The etiology of chemically induced cancer is thought to involve the co
valent binding of carcinogens to DNA (adducts) leading to mutations in
oncogenes or tumor suppressor genes, and ultimately to tumors. Thus,
the DNA-carcinogen adduct has been used as a measurable biochemical en
dpoint in laboratory studies designed to assess carcinogen exposure, c
arcinogen metabolism, mutagenesis, and tumorigenesis. Unfortunately, t
he significance of adducts in the etiology of human cancer is still un
clear. This is partially due to the difficulty detecting adducts at ca
rcinogen exposures relevant to humans, which are often orders of magni
tude lower than animal model exposures. The relationship between adduc
ts and higher biological effects is also not known at low doses. We ha
ve been assessing the DNA damage caused by exposure to heterocyclic am
ine carcinogens in the diet. Using the technique of P-32-postlabeling
in combination with accelerator mass spectrometry, we have determined
that DNA adduction in rodents decreases linearly with decreasing dose
from the high doses used in typical cancer bioassays to the low doses
relevant to human exposures. For a given tissue, adduct levels are cor
related with dose, but the level of DNA modification by carcinogens is
tissue-specific and does not completely correlate with tumor site. Th
is lack of correlation may be due to differences in adduct formation a
nd repair rates among tissues. Comparison of carcinogen metabolism rou
tes between rodents and humans also indicates that species differences
could influence the amount and type of damage resulting from exposure
to these carcinogens. The use of model systems to study dosimetry, sp
ecies differences in adduction, and role of adducts in mutation will u
ltimately lead to a better understanding of the significance of adduct
s in human disease. This should eventually allow the use of adducts as
biomarkers for estimating carcinogen exposure and individual suscepti
bility. (C) 1993 Wiley-Liss, Inc.