N02018: Vitamins C, E, DNA malonaldyde adducts in rectal biopsies and lymphocytes

Study Duration: February 2001 to January 2003

Contractor: MRC Dunn Human Nutrition Unit, Cambridge

Diet is thought to be important in protecting against colorectal cancer, but the mechanisms underlying why diet should damage DNA leading to cancer are uncertain. Markers of DNA damage relevant to cancer are needed to strengthen epidemiological studies that aim to estimate risk of cancer from diet. DNA adducts is one such group of biomarkers. In colonic biopsies we had previously measured a DNA adduct that forms in response to oxidised lipids, and which seemed to be associated with increased risk of colonic polyps (a precurser of colon cancer). We wanted to see if these levels in the colon were related to levels in blood, and if the blood levels were related to consumption of antioxidant vitamins C and E.

Rationale and objectives:
To determine if levels of the DNA damage marker (M1dG) in blood samples were related to levels in the colon, in order that blood levels of this adduct replace colonic biopsies in epidemiological studies. If blood and biopsy levels were related, blood levels in an ongoing prospective study of diet and cancer, could have been measured in subjects who developed cancer, and compared with subjects who did not. The ultimate aim was to link dietary intake to the marker, and then risk of colorectal cancer.

A method for measuring M1dG in blood was set up, and levels measured in individuals who had previously had their levels measured in colonic biopsies. The biopsies were obtained as part of a screening test for the presence of polyps in the large bowel The diet of these people was also assessed from food diaries, and the food tables associated with the food diaries were updated for vitamins E and C by 'infilling' missing values. Vitamin E and C, are antioxidants, thought to protect against lipid peroxidation and hence the formation of the M1dG adduct.

• Updating of the food tables to infill missing vitamin E and C data reduced the statistical significance of a previously found protective effect of vitamin E against the development of colorectal adenomas.
• Although the method for M1dG seemed to work for standards and quality controls, levels in blood were very low and hence difficult to measure accurately. A blind Inter-laboratory trial of blood samples between three laboratories using this method showed that there was no agreement in results. There was also no correlation between levels introduced in a cell culture line and those measured, despite the fact that existence of the adduct was shown by immunohistochemistry in the cell line.
• There was no agreement between levels of M1dG in the blood and biopsy samples.
• There were no associations with blood M1dG and diet and plasma levels of vitamin C probably due to the low levels present.

M1dG occurs at very low levels in human blood so that in blood it is an unreliable indicator of DNA damage and is unlikely to be a suitable risk marker for cancer. The infilling has shown that incomplete food table data can lead to biased results when comparing the risks of disease in relation to diet. When evaluating risks indicated by epidemiological studies, FSA should take this into account.

Contact: Peter Sanderson
Tel: 020 7276 8920
Email: peter.sanderson@foodstandards.gsi.gov.uk