Toxicology Letters 232 (2015) 429–437
Contents lists available at ScienceDirect
Toxicology Letters journal homepage: www.elsevier.com/locate/toxlet
Tobacco smoking-response genes in blood and buccal cells Hyun-Kyung Na a , Minju Kim a , Seong-Sil Chang b , Soo-Young Kim b , Jong Y. Park c , Myeon Woo Chung d, Mihi Yang a, * a
Research Center for Cell Fate Control, College of Pharmacy, Sookmyung Women’s University, Seoul, Republic of Korea Department of Occupational and Environmental Medicine, Eulji University Hospital, Daejeon, Republic of Korea c Department of Cancer Epidemiology, Moffitt Cancer Center, Tampa, FL, U.S.A d Laboratory Animal Resources Division, National Institute of Food and Drug Safety Evaluation, Ministry of Food and Drug Safety, Osong, Republic of Korea b
H I G H L I G H T S
G R A P H I C A L A B S T R A C T
A biological monitoring was performed to find tobacco smokingspecific toxic mechanisms in surrogate and non-invasive tissues. The Fcg-receptor mediated phagocytosis and leukocyte transendothelial migration pathways were differentially expressed between smokers and nonsmokers. The ACTG1, involved in the maintenance of actin cytoskeleton, cell migration and cancer metastasis, was highly expressed in both of blood and buccal cells. Smokers showed high levels of urinary malondialdehyde (MDA) and down-regulation of expressions of antioxidant related genes including TPO, MPO, GPX2, PTGR1, and NUDT1.
A R T I C L E I N F O
A B S T R A C T
Article history: Received 21 July 2014 Received in revised form 1 October 2014 Accepted 2 October 2014 Available online 13 November 2014
Tobacco smoking is a well-known cause of various diseases, however, its toxic mechanisms for diseases are not completely understood, yet. Therefore, we performed biological monitoring to find tobacco smoking-responsive mechanisms including oxidative stress in Korean men (N = 36). Whole genome microarray analyses were performed with peripheral blood from smokers and age-matched nonsmokers. We also performed qRT-PCR to confirm the microarray results and compared the gene expression of blood to those of buccal cells. To assess the effects of tobacco smoking on oxidative stress, we analyzed urinary levels of malondialdehyde (MDA), a lipid peroxidation marker, and performed PCR-based arrays on reactive oxygen species (ROS)-related genes. As results, 34 genes were differently expressed in blood between smokers and nonsmokers (ps < 0.01 and >1.5-fold change). Particularly, the genes involved in immune responsive pathways, e.g., the Fcg-receptor mediated phagocytosis and the leukocyte transendothelial migration pathways, were differentially expressed between smokers and nonsmokers. Among the above genes, the ACTG1, involved in the maintenance of actin cytoskeleton, cell migration and cancer metastasis, was highly expressed by smoking in both blood and buccal cells. Concerning oxidative
Keywords: Tobacco smoking Reactive oxygen species (ROS) Microarray Expression Network analysis Buccal cells
* Corresponding author at: Hyochangwon-gil 52, Yongsan-gu, Seoul, 140-742, Republic of Korea. Tel.: +82 220777179. E-mail address: [email protected] (M. Yang). http://dx.doi.org/10.1016/j.toxlet.2014.10.005 0378-4274/ ã 2014 Elsevier Ireland Ltd. All rights reserved.
