Mining for Cancer Biomarkers in the Glycoproteome of Ovarian Proximal Fluids

The high mortality rate associated with a positive diagnosis of ovarian cancer, combined with the observation that ovarian malignancies are accompanied by a significant upregulation of sialylated proteins specific to the ST6GAL1 gene, makes the glycoproteome of ovarian proximal fluids a prime area for biomarker mining. Recently, researchers embarked upon a full-scale proteomic evaluation of the biomarker potential of this area.

Kuzmanov et al. (2012) analyzed samples from patients currently diagnosed with ovarian cancer, as well as supernatants from four cell lines derived from ovarian cancer cells. The researchers collected fluids from 13 ovarian ascites (three serous, three endometrioid, three mucinous, and four undifferentiated) and 14 malignant cysts (four serous, four endometrioid, three mucinous, and three undifferentiated), derived from patients with ovarian malignancies in FIGO stages III and IV. For non-malignant controls, they gathered fluid from benign ovarian cysts and peritoneal effusions. Kuzmanov et al. subjected these clinical samples, as well as conditioned media from the four ovarian cancer cell lines (OVCAR3, OVCAR5, ES-2 and TOV-112D), to sialylated glycoprotein/peptide enrichment using elderberry lectin affinity and hydrazide chemistry. Using an LTQ Orbitrap (Thermo Scientific) for tandem mass spectrometry, scientists filtered nearly 700 samples for high abundance proteins and other contaminants, leaving 333 identified proteins and 579 identified glycosylation sites for sialylation.

These results break down as follows: 151 proteins and 291 sites were found in cancer-associated fluids, 174 proteins and 315 sites were found in peritoneal effusions and benign cysts, and 222 proteins and 329 sites were identified in conditioned media from ovarian cancer cell lines. When comparing the proteins identified across all three sample types, the researchers isolated 21 candidate proteins derived from the cell line samples/cancer ascites and cyst fluid, or only in the cancer-associated fluids. Because these proteins are unique to fluids derived from patients with ovarian cancer, they stand out as potential biomarkers ripe for further evaluation.

The researchers subjected the identified proteins to further analysis for subcellular localization, secretion status, and presence of transmembrane domains, with Protein Centre software (Thermo Scientific). The results indicate that 288 of the proteins contain a signal peptide known to utilize the classical secretory pathway that ends in glycosylation; 175 of the identified proteins contained at least one transmembrane domain. The researchers observed that a striking 98.8% of the proteins (329 of 333) had a signal peptide and/or subcellular localization to the membrane, cell surface or extracellular space. This increases the protein’s likelihood of circulation and is thus a favorable characteristic for potential biomarkers.

The results of this study were in keeping with previous observations that only one gene, ST6GAL1, is consistently upregulated for various pathological tumours in comparison with both normal tissue and benign cysts. Most other sialyltransferase genes demonstrated either no change in expression or downregulation in pathological samples. Although researchers remain unclear on the exact role that ST6GAL1 plays in the progression of malignancy and metastatic activity, it is evident that malignant conditions are linked to measurable sialylation. The researchers assert that this information could lead to the development of hybrid, binary assays to measure the quantities of both proteins and glycans in ovarian samples, resulting in greater efficiency and accuracy in detecting and monitoring malignant transformation and progress for this tissue type. The candidate list produced in this study shows distinct promise in leading to specific biomarkers not unlike those already in use for prostate cancer.

References

Kuzmanov, U., et al. (2012) “Glycoproteomic identification of potential glycoprotein biomarkers in ovarian cancer proximal fluids,” Clinical Chemistry and Laboratory Medicine, doi: 10.1515.