|Home | About | Journals | Submit | Contact Us | Français|
The metabolic pathways utilized by higher eukaryotic organisms to deal with potentially carcinogenic xenobiotic compounds from tobacco smoke have been well characterized. Carcinogenic compounds such as polycyclic aromatic hydrocarbons are metabolized sequentially in two-phases: in phase I CYP1A1 catalyzes conversion into harmful hydrophilic DNA adducts, while in phase II GSTT1 enables excretion via conjugation into polar electrophiles. In an effort to understand susceptibility to in utero tobacco exposure, we previously characterized known metabolic functional polymorphisms and demonstrated that while deletion of fetal GSTT1 significantly modified birth weight in smokers, no polymorphism fully accounted for fetal growth restriction. Since smoking upregulates CYP1A1 expression, we hypothesized that non-allelic (epigenetic) dysregulation of placental CYP1A1 expression via alterations in DNA methylation (meCpG) may further modify fetal growth. In the present manuscript, we compared placental expression of multiple CYP family members among gravidae, and observed significantly increased CYP1A1 expression among smokers relative to controls (4.4-fold, p<0.05). To fully characterize CYP1A1 meCpG status, bisulfite modification and sequencing of the entire proximal 1 kb promoter (containing 59 CpG sites) was performed. CpG sites immediately proximal to the 5′-XRE transcription factor binding element were significantly hypomethylated among smokers (55.6% vs 45.9% meCpG, p=0.027), a finding which uniquely correlated with placental gene expression (r=0.737, p=0.007). Thus in utero tobacco exposure significantly increases placental CYP1A1 expression in association with differential methylation at a critical XRE element.
The metabolic pathways utilized by higher eukaryotic organisms to deal with potentially carcinogenic xenobiotic compounds from tobacco smoke have been well characterized (1–4). While differing susceptibilities to in utero tobacco exposure are not understood, genetic epidemiologic studies have linked cancer risk due to an altered ability to metabolize carcinogenic polycyclic aromatic hydrocarbons (PAH), or addictive (nicotine) compounds into less harmful intermediates along these characterized pathways (1–4).
Of the over 4000 substances in tobacco smoke, PAH compounds together with nitrosamines comprise the likely carcinogenic species (2–4). The majority of chemical carcinogens are metabolized in a sequential series of two-phase enzymatic metabolic reactions (Figure 1A). Phase I enzymes, such as the cytochrome P450 arylhydrocarbonhydroxylases (AHH), catalyze the conversion of PAH into reactive hydrophilic intermediates which have the potential to form harmful DNA adducts. Following high-affinity binding of PAH compounds to their intracellular aryl hydrocarbon (AH) ligands, the complex is translocated to the nucleus where it dissociates then heterodimerizes to form a DNA binding complex (AH:ARNT) to modulate chromatin disruption and regulate induction of CYP1A1 expression (2–4). Induced CYP1A1 thereby drives conversion of PAH into the hydrophilic intermediates to increased PAH-DNA adducts. In turn, these reactive electrophilic intermediates are made excretable by phase II enzymes, such as the GSTT1 gene product, via conjugation into excreted polar electrophiles. Theoretically, any association of increased Phase I activity (e.g., increased expression of CYP1A1) in combination with decreased Phase II activity (e.g., decreased GSTT1 expression) may yield increased susceptibility to tobacco-related adverse outcomes.
Maternal tobacco use has long-been identified as a major independent risk factor for intrauterine growth restriction (IUGR; 5–9). However, not all fetuses exposed to tobacco smoke experience growth restriction (5). This discrepancy cannot be accounted for by dose effect alone (10–12). Thus, current efforts aimed at understanding the potential genetic, epigenetic, and metabolic bases of this variable susceptibility to tobacco smoke exposure are of importance in perinatal medicine. Wang et al first reported in 2002 that an association between maternal cigarette smoking and infant birth weight differs by polymorphisms of two maternal metabolic genes: CYP1A1 and GSTT1 (13). These authors’ analysis was limited to maternal genotype alone, and a test of interaction between maternal smoking and maternal CYP1A1 and GSTT1 genotypes on birth weight ratio was not statistically significant.