430
H.-K. Na et al. / Toxicology Letters 232 (2015) 429–437
stress, smokers showed high levels of urinary MDA and down-regulation of expressions of antioxidant related genes including TPO, MPO, GPX2, PTGR1, and NUDT1 as compared to nonsmokers (ps < 0.05). In conclusion, these results suggest that systemically altered immune response and oxidative stress can be tobacco-responsive mechanisms for the related diseases. Based on consistent results in blood and buccal cells, expression of the ACTG1 can be a tobacco smoking-responsive biomarker. ã 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Tobacco smoking is the leading avoidable cause of morbidity and mortality worldwide, contributing to approximately 6 million deaths per year (WHO, 2011): it is a major risk factor for cardiovascular diseases (Prasad et al., 2009), airway inflammatory diseases (Arnson et al., 2010) and many types of cancer (Humans, 2012; Jemal et al., 2011). Toxic components in tobacco or tobacco smoking-induced chemicals, e.g., benzene, 4-(methylnitrosamino)-1-(3-pyridyl)-1butanone (NNK), and polycyclic aromatic hydrocarbons (PAHs), have been emphasized as causes of tobacco-related toxicity. In addition, bioproduced reactive oxygen species (ROS) by tobacco smoking have been suspected to play a significant role as tobaccorelated toxic mechanisms (Yao and Rahman, 2011). However, specific mechanisms of tobacco smoking-related diseases are not completely understood yet (Steiling et al., 2009). The health risks of tobacco smoking are highly complicated, because tobacco contains more than 4000 chemicals (Wogan et al., 2004), which induce a variety of physiological responses. For example, many constituents of tobacco, such as nicotine, benzo[a] pyrene, and hydroquinone, have been reported to induce immune alteration (Stampfli and Anderson, 2009). Therefore, we decided to integrate scattered and complicated toxic information of tobacco smoking for clear understanding of its toxic mechanisms. For this purpose, genome-wide expression profiling has emerged as a powerful approach for identification of molecular responsive markers, prediction and relation to clinical outcome of tobacco smoking (Gower et al., 2011). Previous genome wide studies suggested that tobacco smoking altered pathways involved in regulation of oxidative stress (Wang et al., 2010), carcinogenesis (Charles et al., 2008), or host defense (Doyle et al., 2010) with microarray based gene expression profiling. However, these results are not consistent. It may be due to multiple reasons: first, ethnic variations, e.g., in Caucasians, Asians, or Hispanics, etc., may be one of causes of inconsistent results (Boyle et al., 2010; Charlesworth et al., 2010; Shahdoust et al., 2013). Secondly, lack of accurate biological monitoring of tobacco smoking may be a reason. Thirdly, the different types of tissues, e.g., blood, buccal, or bronchial epithelium, may generate inconsistent results (Beineke et al., 2012; Boyle et al., 2010; Wang et al., 2010). In order to identify reliable targets of tobacco smoking, surrogate biospecimen or multiple tissues are needed. In the present study, we performed whole genome expression analyses between Korean male smokers and nonsmokers to understand tobacco smoke-related pathological mechanisms and clarify the gene expression profiles in surrogate and noninvasive tissues. At first, we performed whole genome microarrays to estimate the systemic influence of tobacco smoking. Based on the microarray results, we confirmed microarray results on selected genes in both blood and buccal epithelium cells to evaluate their appropriateness as responsive or effective biomarkers for tobacco smoking. To further investigate the effect of tobacco smoking on the oxidative stress response pathway, we analyzed tobacco smoking-
induced expression profiles of ROS-related genes and urinary malondialdehyde (MDA), a typical biomarker for oxidative stress. 2. Materials and methods 2.1. Study subjects We recruited 36 healthy Korean men (age = 40.2 5.9 years) at Eulji University Hospital (Daejeon, South Korea). The study scheme is shown in Fig. 1. All subjects provided a written informed consent and completed extensive questionnaires including medical and smoking history, dietary patterns, alcohol drinking, environment of residency, etc. Carbon monoxide (CO) in blood was analyzed in all the subjects with MicroCO breath CO Monitor (CareFusion, San Diego, CA). All of the study protocol was approved by Institutional Review Board of Eulji University Hospital. 2.2. Sample collection We collected buccal cell samples from the subjects with sterile cotton swabs: the subjects rinsed their mouth twice with 200 ml of drinking water and the swabs were scraped gently more than 10 times against the buccal mucosa on the inside of the cheek. The cells in swab head were immediately immersed in 500 ml of RNAlater1 stabilization reagent (QIAGEN, Valencia, CA) and stored at 80 C until analyses. We also collected 40 ml of urine and 10 ml of peripheral blood into 50 ml of conical tubes and PAXgeneTM Blood RNA tubes (QIAGEN), respectively, and stored them at 80 C until analyses. 2.3. Analyses of urinary cotinine Urinary cotinine was analyzed by ion-pair HPLC/UVD method (Yang et al., 2001) with minor modifications. In brief, 900 ml of each urine sample was mixed with 100 ml of 80 mM 2-phenylimidazole, as an internal standard, and 330 ml of 3 M NaOH. The mixture was twice extracted with 3 ml of CH2Cl2 per each. After evaporating CH2Cl2extract, we dissolved the residue in 1 ml of water and injected 20 ml of its supernatant fraction to HPLC. The HPLC system consisted of dual Younglin SP930D pumps (Younglin, Seoul, Korea), the MIDAS COOL auto sampler (Spark Holland, Emme, The Netherlands), the SPD-10A UV–vis detector (Shimadzu, Kyoto, Japan), and the TSK gel ODS-80TM column (5 mm, 4.6 mm 150 mm, Toyo Soda Co., Toyko, Japan). The analyses were carried out with the gradient mode: mobile phase A, a mixture of acetonitrile/water (15/85) containing 20 mM KH2PO4 and 3 mM sodium 1-octanesulfonate (pH 4.5); B, methanol; flow rate was 0.7 ml/min, ratio of A to B, 0–20 min, 100:0; 20–25 min, 100:0 to 50:50; 25–30 min, 50:50; 30–35 min, 50:50 to 100:0; 35–45 min, 100:0. Column was kept at 50 C and the absorbance was observed at 254 nm. Urinary cotinine was adjusted for urinary creatinine, measured with ion-pair HPLC/UVD method, as described by Ogata and Taguchi (Ogata and Taguchi, 1987) with minor modifications (Yi et al., 2011).
