The role of methylated DNA biomarkers in detecting and monitoring colorectal cancer

Abstract

Colorectal cancer (CRC) is one of the commonest cancers in the world with rising incidence and mortality. Non-invasive early diagnosis and monitoring during treatment are of major importance, demonstrating the need for tests that could be repeatedly performed with high accuracy. DNA methylated biomarkers have been increasingly used for the early detection and monitoring of cancer. The methylated septin 9 (SEPT9) has been approved by the Food and Drug Administration for the screening of CRC. This study evaluated the role of detecting methylated SEPT9 in blood, for detection of early colorectal neoplasm including adenoma, and post-operative monitoring of CRC patients after surgery. In addition to conventional quantitative PCR, we specifically explore the role of the more sensitive droplet digital PCR (ddPCR) for this purpose with comparison to the conventional carcinoembryonic antigen (CEA). Other promising methylated biomarkers were also explored.

Through conventional qPCR, SEPT9 with 1/3 algorithm (at least one positive result of three PCR reactions) had higher sensitivity and accuracy than CEA in detecting CRC (73.2% versus 48.2% and 72.4% versus 58.9%, both P < 0.01). The combination with CEA could improve the sensitivity and accuracy of SEPT9 with 1/3 algorithm to 85.7% (P = 0.039) and 74.7% (P = 0.532), respectively. Both higher SEPT9 and CEA were associated with higher tumor staging (P < 0.05). On post-operative surveillance, the 2/3 algorithm of SEPT9 had higher rates of turning from positive to negative than the 1/3 algorithm and CEA (P < 0.01) in 6-month. Higher pre-operative positive rates of SEPT9 with 2/3 algorithm could predict new metastases after surgery (P = 0.042). Increasing quantitative SEPT9 levels between two post-operative time points were also associated with new metastases and cancer-associated mortality (both P < 0.05).

With the use of ddPCR, SEPT9 methylated ratio (methylated to unmethylated concentration/copies) and abundance (fraction of methylated concentration/copies in total) were higher in CRC patients than normal subjects (both P < 0.05). Notably, SEPT9 abundance increased in patients with proximal than distal cancer (P = 0.017). In 3-month post-operative samples, there was a decrease in methylated ratio and abundance (P = 0.053 and 0.005, respectively), especially the SEPT9 abundance in stage III cancer and left-sided cancer (both P < 0.01).

In addition to SEPT9, through review of published articles and analysis of The Cancer Genome Atlas (TCGA), there were many other potential methylated biomarkers for CRC identified, including EYA4, TWIST1, and BCL2. Among all, BCL2 had the best performance by ddPCR, with higher methylation in CRC cell lines than cancer-adjacent normal tissues and decreased methylated ratio and abundance in plasma samples at 3-month after CRC surgery.

In conclusion, SEPT9 could be used in both screening and monitoring through conventional PCR. The newly developed ddPCR could also be used for methylation detection with a lower amount of plasma samples, but more clinical studies are needed to verify its accuracy for this purpose.