We have recently and significantly expanded the scope of these authors’ original analysis with paired maternal and fetal samples from a large, prospective study conducted through the NICHD Maternal-Fetal Medicine Units Network (Aagaard-Tillery et. al. in press). Specifically, we performed blinded genotyping for known functional allelic variants of CYP1A1 (Ile462ValAA>AG/GG), GSTT1 (deletion), and CYP2A6 (Lys160HisT>A) in smokers and their offspring alongside 1:1 matched controls. In our analysis, deletion of fetal GSTT1 was singularly observed to significantly reduce the fetal birth weight ratio among smokers (p for interaction of 0.02). However, our study failed to fully account for susceptibility to fetal growth restriction per se. Thus, it remained a formal possibility that dysregulation in the expression of these integral genes (or their metabolic pathways) may play a significant role in modifying fetal growth in response to maternal tobacco use.
In considering potential candidates, two lines of evidence led us to focus on regulation of expression of the Phase I CYP1A1 gene. First, aromatic hydrocarbon emissions are derived from both combustion of fossil fuels (coal, diesel, and gasoline) and environmental tobacco smoke. Multiple population-based analyses have demonstrated that the risk of fetal death, premature birth, and low birth weight is significantly higher for those with high prenatal ambient PAH exposure from all sources (14–20). Extension of these studies to include direct exposure measures (i.e., quantitation of PAH level by personal air monitoring) reveal significant interactions between maternal CYP1A1 haploytpe and exposure to hydrocarbons on the detected level of PAH-DNA adducts present in cord blood (21). Second, in both human and animal models, environmental tobacco smoke induces aryl hydrocarbon hydroxylase activity and placental expression of CYP1A1 (22, 23).
Emerging evidence has shown that in addition to genomic base pair differences, gene expression can be silenced by non-allelic mechanisms including epigenetic influences such as covalent modifications of histones and DNA methylation (recently reviewed in 24, 25). Other authors have previously observed that CYP1A1 is inducible in its placental expression among smokers and that well-characterized xenobiotic response elements (XREs) in the proximal promoter are differentially methylated at CpG islands in lung tissue of smokers (26–29). Given these published observations of others alongside our prior observations we hypothesized that non-allelic modulation of CYP1A1 expression may contribute to risk of adverse pregnancy outcomes (5,30,31). Since hypermethylation in key gene regulatory sequences at CpG islands is generally associated with gene silencing, we sought to compare placental gene expression of multiple CYP family members among gravidae, and determine methylation status of the proximal promoter region of the CYP1A1 gene.
We report that placental CYP1A1 expression was upregulated 4.4 fold (p=0.003) in smokers compared with non-smokers. Consistent with hypomethylation of crucial promoter regions increasing gene transcription, the CpG sites proximal to the XRE transcription factor binding element in Region I (CYP1A1-I) was specifically and significantly hypomethylated among smokers (55.6% vs 45.9% meCpG, p=0.027). Thus maternal smoking and aromatic hydrocarbon exposure leading to observed increases in DNA adduct formation in cord blood occurs via an epigenetic dysregulation of the CYP1A1 promoter at a region surrounding a critical XRE element. We speculate that such hypomethylation at critical regulatory regions may serve as an epigenetic signature to leave a transcriptional memory of fetal exposure to maternal smoking, thereby potentially predisposing an individual to the further generation of reactive and potentially carcinogenic DNA adducts.
Outcome data and placental samples for this study were obtained from a cohort of gravidae who had smoked during their pregnancy alongside non-smoking controls (n=34 total). This study was conducted for the purpose of evaluating differential gene expression and methylation in association with in utero tobacco exposure. The Institutional Review Board of Baylor College of Medicine and its affiliated institutions approved this study, and written informed consent was obtained from each participant at the time of enrollment. Inclusion criteria included singleton gestations at term (>35 weeks gestation) with self-admitted tobacco use or non-use controls (10, 32–35). Exclusion criteria included multiple gestations, a priori known fetal anomalies or aneuploidy, chorioamnionitis, and maternal hepatic, hypertensive, or endocrine disorders. Data abstraction occurred at the time of enrollment, and included potential maternal comorbidities. Univariate analyses of our cohort failed to reveal significant differences among the groups by virtue of maternal age, BMI, gestational age, or fetal weight at delivery. As anticipated, fetal length was influenced by in utero tobacco use (Table I).
We have previously demonstrated that among the known functional polymorphisms for CYP1A1 and GSTT1, only deletion of fetal GSTT1 significantly and specifically modified the effect of smoking on gestational age corrected birth weight. It remained a formal possibility that non-allelic regulation of CYP1A1 expression could be further rendering adverse events in response to in utero tobacco exposure. We therefore analyzed placental expression of four CYP family genes (CYP2A6, CYP2E1, CYP1B1 and CYP1A1) and observed that CYP1A1 expression was specifically and significantly increased 4.4 fold among smokers compared with non-smokers (p=0.003; Figure 1B). In contrast, neither expression of the other CYP genes nor GSTT1 differed by virtue of maternal smoking behavior (Figure 1B).