H.-K. Na et al. / Toxicology Letters 232 (2015) 429–437
431
Fig. 1. Study scheme: thirty six Korean men were recruited and were equally divided into smokers and nonsmokers, according to urinary cotinine and questionnaires. In the first part, whole genome microarray was performed for 12 subjects. The results of the microarray were confirmed by qPCR in the dependents (N = 12) and independent subjects (N = 24). In the second part, we measured urinary MDA and performed PCR-based array on ROS-related genes in the 6 subjects of the 12 subjects, who were analyzed in whole genome microarray.
2.4. Analyses of urinary MDA MDA was determined as adducts of 2-thiobarbituric acid (TBA, CAS number: 504-17-6) (Lee et al., 2013; Tagesson et al., 1996). Twenty three mM of TBA (Sigma, St. Louis, MO, USA) was freshly prepared in water. Standard MDA solutions [0.625–10 mM of 1,1,3,3,-tetraethoxypropane (TEP, CAS number: 122-31-6) (Sigma) in 40% of ethanol/water] were also prepared before the experiments. At first, 50 ml of each urine sample was mixed with 300 ml of 0.5 M phosphoric acid and 150 ml of the TBA solution. The mixtures were incubated at 95 C for 1 h in capped tubes and then chilled on ice for 5 min. Five hundred microliters of methanol was added to each mixture. After centrifugation (13,000 rpm for 10 min), we injected 20 ml of its supernatant to the above HPLC. The TBA–MDA adducts were detected at 532 nm with the isocratic mode: mobile phase, a mixture of 50 mM potassium phosphate buffer (pH 6.8) and methanol (58:42, v/v); flow rate, 0.6 ml/min. 2.5. Whole-genome microarray analyses RNA was extracted from blood and buccal cells using PAXgeneTM Blood RNA kit (QIAGEN) and RNeasy mini kit (QIAGEN), respectively, according to the manufacturers’ protocol. The integrity of the total RNA was evaluated with the Bioanalyzer 2100 (Agilent Technologies, Palo Alto, USA): only samples with RNA integrity number (RIN) >7.0 were used for the following microarray analyses. We applied 12 RNA samples to the Illumina HT-12 v. 4 BeadChip (Illumina, Inc., San Diego, CA, USA) with 47,323 transcript probes, as previously described with minor modifications (Zeller et al., 2010). Briefly, 500 ng of total RNA was reverse-transcribed, amplified, and biotinylated with the
TargetAmpTM-Nano Labeling Kit (Epicentre, Madison, WI, USA). Seven hundred fifty nanograms of each biotinlyated cRNA was hybridized to a single BeadChip for 16–18 h at 58 C. BeadChips were scanned with an Illumina Bead Array Reader (Illumina). For data analyses of whole genome microarray, raw data were extracted by Illumina GenomeStudio v. 2011.1 (Illumina). Array probes were transformed by logarithm and normalized by quantile method. False discovery rate (FDR) was controlled by adjusting p value using Benjamini–Hochberg algorithm. Hierarchical clustering was performed with average linkage and Euclidean distance. Logistic regression analysis was used to calculate the gene expression score (GES) based on age and the expression levels of 6 genes which were significantly associated with smoking (p < 0.05 and >2-fold change). The formula for the GES algorithm was: pr (smoker)/(1-pr (smoker)) = 138.1 + 0.577 age 0.2556 DEFA1B 13.7 LOC388588 0.5431 DEFA4 8.785 VWCE + 28.39 LRRN3 + 7.461 HS.137971 + 3.896 VAV3. To investigate pathways and gene networks related to smoking, we performed the ‘gene-enrichment and functional annotation analyses’ with significant probe set, using the gene ontology (GO) terms and KEGG pathways on DAVID (http://david.abcc.ncifcrf.gov/home. jsp). All analyses and visualization of whole genome microarray results were conducted with R software package v. 2.15.1. 2.6. Quantitative real-time PCR (qPCR) analyses To confirm the results of whole genome microarray, qPCR analyses were performed. Blood and buccal cDNA were prepared from 150 ng and 25 ng of total RNA, respectively, with QuantiTect Reverse Transcription Kit (QIAGEN). According to the manufacturer’s instructions, we performed qPCR with TaqMan Gene
