A little more than two years after its launch, the second phase of the Clinical Proteomic Tumor Analysis Consortium held its first annual scientific symposium this week at the National Institutes of Health's headquarters in Bethesda, Md.
Participating researchers presented on various portions of the project, including initial findings from proteomic studies of cancer samples originally characterized by the National Cancer Institute-funded Cancer Genome Atlas.
Broadly speaking, this data demonstrated the ability of mass spec-based proteomics to detect evidence of genetic mutations at the protein level and to provide additional cancer typing information above and beyond that offered by genomics and transcriptomics.
The symposium also highlighted the continuing advance of several proteomic techniques, most prominently multiple-reaction monitoring mass spec and the use of transcriptomic data to generate sample-specific mass spec reference databases.
Launched in August of 2011, CPTAC 2 – a five-year project slated to cost between $75 million and $120 million – aims to combine protein biomarker discovery and verification studies in tumor tissue samples with genomic characterizations of those same samples done by TCGA (PM 8/26/2011).
The effort builds on the initial five-year, $104 million CPTC initiative launched in 2006, which worked to build a foundation of technologies and standards to advance the application of proteomics to cancer research. That project established five multidisciplinary, multi-institution research centers and developed collaborations with more than 60 public and private institutions around the world.
The second phase of the program established research centers at eight institutions including Washington University in St. Louis, the University of North Carolina, Boise State University, Pacific Northwest National Laboratory, the Broad Institute, Fred Hutchinson Cancer Research Center, Johns Hopkins University, and Vanderbilt University.
The groups undertook analysis of three tumor types – breast, colorectal, and ovarian – with the aim of profiling around 100 samples of each, and at this week's meeting they shared initial results from these analyses.
Among the findings to emerge from these results was a general picture of the current capabilities of shotgun mass spec workflows and instrumentation. In recent years, a number of researchers have performed experiments in which they essentially "saturated" the proteome that can be identified by standard mass spec methods, arriving at a figure of around 12,000 proteins detectable in a given cell.
The CPTAC researchers, largely using Thermo Fisher Scientific Q Exactive or Orbitrap Velos or AB Sciex TripleTOF 5600 instrumentation, arrived at similar numbers in their analyses, with most identifying between 10,000 and 12,000 proteins per experiment.
Beyond that, the project's tumor analyses identified a number of new potential cancer subtypes based on proteomic data, demonstrating that proteomic analysis can, in fact, add information to that provided by genomic techniques. For instance, Vanderbilt researcher Dan Liebler's analysis of 95 colorectal cancer tumors identified five proteomic-based subtypes for the disease. Likewise, an analysis of 105 breast tumors by Broad Institute scientists found that proteomic and phosphoproteomic-based subtypes recapitulated RNA-based classification while also adding a new subtype, said Broad researcher Phillip Mertins. An analysis of 96 ovarian cancer tumors led by PNNL researcher Karin Rodland found four proteomic subtypes and four phosphoproteomic subtypes that appear to correlate well with transcriptomic subtypes, as well as one apparently new proteomic subtype and two new phosphoproteomic subtypes.
These results track with previously published proteomic and phosphoproteomic analyses of TCGA tumors, particularly those done by MD Anderson researcher Gordon Mills as part of the TCGA project. For instance, in a 2012 Nature paper, Mills and his colleagues identified a distinct proteomic-based cancer subtype not detected by the study's genomic analyses, that, he recently told ProteoMonitor, has since been shown to "be extremely robust across multiple datasets" (PM 7/8/2013).
One caveat Rodland noted with regard to the phosphoproteomic subtypes her group identified is that in one the p38 pathway is prominently involved. This, she said, is cause for caution because p38 is among the phosphoproteins identified by Broad researcher and CPTAC participant Steven Carr as being highly susceptible to alteration due to delayed sample freezing time.
To address concerns about whether the TCGA samples were frozen quickly enough upon collection for use in phosphoproteomic work, Carr has undertaken a set of experiments looking at the effect of warm and cold ischemia on the phosphoproteome. In work presented at this week's meeting, he found that while ischemia caused no change in the proteome itself at time points up to 60 minutes, between 5 percent and 24 percent of the phosphoproteome fluctuated up to six-fold under such conditions. As he noted, this means that the majority of the phosphoproteome is apparently stable, and so the consortium has decided to proceed with phosphoproteomic analysis of the TCGA samples. However, as Rodland observed with regard to her p38-related results, caution is warranted, particularly in the case of proteins Carr's work identified as affected.
Rodland's group also extensively investigated whether they could detect evidence of genomic variants at the protein level. Ultimately, they identified several thousand variants at the peptide level, some of which they now plan to build MRM-MS assays to, which will allow them to look for these variants in other tumor sets.
Rodland highlighted one such assay her team built to two isoforms of the Erg protein that she said could detect the isoforms in as few as 3,000 cells against a background of a million wild-type cells.
The other tumor characterization centers, including those at Vanderbilt, Johns Hopkins, Washington University, and the Broad Institute likewise found evidence of genomic variants in their proteomic data. Key to this effort was the use of tumor-specific reference databases built using the transcriptomic data previously generated by TCGA.
While proteins are, of course, the ultimate concern, mass spec-based proteomics relies on genomic reference databases for matching experimental spectra and peptide sequences to their corresponding proteins.
These databases are frequently updated, but they are nonetheless incomplete given the vast number of different protein forms in the human proteome and the fact that not all of these forms are necessarily expressed in every cell or tissue type or in every sample, as in the case of altered protein forms arising from genetic mutations.
These limitations, combined with the rise of next-generation sequencing, have, as in the case of CPTAC, led some researchers to create custom search databases specific to the samples they are investigating. Indeed, in a May interview, University of Wisconsin-Madison researcher Lloyd Smith, who is not associated with CPTAC, told ProteoMonitor that he thought the method could in the future become the standard for mass spec-based proteomics (PM 5/10/2013). Based on the research presented this week, the CPTAC effort represents one of the more extensive examples of this still relatively new, but growing, trend.
The symposium also highlighted a somewhat more established, but likewise growing, technique in MRM-MS, with Fred Hutchinson researcher Amanda Paulovich presenting work by her team building MRM-MS assays for monitoring phospho-signaling in the DNA damage repair pathway. She and her colleagues have developed a 75-plex immuno-MRM panel for measuring the pathway with a median assay coefficient of variation of 8 percent and median lower limits of quantitation of 12.6 femtomoles.
In addition to presenting her team's MRM panel, Paulovitch argued for the utility of MRM-MS more generally, echoing other targeted proteomics proponents like Swiss Federal Institute of Technology Zurich researcher Ruedi Aebersold (PM 9/27/2013) in asserting that the technique was well suited as a replacement for antibody-based protein quantitation.
As part of that goal, CPTAC is developing an MRM assay portal that researchers will be able to use to find MRM assays to proteins of interest. Slated to launch within the next few months, the portal currently features 596 well characterized MRM assays with another 39 immuno-MRM assays on the way, Paulovitch's Fred Hutchinson colleague Jeff Whiteaker said.