Author Manuscript Published OnlineFirst on August 29, 2017; DOI: 10.1158/1940-6207.CAPR-17-0198 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
1
Tobacco-specific carcinogens induce hypermethylation, DNA adducts and
2
DNA damage in Bladder Cancer
3
Feng Jin1#, Jose Thaiparambil2#, Sri Ramya Donepudi1, Venkatrao Vantaku3, Danthasinghe Waduge
4
Badrajee Piyarathna3, Suman Maity1, Rashmi Krishnapuram3, Vasanta Putluri1, Franklin Gu4, Preeti
5
Purwaha3, Salil Kumar Bhowmik3, Chandrashekar R Ambati1, Friedrich-Carl von Rundstedt
6
Roghmann7, Sebastian Berg7, Joachim Noldus7, Kimal Rajapakshe3, Daniel Gödde8, Stephan Roth9,
7
Stephan
8
Balasubramanyam Karanam12, Martha K. Terris13, Shyam M Kavuri14, Seth P. Lerner4, Farrah
9
Kheradmand15, Cristian Coarfa1,3, Arun Sreekumar1,3,4, Yair Lotan16, Randa El-Zein2, Nagireddy
Störkel8,
Stephan
Degener9,
Michailidis10,
George
Benny
Abraham
5,6
, Florian
Kaipparettu11,
10
Putluri1,3*
11
1
12
Baylor College of Medicine, Houston, TX, USA; 2Department of Radiology, Houston Methodist Research
13
Institute, Houston, TX, USA; 3Department of Molecular and Cell Biology, 4Verna and Marrs McLean
14
Department of Biochemistry, Baylor College of Medicine, Houston, TX, USA; 5Scott Department of
15
Urology, Baylor College of Medicine, Houston, TX, USA; 6Department of Urology, Jena University
16
Hospital, Friedrich-Schiller-University, Jena, Germany;
Dan L. Duncan Cancer Center, Advanced Technology Core, Alkek Center for Molecular Discovery,
7
Department of Urology, Marien Hospital,
8
17
Ruhr-University Bochum, Herne, Germany; Department of Pathology, 9Department of Urology Helios
18
Klinikum, Witten-Herdecke University, Wuppertal, Germany;
19
Florida, Gainesville, FL, USA;
20
Medicine, TX, USA;
21
Tuskegee, AL, USA; 13Augusta University, Augusta, GA, USA;
12
11
10
Department of Statistics, University of
Department of Molecular and Human Genetics, Baylor College of
Department of Biology and Center for Cancer Research, Tuskegee University, 14
Lester and Sue Smith Breast Center,
15
22
Baylor College of Medicine, Houston, TX, USA;
23
Research in Inflammatory Diseases, Michael E. DeBakey VA, Baylor College of Medicine, Houston, TX,
24
USA;
16
Department of Medicine & Center for Translational
Department of Urology, University of Texas Southwestern, Dallas, TX, USA. # Authors
25
contributed equally to the generation of the manuscript.
26
*
27
Nagireddy Putluri, Ph.D.
28
Department of Molecular and Cellular Biology
29
Baylor College of Medicine, Houston, TX 77030, Tel.: (713) 798 3139: Email:
[email protected]
30
Conflict of interest statement: Authors do not have any conflict of interest
Corresponding author:
1
Downloaded from cancerpreventionresearch.aacrjournals.org on September 5, 2017. © 2017 American Association for Cancer Research.
Author Manuscript Published OnlineFirst on August 29, 2017; DOI: 10.1158/1940-6207.CAPR-17-0198 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
31
Abstract
32
Smoking is a major risk factor for the development of Bladder Cancer (BLCA); however, the
33
functional consequences of the carcinogens in tobacco smoke and BLCA-associated metabolic
34
alterations remains poorly defined. We assessed the metabolic profiles in BLCA smokers and
35
non-smokers, and identified the key alterations in their metabolism. Liquid Chromatography –
36
Mass Spectrometry (LC-MS), and bioinformatic analysis were performed to determine the
37
metabolome associated with BLCA smokers and were further validated in cell line models.
