Skip to main content
Premium Trial:

Request an Annual Quote

Cancer Methylome Project Seeks to Harness DNA Methylation for Prognostic Testing

Premium

NEW YORK (GenomeWeb) – Researchers from the University of Texas Health Science Center are looking to better understand the role of methylation in cancer progression, prognosis, and survival through a 1,000-patient sequencing study called the Cancer Methylome System.

The team, which recently described its methylome sequencing protocol in the Journal of Visualized Experiments, uses a technique known as methyl-binding DNA capture sequencing (MBDCap-seq).

The protocol, which targets DNA bound to methyl binding domain proteins for sequencing, has evolved over the years from its original description in 2005, in which methylation was analyzed via PCR, to more recent iterations that use next-generation sequencing.

The University of Texas Health Science Center team, in collaboration with researchers at Indiana University and elsewhere, has now used the technique to analyze methylation in more than 700 patient samples across breast, prostate, liver, ovarian, and other cancers, lead author Rohit Jadhav told GenomeWeb.

The goal of the project is to "investigate how DNA methylation can be used for predictive or prognostic testing," he said. The Cancer Methylome System project is coordinated by the San Antonio-IU Integrative Cancer Biology Program and receives funding from the National Cancer Institute and the National Institutes of Health. Tim Huang, chair of molecular medicine at the University of Texas Health Science Center, is the principal investigator of the project.  

The researchers are focusing especially on understanding the epigenomics of hormone resistant cancers. Thus far, Jadhav said, the group's work has identified prognostic methylation signatures in endometrial cancers, prostate cancer, and estrogen receptor positive breast cancer.

The researchers are continuing to look across patient cohorts for additional methylation signatures, validating their findings, and looking for ways to apply their findings to patient care.

MBDCap-seq is similar to ChIP-seq except that it targets the methyl-CpG binding domain protein, MBD2. The protein binds to methylated CpGs, in particular methylated cytosines that are in close proximity to each other, and therefore biologically relevant, making these regions a good target for sequencing to identify important areas of methylation. In addition, although bisulfite sequencing is the gold standard for analyzing methylation, it is cost prohibitive for large numbers of patient samples, Jadhav said. However, the group does use targeted bisulfite sequencing as a means to validate at a single-base resolution the methylated sites identified from MBDCap-seq, he said.

MBDCap-seq has a resolution of around 100 to 200 base pairs, which is good for a broad survey of the genome, but bisulfite sequencing can validate at the single-base level, he said. Other techniques, like array-based bisulfite analysis, are even more cost effective than MBDCap-seq, but have lower coverage. As such, MBDCap-seq is a good balance between cost and coverage, according to Jadhav.

The researchers have also developed a customized bioinformatics pipeline for MBDCap-seq data, using a technique known as LONUT, for locating non-unique matching tags. The team originally developed the pipeline for ChIP-seq experiments but has now applied it to its MBDCap-seq experiments. The incorporation of LONUT increases the number of DNA reads that are analyzed, Jadhav said. Nearly 60 percent of raw reads are thrown out because the algorithms ignore non-unique tags. LONUT reanalyzes those non-unique tags and tries to uniquely match them to a location in the genome to improve detection of the methylated sites.

In the JoVE publication, the researchers described using MBDCap-seq to study 77 breast cancer tumor samples, 38 breast cancer cell lines, and 10 normal breast tissue samples. They identified more than 13,000 CpG islands in promoters and found that nearly 20 percent were differentially methylated in tumors compared to normal tissue. More than half of the 6,959 CpG islands in intragenic regions were differentially methylated and 28 percent of the 4,847 gene promoters without CpG islands were differentially methylated.

Jadhav said the researchers are following up on a number of their findings. For example, the team has identified a methylation signature in prostate cancer patients and is now looking whether it can identify the signature in circulating tumor cells or circulating tumor DNA from either urine or blood to better predict disease recurrence. In order to design a specific test, Jadhav said, the researchers would likely not use MBDCap-seq, which analyzes the whole genome, but instead target the specific methylation signature identified in the research studies. So far, their work has found that methylation signatures tend to be unique to a specific cancer type, which could potentially make methylation a better marker for noninvasive detection than DNA variants.

Also,  it is currently difficult to determine what body site ctDNA originated from. But if tumors from different body sites have different methylation signatures, looking at methylation could add a level of confidence to the analysis, he said.

Other researchers have also found that cancer methylation signatures are specific to the tumor's site of origin, including Dennis Lo's group from the Chinese University of Hong Kong. In addition, a group from the Hebrew University-Hadassah Medical School recently demonstrated that it could use methylation signatures to trace cell-free DNA back to its tissue of origin, which it said could be used for a number of applications, including assessing a tissue's response to injury and even early diagnosis of diseases such as type 1 diabetes or neurodegenerative disease.