenge.”
At another panel discussion, Sam Hanash, a principal investigator of the Fred Hutchinson Cancer Research Center and the founding president of HUPO, called the biomarker field “the Wild West,” noting that “people make claims; it’s really chaotic.” [See PM 10/11/07]
The field, he said, is overrun with people who aren’t properly equipped to do the research, due to a shortage in funding and other resources.
Furthermore, Martin McIntosh, principal investigator in the computational proteomics laboratory at Fred Hutchinson, said that sample collection is extremely important for biomarker discovery, but undervalued. As a result, the findings from many studies may be compromised early on.
Helmut Meyer, director of the Medical Proteome Center at Ruhr-Universität Bochum, Germany, said that just collecting data is no longer sufficient. “We have to learn more about diseases, and we have to learn more [about] how to use the technology,” he said.
Sudhir Srivastava, chief of the Biomarkers Research Group in the National Cancer Institute’s Division of Cancer Prevention, said that while there is “no dearth of biomarkers,” there is a need for clinically relevant biomarkers, calling it the major bottleneck in biomarker research.
Specifically, the shortage is in clinically validated biomarkers, he said, largely due to a lack of platforms for such work.
His prescription for improving the discovery-to-diagnostic pipeline includes developing technology qualifications for clinical proteomics, establishing team-based disease-specific biomarker-related projects, and developing standards for reagents and reference materials.
During the conference, Henry Rodriguez, director of the Clinical Proteomic Technologies Initiative at the NCI, cited a lack of “high quality” affinity reagents as a major barrier in proteomics research. While the market is flooded with commercially available reagents, few are well characterized, he said.
To meet that challenge, his office is developing a pipeline in partnership with Argonne National Laboratory, which is producing antigens, and various players in the private sector involved in producing monoclonal antibodies.
While HUPO goes about figuring out what it can do to push biomarker research and clinical applications — the organization is preparing a paper on the challenges in the field and what steps it will take to address them — Srivastava had some recommendations.
First, HUPO should create a database of biomarkers discovered by its members. The organization also needs to do a better job of publicizing funding opportunities for biomarker research, he said. He recommended it solicit private-public partnerships on specific biomarker projects. And finally, Srivastava said, reagents and tools developed through HUPO initiatives need to be made available to the scientific community.