NEW YORK (360Dx) – A team led by researchers at the Australian Prostate Cancer Research Centre (APCRC) has found that a single-nucleotide polymorphism in the kallikrein-related peptidase 3 (KLK3) gene that codes for prostate-specific antigen (PSA) could impact the accuracy of PSA tests.
In a study published last month in Clinical Chemistry, the scientists found that the SNP affected PSA stability, enzymatic activity, and glycosylation, and might also affect binding of the protein to some of the antibodies used in clinical PSA tests. They also found that subjects carrying the variant had significantly lower total PSA levels and higher free-to-total-PSA ratios than subjects with only the wildtype form of the gene.
Jyotsna Batra, a researcher at the APCRC and senior author on the study, said that she and her colleagues are now interested in developing assays that could account for this variant and improve the accuracy of PSA tests.
The variant in question, the rs61752561 SNP, has been linked to prostate cancer risk by previous genome-wide association studies, but the biology underlying the variant's link to prostate cancer had been little explored, Batra said.
To get at the mechanisms by which the variant might affect a person's risk of developing prostate cancer, the researchers performed a series of structural and functional analyses to compare the variant form of the protein to its wildtype form.
Among their main findings was the fact that the SNP created a new glycosylation site that could affect its function and might also alter its binding to certain antibodies used in commercial PSA tests. Additionally, the SNP appeared to alter the interaction of PSA with known binding partners, which, Batra noted, could affect measurements of free versus total PSA, the ratio of which is often used by clinicians to assess a patient's risk of having prostate cancer following an elevated PSA reading.
Batra added that changes in PSA glycosylation due to the SNP could also change the protein's localization, which might also impact the accuracy of PSA testing, though she noted that the Clinical Chemistry study didn't address this question.
"We haven't shown that this [additional] glycosylation specifically changes [PSA's] localization, but our hypothesis would be that if the localization is changing in such as way that [the protein] is not secreted as much as it should be secreted, then the PSA test would be affected," she said.
She and her colleagues did demonstrate that the additional glycosylation could potentially affect binding of the variant to anti-PSA antibodies. Examining crystal structures of PSA and three antibodies targeting the protein, they found that one of the three binds to PSA near the glycosylation site created by the SNP, which, they noted, "may affect antibody binding." The other two antibodies bind at locations farther from this glycosylation site, "suggesting that glycosylation at this site would not affect antibody binding."
They also demonstrated differences in levels of free-to-total PSA using a series of experiments in which they spiked variant and wild-type protein into female serum and measured them with commercial tests from Siemens and Beckman Coulter.
Using the Siemens Immulite assay, the researchers found that for the wildtype protein, free-to-total PSA ranged from 62 percent to 72 percent across the three concentrations tested. The same measurement ranged from 34 percent to 37 percent for the variant PSA. Using Beckman Coulter assays for free and total PSA, the researchers found that ratios ranged from 69 percent to 71 percent for wildtype PSA and 46 percent to 49 percent for the PSA variant. The authors suggested these differences might stem from differences in the binding of the wildtype and variant forms to their interaction partners, which could alter the levels of free versus total PSA present in a sample.
Using patient samples from the Malmö Diet and Cancer cohort, the researchers looked at the PSA levels of 2,647 patients without a diagnosis of prostate cancer, 2,415 of whom were homozygous for wildtype PSA, 227 of whom carried both a wildtype and variant allele, and five of whom were homozygous for the variant PSA. Total PSA levels were significantly lower in the men carrying the variant than in the men with two wildtype alleles, while free-to-total PSA levels were significantly higher in the men carrying the variant.
As the authors noted, the directionality of the variant's effect on free-to-total PSA differed between the recombinant protein spike-in experiment (where the variant was associated with lower ratios) and in the Malmö cohort (where it led to higher ratios). They suggested that this could be due to "the different forms of PSA and their complexed forms that exist in blood," adding that this observation "needs further clarification."
Regardless, though, they noted, both findings indicate that the rs61752561 SNP could "affect clinically measured PSA concentrations."
Batra said that previous work suggests that PSA variants might also be present among PSA sold by reagent vendors for use as standards and calibrators in PSA assays, which, she noted, could affect the performance of those assays.
In a 2015 study in the Journal of Proteome Research, scientists at Indiana University and Texas Tech University used LC-MS/MS to characterize PSA standards produced by LeeBio, Sigma-Aldrich, and EMD Millipore. They found that each of these PSA standards contained some level of the rs61752561 SNP variant PSA, with the Sigma product made up of 0.2 percent of the variant form, the LeeBio product made up of 17 percent of the variant form, and the EMD product consisting of 29 percent variant PSA.
"Even in PSA standards there is a different mix of PSA [forms]," Batra said. "So, we really need to better standardize the baselines of the clinical PSA measurements."
She said that she and her collaborators now plan to develop glycan-based assays to help distinguish between the two forms of PSA in patient samples, which could help improve testing. While the glycosylation site on wildtype PSA is typically fucosylated, the glycosylation site created by the rs61752561 SNP is modified by mannose glycans.
This means the researchers can potentially use lectin-based assays to measures the two forms of glycosylation and distinguish between wildtype and variant PSA in patient samples.
"Measuring [those] glycosylations… could improve the sensitivity and specificity of the current PSA test and could be achieved very quickly," she said.