Tobacco-Specific Carcinogens Induce Hypermethylation, DNA Adducts, and DNA Damage in Bladder Cancer

Abstract

Smoking is a major risk factor for the development of bladder cancer; however, the functional consequences of the carcinogens in tobacco smoke and bladder cancer-associated metabolic alterations remain poorly defined. We assessed the metabolic profiles in bladder cancer smokers and non-smokers and identified the key alterations in their metabolism. LC/MS and bioinformatic analysis were performed to determine the metabolome associated with bladder cancer smokers and were further validated in cell line models. Smokers with bladder cancer were found to have elevated levels of methylated metabolites, polycyclic aromatic hydrocarbons, DNA adducts, and DNA damage. DNA methyltransferase 1 (DNMT1) expression was significantly higher in smokers than non-smokers with bladder cancer. An integromics approach, using multiple patient cohorts, revealed strong associations between smokers and high-grade bladder cancer. In vitro exposure to the tobacco smoke carcinogens, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone and benzo[a]pyrene (BaP) led to increase in levels of methylated metabolites, DNA adducts, and extensive DNA damage in bladder cancer cells. Cotreatment of bladder cancer cells with these carcinogens and the methylation inhibitor 5-aza-2'-deoxycytidine rewired the methylated metabolites, DNA adducts, and DNA damage. These findings were confirmed through the isotopic-labeled metabolic flux analysis. Screens using smoke-associated metabolites and DNA adducts could provide robust biomarkers and improve individual risk prediction in bladder cancer smokers. Noninvasive predictive biomarkers that can stratify the risk of developing bladder cancer in smokers could aid in early detection and treatment. Cancer Prev Res; 10(10); 588-97. ©2017 AACR.

Figure 1. Metabolic profiling of BLCA tissue and urine samples from smokers and non-smokers

A) An overview of the strategy used to profile and characterize the metabolome of bladder cancer tissues from non-smokers (n=41) and smokers [(n=78; (36 current and 42 former)]. B) Heat map of hierarchical clustering of 90 differential metabolites across 119 BLCA tissues. Columns represent individual tissue samples and rows represent distinct metabolites. Shades of yellow and blue represents higher and lower levels of metabolites, relative to the median metabolite levels respectively (FDR<0.1). C) Pathway analysis of the metabolic profiles in the “smoking-associated BLCA metabolic signature”, the node size is proportional to the number of metabolites in the pathway and colored node represents a statistically significant enrichment. D) A ROC curve generated using 23 of the smoke associated BLCA metabolites in urine to delineate smokers from non-smokers with BLCA (AUC = 0.87, p-value <0.0001).

Figure 2. Levels of DNA adducts, xenobiotic enzymes, DNA damage markers in bladder cancer (BLCA) tissues between smokers and non-smokers

A) Box plots showing higher expression of NNK, methyl and other DNA adducts in tissues of smokers (S; n = 15) compared to non-smokers (NS; n = 15; p < 0.005). B) Box plots showing mRNA expression of xenobiotic enzymes by qPCR of BLCA tissues from non-smokers (n = 6) and smokers (n = 6; p < 0.005). C) Western blots and quantification of γ-H2AX and Chk2 protein levels in tissues of smokers (n=15) and non-smokers (n=15) with BLCA (p < 0.0008). β-actin was used as the loading control. D) Immunohistochemical staining of γ-H2AX.

Figure 3. Tobacco-specific carcinogens induce methylation, DNA adducts, DNA damage and activation of DNMT1

A) Heat map of metabolites from J82 cells that were treated with NNK (100 μM), BaP (10 μM), with NNK followed by AZA (5μM) (NNK +AZA) and BaP + AZA (FDR<0.25). B) DNA adducts in untreated and treated cells with NNK, BaP, NNK +AZA and BaP + AZA were measured by LC-MS/MS (p-value <0.005). C) Confocal microscopy and quantification analysis of γ-H2AX (green) in J82 cells treated with NNK, NNK+AZA, BaP and BaP+AZA (p<0.0001). D) Western blots and quantification of γ-H2AX protein levels in J82 cells treated with aphidicolin, NNK, and AZA (p<0.0001). β-actin was used as the loading control. Comparison of DNMT1 protein (E) and mRNA (F) expression in tissues of smokers (n=15) and non-smokers (n=15) and their quantification (p<0.05). G) Immunohistochemical analysis of DNMT1 expression in BLCA tissues from smoker and non-smoker.

Figure 4. NNK induced DNMT1 expression, methionine flux and Epithelial-Mesenchymal Transition (EMT) markers

A) Box plots showing mRNA and protein expression of DNMT1, in untreated and treated J82 cells with NNK (p<0.005) and their quantification. B) The role of DNMT1 in methionine pathway. C) Isotopic distributions of 13C labeled methionine pathway metabolites showing down regulation of SAM and upregulation of SAH upon treatment with NNK and reverse with NNK followed by AZA in J82 cells. D) Correlation of DNMT1 expression with EMT scores from multiple public cohorts. F, G) mRNA and protein expression of N-cadherin, E-Cadherin and vimentin in J82 cells treated with NNK and BaP (p <0.005) and their quantification. β-actin was used as the loading control. I) Working model showing the alterations in BLCA associated with smoking.