The methylation sites of the CYP1A1 functional promoter in lung tissue from smokers and non-smokers have been previously characterized by other investigators (26). Based on relative proximity of CpG dinucleotides to critical transcriptional binding elements (i.e., XRE sites), these authors designated 5 regions (Regions I-V) and subsequently determined the capacity for differential methylation among these regions with bisulfite modification and sequencing (26). Employing such an approach, these authors observed that the CpG dinucleotides proximal to XRE binding elements were differentially methylated in bronchial lavage samples between active smokers and long-term ex-smokers, while those distal from such elements did not demonstrate differential methylation (26). In order to characterize differential placental methylation in the CYP1A1 promoter in response to maternal tobacco use, we employed two similar approaches.
As a first step in our analysis of the functional promoter, we bisulfite modified then amplified CpG dinucleotides in the CYP1A1 promoter by a priori designated regions as previously described (26): Region I (−1411 to −1295), Region II (−1295 to −1006), Region III (−583 to −395) and Region IV (−395 to −228). In total, these regions contains a total of 59 CpG sites which have the potential to undergo differential methylation and include XRE transcriptional binding elements in Regions I and II (Figure 2A).. The total percent methylation for each region was calculated for both the smokers and non-smokers, and differences among regions were compared by the two-tailed Student t-test. As shown in Table II, the first core primed region (Region I) which contains an XRE transcriptional binding element, was unique in significant rate of methylation in smokers compared with non-smokers (55.6% vs 45.9% meCpG, p=0.027; Table II). In support of previously published data utilizing primary lung tissue (26) partial or no methylation was observed in placentas from smokers versus nonsmokers in primed regions II-IV (Figure 2B; Table II). Interestingly Region I contains a critical XRE which is known to regulate transcription of CYP1A1 in lung (26).
As a second and alternate method of analyzing differential methylation across the CYP1A1 promoter in relation to CpGs which surround an XRE, we alternately designated the CpGs within the functional promoter as belonging to one of 11 sequential (5′ to 3′) groups each containing 5 CpG dinucleotides. In such a manner, we were able to assess whether only those dinucleotides most proximal to and surrounding the XRE sites underwent differential methylation. As presented in Table III, among the 11 groups of 5 CpGs, only the two XRE-containing groups within these 55 CpGs demonstrated significant differential methylation (CpGs 1 to 5; p = 0.027 and CpGs 21 to 25; p = 0.006). Of note, the only group of dinucleotides with hypomethylation in association with smoking status was CpG 1 to 5, which was contained within Region I (Table 2 and and33).
Differential methylation of CpG dinucleotides in genomic DNA is generally considered to correlate with altered transcription (24,25). However, the direct evidence for true correlations in complex mammalian systems is limited. Given our differential methylation surrounding the XRE element in Region I, we therefore sought to better correlate placental CYP1A1 expression with the level of site-specific methylation of Region I. To do so, we plotted the relative expression level of CYP1A1 against the percent methylation for each region of the CYP1A1 promoter in both smokers and non-smokers and interrogated the relationship with bivariate correlations (Pearson’s correlation for variance) employing a two-tailed test for significance. As demonstrated in Figure 3, we found that the percent methylation of Region I inversely correlates with expression level (r = −0.737, p=0.007); this correlation held true regardless of maternal smoking behavior. In a linear regression model controlling for the potential covariates of fetal gender and maternal comorbidities, percentage CpG methylation in Region I independently predicted CYP1A1 expression (d.n.s). Moreover, there was no correlation between methylation status of Regions II-IV and placental CYP1A1 expression (d.n.s).
The metabolic pathways utilized by higher eukaryotic organisms to deal with potentially carcinogenic xenobiotic compounds from tobacco smoke have been well characterized (1–4, 28). Several CYP family members are involved in the Phase I processing of these xenobiotics into reactive oxygenated intermediates (ROIs), which are further processed by the Phase II enzymes to form hydrophilic excreted conjugated compounds. The contribution of the CYP family members in the production of ROIs (which can not only form harmful DNA adducts but also set the “tumorigenesis machinery in motion”) has also been intensively studied (29). An increase in Phase I activity, which increases cellular levels of ROIs, without a subsequent increase in Phase II activity to rid the cell of the ROIs, leads to the accumulation of DNA adducts. With respect to fetal development and maternal tobacco use, it has been recently appreciated that aryl hydrocarbon activity and DNA adducts are concomitantly increased in the cord blood of smokers (15, 21, 23). The molecular mechanisms underlying such observations have been poorly understood.
