Researchers at the Max Planck Institute of Biochemistry have completed a proteomic analysis of normal colon tissue, primary colorectal tumor tissue, and colorectal cancer metastases, finding significant differences between protein expression in normal and cancer tissue but little difference between primary and metastatic tumor tissue.
For the study, published in the current issue of Molecular Systems Biology, the researchers drew on a series of recent developments in mass spec techniques that allowed them to combine membrane protein measurement, laser microdissection, and the use of formalin-fixed paraffin-embedded samples in their analysis. They also developed and validated a new standards-free approach for estimating absolute protein copy numbers.
In the work, the researchers quantified roughly 7,500 proteins in patient-matched normal mucosa tissue, primary carcinoma, and nodal metastases. They found significant changes in expression of 1,808 proteins between the normal and primary cancer tissue.
As the authors noted, however, this figure likely understates the level of proteome remodeling between the tissue types given that quantitative differences between low-abundance proteins were likely underreported due to technical limitations. For high-abundance proteins, roughly half underwent significant changes in expression – suggesting, the researchers said, "that about 50 percent of the proteome is changed between normal and cancer cells."
According to study author Jacek Wisniewski, a professor and researcher in Matthias Mann's Proteomics and Signal Transduction Group at Max Planck, this finding was initially surprising given that it showed considerably more remodeling than had been found in past transcriptomics-based studies.
However, he said, in a follow-up experiment comparing different types of CaCo-2 cells – a cell line derived from epithelial colorectal adenocarcinoma that can also be cultured to resemble the normal enterocytes of the small intestine – the researchers found similarly extensive proteomic differences between cancer-like and enterocyte-like cells, supporting their tissue-based findings.
As their CaCo-2 analyses demonstrated, "normal enterocytes and cancer [cells] are in fact totally different cells," Wisniewski told ProteoMonitor. "Why shouldn't they be different by [a large percentage] of their proteins like this?"
Perhaps more surprising was the fact that the researchers found very little difference in protein expression between the primary colon tumors and metastases.
"We expected to find a lot of differences" between these tissue types, Wisniewski said, "but we were unable to do it."
"There is a lot of literature [on this question]," he added. "Some people have observed some differences and others not. We don't claim that there is no difference, but using our approach we were not able to detect changes."
Wisniewski said, however, that the workflows and platform used in the study were still being optimized. The work made use of a number of techniques recently developed by Mann's group including sample preparation steps to improve analysis of membrane proteins; methods for mass spec identification and quantitation of large numbers of proteins from relatively small amounts of laser-microdissected sample; and techniques enabling mass spec analysis of proteins and post-translational modifications in FFPE tissue.
Wisniewski noted, in particular, that the laser microdissection – which enabled the researchers to isolate well-defined populations of tumor cells – and mass spec instrumentation were currently the limiting factors of the workflow.
"These are not so easy tasks," he said. "We are … now continuing our analysis with a larger number of samples, and we hope maybe in a year to show" a more extensive analysis.
"Because, of course, the 7,500 proteins we quantified in this study is not the complete proteome," Wisniewski added. "The complete proteome is larger and we hope to achieve those numbers in the future."
Improvements in mass spec platforms are likely to aid in the researchers' pursuit of expanded proteome coverage is. Wisniewski and his colleagues performed the tissue-based portion of the study on a Thermo Fisher Scientific Orbitrap Velos, but for their work in CaCo-2 cells, they used the newer Thermo Scientific Q Exactive.
"The Q Exactive gave us much better results," Wisniewski said, noting that it significantly upped the number of peptide and protein IDs that his team was able to make.
In the paper, the researchers also presented a new method for estimating absolute copy numbers of proteins per cell by comparing "individual [label-free quantitation] intensities with the total MS signal of the measured proteome" – a technique they titled total protein approach, or TPA. The approach eliminates the need for added standards, which are typically used to make protein copy number estimations.
To validate their technique, the researchers compared the copy numbers of 23 proteins calculated using both TPA and a conventional approach using labeled standards, finding good correlation between the two. The correlation, the authors said, was particularly striking given the fact that the two data sets were generated on two different mass spec instruments – an Orbitrap Velos for the standards-based method and a Q Exactive for the TPA data. This, they noted, suggests the TPA method could be used to estimate per-cell copy numbers in previously generated proteomic data sets.
Wisniewski said that he and his colleagues plan to continue to focus on colorectal cancer with the aim of optimizing their workflow and expanding their proteome coverage.
"What we'd like to do is make [the analysis] deeper and more profound," he said, and "to really complete the analysis of the proteomes of the enterocytes and cancer cells. That is where we're going now."
Wisniewski noted that, while the study's coverage was more extensive than previous proteomics efforts, the particular protein changes observed were not novel, for the most part. "We published a list of the [expression] changes, and we are pretty sure that all the changes we saw have been seen at least once" in previous studies, he said.
"So we are not really discovering [colorectal cancer] biomarkers or anything like this in this case," he said. "This is more a systematic point of view… a description of the system. And I think an understanding of the system is probably more important than the short-term benefit of finding [potential] biomarkers."