Researchers at the Memorial Sloan-Kettering Cancer Center have found that the degradation products of relatively high-abundance blood proteins, which had previously been dismissed by some as waste products that serve no diagnostic purpose, may actually serve as good biomarkers for disease.
The peptides that are good markers for disease may be formed when proteases specific for disease cleave the relatively abundant blood proteins, which serve as substrates, explained Paul Tempst, the leader of a study published in this month's issue of the Journal of Clinical Investigation.
While the disease-specific proteases themselves are not abundant enough in blood to be seen by regular proteomics experiments, the serum peptidome, which contains disease-specific cleavage fragments of known proteins and peptides, could serve as a new type of biomarker, Tempst said.
"Basically, we were lucky to pick up this activity," said Tempst. "There are a lot of low-concentration enzymes you can't see, but if you give them enough time and enough substrate, you can get a catalytic product that's measurable with MALDI-TOF."
Lance Liotta and Emanuel Petricoin wrote in a commentary that accompanied the JCI paper that while some researchers have dismissed the serum peptidome as "biological trash," Tempst's work shows that the peptidome may, in fact, contain "diagnostic gold."
"Basically, we were lucky to pick up this activity. There are a lot of low-concentration enzymes you can't see, but if you give them enough time and enough substrate, you can get a catalytic product that's measurable with MALDI-TOF."
"Regardless of whether the peptidome is derived from the tissue/microenvironment in vivo or ex vivo, the implications for the diagnostics arena are enormous," wrote Liotta and Petricoin, both researchers at George Mason University who themselves pioneered studies that used patterns of mass spectra as biomarkers for disease. "Fingerprints of MS ions can be replaced with panels of named protein biomarkers."
Benjamin Cravatt, a proponent of "activity-based proteomics" at the Scripps Research Institute (see ProteoMonitor 11/18/2005) concurred, saying that Tempst's work suggests that a serum peptide profile could be used as a new type of pattern-based disease biomarker.
"This work is highly significant because it provides evidence that different tumor types secrete and/or shed distinct sets of proteases that, through their catalytic activity, generate unique serum peptide profiles," he wrote to ProteoMonitor in an e-mail. "These peptide profiles can in turn be used as patterns to accurately predict the presence of a particular tumor type in patients."
Tempst's research team first delved into biomarker research about three and a half years ago, when the biomarker field became hot after Liotta and Petricoin published their landmark Lancet paper (see PM 2/18/2002), Tempst said. The researchers decided to concentrate on peptides, rather than proteins in blood because Tempst saw evidence that biomarkers that had been approved for diagnostics were nearly always peptides, rather than proteins.
As the researchers used MALDI-TOF mass spectrometry to conduct proteomic experiments to try to find biomarkers for prostate, breast, and bladder cancer, they saw more complex mass spectra patterns when samples were allowed to sit out of the ice bucket or refrigerator for longer periods.
"We wanted to see as many peaks as possible, so we did some time-course experiments," Tempst told ProteoMonitor. "We also added protease inhibitors to some samples, and found that they didn't get as complex of patterns."
After spending much effort to optimize and standardize serum collection protocols, the researchers conducted a prospective study. They analyzed the serum peptide profiles of 32 patients with advanced prostate cancer, 21 patients with breast cancer, and 20 patients with bladder cancer. Eventually, they came up with a large number of mass spectra profiles for both cancer and control samples.
The researchers then used Agilent's GeneSpring software to compare 650 different features of the mass spectra. After the analysis, the researchers selected 69 peptides that had mass spectra with significant differences between cancer and control. The researchers then sequenced those peptides.
When the researchers analyzed the peptide sequences, they found that the peptides were cleavage byproducts of about a dozen "founder peptides," including FPA, fibrinogen, C3f, C4a, and ITIH4. These larger peptides were themselves cleavage products of relatively abundant blood proteins.
Tempst pointed out that it was fortuitous that the serum samples they analyzed were never placed in tubes with protease inhibitors. He also noted that in this study, serum was better than plasma because the anti-coagulant factors used during plasma collection may suppress disease-associated proteases.
"Usually these things are the result of some sort of accident," said Tempst. "If those tubes had contained protease inhibitors, we would never have found [our results]."
The next step after finding that disease biomarkers are by-products of broken-down blood proteins is to try to identify the currently invisible disease-related proteases themselves, Tempst said. This can be done using old-fashioned biochemistry, or by doing "targeted proteomics," in which one looks specifically for proteases using protease inhibitors as capture baits, he added.
Another project that Tempst's team is working on is to develop an assay to absolutely quantify the founder peptides. The assay would allow researchers to determine how much of the founder peptides were present before degradation by proteases. That way, researchers could better tell if differences in mass spectra and cleavage profiles were due to differences in starting abundance of founder peptides, or due to other reasons.
Gil Omenn, the leader of the Human Proteome Organization's Plasma Proteome Project, noted during a discussion of Tempst's work at this week's PEPTALK conference in San Diego that the recent study shows that "the action does not have to be at the hardest-to-detect, lowest-abundance proteins."
HUPO PPP recommends that plasma be used over serum because it appears to give better results in PPP experiments. However, the way that researchers decide to handle their samples depends ultimately on what they are trying to determine, Omenn noted.
"Obviously, if you're doing a study of proteases, you can't use protease inhibitors," said Omenn.
Tempst said he has filed for a few patents based on his serum peptidome work, but he has not spoken to anyone yet regarding commercialization of the work.
"I think this is going to lead to something down the road," he said. "Whether it will be in the public domain or not, I don't know."
Other researchers who were involved in the serum peptidome study include Josep Villanueva, the first author of the study; David Shaffer; Carols Chaparro; Hediye Erdjument-Bromage; Adam Olshen; Martin Fleisher; Hans Lilja; Edi Brogi; Jeff Boyd; Marta Sanchez-Carbayo; Eric Holland; Carlos Cordon-Cardo; and Howard Scher.
— Tien-Shun Lee ([email protected])