Although there exists a number of acknowledged genetic and environmental factors which further influence birth weight, an established causal relationship with tobacco use and delivery of SGA infants exists (5–12). In 1957, Simpson and Linda reported their observations that infants born to mothers who smoked 10 cigarettes or more per day weighed an average 200 grams less than those delivered by reported non-smokers (36). In the interval since, multiple authors have repeatedly demonstrated the persistence of this association with Relative Risk estimates ranging from 1.5 to 2.9 (reviewed in 12). Moreover, causality has been implicated by repeated observations of a dose-response relationship as well as a positive effect of smoking cessation on fetal weight. Despite this causal relationship, an essential observation remains: not all infants exposed to tobacco are small for gestational age. It is therefore likely that the effect of smoking on fetal growth involves interactions between multiple epidemiologic, genetic or epigenetic, and sociodemographic factors.
Other investigators had previously demonstrated that CYP1A1 expression increases in the lung tissue of smokers (reviewed in 30). We similarly find a significant increased CYP1A1 expression in placental tissue from smokers (4.4 fold). This observed increase in CYP1A1 expression was not accompanied by an increase in the Phase II enzyme, GSTT1 nor other CYP enzymes. This is of probable true biologic relevance for a number of reasons.
First, we have previously demonstrated that fetal homozygous deletion of the singular phase II PAH gene integral to excretion of DNA adduct forming reactive intermediates (GSTT1) significantly and specifically modifies fetal growth patterns in response to maternal smoking (Aagaard-Tillery et al., in press Obstet Gynecol). These findings persisted in multiple allelic interaction models to suggest an interaction between the fetal metabolic gene GSTT1, maternal smoking, and modification of birth weight. Of note, 18–22% of the population carries a homozygous deletion of GSTT1. As discussed, phase I gene-products, such as CYP1A1, are integral in metabolic activation of PAH compounds into oxidized derivatives, resulting in reactive oxygen intermediates capable of covalently binding DNA to form adducts; as a balance to such intermediary forming reactions, conjugation with endogenous species to form hydrophilic glutathione conjugates which are then readily excreted occurs. Our data support the notion that the discrepant variation in fetal susceptibility to smoking-related growth restriction may result from the diminished ability of the fetus to excrete these reactive intermediates (fetal phase II GSTT1(del)). However, the placenta is uniformly capable of delivering the first “hit” to such fetuses by virtue of tobacco-induced increased CYP1A1 expression.
Second, we have built on these observations and demonstrated that increased placental CYP1A1 expression was specifically and significantly associated with hypomethylation of the CYP1A1 promoter region in smokers compared with non-smokers (Figure 2 and Table II). Region I, which contains an XRE element that is involved in regulation of CYP1A1 expression, was the only region which demonstrated significant differential methylation within the proximal promoter (Table II). This association held true within an individual, as there was a significant correlation between CYP1A1 expression and Region I hypomethylation (Figure 3).
Thus, although our observations are unique in collective concept and findings, they are not without biologic plausibility and merit. First, mechanisms leading to IUGR following in utero tobacco exposure have generally often been attributed to chronic fetal hypoxia, yet, nicotine, cotinine and DNA adducts are known to cross or collect in the placenta of smokers. As discussed, phase I gene-products, such as CYP1A1, are integral in metabolic activation of PAH compounds into oxidized derivatives, resulting in reactive oxygen intermediates capable of covalently binding DNA to form adducts; as a balance to such intermediary forming reactions, conjugation with endogenous species to form hydrophilic glutathione conjugates which are then readily excreted occurs. Thus while it is possible that chronic hypoxia is a primary mediator of IUGR in response to in utero tobacco exposure, our data support the notion that the discrepant variation in fetal susceptibility to smoking-related growth restriction results from the diminished ability of the fetus to excrete these reactive intermediates (fetal phase II GSTT1(del)).
