Amplify, Amplify: Shotgun Proteomics Boosts the Signal for Biomarker Discovery

Disease processes or exposure to environmental toxicants can produce tiny modifications (called adducts) on proteins in the blood. A clinical assay that reliably detects those modifications in plasma or serum could confirm environmental exposures or speed diagnosis of diseases such as cancer. But scientists have identified so many protein adducts that might serve as candidate biomarkers that finding the best ones to advance into further testing presents a major challenge. Advancements in an approach known as shotgun proteomics now promise to streamline the discovery process by identifying the most promising biomarkers for exploration. 
 
Scientists are still laying the rudiments of a systematic “pipeline” for biomarker discovery, says Dan Liebler, a professor of biochemistry, pharmacology, and biomedical informatics at Vanderbilt University School of Medicine. Liebler has for almost 10 years used proteomics to study protein damage caused by oxidative stress or chemical toxicity induced by reactive endogenous chemicals. “We have knowledge of [reactions or changes] that could be advanced at some point to biomarkers, but there are so many candidate changes in tissues or living systems exposed to environmental stressors, and we don’t have efficient mechanisms for identifying the best possible markers to move forward in the pipeline,” Liebler says. 
 
In the last 10 years proteomics technologies involving mass spectrometry have made it possible to quickly identify proteins and quantify adducts, down to the very amino acid site of modification, so scientists can more quickly screen potential biomarkers. In a typical shotgun proteomics approach, a scientist would take a biologic sample, add enzymes to digest all the proteins to peptides, fractionate the peptides, then analyze them on an ion trap mass spectrometer. The resulting spectra can be compared against peptide databases to determine which proteins are present in the sample. 
 
Scientists have used the shotgun pro−teomics approach to discover many protein modifications of interest. But the list of biomarker candidates must be narrowed by determining which ones can be reproducibly measured in large numbers of clinical samples, such as blood samples from unexposed and exposed people. Performing such tests of candidate bio−markers currently requires development of targeted immunoassays, which is difficult and expensive because scientists have to develop an antibody for use in the assay that is specific to the protein of interest. 
 
Improvements in mass spectrometry assays have the potential to change that. Proteomics researchers now predict that in as little as three to four years hybrid immuno–mass spectrometry assays, which combine some elements of traditional immunoassays with some elements of shotgun proteomics, will be used to evaluate candidate bio−markers in a large number of samples.

proteomics to study protein damage caused by oxidative stress or chemical toxicity induced by reactive endogenous chemicals. "We have knowledge of [reactions or changes] that could be advanced at some point to biomarkers, but there are so many candidate changes in tissues or living systems exposed to environmental stressors, and we don't have efficient mechanisms for identifying the best possible markers to move forward in the pipeline," Liebler says.
In the last 10 years proteomics technologies involving mass spectrometry have made it possible to quickly identify proteins and quantify adducts, down to the very amino acid site of modification, so scientists can more quickly screen potential biomarkers. In a typi cal shotgun proteomics approach, a scientist would take a biologic sample, add enzymes to digest all the proteins to peptides, fraction ate the peptides, then analyze them on an ion trap mass spectrometer. The resulting spectra can be compared against peptide databases to determine which proteins are present in the sample.
Scientists have used the shotgun pro teomics approach to discover many pro tein modifications of interest. But the list of biomarker candidates must be nar rowed by determining which ones can be reproducibly measured in large numbers of clinical samples, such as blood sam ples from unexposed and exposed people. Performing such tests of candidate bio markers currently requires development of targeted immunoassays, which is difficult and expensive because scientists have to develop an antibody for use in the assay that is specific to the protein of interest. Improvements in mass spectrometry assays have the potential to change that. Proteomics researchers now predict that in as little as three to four years hybrid immunomass spectrometry assays, which combine some elements of traditional immunoassays with some elements of shotgun proteomics, will be used to evaluate candidate bio markers in a large number of samples.

An Assay for Organophosphate Exposure
One common proteomic approach to identifying biomarkers involves exposing an in vitro system to a source of damage such as a particular chemical, then using a mass spectrometer to identify the protein modifications. That is the basic approach used by Mike MacCoss, an assistant pro fessor of genome sciences at the University of Washington, in his research with pro fessor of medicine and genome sciences Clement Furlong. MacCoss and Furlong are searching for biomarkers of exposure to organophosphates such as tricresyl phosphate, a chemical used as a lubricant in jet engines that has been measured in aircraft cabin air. When tricresyl phos phate is inhaled, it is metabolized into toxic cyclic saligenin phosphate. A defini tive measure of exposure to this chemical could help determine whether such expo sure is the source of neurologic symptoms such as tremor and memory loss that have been reported among airline workers.
In an attempt to develop an assay for organophosphate exposure, MacCoss and colleagues use a combination of protein biochemistry and proteomics. "The way protein biochemistry was done previously is that you would have an activity you were interested in, or you had an antibody rec ognizing a given protein, and as you went about purifying your protein, you'd use the activity to trace [the protein], and then you'd go about determining what your pro tein sequence is," MacCoss says. "We're doing the same thing," he explains. "It's just that we're using a mass spectrometer on the back end to speed that process up." Organophosphate compounds are known to inhibit a family of proteins called carboxyl esterases. To find specific proteins of interest, MacCoss and colleagues first perform activity assays with serum samples in vitro. An organophosphate compound is added to the sample, then a traditional bio chemical activity assay is used to track pro teins that are modified by the compound. The scientists continue to track activity and fractionate the sample until they have purified it, meaning it contains only one protein. At this point they do not know the identity of the protein, but they do know that its activity has been modified by the addition of the organophosphate.
Next the researchers use the puri fied sample to perform a more targeted experiment using microcapillary liquid chromatography-tandem mass spectrom etry to identify the protein of interest and to measure the modifications induced by the original organophosphate addition. To perform this experiment, they add an enzyme such as trypsin to digest the pro tein to peptides. "That produces a series of over lapping peptides that span the entire protein sequence," MacCoss says. The researchers can identify both the pro tein and the exact sites of modification because a modification will cause a slight shift in the pep tide's mass. Using this process, MacCoss and colleagues have identi f ied t hree c a ndidate proteins that are modi fied by organophosphate compounds, and they have identified the site of modification on each protein. Now MacCoss-and other scien tists who have similarly found candidate biomarkers-must test them in a larger number of samples, such as tissue or serum from exposed individuals.

