Addressing one of the most pressing bottlenecks in protein biomarker research, the National Cancer Institute is creating an antibody-characterization laboratory to test antibodies used by the research community.
While many companies sell antibodies, the research community has consistently voiced concerns about their quality. The NCI laboratory, which will be based in the institute’s Frederick, Md., site, will be responsible for characterizing antibodies that will be manufactured by the private sector and thereby improving their quality.
Last week, the NCI issued a solicitation for applicants to make the antibodies with the goal of awarding contracts in mid-January 2008, Henry Rodgriguez, director of the NCI’s Clinical Proteomic Technologies for Cancer program, told ProteoMonitor this week.
Antibodies made by companies awarded contracts will need to meet minimum characterization requirements. The antibodies will then be sent to NCI-Frederick for screening and further characterization.
The institute is also in discussions with the Royal Institute of Technology in Stockholm, Sweden, and the Harvard Medical School’s Institute of Proteomics to additionally characterize the antibodies. The University of Iowa will also play a role in the project by selling the resulting antibodies to the research community.
Initially, each antibody manufacturer will receive 42 antigens produced by Argonne National Laboratory and be asked to produce three monoclonal antibodies for each of them. The goal is to have those antibodies available to the research community by next September. Long term, the NCI would like to have antibodies for more than 150 targets available within five years, Rodriguez said.
The initiative is the latest effort under the NCI’s Clinical Proteomic Technologies Initiative for Cancer program, a five-year, $104 million project to develop new proteomics tools and technologies for cancer research.
A component of CPTI is the Clinical Proteomic Reagents Resource initiative, which receives $2.5 million per year in funding. The antibody project is being developed as part of the reagents resource.
The antibody project sprung from a workshop the NCI held two years ago with researchers to discuss bottlenecks in proteomics. One of the biggest problems identified by participants there was the lack of high quality, “extremely well-characterized reagents,” Rodriguez said, “and specifically, they wanted those antibodies.”
At the Human Protein Atlas, commercial vendors have provided about 5,000 antibodies for testing. Of that, however, only about one-third actually work, said Matthias Uhlén, a professor at the Royal Institute of Technology, who spearheaded the HPA.
“They don’t work in our hands, and therefore they would probably not work in other researchers’ hands,” he said. “So two-thirds of the antibodies that are being sold today don’t work in our hands, and, obviously, this is a problem.”
Another issue is that while there are thousands of antibodies commercially available, many of them work against the same targets, which are the most commonly used antigens. The result is that there is a vast library of antigens for which no antibody exists, according to Uhlén, who estimates that antibodies have been manufactured for about 4,000 protein targets, or only 20 percent of known human targets.
The research community has been working to increase that figure. The HPA currently has 3,015 antibodies, and sometime next year plans to raise that number to 4,000. Also, ProteomeBinders, a European-US consortium, hopes to catalog and eventually produce binding molecules that can detect all human proteins [See PM 03/22/07].
While the NCI-Frederick project is significantly smaller in scale than these two other goals, it nonetheless represents an important step, according to Joshua LaBaer, director of Harvard’s proteomics institute. LaBaer, who recently called for the creation of a repository for antibodies for every protein [See PM 06/28/07], said that while the NCI effort won’t cover the entire human proteome, it could serve as a template for such eventual efforts.
“So two-thirds of the antibodies that are being sold today don’t work in our hands, and obviously, this is a problem.”
“This will, to some extent, represent the first step toward this idea of building a complete collection of analyte-specific reagents against proteins of interest,” he told ProteoMonitor this week. “I think the most important thing will be that, in a sense, they’ll start setting standards for how this type of project ought to be done. In addition to making the antibodies, building a database [that] includes information about their characterization is the key.”
He and Uhlén said that NCI’s decision to go with only monoclonal antibodies was an important strategy because they are renewable. Polyclonal antibodies, in contrast, are not.
“The thought process there was that if you’re going to do all this characterization, then it will be really important to have the affinity captured reagents that are renewable such that when you produce a new batch, you can always go back to the original characterization,” Rodriguez said.
The NCI is asking for three monoclonal antibodies for each antigen because of the different methods that can be used to characterize them.
“People do either ELISAs [or] … Western [blots], so different antibodies have different performance characteristics, depending on the assay that you’re using,” Rodriguez said. “Having more antibodies that go through a lot more characterization … gives back to the community reagents that have a lot more broader use to it.”
Also, having multiple antibodies for each antigen enables them to target different regions of the protein, he said.
‘Raising the Bar’
As part of the NCI-Frederick project, each manufacturer will initially screen and characterize the candidate antibodies and send them to Maryland, where NCI researchers will reproduce the experiment using standard protocols.
NCI-Frederick researchers will use a combination of ELISAs, Western blots, surface plasmon resonance, immunohistochemistry, immunoprecipitation, immunofluorescence, and immuno-mass spectrometry to further characterize the antibodies. If the Royal Institute of Technology and Harvard agree to be part of the project, the antibodies would be sent to them for final characterization.
In contrast, the typical commercial antibody is characterized by ELISA and Western blot only, Rodriguez said.
“We’ll take existing characterization that is done today, but we’re also going to throw into that reagent additional characterization that even adds more value to an investigator [so they can] say, ‘This material that I just purchased, now I know how it performs for these different types of assays,’” he said.
After all the screening and characterization, the final antibodies and their antecedent hybridomas will be made available to researchers, for a charge, through the Developmental Studies Hybridoma Bank at the University of Iowa.
The antibodies will have either no or minimal commercial intellectual property rights attached to them, a major concern that was expressed at the 2005 workshop, Rodriguez said.
“At the end of the day, if you’re still [in the] private sector, you can easily acquire the hybridoma, you can easily get access to these antibodies, and then you can further modify them to make them specialized for your platform,” he said.
As of the time of publication, the Royal Institute of Technology had not officially agreed to participate in the NCI antibody project. However, Uhlén said that cancer-related proteins are a good area to start such a project because “they are, of course, of very big interest.
“I hope that the biggest benefit of the NCI project … is that it raises the bar [for what makes] a good antibody,” he said.