Esophageal adenocarcinoma: when DNA methylation informs the treatment

from the Grady lab, Clinical Research Division

Barrett’s esophagus (BE) is an inflammatory condition of the upper gastro-intestinal tract observed as a consequence of gastro-esophageal reflux when metaplastic columnar epithelium replaces the stratified squamous epithelium in the esophagus. BE strongly predisposes to the development of esophageal adenocarcinoma (EAC), the sixth most common cause of cancer death worldwide according to the Mayo Clinic. Accumulation of genetic and epigenetic modifications play a significant role in tumor progression, and tumor heterogeneity between individuals is associated with various responses to treatment. Specifically, unique DNA hypermethylation profiles have been identified in EAC; however, it remains unclear whether these patterns already exist in BE. Understanding tumor heterogeneity based on genetic and epigenetic profiles, as well as evolution of these profiles between BE and EAC, would help refine treatment options for individual patients.

With this goal in mind, Dr. Ming Yu and Mr. Sean Maden working in the group of Dr. William Grady (Clinical Research Division) and in collaboration with Dr. Stachler from Dana-Farber Cancer Institute and Brigham and Women's Hospital at Harvard Medical School, characterized genetic and epigenetic profiles in patients with BE or EAC. According to Dr. Yu, “EAC at diagnosis is often already at an advanced stage that is not responsive to currently used therapies. Thus, there is an urgent need for the effective identification of optimal treatment strategies for EAC patients.” The results of their study were recently published in the journal Gut.

The study used integrative analysis of gene expression and methylation data, along with cell line experiments, to characterize subtypes of BE and EAC. “Our integrative analysis combined with functional in vitro studies provides novel insights into the biological mechanisms of EAC carcinogenesis. Our study in EAC cell lines reveals that the DNA methylation subtypes have distinct susceptibilities to conventional and targeted therapies, which, if validated in clinical studies, may be used to improve precision medicine of EAC patients in the future”, explained Dr. Ming. Samples were extracted from fixed tissues obtained from patients with confirmed non-dysplastic cancer-free BE (n=59), EAC (n=23), normal squamous esophagus (n=33) or normal fundus (n=9). DNA methylation analyses of 1515 of most variables methylation loci revealed four classes of EAC: high (HM; n=981 probes), intermediate (IM; n=181), low (LM; n=58) and minimal (MM; n=39) methylation patterns. Methylation levels in normal tissue samples were similar to the MM subtype. These observations were in accordance with results from EAC cohorts available from The Cancer Genome Atlas (TCGA) database. These four classes of methylation were also observed in BE samples. However, no correlation between the methylation status and gender, age or tobacco use could be found for BE or EAC.

Four subtypes of DNA methylation presenting distinct molecular characteristics were identified in EAC samples. Functional validation in EAC cell lines demonstrated the biological and clinical relevance of EAC methylation subtypes.
Four subtypes of DNA methylation presenting distinct molecular characteristics were identified in EAC samples. Functional validation in EAC cell lines demonstrated the biological and clinical relevance of EAC methylation subtypes. Figure provided by Mr. Sean Maden.

Specific mutations were analyzed by targeted sequencing of the 243 most commonly mutated genes known in gastro-esophageal cancer. Unique characteristic mutations were observed per methylation subtype. For example, ERBB2 mutations and amplifications were enriched in the HM EAC subtype, while MDM2 amplifications were observed exclusively in MM. Additionally, higher mutation rates were observed in the HM subtype (10.9 mutations per MB), relative to IM (6.87 mutations per MB), MM (5.54 mutations per MB) and LM (5.19 mutations per MB). Such characteristics were not observed in BE samples.

Several tumor suppressors were epigenetically silenced in the HM subtype, including PTPN13, a non-receptor type protein tyrosine phosphatase. To investigate the consequences of such gene repression on cell transformation, EAC cell lines with aberrant PTPN13 promoter methylation were treated with 5-azacytidine, a demethylating agent. The treatment decreased promoter methylation and increased PTPN13 mRNA expression, confirming the epigenetic control of PTPN13 gene expression. Decreased expression of PTPN13 using shRNA led to increased cell growth, colony formation and migration potential, all characteristics of transformed cells, confirming that PTPN13 repression could lead to increased transformation capabilities in vitro. Interestingly, this repression was associated with increased signaling of downstream oncogenic pathways Src1 and EGFR/ERBB2. These data suggest a novel molecular mechanism of activating oncogenic kinase pathways in EAC by EGFR/ERBB2 genomic alterations and promoter DNA hypermethylation of PTPN13, which is characteristic of the HM subtype cancers.

The results of this study strongly suggest that tumor subtypes based on epigenetic profiles are of clinical importance and should be considered for treatment optimization of EAC patients. When asked about future developments, Dr. Yu answered, “in order to develop highly specific and sensitive DNA methylation biomarkers to identify methylation subtype in BE and EAC, we plan to conduct a set of studies to determine if a small panel of CpGs could be used to accurately classify the methylation subtype status of an EAC case, which would be useful low-cost assay for clinical studies with large sample sets. Additionally, we would like to investigate the functional role of DNA methylation during EAC tumorigenesis and are performing functional studies on other gene candidates with the state-of-the-art CRISPR-Cas9 gene editing technology to gain deeper understanding of how aberrant DNA methylation affects EAC biology.”


Funding for this study was provided by the National Institutes of Health, the Degregorio Family Foundation, the Price Family Foundation and the Veterans Administration Puget Sound Health Care System.

Fred Hutch/UW Cancer Consortium faculty members Drs. William Grady and Andrew Kaz contributed to this research.


M Yu, SK Maden, M Stachler, AM Kaz, J Ayers, Y Guo, KT Carter, A Willbanks, TJ Heinzerling, RM O’leary, X Xu, A Bass, AK Chandar, A Chak, R Elliott, JE Willis, SD Markowitz, WM Grady. 2018. Subtypes of Barrett's oesophagus and oesophageal adenocarcinoma based on genome-wide methylation analysis. Gut journal. [Epub ahead of print].