Metabolomics marks out prostate cancer
Friday, 13 February, 2009
The metabolite sarcosine has been identified as a potential urine-based biomarker for prostate cancer progression.
In a study published in Nature, researchers from the University of Michigan Medical School screened 1126 metabolites from clinical samples of prostate cancer from tissue, urine and serum.
The metabolomic screen was able to distinguish between benign cancer, clinically localized cancer and metastatic cancer, the researchers say. They have identified 87 metabolites that are able to distinguish between benign tissue and cancer, a subset of six of which were significantly increased in metastatic cancer.
Of those six – including uracil, kynurenine, glycerol-3-phosphate, leucine and proline – sarcosine was chosen for further study. Levels of sarcosine were markedly higher in metastatic samples than in localised cancer samples, and were undetectable in benign samples.
Further study found that sarcosine levels correlated with cancer cell invasiveness and, remarkably, that the mere addition of sarcosine to benign prostate epithelial cells changed their phenotype to an invasive one.
They also found that sarcosine levels in urine had a predictive value, meaning urine metabolite analysis could be used in addition to the traditional PSA blood test to monitor disease progression.
The researchers, led by Arul Chinnaiyan, plan to investigate the other potential biomarkers in the study, in the hope of developing a panel of metabolites for diagnostic use.
Pending patents on the technologies have been licensed to a metabolomics company called Metabolon, in which several of the authors have a financial interest.
Could this 'nanobone' material allow children to regrow bone?
Australian scientists believe they have found an alternative to invasive bone graft procedures in...
The last mile of life sciences innovation
How do we translate technological possibility into meaningful real-world impact?
Machine learning uncovers patterns in DNA methylation
A new machine-learning method identifies differentially methylated DNA regions without sample...
