N12009 and N12011: Developing biomarkers of colorectal neoplasia - quantification of aberrant CpG island methylation in human faecal DNA

Study Duration: October 2002 to September 2005

Contractor: Institute of Food Research and Newcastle University

Colorectal cancer develops from adenomatous polyps that emerge gradually from aberrant fields of mucosal epithelial cells. Biomarkers are needed for research into the role of diet in the early stages of this common condition. In a recent 12-month FSA project (N12001 & N12005), we set out to assess the feasibility of using faecal DNA residues to detect and measure aberrant methylation of CpG islands in the promoter regions of genes derived from colorectal mucosal cells. This approach has been successful. We have developed fast and inexpensive techniques that utilise small (ca.200 mg) faecal samples to detect CpG island methylation in DNA sequences derived from the epithelial cells of the colorectal mucosa.

This study will seek to build upon the previous work. New techniques will be used to assess the distribution of abnormal CpG island methylation of selected genes in a much larger group of volunteers from the general population, and from out-patients attending a gastroenterology clinic. In parallel with this task, it is proposed to use a subtractive hybridisation PCR technique to identify previously unknown CpG islands that become methylated during the earliest stages of the pathogenic process leading to colorectal neoplasia. The feasibility of using patterns of such methylated genes to characterise the various stages of the adenoma-sequence in otherwise healthy individuals will be assessed. The final objective is to provide FSA and other agencies with a protocol for the characterisation of incipient colorectal neoplasia by faecal analysis of mucosal CpG island methylation, for use in future epidemiological and dietary intervention studies.

Outcome/key results: Of the genes studied in patients from the endoscopy clinic, one in particular, called HPP1, proved to be of great interest because of the relatively high levels of methylation detectable in stool, and because the extent of methylation detected varied with the health of the patient, in a statistically significant manner. There was a higher level of methylation of HPP1 in the stools of patients with cancer, or with the precancerous polyps, which are common in the general population. Thus, HPP1 is a strong candidate for use as a faecal methylation marker, both to detect the presence of disease and to identify healthy individuals at increased risk of disease.

In the healthy volunteers, varying levels of methylation were detected in stool samples for all of the chosen genes, and a 'DNA methylation fingerprint' was prepared for each subject. Relatively high and statistically significant 'between-gene' correlations in the levels of methylation for some, but not all, genes in the data set were observed. A statistical technique called principal components analysis was used to simplify the data and explore the underlying relationships between genes. Some individuals had uniformly low levels of methylation whereas others showed substantial methylation for all the genes studied.

What it means and why it�s important

Human DNA is only present in very small quantities in stool and it is very difficult to extract in a form that can be analysed accurately. In this project, both faecal samples and tissues were analysed using two methods that, although well-established for conventional sources of DNA, have rarely been used for DNA from stools. The study confirmed the feasibility of using human DNA residues from stool samples to detect and quantify aberrant methylation of the promoter regions of several genes known to be involved in the development of bowel cancer. In particular, one gene, HPP1, may be particularly valuable for use as a biomarker, both to detect the presence of disease, and to identify healthy individuals at increased risk of disease.

In the second stage of the project, the team demonstrated that methylation of a range of different genes could be detected using stool samples from healthy individuals, and the feasibility and potential power of obtaining a methylation fingerprint for these genes. This approach may be used to establish a picture of the level of methylation across the whole colorectal mucosal field, which is likely to be different for each person and may provide information on individual risk of bowel cancer.