There are a limited number of methodological limitations which ought to be considered when broadly interpreting our findings. First, our samples were obtained from a relatively small cohort of smokers. Nevertheless, statistical significance was readily reached. Second, tobacco smoke contains multiple compounds, and while we have previously established functional polymorphisms along well-established metabolic pathways related to varying cancer susceptibility, the relative role of these genes in perinatal outcomes is unknown. Third, we cannot ascertain a dose-dependent effect nor have we directly assayed for serum or urine cotinine levels. However, multiple previous studies have validated maternal self-reporting of smoking behavior with 96% of “nonsmokers” having nondetectable serum cotinine levels at <10 ng/ml, in accordance with other authors findings (31–35).
Despite these limitations, the strengths of our study are several. First, we have provided an exhaustive characterization of the near entirety of the CYP1A1 proximal promoter. This encompasses more than 1 kb, and 59 CpG sites. In total, we examined over 250 clones to encompass over 5000 CpG site reads to determine differential methylation. Second, we have attempted to directly correlate gene expression with site specific (rather than “global” or “net”) DNA methylation as a measure of causality. Third, we have performed these analyses in the most relevant tissue with relation to uteroplacental insufficiency.
Recent advances in human genome research, pharmacogenetics, medical genetics, and the evolving field of epigenetics have furthered our understanding of the interactions of the heritable genome and chromatin structure in the causal pathways employed in the development of human disease (24, 25). The concept of epigenetic mechanisms providing a “memory” of previous transcriptional activation has also been proposed from yeast to man (37, 38). The “developmental origins of adult disease” hypothesis builds on these observations to postulate that fetal exposures in utero can cause an epigenetic change within the fetus which can manifest in adulthood as an increased potential for disease (39, 40).
Reports of the changes in DNA methylation patterns of specific genes attributed to gestational milieu are emerging in recent literature. In human population-based analyses, targeted genomic methylation profiling in known differentially methylated regions (DMRs) from peripheral mononuclear cells among adults which were exposed in utero to calorie restriction during the Dutch Hunger Winter demonstrated differential methylation at the imprinted IGF2 region (39). In animal models, alterations in methylation patterns due to in utero exposure to maternal protein restriction can be seen even in the F2 generation (40).
Here we present the first evidence that maternal smoking alters DNA methylation levels in utero in human placenta. Because of the limitations of biologic material we did not assay for CYP1A1 levels nor characterize its promoter directly in fetal blood nor hepatic tissue. Yet based on the reports of Jauniaux and Burton we would anticipate that CYP1A1 levels in the fetus may similarly differ and would do so with a similar mechanism (41). If indeed this where observed to be true, then differential methylation of the CYP1A1 promoter would serve as an epigenetic mechanism allowing for a molecular “memory” of fetal exposure to maternal tobacco smoke. In turn, based on the available evidence from both population based epigenetic studies (39) and animal models (40, 42, 43), this may function as a mechanism by which in utero tobacco exposure might potentially manifest in adulthood as an increased susceptibility to disease.
Along these lines, not only has a correlation between maternal smoking and a myriad of fetal morbidities has been established, but the susceptibility of the tobacco-exposed fetus to later childhood disorders has been well described. Maternal smoking has been linked not only to reduced birth weight, but a predisposition to asthma, reduced fecundity, and obesity in adulthood (44–46). Therefore, the potential exists that altered DNA methylation patterns established in utero can influence the etiology of adult diseases. Importantly, as established in animal models, in cases where an in utero exposure leads to hypomethylation of specific regions of the genome, supplementation of the diet with cofactors in the one carbon metabolism pathway alleviate the differential methylation (47–49). Such cofactors are found in standard prenatal vitamins (49).
Our collective data provide likely potential insights into such a mechanism. The implications of our findings are two-fold. First, our prior data illustrate that a fetal metabolic gene (GSTT1) which is integral in the excretion of reactive intermediates of aromatic hydrocarbons modifies fetal growth specifically in response to in utero tobacco exposure. These findings imply that tobacco metabolites may reach the fetus and thus modify fetal growth if not excreted. Second, our current data provide the epigenetic correlation for the ubiquitous increased expression of placental CYP1A1 and subsequent DNA adduct accumulation which has been previously observed (14–17, 32). Future studies aimed at illuminating the complex interplay of genomic-epigenomic-environmental interactions may help dissect multifactorial etiologies and identify at-risk populations for the common adverse pregnancy outcomes.