432
H.-K. Na et al. / Toxicology Letters 232 (2015) 429–437
Expression Assays on the 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA). Taqman qPCR primers and probes were as followed: DEFA4 (Hs00157252_m1), CYP1B1 (Hs00164383_m1), FCGR3A (Hs02388314_m1), VAV3 (Hs00916814_m1), ACTG1 (Hs03044422_g1), CXCR4 (Hs00237052_m1). We prepared reaction mixtures with 4 ml and 9 ml of blood and buccal cDNA, respectively, in 20 ml final reaction volume, which included 2 Taqman universal PCR master mix (Applied Biosystems) with AmpErase UNG (uracil-N-glycosylase), 3 pmol of each Taqman primer, and 5 pmol of each probe labeled with FAM dye. The PCR conditions used were as follows: 2 min at 50 C for AmpErase UNG activation and 10 min at 95 C for the UNG inactivation, followed by 40 cycles of 15 s at 95 C for denaturing and 1 min at 60 C for annealing and extension. Commercial Raji cDNA was used to make calibration curves. A reference gene, 18S rRNA, was also amplified in each sample to normalize the expression of each gene with commercialized 18S rRNA primers and probes (Hs99999901_s1). 2.7. PCR-based microarray analyses on ROS-related genes To assess the effects of tobacco smoking on oxidative stress, PCR-based microarray analyses were performed on 3 pairs of smokers and nonsmokers. Briefly, RNA (1 mg) from each sample was reverse transcribed into cDNA with the RT2 First Strand Kits (QIAGEN). We mixed 0.94 ml of the cDNA with 12.50 ml of RT2 qPCR Master Mix SYBR Green and 11.56 ml of distilled water. Twenty five microliters of a final reaction was loaded into 96-well RT2 Profiler PCR Array. The real time PCR was performed on the Human Oxidative Stress Plus RT2 Profiler PCR Array (QIAGEN), which targets 84 core ROS-related genes for oxidative stress response. The reaction was run with the following conditions: 10 min at 95 C for the activation of Hot Start DNA polymerase, followed by 40 cycles of 15 s at 95 C for denaturation, and 60 s at 60 C for annealing (fluorescence detection). Relative changes in gene expression were calculated by the delta Ct method with the adjustment for the average expression of a housekeeping gene, B2M (beta-2-microglobulin) to normalize RNA input. The data were analyzed with SABiosciences’ PCR array data analysis software (http://pcrdataanalysis.sabiosciences.com/pcr/arrayanalysis.php).
Fig. 2. Histogram of urinary cotinine (A) and MDA (B) levels in all subjects (N = 36): X-axes and Y-axes indicate urinary levels of cotinine [(A) or MDA (B)] and the probability of subjects, respectively. Upper parts of the figures show outlier box plots with the squares in the boxes showing the interquartile range.
used to compare the levels of urinary metabolites in smokers to those in nonsmokers. Regression analyses were performed to determine the association between levels of urinary cotinine and gene expression and between gene expressions in blood and buccal samples. All statistical analyses were conducted using the package of JMP v. 4.0.2. (SAS Institute, Cary, NC).
