Exposure assessment in cancer epidemiological studies relies on measurable
intermediate molecular biomarkers with high sensitivity and specificity in order to
prevent common problems due to misclassification of exposure. Studies on the
early stages of carcinogenesis have helped to identify molecular changes that
are detectable in pre-cancerous lesions and that are thought to occur as the
result of specific exposures such as tobacco smoking. More recently, in vitro
evidence started to support the potential cancer-protective role of various micronutrients
acting through epigenetic and genetic mechanisms. Somatic mutations
in “master” cancer genes and modifications of epigenetic patterns in the promoter
region of specific genes involved in cell cycle, apoptosis or DNA repair may prove
good candidates of carcinogenic and dietary exposure even if the evidence that
these changes may be present and detectable in “normal” tissue are still scarce
(due in part to the practical and ethical difficulty to conduct experimental
prospective studies in healthy individuals).
In this thesis, I have developed two projects exploring the application of TP53,
KRAS, EGFR mutations and of DNA methylation changes as biomarkers of
exposure to tobacco smoking, in experimental and observational study designs.
Somatic mutations were analysed by dHPLC, ME-PCR, RFLP and sequencing
and DNA-methylation analysis was performed by pyrosequencing. Moreover,
somatic mutations were analysed in a prospective context of lung cancer
recurrence; also the capacity of dietary polyphenols and isothiocyanates to
modify methylation patterns in smokers was assessed in an intervention trial.
The results show that somatic mutations are good markers of different forms of
tobacco-related lung cancers but have limited short-term prognostic value, with
the exception of KRAS mutations in adenocarcinoma. Methylation data
suggested that a specific short-term dietary intervention may stabilize global
epigenetic (LINE1 DNA methylation) patterns in peripheral white blood cells.