Outcome data and placental samples for this study were obtained from a cohort of gravidae who had smoked during their pregnancy alongside non smoking controls. Patients were admitted to either Ben Taub General Hospital or St. Luke’s Episcopal Hospital (Houston, TX). Inclusion criteria included singleton gestations at term (>35 weeks gestation) with self-admitted tobacco use or non-use controls (10, 31–33). Exclusion criteria included multiple gestations, a priori known fetal anomalies or aneuploidy, chorioamnionitis, secondary substance exposure (e.g., cocaine, heroine, marijuana, or others as detected in routine urine toxicology assessment), and maternal hepatic, hypertensive, or endocrine disorders. Data abstraction occurred at the time of enrollment, and included potential maternal comorbidities. Toxicology screens were administered when clinically indicated.
Within 30 minutes of delivery, equal full-thickness sections of placenta were uniformly collected and stored in 50mL conical tubes which were immediately placed on dry ice. Samples were stored at −80°C until use in accordance with IRB approved protocols.
Approximately 50mg of tissue was lysed using a chaotropic buffer, and RNA was extracted from each placenta using the Machery Nagel NucleoSpin kit (Bethlehem, PA). RNA was quantitatively converted to cDNA from each sample using the Applied Biosystems High Capacity cDNA Reverse Transcription Kit (Foster City, CA). Real-time quantitative RT-PCR (qRT-PCR) analyses were performed using 2μL cDNA, and 2μM final concentration of forward and reverse primers and TaqMan probes in a total reaction volume of 5μL. We used the iQ5 Real-Time PCR Detection System from BioRad. Relative quantification of each gene was calculated after normalization to GAPDH by using the comparative Ct method (50).
Sodium bisulfite treatment converts unmethylated cytidine into thymidine. Methylation of CpG sites can thus be detected by recognition of unmodified cytidine at CpG sites by sequencing. The EZ DNA methylation kit (Zymo Research, Orange, CA) was used to bisulfite modify and purify sample DNA. In a modification of previously published methods, four fragments of the CYP1A1 promoter region 1400-bp upstream of the gene were amplified: Region I (−1411 to −1295), Region II (−1295 to −1006), Region III (−583 to −395) and Region IV (−395 to −228). Primer sequences are as follows:: Region I 5′-GTTAGTTGGGGTTAGGTTGAG-3′ (sense) and 5′-CATAACCTAACTACCTACCTCC-3′ (antisense); Region II 5′-GTTAGTTGGGGTTAGGTTGAG -3′ (sense) and 5′-AAACCCCCACCCTACCCCCC-3′ (antisense); Region III 5′-GGGTTTTGGGGGATAGGTTT-3′ (sense) and 5′-CG/ATACAAATACCTCCCCAAC-3′ (antisense) Region IV 5′-GGAAGGAGGTTATTAA/TGGGG-3′ (sense) and 5′-CACCTAAAAATCCCAATTCCAA-3′ (antisense). PCR conditions were as follows: 5 μl template DNA, 5 μl 10X ABI (PE Applied Biosystems, Warrington, UK) Buffer, 4.5 μM MgCl2, 4 μl 2.5μM dNTP, 5 μl 10 μM sense and antisense primers, and 0.5 μl TaqGold (ABI) Polymerase. Reactions were run under conditions of 95°C 15 min, with 40 cycles (95°C 30 sec, 56°C 30 sec, 72°C 1 min) with 10 min 72°C elongation for adenylation. The amplified PCR products were subsequently cloned, and a minimum of four clones from each reaction were sequenced. Samples were sequenced by Lone Star Labs (Houston, TX) and aligned using CLUSTAL W (Kyoto University). Figure 2A, which depicts the CpG sites, was created by the ABI Methyl Prime software.
Univariate comparisons were performed using Chi-square or Fisher’s exact test for discrete variables and the Wilcoxon test for the continuous variables. Real Time PCR was analyzed using the 2−ΔΔCT method as previously described (50). Results are displayed as fold change of smokers compared with non-smoking control samples. An independent samples students T-test was performed for each gene. An independent samples, one-tailed students T-test was performed to analyze the differential methylation between smokers and non-smokers given the a priori assumption of hypomethylation. For analysis of correlation, relative gene expression was calculated using the ΔCT from each subject and plotted against the percent methylation for that individual with analysis of variance by Pearson’s coefficient. SPSS (SPSS, INC) was uses for analyses and a nominal p-value <0.05 was considered significant. No adjustment was made for multiple comparisons.
This work was supported by the NIH Director New Innovator Award (DP2120OD001500-01 K.A.T.) and the NIH REACH IRACDA (K12 GM84897 M.S.).
The authors have no conflicts of interest nor financial disclosures.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.