2.8. Statistical analyses 3. Results Shapiro–Wilk W test was used to test the distributional normality in levels of urinary cotinine, MDA and gene expression. T-test was used to compare the differences in characteristics between smokers and nonsmokers. In addition, Wilcoxon test was Table 1 Characteristics of study participants. Variables
Subjects Nonsmoker (N = 18)
Age (years) BMI (kg/m2) Cigarettes per day Pack-year CO (ppm) Urinary cotinineb (mg/g creatinine) Urinary MDA (umol/g creatinine)
40.50 5.95 25.16 2.61 0 0 0.06 0.24 0.01 0.01a 1.25 0.69
p value a
Smoker (N = 18) 39.83 23.90 19.56 21.76 20.00 1.69
b
5.96 2.69 5.74 9.35 8.26 0.73
0.74 0.19
Contents lists available at ScienceDirect
Toxicology Letters journal homepage: www.elsevier.com/locate/toxlet
Tobacco smoking-response genes in blood and buccal cells Hyun-Kyung Na a , Minju Kim a , Seong-Sil Chang b , Soo-Young Kim b , Jong Y. Park c , Myeon Woo Chung d, Mihi Yang a, * a
Research Center for Cell Fate Control, College of Pharmacy, Sookmyung Women’s University, Seoul, Republic of Korea Department of Occupational and Environmental Medicine, Eulji University Hospital, Daejeon, Republic of Korea c Department of Cancer Epidemiology, Moffitt Cancer Center, Tampa, FL, U.S.A d Laboratory Animal Resources Division, National Institute of Food and Drug Safety Evaluation, Ministry of Food and Drug Safety, Osong, Republic of Korea b
H I G H L I G H T S
G R A P H I C A L A B S T R A C T
A biological monitoring was performed to find tobacco smokingspecific toxic mechanisms in surrogate and non-invasive tissues. The Fcg-receptor mediated phagocytosis and leukocyte transendothelial migration pathways were differentially expressed between smokers and nonsmokers. The ACTG1, involved in the maintenance of actin cytoskeleton, cell migration and cancer metastasis, was highly expressed in both of blood and buccal cells. Smokers showed high levels of urinary malondialdehyde (MDA) and down-regulation of expressions of antioxidant related genes including TPO, MPO, GPX2, PTGR1, and NUDT1.
A R T I C L E I N F O
A B S T R A C T
Article history: Received 21 July 2014 Received in revised form 1 October 2014 Accepted 2 October 2014 Available online 13 November 2014
Tobacco smoking is a well-known cause of various diseases, however, its toxic mechanisms for diseases are not completely understood, yet. Therefore, we performed biological monitoring to find tobacco smoking-responsive mechanisms including oxidative stress in Korean men (N = 36). Whole genome microarray analyses were performed with peripheral blood from smokers and age-matched nonsmokers. We also performed qRT-PCR to confirm the microarray results and compared the gene expression of blood to those of buccal cells. To assess the effects of tobacco smoking on oxidative stress, we analyzed urinary levels of malondialdehyde (MDA), a lipid peroxidation marker, and performed PCR-based arrays on reactive oxygen species (ROS)-related genes. As results, 34 genes were differently expressed in blood between smokers and nonsmokers (ps < 0.01 and >1.5-fold change). Particularly, the genes involved in immune responsive pathways, e.g., the Fcg-receptor mediated phagocytosis and the leukocyte transendothelial migration pathways, were differentially expressed between smokers and nonsmokers. Among the above genes, the ACTG1, involved in the maintenance of actin cytoskeleton, cell migration and cancer metastasis, was highly expressed by smoking in both blood and buccal cells. Concerning oxidative
Keywords: Tobacco smoking Reactive oxygen species (ROS) Microarray Expression Network analysis Buccal cells
* Corresponding author at: Hyochangwon-gil 52, Yongsan-gu, Seoul, 140-742, Republic of Korea. Tel.: +82 220777179. E-mail address: [email protected] (M. Yang). http://dx.doi.org/10.1016/j.toxlet.2014.10.005 0378-4274/ ã 2014 Elsevier Ireland Ltd. All rights reserved.
