Mass spectrometry-based proteomics is a molecular biology tool that precisely quantifies thousands of proteins from complex samples and can be used to generate quantitative protein profiles. It is important to consider the proteomic landscape of cells, with particular interest in cancer cells, in order to fully characterize molecular pathways.

Downstream transcriptomics often misses the capture cancer mutations, a phenomenon that occurs because while gene transcript or messenger RNA (mRNA) levels often correspond to total protein levels for a given gene, there is also widespread decoupling between proteins and their mRNAs due to differences in translation. Moreover, proteins are further modified through post-transcriptional modifications -- such as the phosphorylation of kinases that regulate cancer-signaling pathways.

In the new study, researchers from Baylor College of Medicine sought to combine proteomic data with transcriptomic data to obtain a more comprehensive view of the complex regulatory mechanisms that underlie cancer. They hypothesized that aggressive cancers differ due to altered pathways manifested at the protein level, and that the differences can help in identifying functional therapeutic targets.

"We analyzed protein data that included tens of thousands of proteins from about 800 tumors including seven different cancer types -- breast, colon, lung, renal, ovarian, uterine and pediatric glioma -- made available by the Clinical Proteomic Tumor Analysis Consortium (CPTAC) mass-spectrometry-based proteomics datasets," said author Chad Creighton, PhD, professor of medicine and co-director of Cancer Bioinformatics at the Dan L. Duncan Comprehensive Cancer Center at Baylor.

In total, 14,586 proteins and 44,763 phosphoproteins were detected across the seven cancers, an average of 9,700 proteins per cancer. Generally, the researchers found hundreds of differentially expressed proteins and phosphoproteins in higher grade (histologic parameter assigning the degree of differentiation of the cancer cells) or stage (clinical parameter indicating how extensively the tumor has spread outside of its site of origin) cancers.

They found that the results from protein analysis were often in alignment with corresponding mRNA analysis. However, proteomic patterns were often not observable in transcriptomic (gene transcript) patterns or vice versa. This indicates that proteomics data may capture biological information that is missed in gene expression profiling studies (quantification of RNA transcripts).

Cancers are the result of DNA damage arising from single-nucleotide variants. Some of the proteins associated with grade or stage identified in the analysis reflected changes in somatic copy number alterations in the cancer genome.

The associations of a subset of the proteomic-grade signatures with copy number alteration patterns may represent proteins that are selectively amplified or lost in the more aggressive cancers. For instance, lower-expression proteins associated with higher-grade cancers were characterized with genes frequently lost.

In contrast, higher-grade cancers had protein levels associated with increased metabolic pathways including glycolysis, lipid synthesis, the Krebs cycle, and the Warburg effect.

Functional study of protein kinases in aggressive cancers

The team also sought to provide proof of concept that identifying overexpressed kinases using proteomic analysis of tumor tissues is a useful approach for the identification of novel therapeutic drug targets.

"That's exactly what we were able to do with this new, very powerful dataset," said author Diana Monsivais, PhD, assistant professor of pathology and immunology at Baylor. "We focused on the uterine cancer data for which the computational analysis identified alterations in a number of proteins that were associated with aggressive cancer. We selected protein kinases, enzymes that would represent stronger candidates for therapeutics."

The team examined the uterine data for potential targets for functional studies in uterine endometrial cell lines. From a set of 347 protein kinases, they selected four to investigate in functional studies.

These protein kinases, along with more than 20 others, were associated with higher grade uterine cancer as well as other types of aggressive cancers. They found that manipulating the expression of some of the kinases reduced the survival or the ability to migrate for some uterine cancer cells.

Specifically, they found that mitogen-activated protein kinase kinase kinase 2 (MAP3K2), which activates the mitogen-activated protein kinase kinase 5-extracellular signal-regulated kinase 5 (MEK5/ERK5) cell signaling pathway, has a unique role in cell migration, a process by which cancer spreads to other tissues.

"Our experiments provided proof-of-concept that proteomics analysis is a useful strategy not only to better understand what drives cancer, but to identify new ways to control it or eliminate it," Monsivais said.

Creighton said that historically researchers have only generated transcriptomic data.

"Looking at the protein data itself, which is made available by the CPTAC, enables researchers to extract a new layer of information from these cancers," Creighton said. "In this study, we compared mRNA and protein signatures and, although in many cases they overlapped, about half the proteins in the proteomic signatures were not included in the corresponding mRNA signature, suggesting the need to include both mRNA and protein data in cancer studies."

Do you have a unique perspective on your research related to proteomics or molecular biology? Contact the editor today to learn more.