EPIGENETICS 2023, VOL. 18, NO. 1, 2214392 https://doi.org/10.1080/15592294.2023.2214392 RESEARCH ARTICLE Impact of tobacco, alcohol, and marijuana on genome-wide DNA methylation and its relationship with hypertension Natàlia Carreras-Gallo a, Varun B. Dwarakab, Alejandro Cáceres a,c,d , Ryan Smithb, Tavis L. Mendezb, Hannah Wentb, and Juan R Gonzalez a,c,e a Epidemiology, Barcelona Institute for Global Health (ISGlobal), Barcelona, Spain; bTruDiagnostic, Lexington, KY, USA; cCentro de Investigación Biomédica en Red en Epidemiología y Salud Pública (CIBERESP), Madrid, Spain; dDepartment of Mathematics, Escola d’Enginyeria de Barcelona Est (EEBE), Universitat Politècnica de Catalunya, Barcelona, Spain; eDepartment of Mathematics, Universitat Autònoma de Barcelona, Barcelona, Spain ABSTRACT ARTICLE HISTORY Tobacco, alcohol, and marijuana consumption is an important public health problem because of Received 18 October 2022 their high use worldwide and their association with the risk of mortality and many health Revised 13 April 2023 conditions, such as hypertension, which is the commonest risk factor for death throughout the Accepted 9 May 2023 world. A likely pathway of action of substance consumption leading to persistent hypertension is KEYWORDS DNA methylation. Here, we evaluated the effects of tobacco, alcohol, and marijuana on DNA Tobacco; Alcohol; Marijuana; methylation in the same cohort (N = 3,424). Three epigenome-wide association studies (EWAS) DNA methylation; were assessed in whole blood using the InfiniumHumanMethylationEPIC BeadChip. We also Epigenome-wide association evaluated the mediation of the top CpG sites in the association between substance consumption study; Hypertension and hypertension. Our analyses showed 2,569 CpG sites differentially methylated by alcohol drinking and 528 by tobacco smoking. We did not find significant associations with marijuana consumption after correcting for multiple comparisons. We found 61 genes overlapping between alcohol and tobacco that were enriched in biological processes involved in the nervous and cardiovascular systems. In the mediation analysis, we found 66 CpG sites that significantly mediated the effect of alcohol consumption on hypertension. The top alcohol-related CpG site (cg06690548, P-value = 5.9·10−83) mapped to SLC7A11 strongly mediated 70.5% of the effect of alcohol consumption on hypertension (P-value = 0.006). Our findings suggest that DNA methyla­ tion should be considered for new targets in hypertension prevention and management, parti­ cularly concerning alcohol consumption. Our data also encourage further research into the use of methylation in blood to study the neurological and cardiovascular effects of substance consumption. KEY POLICY HIGHLIGHTS ● The consumption of tobacco, alcohol, and marijuana is very high worldwide and is associated with common diseases, like cardiovascular and neurological disorders. ● This study found that tobacco and alcohol have large effects on genome wide DNA methyla­ tion while marijuana consumption has nonsignificant effects. ● The genes differentially methylated were enriched in pathways related to neurodevelopment, suggesting the mediation between recreational drug consumption and neurological disorders. ● More remarkably, 66 alcohol related CpG sites significantly mediated the association between heavy drinking and hypertension. ● Our findings suggest that DNA methylation changes should be considered for new targets in disease prevention for recreational drug consumers. Introduction illicit drug globally [2]. The consumption of these Tobacco, alcohol, and marijuana are the most substances is increasing, mainly among adoles­ commonly used drugs of abuse in the United cents, and the health and social problems asso­ States [1]. While tobacco and alcohol consumption ciated with them are an important public health is legal, marijuana is the most commonly used concern. CONTACT Juan R Gonzalez juanr.gonzalez@isglobal.org Barcelona Institute for Global Health (ISGlobal), Barcelona, Spain Supplemental data for this article can be accessed online at https://doi.org/10.1080/15592294.2023.2214392 © 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent. 2 N. CARRERAS-GALLO ET AL. Addiction is a major consequence of consuming 9-tetrahydrocannabinol (THC), which is psy­ tobacco, alcohol, and marijuana, which results in choactive, and cannabidiol (CBD), which is non- the strengthening of all the other health problems psychoactive. These compounds participate in the associated with them [3]. endocannabinoid system [13]. Cigarette smoking is the leading cause of pre­ In summary, the consumption of tobacco, alco­ ventable death and disease in the US and is hol, and marijuana can lead to similar adverse responsible for approximately 8 million worldwide effects on health, with the neurological and cardi­ deaths every year [4]. Most of these deaths arise ovascular systems being particularly affected. from cancers (mainly lung cancer), respiratory Recent research suggests that epigenetics is disease, and cardiovascular disease [5]. Cigarette a potential mediator between the