430
H.-K. Na et al. / Toxicology Letters 232 (2015) 429–437
stress, smokers showed high levels of urinary MDA and down-regulation of expressions of antioxidant related genes including TPO, MPO, GPX2, PTGR1, and NUDT1 as compared to nonsmokers (ps < 0.05). In conclusion, these results suggest that systemically altered immune response and oxidative stress can be tobacco-responsive mechanisms for the related diseases. Based on consistent results in blood and buccal cells, expression of the ACTG1 can be a tobacco smoking-responsive biomarker. ã 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Tobacco smoking is the leading avoidable cause of morbidity and mortality worldwide, contributing to approximately 6 million deaths per year (WHO, 2011): it is a major risk factor for cardiovascular diseases (Prasad et al., 2009), airway inflammatory diseases (Arnson et al., 2010) and many types of cancer (Humans, 2012; Jemal et al., 2011). Toxic components in tobacco or tobacco smoking-induced chemicals, e.g., benzene, 4-(methylnitrosamino)-1-(3-pyridyl)-1butanone (NNK), and polycyclic aromatic hydrocarbons (PAHs), have been emphasized as causes of tobacco-related toxicity. In addition, bioproduced reactive oxygen species (ROS) by tobacco smoking have been suspected to play a significant role as tobaccorelated toxic mechanisms (Yao and Rahman, 2011). However, specific mechanisms of tobacco smoking-related diseases are not completely understood yet (Steiling et al., 2009). The health risks of tobacco smoking are highly complicated, because tobacco contains more than 4000 chemicals (Wogan et al., 2004), which induce a variety of physiological responses. For example, many constituents of tobacco, such as nicotine, benzo[a] pyrene, and hydroquinone, have been reported to induce immune alteration (Stampfli and Anderson, 2009). Therefore, we decided to integrate scattered and complicated toxic information of tobacco smoking for clear understanding of its toxic mechanisms. For this purpose, genome-wide expression profiling has emerged as a powerful approach for identification of molecular responsive markers, prediction and relation to clinical outcome of tobacco smoking (Gower et al., 2011). Previous genome wide studies suggested that tobacco smoking altered pathways involved in regulation of oxidative stress (Wang et al., 2010), carcinogenesis (Charles et al., 2008), or host defense (Doyle et al., 2010) with microarray based gene expression profiling. However, these results are not consistent. It may be due to multiple reasons: first, ethnic variations, e.g., in Caucasians, Asians, or Hispanics, etc., may be one of causes of inconsistent results (Boyle et al., 2010; Charlesworth et al., 2010; Shahdoust et al., 2013). Secondly, lack of accurate biological monitoring of tobacco smoking may be a reason. Thirdly, the different types of tissues, e.g., blood, buccal, or bronchial epithelium, may generate inconsistent results (Beineke et al., 2012; Boyle et al., 2010; Wang et al., 2010). In order to identify reliable targets of tobacco smoking, surrogate biospecimen or multiple tissues are needed. In the present study, we performed whole genome expression analyses between Korean male smokers and nonsmokers to understand tobacco smoke-related pathological mechanisms and clarify the gene expression profiles in surrogate and noninvasive tissues. At first, we performed whole genome microarrays to estimate the systemic influence of tobacco smoking. Based on the microarray results, we confirmed microarray results on selected genes in both blood and buccal epithelium cells to evaluate their appropriateness as responsive or effective biomarkers for tobacco smoking. To further investigate the effect of tobacco smoking on the oxidative stress response pathway, we analyzed tobacco smoking-
induced expression profiles of ROS-related genes and urinary malondialdehyde (MDA), a typical biomarker for oxidative stress. 2. Materials and methods 2.1. Study subjects We recruited 36 healthy Korean men (age = 40.2 5.9 years) at Eulji University Hospital (Daejeon, South Korea). The study scheme is shown in Fig. 1. All subjects provided a written informed consent and completed extensive questionnaires including medical and smoking history, dietary patterns, alcohol drinking, environment of residency, etc. Carbon monoxide (CO) in blood was analyzed in all the subjects with MicroCO breath CO Monitor (CareFusion, San Diego, CA). All of the study protocol was approved by Institutional Review Board of Eulji University Hospital. 2.2. Sample collection We collected buccal cell samples from the subjects with sterile cotton swabs: the subjects rinsed their mouth twice with 200 ml of drinking water and the swabs were scraped gently more than 10 times against the buccal mucosa on the inside of the cheek. The cells in swab head were immediately immersed in 500 ml of RNAlater1 stabilization reagent (QIAGEN, Valencia, CA) and stored at 80 C until analyses. We also collected 40 ml of urine and 10 ml of peripheral blood into 50 ml of conical tubes and PAXgeneTM Blood RNA tubes (QIAGEN), respectively, and stored them at 80 C until analyses. 2.3. Analyses of urinary cotinine Urinary cotinine was analyzed by ion-pair HPLC/UVD method (Yang et al., 2001) with minor modifications. In brief, 900 ml of each urine sample was mixed with 100 ml of 80 mM 2-phenylimidazole, as an internal standard, and 330 ml of 3 M NaOH. The mixture was twice extracted with 3 ml of CH2Cl2 per each. After evaporating CH2Cl2extract, we dissolved the residue in 1 ml of water and injected 20 ml of its supernatant fraction to HPLC. The HPLC system consisted of dual Younglin SP930D pumps (Younglin, Seoul, Korea), the MIDAS COOL auto sampler (Spark Holland, Emme, The Netherlands), the SPD-10A UV–vis detector (Shimadzu, Kyoto, Japan), and the TSK gel ODS-80TM column (5 mm, 4.6 mm 150 mm, Toyo Soda Co., Toyko, Japan). The analyses were carried out with the gradient mode: mobile phase A, a mixture of acetonitrile/water (15/85) containing 20 mM KH2PO4 and 3 mM sodium 1-octanesulfonate (pH 4.5); B, methanol; flow rate was 0.7 ml/min, ratio of A to B, 0–20 min, 100:0; 20–25 min, 100:0 to 50:50; 25–30 min, 50:50; 30–35 min, 50:50 to 100:0; 35–45 min, 100:0. Column was kept at 50 C and the absorbance was observed at 254 nm. Urinary cotinine was adjusted for urinary creatinine, measured with ion-pair HPLC/UVD method, as described by Ogata and Taguchi (Ogata and Taguchi, 1987) with minor modifications (Yi et al., 2011).
