Assessing epigenetic silencing of a metastasis-suppressing gene in cell-free DNA provides prognostic information for non-small cell lung cancer, according to a recent study led by researchers from the University of Athens in Greece.
Published this month in the British Journal of Cancer, the study examined methylation in CpG islands of the promoter region in breast cancer metastasis suppressor-1 gene, or BRMS1. In plasma-derived cfDNA, but not in biopsy tissue, there was an association between methylated BRMS1 and reduced overall survival in NSCLC patients, while healthy individuals showed a complete lack of BRMS1 promoter methylation.
The study also demonstrated the use of two novel methylation-specific PCR methods to assess this methylation status from plasma samples.
In an email with PCR Insider, lead author Eli Lianidou, a professor of chemistry at the University of Athens, said this result was in harmony with the literature. Although the vast majority of patient primary tissue was highly methylated, "in plasma, not all samples were positive for methylation," and those that were correlated with worse overall survival, said Lianidou. "In many studies, so far, that compare information taken by using primary biopsies and cfDNA in plasma, there is not a 100 percent agreement," she said.
Lianidou and her colleagues examined methylation status of 57 NSCLC tumors and adjacent non-cancerous tissue, plasma cfDNA from 48 of these patients, as well as cfDNA from 74 advanced NSCLC cases and 24 healthy controls.
For bisulfite conversion they used a kit from Zymo Research, which Lianidou said gave them better results than other kits in prior comparison studies. The group then used two different methylation-specific PCR (MSP) methods. One was a Promega GoTaq hot start, visualized with ethidium bromide. The other was real-time MSP with locked nucleic acid (LNA) hydrolysis probes run on Roche's LightCycler 2.0.
"We decided to proceed with real-time MSP because this approach is much more sensitive, high throughput, and much better for clinical samples," said Lianidou. "In the LightCycler we are using closed tubes and there is no contamination danger, since we are not running PCR products in a gel electrophoresis."
The study used probes specific for the methylated and unmethylated BRMS1 promoter region. Notably, the researchers also used dual-labeled locked nucleic acid hydrolysis probes in their real-time MSP.
LNA probes are the proprietary formulation of Danish life sciences and diagnostics company, Exiqon, which now also provides custom oligos. Proligo, IDT and TIB-MolBiol also custom-design dual-labeled oligo probes containing LNAs, and Roche obtained a patent last year for probes to detect nucleic acid differences using LNA conjugated to fluorophores and quenchers. These types of probes have a fluorophore, such as fluorescein or Cy3, on one end, and a quencher, such as Iowa Black, Black Hole Quencher, or BlackBerry Quencher (BBQ), on the other. The LNA probe used by Lianidou had a 5' fluorescein tag and a 3' BBQ, a quencher manufactured by Michigan-based Berry & Associates. Lianidou's probe also contained an LNA modification in which the ribose ring is locked by a methylene bridge connecting the 2' oxygen atom and the 4' carbon atom of the nucleic acids, according to Exiqon's website.
After determining methylation status by MSP, the authors used the Kaplan-Meier method to correlate these data with disease-free interval, progression-free survival, and overall survival. The study showed no significant relationship between BRMS1 promoter methylation and prognosis when comparing NSCLC biopsy tissue and healthy tissue from the same patient. However, non-methylated BRMS1 promoter in cfDNA had significantly different Kaplan-Meier estimates for disease-free interval and overall survival in operable NSCLC, and overall survival in advanced cases.
Lianidou's group has published extensively on cancer biomarkers. Last month, the group authored a review in Clinical Chemistry where it contrasted circulating tumor cell DNA and cfDNA as potential non-invasive markers. Her group also published recently in Molecular Cancer Research on BRMS1 methylation in CTCs from plasma of patients with breast cancer. Finally, also in Clinical Chemistry, Lianidou and her team directly compared methylation of another gene, SOX14, in CTCs versus cfDNA in patients with solid tumors. In that study, they found SOX17 promoter methylation in CTCs and in matched cfDNA to be highly correlated.
In the current study, Lianidou and her colleagues isolated cfDNA from plasma using Roche's High Pure Viral Nucleic Acid kit. This kit does not have a lysis component, so it isolates and purifies only freely circulating nucleic acids, some of which are presumed to come from degraded tumor cells. "We use this kit since DNA isolated is not contaminated with DNA from circulating cells in plasma," Lianidou said.
In the long run, Lianidou said that it is also possible to obtain quantitative data about methylation levels using high-resolution melting analysis. "We and others have already shown that by using a combination of methylation-independent PCR and HRMA we can semi-quantify the percentage of methylated DNA in the presence of unmethylated sequence. We are now working on the development of a similar assay for BRMS1," she said.
Interestingly, the epigenetic signature revealed in this study may directly decrease the metastasis-suppressing gene expression in patients, providing a direct biological link to prognosis. Hypermethylation in a promoter CpG island has been shown to diminish BRMS1 mRNA and protein levels in NSCLC compared to non-cancerous lung tissue, according to a 2010 study in the Journal of Pathology. How this prognostic information may prove clinically useful is as yet unknown, but Lianidou suggested that further studies of BRMS1 as a non-invasive biomarker are warranted.