38
Smokers with BLCA were found to have elevated levels of methylated metabolites, polycyclic
39
aromatic hydrocarbons (PAHs), DNA adducts and DNA damage. DNA methyltransferase 1
40
(DNMT1) expression was significantly higher in smokers than non-smokers with BLCA. An
41
integromics approach, using multiple patient cohorts, revealed strong associations between
42
smokers and high-grade BLCA. In vitro exposure to the tobacco smoke carcinogens, 4-
43
(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) and benzo[a]pyrene (BaP) led to increase
44
in levels of methylated metabolites, DNA adducts, and extensive DNA damage in BLCA cells.
45
Co-treatment of BLCA cells with these carcinogens and the methylation inhibitor 5-aza-2'-
46
deoxycytidine (AZA) rewired the methylated metabolites, DNA adducts, DNA damage. These
47
findings were confirmed through the isotopic labeled metabolic flux analysis. Screens using
48
smoke associated metabolites and DNA adducts could provide robust biomarkers and improve
49
individual risk prediction in BLCA smokers. Non-invasive predictive biomarkers that can stratify
50
the risk of developing BLCA in smokers could aid in early detection and treatment.
51
Introduction
52
Bladder Cancer (BLCA) is a leading cause of cancer-related deaths globally [1-4] and
53
approximately 50% of BLCA patients are cigarette smokers [5-7]. In addition, patients with a 2
Downloaded from cancerpreventionresearch.aacrjournals.org on September 5, 2017. © 2017 American Association for Cancer Research.
Author Manuscript Published OnlineFirst on August 29, 2017; DOI: 10.1158/1940-6207.CAPR-17-0198 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
54
current or past history of smoking have a threefold higher chance of developing BLCA [8-11].
55
Moreover, the intensity and duration of smoking has been shown to affect the grade and stage of
56
the BLCA and high-dose smokers have more aggressive BLCA phenotypes [8, 12]. Smokers
57
with BLCA are also more likely to have resistance to chemotherapy [13]. At least 70 tobacco
58
smoke compounds, including 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) and
59
benzo[a]pyrene (BaP) are carcinogens [14, 15]. NNK has been identified as a potent carcinogen
60
which induces DNA adducts, mutations and promotes tumor growth through receptor-mediated
61
effects [14, 16]. Thus, prognostic and predictive biomarkers that can stratify the risk of
62
developing BLCA based on smoking habits are urgently needed. Recently, we found that hyper
63
methylation of xenobiotic enzymes was associated with BLCA progression [17]. However, the
64
metabolic signature of smoke-induced BLCA and its downstream effects remain to be
65
elucidated.
66
Using a metabolomics approach, we compared the metabolism in smoking and non-
67
smoking patients to identify metabolic signatures and determined the underlying mechanism
68
associated with smoking-induced BLCA. Additionally, we detected methylated metabolites in
69
the urine of patients, which could be used to develop a non-invasive, urine-based assay to detect
70
BLCA in smokers.
71 72
Methods
73
A total number of 119 pathologically-verified BLCA tissues (78 smokers and 41 non-
74
smokers), normal bladder (n= 10 smokers) or benign tissues (n= 14 smokers), and 108 BLCA
75
urine samples (71 smokers and 37 non-smokers) were obtained from different tumor banks
76
(Supplementary Table 1). We procured sample cohorts of smokers and non-smokers, which 3
Downloaded from cancerpreventionresearch.aacrjournals.org on September 5, 2017. © 2017 American Association for Cancer Research.
Author Manuscript Published OnlineFirst on August 29, 2017; DOI: 10.1158/1940-6207.CAPR-17-0198 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
77
include early to late stages (ta, t1, t2, t3 and t4), distant lymph nodes (N0, N1, N2, N3 and Nx)
78
and Lympho Vascular Invasion (LVI) status of BLCA (Supplementary Table 1). Metabolites
79
and DNA adducts were examined using a 6490 QQQ equipped with a 1290 LC-MS (Agilent
80
Technologies, USA). Receiver Operator Characteristic (ROC) curves and logistic regression
81
modelling were used to assess BLCA in urine samples. An integrative approach was used to
82
identify transcriptomic signatures differentiating high-grade from low-grade BLCA in public
83
data sets and evaluated the correlation between the BLCA grade and smoking status of the
84
patients. In vitro models were established to study the tobacco smoke compounds effects.
