NEW YORK (GenomeWeb) – A team led by investigators at Baylor College of Medicine, the Pacific Northwest National Laboratory, and Vanderbilt University has gained new insights into the biology and treatment outcomes observed in some colon cancers using a combination of proteomic and genomic profiling.
"We were able to not only confirm previously described colon cancer molecular markers but also to uncover new differences between proteins produced by tumors and normal tissue that may be worth further study," co-senior and co-corresponding author Bing Zhang, a molecular and human genetics researcher at BCM's Lester and Sue Smith Breast Center, said in a statement.
With other members of the National Cancer Institute's Clinical Proteomic Tumor Analysis Consortium (CPTAC), Zhang and colleagues used exome sequencing, arrays, RNA sequencing, microRNA sequencing, and proteomics to prospectively assess mutations, copy number, gene expression, miRNA profiles, protein expression, and/or phosphoproteomic levels in tumor and matched normal samples from up to 110 colon cancer patients — data they used to search for drug targets, potential tumor markers, and more.
"Analysis of the genes tells us what might possibly go wrong. But we don't know exactly what actually has gone wrong until we analyze the proteins," co-senior and co-corresponding author Tao Liu, a biological sciences researcher at PNNL, said in a statement.
The team's findings, published online today in Cell, pointed to previously unappreciated cancer drivers, along with candidate markers for the disease and potential avenues for targeted treatment. Using protein expression patterns, the group also got clues about why some tumors with high levels of microsatellite instability (MSI-H) do not respond to checkpoint immune blockade immunotherapy treatments as expected.
In particular, the results highlighted enhanced glycolysis pathways activity in the MSI-H tumors with relatively little infiltration by CD8+ T cells, the researchers explained, hinting that it may be possible to thwart checkpoint blockade immunotherapy resistance in some MSI-H colon tumors by targeting glycolysis.
For their analysis, the researchers focused on tumor and matched colon tissue and/or blood samples from 110 individuals with colon cancer, using exome sequences for 106 of the tumor-normal pairs to search for single nucleotide changes, small insertions and deletions, microsatellite instability, and other genomic changes expected to shift the tumors' protein output.
For the subsequent analyses, the team set those changes alongside proteomic/phosphoproteomic profiles produced by label-free shotgun proteomics and liquid chromatography-tandem mass spectrometry, as well as array-based copy number profiles for the exome-sequenced samples, and RNA and miRNA sequences.
With the genomic data, for example, the researchers saw recurrent retinoblastoma (RB1) gene amplification in tumor samples compared to matched normal tissue. Incorporating phosphoproteomic data pointed to increased phosphorylation of the protein encoded by RB1, which in turn interfered with the protein's ability to inhibit transcription factors that regulate CDK2 activation.
Based on those findings, they suggested that colon cancers driven by enhanced retinoblastoma phosphorylation might respond to targeted treatments aimed at dialing down the activity of the CDK2 protein product or related kinase enzymes.
Analyzing genomic and proteomic data in parallel also made it possible to clarify potential roles for SOX9 in colon cancer, the team reported. The gene is frequently mutated in colon cancer and was suspected of acting as a tumor suppressor. But the SOX9 protein product turned up at high levels in the tumor samples assessed for the study, suggesting it may drive rather than suppress tumor growth.
In addition to looking at features in colon cancer subtypes and uncovering a potential role for the glycolysis pathway in checkpoint blockade resistance, the investigators used the comprehensive datasets to get a look at other tumor alterations that might prompt immune cell infiltration or help in developing personalized vaccine-based immunotherapy in the future.
The team is making processed data from the study available to other research teams — along with computational tools for analyzing them further — through an interactive LinkedOmics web site. Raw genome sequence and/or proteomic data from the study have also been deposited in the Sequence Read Archive and a CPTAC Data Portal.
"[W]e anticipate broad usage of these datasets for new biological discoveries and therapeutic hypothesis generation," the authors concluded.