Skip to main content
Premium Trial:

Request an Annual Quote

Non-Cancerous Prostate Disease GWAS Uncovers PSA Level-Linked Variants

NEW YORK (GenomeWeb) - A team led by investigators at Amgen subsidiary Decode Genetics and the University of Iceland has identified more than a dozen genetic loci involved in non-cancerous prostate conditions that are increasingly common in aging men — including sites with additional ties to prostate-specific antigen (PSA) levels. 

As they reported in Nature Communications yesterday, the researchers did a genome-wide association study involving 20,621 affected individuals of European ancestry and more than 280,500 unaffected controls from Iceland and the UK, searching for variants that coincided with "benign prostatic hyperplasia" and related lower urinary tract symptoms (BPH/LUTS). In the process, they narrowed in on 23 variants with genome-wide significant ties to BPH/LUTS.

When the team took a closer look at these potential BPH/LUTS risk variants, which fell at 14 sites in the genome, it did not see genetic links to prostate cancer itself. Even so, the results indicated that 15 of the variants were implicated in serum PSA levels, providing a potential explanation for altered PSA levels in some prostate cancer-free men.

"We show that genetic correlation between BPH/LUTS and PSA levels is of a similar magnitude to the genetic correlation between prostate cancer and PSA levels," co-corresponding and senior author Kari Stefansson, president, chairman, CEO, and co-founder of Decode and a medicine researcher at the University of Iceland, and his colleagues wrote. "This underlines the complexity of interpreting the commonly applied PSA test, intended to screen for prostate cancer."

For the GWAS, the investigators considered 20,621 cases and 280,541controls using a combination of Illumina or Affymetrix array-based genotyping, variant imputation, and/or long-range phasing done with the help of 15,220 whole-genome sequences from Icelanders and/or data from the UK10K and Haplotype Reference Consortium efforts.

The group included 9,443 symptomatic men and 104,000 unaffected controls from Iceland, along with 11,178 BPH/LUTS cases and 176,541 undiagnosed controls from the UK Biobank.

After identifying one variant apiece at 14 loci with apparent ties to BPH/LUTS, the researchers used conditional analyses to unearth nine more secondary associations at eight of those loci.

From there, they dug into the BPH/LUTS-linked sites in the genome using functional annotation clues, bioinformatics, and available expression quantitative trait locus data. They also considered the genetic overlap between the BPH/LUTS-associated variants and predicted PSA levels relative to the prostate cancer, given the previously documented interactions between 15 of the BPH/LUTS-related variants and PSA levels in prior analyses.

"Our results demonstrate a strong genetic correlation between serum levels of PSA and prostate cancer, but still it is much weaker than for BPH/LUTS and PSA levels," the authors noted.

That finding prompted further efforts to come up with polygenic risk scores to tease out the interactions between serum PSA variation and the genetic variants that contribute to either BPH/LUTS or prostate cancer, using actual serum PSA measurements for more than 18,900 men from Iceland — analyses that underlined the largely independent genetic contributors to BPH/LUTS and prostate cancer, despite shared impacts on PSA in the blood.

Moreover, the authors noted that the current results might help in coming up with new treatment options for BPH/LUTS, since the standard treatment approaches come with the risk of side effects, and address symptoms of the conditions without offering a potential cure.

"In order to improve treatment, a better understanding of the basic disease-causing factors is needed," Stefansson and his co-authors concluded. "Our results provide several potential focus points for future research within this field."