H.-K. Na et al. / Toxicology Letters 232 (2015) 429–437
431
Fig. 1. Study scheme: thirty six Korean men were recruited and were equally divided into smokers and nonsmokers, according to urinary cotinine and questionnaires. In the first part, whole genome microarray was performed for 12 subjects. The results of the microarray were confirmed by qPCR in the dependents (N = 12) and independent subjects (N = 24). In the second part, we measured urinary MDA and performed PCR-based array on ROS-related genes in the 6 subjects of the 12 subjects, who were analyzed in whole genome microarray.
2.4. Analyses of urinary MDA MDA was determined as adducts of 2-thiobarbituric acid (TBA, CAS number: 504-17-6) (Lee et al., 2013; Tagesson et al., 1996). Twenty three mM of TBA (Sigma, St. Louis, MO, USA) was freshly prepared in water. Standard MDA solutions [0.625–10 mM of 1,1,3,3,-tetraethoxypropane (TEP, CAS number: 122-31-6) (Sigma) in 40% of ethanol/water] were also prepared before the experiments. At first, 50 ml of each urine sample was mixed with 300 ml of 0.5 M phosphoric acid and 150 ml of the TBA solution. The mixtures were incubated at 95 C for 1 h in capped tubes and then chilled on ice for 5 min. Five hundred microliters of methanol was added to each mixture. After centrifugation (13,000 rpm for 10 min), we injected 20 ml of its supernatant to the above HPLC. The TBA–MDA adducts were detected at 532 nm with the isocratic mode: mobile phase, a mixture of 50 mM potassium phosphate buffer (pH 6.8) and methanol (58:42, v/v); flow rate, 0.6 ml/min. 2.5. Whole-genome microarray analyses RNA was extracted from blood and buccal cells using PAXgeneTM Blood RNA kit (QIAGEN) and RNeasy mini kit (QIAGEN), respectively, according to the manufacturers’ protocol. The integrity of the total RNA was evaluated with the Bioanalyzer 2100 (Agilent Technologies, Palo Alto, USA): only samples with RNA integrity number (RIN) >7.0 were used for the following microarray analyses. We applied 12 RNA samples to the Illumina HT-12 v. 4 BeadChip (Illumina, Inc., San Diego, CA, USA) with 47,323 transcript probes, as previously described with minor modifications (Zeller et al., 2010). Briefly, 500 ng of total RNA was reverse-transcribed, amplified, and biotinylated with the
TargetAmpTM-Nano Labeling Kit (Epicentre, Madison, WI, USA). Seven hundred fifty nanograms of each biotinlyated cRNA was hybridized to a single BeadChip for 16–18 h at 58 C. BeadChips were scanned with an Illumina Bead Array Reader (Illumina). For data analyses of whole genome microarray, raw data were extracted by Illumina GenomeStudio v. 2011.1 (Illumina). Array probes were transformed by logarithm and normalized by quantile method. False discovery rate (FDR) was controlled by adjusting p value using Benjamini–Hochberg algorithm. Hierarchical clustering was performed with average linkage and Euclidean distance. Logistic regression analysis was used to calculate the gene expression score (GES) based on age and the expression levels of 6 genes which were significantly associated with smoking (p < 0.05 and >2-fold change). The formula for the GES algorithm was: pr (smoker)/(1-pr (smoker)) = 138.1 + 0.577 age 0.2556 DEFA1B 13.7 LOC388588 0.5431 DEFA4 8.785 VWCE + 28.39 LRRN3 + 7.461 HS.137971 + 3.896 VAV3. To investigate pathways and gene networks related to smoking, we performed the ‘gene-enrichment and functional annotation analyses’ with significant probe set, using the gene ontology (GO) terms and KEGG pathways on DAVID (http://david.abcc.ncifcrf.gov/home. jsp). All analyses and visualization of whole genome microarray results were conducted with R software package v. 2.15.1. 2.6. Quantitative real-time PCR (qPCR) analyses To confirm the results of whole genome microarray, qPCR analyses were performed. Blood and buccal cDNA were prepared from 150 ng and 25 ng of total RNA, respectively, with QuantiTect Reverse Transcription Kit (QIAGEN). According to the manufacturer’s instructions, we performed qPCR with TaqMan Gene