Finding the Tiniest Trees in the Forest
One major challenge in testing candidate biomarkers in vivo is that the modified protein will be present in a typical patient sample in a very tiny concentration, even tinier than in the in vitro experiments. "The real challenge is being able to see the smaller, less abundant modified proteins that are going to be more informative, versus the 'redwoods'-or abundant pro teins-that will be everywhere in the forest of blood proteins," says B. Alex Merrick, a staff scientist in the NIEHS Laboratory of Respiratory Biology. Finding biomarkers of specific environmental exposures can be especially difficult compared with identi fying biomarkers of disease. "A lot of times you're down at the noise level looking for changes," Merrick says. "It's more difficult to prove [modification by exposure]." MacCoss tries to address that prob lem by making affinity reagents that will enrich serum samples for the protein of interest without going through the labori ous purification used during the in vitro experiments. "The affinity reagent will that are going to be more informative, versus the "redwoods"-or abundant proteins-that will be everywhere in the forest of blood proteins.

-B. Alex Merrick NIEHS
preferentially bind one protein over other ones, so you effectively make it so that your protein of interest is a greater proportion of the total proteins left over in your mixture," he says. "Then we will follow that by a mass spectrometry step, and that will allow us to make the process much faster." MacCoss pre dicts he will be able to do the affinity enrich ment in parallel for many samples at once, and then move on to the mass spectrometry assay.

Hybrid Assays
Right now, according to Liebler, mass spectrom etry provides a "bridging technology" that can help scientists sort through bio marker possibilities without having to devel op a highquality, robust assay such as an immuno assay to evaluate each potential biomarker. But evaluating biomarker can didates in a large group of patient samples still requires developing a tar geted immunoassay for those specific biomarkers. "It's very hard and expensive to make really good antibodies to do highquality ELISAs-the standard plat form for immunoassay," Liebler says. But L ie ble r pre d ic t s t h a t m a s s spectrometry-based assays could soon advance to the point that they are used in place of standard immunoassays to evaluate biomarkers. "It's likely that mass spectro metry will be able to take a place alongside immunoassays in conducting marker trials, as an analytical platform of choice," Liebler says.
Such a hybrid immuno-mass spec trometry assay would use an antibody (the "immuno" part) to enrich the blood sample for the protein of interest, then the "mass spectrometry" part identifies whether the protein of interest has been modified and to what degree. The hybrid assay would still require development of an antibody, but the antibody would not need to per form at the level required for today's stan dard immunoassays.
"With a hybrid immuno-mass spec trometry assay, the antibody . . . gives the mass spectrometry-based detection a boost by enriching the target from a very complex mixture," Liebler says. He and others are working with such assays in the research setting, but Liebler is not aware of work with such an assay being published yet in the literature. "These assays are really not ready for prime time now," he says. But he predicts that within the next three to four years they'll be used in research studies to validate biomarkers in large numbers of patient samples.
MacCoss has also developed software to improve the ability of mass spectrom etry to detect the lowerabundance pro teins in a sample. In an article published in the 3 April 2009 issue of the Journal of Proteome Research, for example, MacCoss and colleagues described their use of an algorithm that programs the mass spec trometer to preferentially acquire spectra of the peptides that are in lower abun dance, writing, "Our approach uses the high peak capacity of the mass analyzer to resolve and detect peptide features that would not normally be sampled in the presence of more intense interfering sig nals." MacCoss says, "It's not perfect, but it's an improvement. It would work for a case in which a small fraction of your sample contains a modified peptide and you had enough material to run the analy sis multiple times." Another challenge for any future appli cation of clinical assays that identify bio markers of exposure is that proteins-and any modifications that have happened because of exposure-might degrade in the body before a patient is tested. "As soon as a protein becomes degraded, we won't be able to measure the modified form of that protein," MacCoss says. "Some of the pro tein biomarkers that we're following have pretty short halflives. That's something we'll have to work out when we get there." In the meantime, proteomics researchers and their use of hybrid immuno-mass spectrometry technologies for identifying protein adducts as biomarkers in exposed populations may bring us one step closer to linking environmental toxicants with illness and disease.

Angela Spivey
Environmental Health Perspectives • volume 117 | number 5 | May 2009 I t's likely that [mass spectrometry-based assays] will be able to take a place alongside immunoassays in conducting marker trials, as an analytical platform of choice.
-Dan Liebler Vanderbilt University School of Medicine S u g g e s t e d R e a d i n g