NEW YORK (GenomeWeb) – By combining genomic and proteomic characterizations of ovarian cancers, researchers from Pacific Northwest National Laboratory and Johns Hopkins University have gotten a deeper look into what goes awry in the disease.
The PNNL- and Hopkins-led team examined the cancer proteomes of 169 ovarian cancer patients whose tumors had also been analyzed by The Cancer Genome Atlas. As they reported in Cell today, their analysis of the 9,600 proteins they identified and associated clinical data led to the discovery of a number of copy alterations and homed in on key disease pathways, including ones linked to survival. Further, they found that certain protein acetylation events could present a way to stratify patients for treatment.
"Correlating our data with clinical outcomes is the first step toward the eventual ability to predict outcomes that reflect patient survival, with potential applications for precision medicine and new targets for pharmaceutical interventions," Hopkins' Daniel Chan said in a statement.
Researchers at PNNL and Hopkins analyzed 174 ovarian high-grade serous carcinoma samples and clinical data that had been collected by TCGA. Five samples were later re-classified as non-high-grade serous carcinoma.
PNNL researchers selected their 122 samples based on whether they might harbor a homologous recombination deficiency. Hopkins researchers, meanwhile, selected their 84 samples based on survival times, choosing the shortest-surviving and longest-surviving patients. Thirty-two samples were analyzed by both groups. Then through the Clinical Proteomic Tumor Analysis Consortium, the teams combined their efforts.
Using iTRAQ labeling and tandem LC-MS/MS,the researchers identified some 9,600 proteins in all tumors, though they limited their subsequent analyses to the 3,586 proteins found in all 169 ovarian cancer samples.
Since chromosomal instability is common in cancer, the researchers investigated how copy-number alterations affect protein abundance by comparing more than known 29,000 CNA loci to their global proteomic data. From this, they found that regions on chromosomes 2, 7, 20, and 22 influenced the levels of more than 200 proteins. These proteins were commonly involved in cell invasion and motility as well as in immune regulation — functions that are implicated in cancer, noted Chan, co-PI Karin Rodland of PNNL, and colleagues noted.
The researchers also developed models of how each of these four regions influence patient survival. A number of proteins were shared across the signatures including the cell-cell adhesion protein catenin B2 and Rho guanosine diphosphate dissociation inhibitor beta, which is involved in cancer invasion and migration.
The researchers then calculated a chromosomal instability (CIN) index for these samples and identified a set of 128 proteins linked with the CIN index. Proteins upregulated in tumors with a high CIN index were typically involved in chromatin organization while proteins upregulated in tumors with a low CIN index were often linked to cell death, they reported.
By estimating the average net phosphorylation of peptides of 2,324 proteins, the researchers gauged the activity level of various pathways and correlated that as well to patient survival times. They found that patients with short survival times had statistically significantly increased levels of ERK1 phosphorylation.
As homologous recombination deficiency — associated with BRCA1, BRCA2, and PTEN mutations — is linked to both improved survival and PARP inhibitor susceptibility, the researchers sifted through their proteomic data to uncover biomarkers that could be used to stratify patients for treatment. Through their analysis, they uncovered a sub-network of 30 proteins whose co-expression patterns differed in HRD and non-HRD patients.
A number of these proteins are involved in histone acetylation or deacetylation, they noted. In particular, they found that dual acetylation at K12 and K16 of histone H4 exhibited a significant difference between HRD and non-HRD samples. Previous studies, the researchers noted, had implicated H4 acetylation in the choice of DNA double-strand break repair pathways.
They also observed an enrichment of HDAC1, which appears to modulate DSB repair mechanism choice. This, the researchers added, could also help explain discrepancies seen in HDAC inhibitor response in ovarian cancer.
"Adding the information about the proteome on top of the genome provides an entirely new dimension of information that has enabled the discovery of new biological insights to ovarian cancer, while creating a valuable resource that the scientific community can use to generate new hypotheses about the disease and how to treat it," PNNL's Karin Rodland added in a statement.