Skip to main content
Premium Trial:

Request an Annual Quote

Breast Cancer Study Combines Top-Down, Bottom-Up Analyses


NEW YORK (GenomeWeb) – A team including researchers from the National Cancer Institute's Clinical Proteomic Tumor Analysis Consortium has completed an analysis of breast cancer integrating bottom-up and top-down mass spec methods.

Detailed in a study published this week in Molecular & Cellular Proteomics, the effort is among the first to combine quantitative bottom-up and top-down measurements of the same samples, and provides an indication of the potential benefits of top-down analysis, said Neil Kelleher, a researcher atNorthwesternUniversity and senior author on the paper.

Bottom-up, peptide-based proteomics has dominated the field since its inception more than a decade ago, but interest in top-down methods looking at undigested full-length proteins has grown as improvements in instrumentation have made analysis of intact proteins easier and researchers have become increasingly aware of the importance of protein isoforms and post-translational modifications.

In conventional bottom-up proteomics, proteins are digested into smaller peptides prior to analysis, so it can be difficult or impossible to determine, for instance, what different combinations of post-translational modifications exist on single proteins.

Top-down analyses, on the other hand, look at undigested proteins, meaning researchers can profile the intact molecule with all its modifications rather than tying that information together from collections of identified peptides.

Because of the complexity of intact proteins and the difficulties inherent in separating them and introducing them to the mass spectrometer, top-down methods continue to lag behind bottom-up approaches.

Recently, however, work by Kelleher and others has narrowed the gap, somewhat. In particular, his lab published last year a paper in Analytical Chemistry demonstrating the feasibility of large-scale top-down label-free quantitation.

The MCP study represents the next step in that work, Kelleher told GenomeWeb this week, noting that the ability to do such quantitative analyses allows for integration of top-down and bottom-up work.

In the study, the researchers used bottom-up proteomics on an AB Sciex TripleTOF 5600+ instrument and top-down proteomics on a Thermo Fisher Scientific Orbitrap Elite to profile the patient-derived mouse xenograft models of basal and luminal B human breast cancer, WHIM2 and WHIM16.

Looking only at proteins between 0 and 30 kDa, the range in which top-down quantitation is currently possible, the bottom-up approach quantified 3,519 proteins and the top-down approach quantified 982 proteoforms mapping to 358 different proteins.

But while bottom-up, as expected, made significantly more protein IDs and was able to detect more instances of differential protein expression between the two xenograft models, top-down, with its ability to look at intact proteins, detected a variety of protein modifications missed by bottom-up, allowing a better look at the diversity of proteoforms in the sample.

For instance, the authors noted, in the case of the "small, abundant proteins" amenable to top-down quantitation, the technique captured changes in modifications not observed by bottom-up methods around 40 percent of the time.

This finding, Kelleher said, lends support to the larger top-down endeavor as it suggests that, as he and other top-down researchers have posited, the method detects biological information not available at the bottom-up level.

"This is kind of the first peek at the answer to the question of: How many times do you see disparity between bottom-up and top-down?" he said. "If you see great discordance, which we did, then it hints that even for small proteins there is a large amount of biology wrapped up in the proteoforms."

And this biology, he added, is to an extent masked by bottom-up methods, which, even when they look at post-translational modifications, do so at the peptide level as opposed to the proteoform level.

Kelleher cited as an example of top-down's potential the study's finding of a significant upregulation of certain proteoforms of the protein type 2 cytoskeletal keratin 8, which is a component of the plasma membrane that is used for distinguishing between lobular and ductal breast cancers.

The bottom-up analysis identified an upregulation of K2C8 in the more aggressive WHIM16 cancer. The top-down data, however, revealed a pattern not apparent from the bottom-up information.

The top-down analysis showed not just upregulation of K2C8 in WHIM16 but upregulation of eight specific proteoforms generated through proteolytic events cutting off head or tail regions of the protein.

"You are sort of shedding a lot of the structure that holds up the plasma membrane in the more aggressive xenograft model," Kelleher said. "And we have a direct comparison to bottom-up, and there you don't know which proteolytic forms you have, but in top-down you do and you see big upregulation."

"So, could these be functional observations? Do metastatic, highly aggressive tumors have this as a generally observable trait — a big morphology change in the cells due to the shedding of these head and tail regions?" he said.

He and his colleagues are now following up to see if they observe a similar phenomenon in a colon cancer model. They also hope to look at patient samples to see if the process is operative in actual cancer patients. He noted that it has been cited in a few previous, non-proteomics papers that observed it using confocal microscopy.

The finding, along with the generally high level of discordance found between top-down and bottom-up analysis, suggests the great potential of top-down, Kelleher said.

"I always hold out the dream that cancer biomarkers resolved at the proteoform level might completely change the conversation relative to what bottom-up has been able to do," he said. "That is the hope."

He added that the involvement of researchers including CPTAC head Henry Rodriguez on the study indicates interest in top-down methods on the part of the consortium.

"I think there is interest," he said. "They were sort of kicking the tires on top-down and how it might function. I think the question is: What information does it provide and could you have other labs do it; could you export the technology?"

"If you believe that proteoforms and proteoform-resolved measurements can help the discussion both for fundamental mechanistic cancer biology and also biomarkers, then I think now is the time to try to decide," he said. "Because every year the technology is getting easier, more instruments can do it, and more labs can do it."