432
H.-K. Na et al. / Toxicology Letters 232 (2015) 429–437
Expression Assays on the 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA). Taqman qPCR primers and probes were as followed: DEFA4 (Hs00157252_m1), CYP1B1 (Hs00164383_m1), FCGR3A (Hs02388314_m1), VAV3 (Hs00916814_m1), ACTG1 (Hs03044422_g1), CXCR4 (Hs00237052_m1). We prepared reaction mixtures with 4 ml and 9 ml of blood and buccal cDNA, respectively, in 20 ml final reaction volume, which included 2 Taqman universal PCR master mix (Applied Biosystems) with AmpErase UNG (uracil-N-glycosylase), 3 pmol of each Taqman primer, and 5 pmol of each probe labeled with FAM dye. The PCR conditions used were as follows: 2 min at 50 C for AmpErase UNG activation and 10 min at 95 C for the UNG inactivation, followed by 40 cycles of 15 s at 95 C for denaturing and 1 min at 60 C for annealing and extension. Commercial Raji cDNA was used to make calibration curves. A reference gene, 18S rRNA, was also amplified in each sample to normalize the expression of each gene with commercialized 18S rRNA primers and probes (Hs99999901_s1). 2.7. PCR-based microarray analyses on ROS-related genes To assess the effects of tobacco smoking on oxidative stress, PCR-based microarray analyses were performed on 3 pairs of smokers and nonsmokers. Briefly, RNA (1 mg) from each sample was reverse transcribed into cDNA with the RT2 First Strand Kits (QIAGEN). We mixed 0.94 ml of the cDNA with 12.50 ml of RT2 qPCR Master Mix SYBR Green and 11.56 ml of distilled water. Twenty five microliters of a final reaction was loaded into 96-well RT2 Profiler PCR Array. The real time PCR was performed on the Human Oxidative Stress Plus RT2 Profiler PCR Array (QIAGEN), which targets 84 core ROS-related genes for oxidative stress response. The reaction was run with the following conditions: 10 min at 95 C for the activation of Hot Start DNA polymerase, followed by 40 cycles of 15 s at 95 C for denaturation, and 60 s at 60 C for annealing (fluorescence detection). Relative changes in gene expression were calculated by the delta Ct method with the adjustment for the average expression of a housekeeping gene, B2M (beta-2-microglobulin) to normalize RNA input. The data were analyzed with SABiosciences’ PCR array data analysis software (http://pcrdataanalysis.sabiosciences.com/pcr/arrayanalysis.php).
Fig. 2. Histogram of urinary cotinine (A) and MDA (B) levels in all subjects (N = 36): X-axes and Y-axes indicate urinary levels of cotinine [(A) or MDA (B)] and the probability of subjects, respectively. Upper parts of the figures show outlier box plots with the squares in the boxes showing the interquartile range.
used to compare the levels of urinary metabolites in smokers to those in nonsmokers. Regression analyses were performed to determine the association between levels of urinary cotinine and gene expression and between gene expressions in blood and buccal samples. All statistical analyses were conducted using the package of JMP v. 4.0.2. (SAS Institute, Cary, NC).
2.8. Statistical analyses 3. Results Shapiro–Wilk W test was used to test the distributional normality in levels of urinary cotinine, MDA and gene expression. T-test was used to compare the differences in characteristics between smokers and nonsmokers. In addition, Wilcoxon test was Table 1 Characteristics of study participants. Variables
Subjects Nonsmoker (N = 18)
Age (years) BMI (kg/m2) Cigarettes per day Pack-year CO (ppm) Urinary cotinineb (mg/g creatinine) Urinary MDA (umol/g creatinine)
40.50 5.95 25.16 2.61 0 0 0.06 0.24 0.01 0.01a 1.25 0.69
p value a
Smoker (N = 18) 39.83 23.90 19.56 21.76 20.00 1.69
b
5.96 2.69 5.74 9.35 8.26 0.73
0.74 0.19
