PLoS ONE (Jan 2020)

A novel sensitive detection method for DNA methylation in circulating free DNA of pancreatic cancer.

  • Keiko Shinjo,
  • Kazuo Hara,
  • Genta Nagae,
  • Takayoshi Umeda,
  • Keisuke Katsushima,
  • Miho Suzuki,
  • Yoshiteru Murofushi,
  • Yuta Umezu,
  • Ichiro Takeuchi,
  • Satoru Takahashi,
  • Yusuke Okuno,
  • Keitaro Matsuo,
  • Hidemi Ito,
  • Shoji Tajima,
  • Hiroyuki Aburatani,
  • Kenji Yamao,
  • Yutaka Kondo

DOI
https://doi.org/10.1371/journal.pone.0233782
Journal volume & issue
Vol. 15, no. 6
p. e0233782

Abstract

Read online

Despite recent advances in clinical treatment, pancreatic cancer remains a highly lethal malignancy. In order to improve the survival rate of patients with pancreatic cancer, the development of non-invasive diagnostic methods using effective biomarkers is urgently needed. Here, we developed a highly sensitive method to detect DNA methylation in cell-free (cf)DNA samples based on the enrichment of methyl-CpG binding (MBD) protein coupled with a digital PCR method (MBD-ddPCR). Five DNA methylation markers for the diagnosis of pancreatic cancer were identified through DNA methylation microarray analysis in 37 pancreatic cancers. The sensitivity and specificity of the five markers were validated in another independent cohort of pancreatic cancers (100% and 100%, respectively; n = 46) as well as in The Cancer Genome Atlas data set (96% and 90%, respectively; n = 137). MBD-ddPCR analysis revealed that DNA methylation in at least one of the five markers was detected in 23 (49%) samples of cfDNA from 47 patients with pancreatic cancer. Further, a combination of DNA methylation markers and the KRAS mutation status improved the diagnostic capability of this method (sensitivity and specificity, 68% and 86%, respectively). Genome-wide MBD-sequencing analysis in cancer tissues and corresponding cfDNA revealed that more than 80% of methylated regions were overlapping; DNA methylation profiles of cancerous tissues and cfDNA significantly correlated with each other (R = 0.97). Our data indicate that newly developed MBD-ddPCR is a sensitive method to detect cfDNA methylation and that using five marker genes plus KRAS mutations may be useful for the detection of pancreatic cancers.