Gene promoter methylation assayed in exhaled breath, with differences in smokers and lung cancer patients

BACKGROUND There is a need for new, noninvasive risk assessment tools for use in lung cancer population screening and prevention programs. METHODS To investigate the technical feasibility of determining DNA methylation in exhaled breath condensate, we applied our previously-developed method for tag-adapted bisulfite genomic DNA sequencing (tBGS) for mapping of DNA methylation, and adapted it to exhaled breath condensate (EBC) from lung cancer cases and non-cancer controls. Promoter methylation patterns were analyzed in DAPK, RASSF1A and PAX5beta promoters in EBC samples from 54 individuals, comprised of 37 controls [current- (n = 19), former- (n = 10), and never-smokers (n = 8)] and 17 lung cancer cases [current- (n = 5), former- (n = 11), and never-smokers (n = 1)]. RESULTS We found: (1) Wide inter-individual variability in methylation density and spatial distribution for DAPK, PAX5beta and RASSF1A. (2) Methylation patterns from paired exhaled breath condensate and mouth rinse specimens were completely divergent. (3) For smoking status, the methylation density of RASSF1A was statistically different (p = 0.0285); pair-wise comparisons showed that the former smokers had higher methylation density versus never smokers and current smokers (p = 0.019 and p = 0.031). For DAPK and PAX5beta, there was no such significant smoking-related difference. Underlying lung disease did not impact on methylation density for this geneset. (4) In case-control comparisons, CpG at -63 of DAPK promoter and +52 of PAX5beta promoter were significantly associated with lung cancer status (p = 0.0042 and 0.0093, respectively). After adjusting for multiple testing, both loci were of borderline significance (p(adj) = 0.054 and 0.031). (5) The DAPK gene had a regional methylation pattern with two blocks (1) approximately -215--113 and (2) -84-+26; while similar in block 1, there was a significant case-control difference in methylation density in block 2 (p = 0.045); (6)Tumor stage and histology did not impact on the methylation density among the cases. (7) The results of qMSP applied to EBC correlated with the corresponding tBGS sequencing map loci. CONCLUSION Our results show that DNA methylation in exhaled breath condensate is detectable and is likely of lung origin. Suggestive correlations with smoking and lung cancer case-control status depend on individual gene and CpG site examined.