NEW YORK – A Johns Hopkins University research team has collected new evidence that future noninvasive tests for prostate cancer might rely on an analysis of RNA and metabolite patterns in urine, either in lieu of or in addition to the DNA biomarkers that have become so exciting in the field of liquid biopsy.
Writing in the journal Scientific Reports late last month, the investigators described early but promising evidence that a combined RNA and metabolomic signature in urine seems to be able to differentiate cases of prostate cancer from both normal controls and potentially confounding non-cancer conditions.
Although the group views its data so far as preliminary, the fact that its method could discern cancer from non-cancer cases so precisely seems to merit further validation, said Ranjan Perera, the paper's senior author and an associate professor of oncology at the Hopkins Kimmel Cancer Center.
Perera and his colleagues haven't developed a plan to do that yet and he said bringing the research forward may require interest from an outside party, either academic or commercial. But he and his colleagues do think their findings offer a strong proof of concept for future development of a noninvasive prostate cancer diagnostic.
Although there has been a noninvasive test method available for prostate cancer for decades now – the serum protein biomarker prostate specific antigen (PSA) – that analyte's modest sensitivity and specificity have frustrated the medical field, driving an acute need for more accurate but similarly facile diagnostic tools. Updated tests that use molecular methods to detect PSA at the RNA level have offered some enhancements, but there remains much room for improvement, the study authors wrote.
"With [tens of millions of] tests done in a year, that’s a huge amount of money … with a lot of false positives and people having to go through unnecessary biopsies," Perera said.
In recent years, the liquid biopsy field has focused on efforts that rely on the detection of circulating tumor DNA — either via mutations or through analysis of copy number changes or tumor-associated epigenetic signatures. Some industry-leading programs in this vein are even spearheaded by other groups at Hopkins, including the cancerSEEK method developed by Hopkins' Bert Vogelstein and others, which is now being commercialized by spinout Thrive.
Perera said his group wanted to take a different tack and look at RNA, something they are not alone in, at least in the prostate cancer field, which has seen several noninvasive RNA-based assays enter the market in recent years.
In their recent study, Perera and his colleagues analyzed urine samples from 64 patients with prostate cancer, 31 with benign prostatic hyperplasia and prostatitis diseases, and another 31 healthy people with none of these conditions.
First, they used an Illumina targeted RNA-seq panel, TruSeq RNA Access, to identify differentially expressed RNAs in 11 prostate cancer (PCa) urine samples, 12 normal samples, and one pooled set of three normal samples.
Interestingly, RNA alone was not enough to distinguish cancer from non-cancer samples. According to the authors, the 3,825 RNA transcripts that they detected in their initial 20 samples "readily but not perfectly" segregated between the normal and PCa groups. This indicated that "RNA expression analysis of urine liquid biopsies by itself was unlikely to reveal sensitive and specific PCa biomarkers."
However, when the investigators combined cancer-associated RNA with disease-specific metabolites, they could separate the cancer cases in their cohort.
According to the authors, global metabolomic profiling of prostate cancer urine samples versus normal controls revealed profiles that were "distinct and separate." Benign hyperplasia and prostatitis samples clustered together with closely overlapping metabolite signatures.
Further integrated analysis of metabolomic and transcriptomic data revealed a glutamate metabolism and tricarboxylic acid cycle node that was specific to prostate-derived cancer cells and cancer-specific metabolic changes in urine.
"If you look at the figure from the metabolomic studies, we could see really black-and-white results when you compare prostate cancer and health normal samples," Perera said. "This is the beginning, but we believe that definitely there is a lot of value here to move forward." One possibility for that might be obtaining funding for a larger study from the Prostate Cancer Foundation, he said. Alternatively, an outside biotech company might decide to take what the group was able to show in their initial study and advance it further, he suggested.
Importantly, Perera added, the team analyzed single-void urine samples obtained without prostatic massage — an uncomfortable procedure that is required for some of the existing RNA-based methods that are being advanced for noninvasive urine-based cancer detection.
In their research efforts, he and his colleagues now also plan to apply their RNA and metabolomics approach to other cancer types. They just received grant funding for a study investigating the detection of childhood medulloblastomas using cerebral spinal fluid.