85
Comprehensive information of sample preparation protocols, metabolites, DNA adducts, quality
86
control, data acquisition and processing, and statistical analysis is described in the
87
supplementary methods.
88 89
Results
90
Smoker-associated BLCA Metabolic Profiles
91
Metabolic profiling by LC-MS was used to characterize metabolites. A summary of the
92
identified metabolites, DNA adducts, their experimental masses and retention times are shown in
93
Supplementary Tables 2 and 3. The experimental strategy used for profiling is shown in
94
Figure 1A.
95
(Supplementary methods).
96
metabolites out of 300 identified metabolites (Figure 1B), and the former are different from the
97
66 altered metabolites found in smokers of normal bladder and BLCA tissues (Supplementary
98
Figure 1B). Specifically, smokers with BLCA had elevated levels of methylated metabolites,
99
hexosamine biosynthetic pathway intermediates, acetylated metabolites, Poly-cyclic aromatic
We measured the metabolites using different chromatographic methods In comparison to non-smokers, BLCA smokers have 90 altered
4
Downloaded from cancerpreventionresearch.aacrjournals.org on September 5, 2017. © 2017 American Association for Cancer Research.
Author Manuscript Published OnlineFirst on August 29, 2017; DOI: 10.1158/1940-6207.CAPR-17-0198 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
100
hydrocarbons (PAH)/ their aromatic counterparts and hydroxylated derivatives compared to non-
101
smokers with BLCA (Figure 1B).
102
Furthermore, expression of intermediates of the methionine cycle was altered, in which
103
S-(5'-adenosyl)-L-methionine (SAM) and S-adenosyl-L-homocysteine (SAH) were decreased
104
and increased, respectively in smokers compared to non-smokers with BLCA. In addition to
105
changes in the primary metabolites, aniline a xenobiotic compound is known to be involved in
106
bladder carcinogenesis [18-20], was high in smokers compared to non-smokers with BLCA.
107 108
Pathway Analysis
109
To address the deregulated metabolic pathways in smokers with BLCA, we used the
110
online enrichment analysis platform ConsensusPathPDB (CPDB) [21] and mapped the
111
differential metabolites between smokers and non-smokers using KEGG, Reactome, and
112
GOBP(Gene Ontology Biological Process) database pathway analysis. We observed alterations
113
in the levels of metabolites associated with methionine metabolism, DNA methylation, nicotine
114
metabolism, glutathione metabolism, nucleotides and methionine salvage pathways in smokers
115
with BLCA (Figure 1C).
116
Evaluation of Tissue-derived PAHs and Methylated Metabolites in Urine
117
Given the ability to discriminate between smokers and non-smokers based on the expression of
118
BLCA-associated metabolites, we investigated the potential metabolites which can be used as
119
urine biomarkers. We used urine specimens from BLCA patients (smokers and non-smokers) to
120
measure tissue-derived BLCA-associated metabolites using a single reaction monitoring (SRM)-
121
based approach (Supplementary Tables 1, 2).
122
signatures in smokers with BLCA, 52 were detected in the urine samples (Supplementary
Out of the 90 tissue-specific metabolic
5
Downloaded from cancerpreventionresearch.aacrjournals.org on September 5, 2017. © 2017 American Association for Cancer Research.
Author Manuscript Published OnlineFirst on August 29, 2017; DOI: 10.1158/1940-6207.CAPR-17-0198 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.
123
Figure 1A), out of which, 40 were differentially expressed in smokers and non-smokers and
124
were further used to run individually as ROC classifiers. Briefly, to run a ROC curve, a logistic
125
regression model was used to build a classifier which was trained on a randomly selected subset
126
of two-thirds of the urine specimens (n = 72, training set) and its predictive performance was
127
assessed on the remaining one-third of the samples (n = 36, test set) (Figure 1D). Logistic
128
regression based classification model using all 52 metabolites was used to derive the activity
129
score [which has an Area Under the Curve (AUC) of 0.70 with a p-value of 0.03]
130
(Supplementary Figure 1C). Based on the performance of the individual metabolites, 23 (out of
131
the 40 differential compounds) were found to have significant ROC (p-value