Skip to main content
Premium Trial:

Request an Annual Quote

U of Toronto Team Develops Proteomic Test for Prostate Cancer Prognosis


NEW YORK (GenomeWeb) – A team led by researchers at the University of Toronto has developed a six-protein panel capable of distinguishing between organ-confined and extraprostatic prostate cancer.

The panel, presented in a study published this week in Nature Communications, could help physicians distinguish between patients with non-aggressive disease and more aggressive forms that could require more significant interventions, Thomas Kislinger, a University of Toronto researcher and senior author on the paper told GenomeWeb.

The study also demonstrated that urine containing expressed prostatic secretions (EPS) produced by a digital rectal exam is a suitable clinical sample for a mass spec-based proteomic test.

Kislinger and his colleagues began their discovery work in direct EPS, which, they theorized could be a good source for biomarker discovery given that it is fluid secreted by the prostate. In previous work they did deep proteome profiling of this fluid followed by a quantitative analysis to identify proteins differentially expressed in men with organ-confined versus extraprostatic disease.

They identified 133 differentially expressed proteins, after which the question became: Could these same proteins be accurately measured in EPS urine?

"The problem with direct [EPS] is that it is nice as a discovery fluid, but it is not very clinically relevant," Kislinger said. This is because direct EPS can typically only be collected just prior to radical prostatectomy.

EPS is present in urine following a digital rectal exam (DRE), though, making it a relatively straightforward sample to collect. "So the hypothesis was that the DRE expels prostatic secretions into urine," Kislinger said, and EPS proteins can then be measured in the urine.

To measure these EPS urine proteins, the researchers developed selected-reaction monitoring mass spec assays to the 133 differentially expressed proteins identified in previous work. They found that they were able to quantify them in the post-DRE urine. Additionally, Kislinger noted, they were not able to detect most of them in regular urine, which bolstered the hypothesis that DREs did stimulate release of EPS proteins into the urine.

EPS urine is not conventionally collected for biobanking purposes, but that has begun to change in recent years Kislinger said. One driver of this change is the development of the non-coding RNA PCA3 as a marker of prostate cancer. This marker, which has been approved by the US Food and Drug Administration as a tool for helping doctors determine whether to do a repeat biopsy of a patient they suspect of having prostate cancer, is measured in EPS urine.

Using their SRM assays, the Toronto researchers evaluated the 133 differentially expressed proteins in a set of 74 patients, using the results from these patients to identify 34 candidate biomarkers. They then measured these 34 proteins in a separate 207-patient cohort and, using a machine-learning analysis, developed a six-peptide panel that was able to distinguish between organ-confined and extraprostatic prostate cancer with an area under the curve of .74.

This was an improvement over the commonly used protein biomarker prostate-specific antigen, which had an AUC of .66, but, Kislinger said, it still falls short of the performance clinicians would desire.

He noted several approaches he and his colleagues might consider in an effort to boost the panel's performance. For instance, combining it with genomic analysis of biopsy tissue might help.

Another possibility would be to increase the number of peptides per protein used for quantification.

"Right now most of our proteins are quantified by one peptide," he said. "You could add additional peptide sequences and see if by doing that you can improve your quantification and performance."

They might also do a discovery project looking at EPS urine enriched specifically for glycopeptides, Kislinger said, noting that because secreted proteins are often N-glycosylated, glycopeptide enrichment can be an effective approach to protein biomarker discovery.

In fact, this is the approach of a protein biomarker firm currently at work in the prostate cancer space. ProteoMedix, a protein diagnostics company spun out of the lab of Swiss Federal Institute of Technology researcher Ruedi Aebersold, used glycoprotein enrichment combined with mass spec to develop a four-protein panel for determining whether patients with suspected prostate cancer should undergo a biopsy.

The company recently began clinical validation of the test. It is also developing a proteomic test for distinguishing between patients with aggressive and non-aggressive disease.

Kislinger and his colleagues are still in the relatively early stages of development for their assay, he said, noting that next they plan to test the panel in a new cohort of roughly 1,000 patients to try to replicate the results of the Nature Communications study.

He said that for this work the researchers will use parallel-reaction monitoring instead of SRM, adding that his lab has already changed over to PRM for much of its targeted protein quantitation work.

PRM is essentially a variety of data-independent mass spec in which the mass spectrometer, rather than analyzing the full range of a sample, is trained on a more targeted mass and time window. Compared to conventional SRM assays, the approach has various potential advantages.

For instance, because their analyzers are able to collect data on a wide range of ions, high-resolution machines could allow for easier assay development and better specificity. In a triple quad-based SRM assay, the first quadrupole isolates a target precursor ion, which is then fragmented in the second quadrupole, after which a set of preselected product ions are detected in the third quadrupole. By contrast, PRM approaches use the upfront quadrupole of a Q-TOF or Q Exactive machine to isolate a target precursor ion, but then monitor not just a few but all of the resulting product ions.

Traditionally, SRM has been thought to offer higher sensitivity than PRM, but Kislinger said that his lab has found that the two approaches have equivalent sensitivity, particularly when using Thermo Fisher Scientific's Q Exactive HF, the most recent version of the company's Q Exactive instrument, a popular platform for PRM mass spec.

This echoes comments from Bhavinkumar Patel, a senior research scientist in mass spec reagents and protein biology at Thermo Fisher Scientific, who noted in an interview with GenomeWeb earlier this year that he and his colleagues had found PRM's sensitivity to be a match for SRM upon moving to the more powerful Q Exactive HF.

Kislinger noted in particular the ease of PRM assay development compared to SRM.

"This [Nature Communications] paper took four years, and the biggest part of it was building these SRM assays and checking every one to make sure the transitions are fine," he said. In recent ovarian cancer work, meanwhile, his labs was able to "in extremely fast time" build a 392-peptide PRM assay.

"So we get much quicker assay development and much higher multiplexing, and the sensitivity is very similar," he said.