consumption of smoke contains more than 7000 chemical com­ toxic substances and the increase in common dis­ pounds. Among them, 1,3-butadiene is highly ease risk [14–17]. DNA methylation, the most associated with cancer risk, acrolein, and acetalde­ studied epigenetic modulation, consists of the hyde are potential respiratory irritants, cyanide, addition of a methyl group (−CH3) in the cytosine arsenic, and cresols are the primary sources of nucleotide without changing the DNA sequence. It cardiovascular risk, and nicotine is the additive occurs in the context of CpG sites, which are component [6]. In addition, nicotine exposure defined as adjacent cytosine and guanine nucleo­ causes well-characterized neurotoxic effects, tides by a phosphate group. DNA methylation is which are highly important in early develop­ dynamic and can be modified by genetic factors, ment [7]. disease, environmental exposures, and lifestyle Light to moderate alcohol intake is associated [17–19]. Moreover, DNA methylation can change with reduced risks for total mortality, cardio­ during the lifetime and across tissues and cell types vascular disease, and diabetes. However, exces­ [20,21]. sive alcohol is the third leading cause of Although genetic mechanisms have been the premature death in the US [8]. Heavy alcohol focus of understanding human diseases, the dis­ use is associated with a higher risk of cardio­ ruption of the epigenetic balance can result in the vascular disease, diabetes, cirrhosis of the liver, modulation of gene expression. Consequently, epi­ pancreatitis, and cancer [9]. One of the causes genetic disruption can cause several major pathol­ of these health consequences is the toxic effect ogies, including cancer and cardiovascular disease of acetaldehyde, which comes from the metabo­ [14]. Tobacco smoking is one of the exposures lization of ethanol [10]. As an example, alcohol with a higher impact on the DNA methylation of and acetaldehyde act as direct toxins to the smokers [22–25]. Its effect has also been observed cardiac myocytes, resulting in contractile dys­ in the newborns’ cord blood of mothers who function [11]. Furthermore, excessive exposure smoked during pregnancy [26–28]. Many studies to alcohol can lead to severe debilitating dis­ also demonstrate that alcohol consumption pro­ eases of the central and peripheral nervous sys­ duces methylation changes at the CpG site level tems [12]. [29–34]. Conversely, only a few studies have Among marijuana health impacts, disturbances demonstrated the effects of marijuana consump­ in the level of consciousness, cognition, percep­ tion on DNA methylation, all of which have tion, affect or behaviour, and other psychophysio­ shown small effects [35,36]. logical functions and responses are known as Hypertension, also known as high blood pres­ short-term effects. Additionally, long-term mari­ sure, is a medical condition in which the blood juana consumption can increase the risk of depen­ pressure in the arteries is persistently elevated. It dence, cognitive impairment, mental disorders affects one billion people and is the most common (psychoses, depression, anxiety, and suicidal beha­ risk factor for death worldwide [37]. There are viour), and adverse physical health effects such as many factors associated with a higher risk of cardiovascular disease, chronic obstructive pul­ hypertension, including body mass index, tobacco monary disease, and cancer [2]. The main compo­ use, physical activity, and alcohol consumption, nents of the cannabis plant are delta- among others [38]. Light to moderate alcohol EPIGENETICS 3 consumption seems to protect against hyperten­ approved by the IRCM IRB. The patients/participants sion because it decreases systolic and diastolic provided their written informed consent to participate blood pressure. However, excessive intake in this study. accounts for about 16% of cases of hypertension worldwide [39,40]. Cigarette smoking enhances DNA methylation Assessment hypertension by inducing cardiovascular mito­ chondrial oxidative stress [41,42]. On the other Peripheral whole blood was collected by the lancet hand, some studies evaluating the effect of mari­ and capillary method into lysis buffer and DNA juana consumption on blood pressure have extract, and 500 ng of DNA of bisulphite were revealed different results. For instance, Abuhasira converted using the EZ DNA Methylation kit et al. demonstrated the therapeutic effect of mar­ (Zymo Research) according to the manufacturer’s ijuana in reducing blood pressure in hypertensive instructions. Bisulphite-converted DNA samples patients [43]. However, other studies revealed an were randomly assigned to a chip well on the increase in blood pressure after marijuana con­ Infinium HumanMethylationEPIC BeadChip, sumption [44,45]. In light of the above, we amplified, hybridized onto the array, stained, hypothesized that changes in DNA methylation washed, and imaged with the Illumina iScan SQ produced by substance consumption may partially instrument to obtain raw image intensities. explain its relationship with hypertension. Meffil R package [48] was used for the pre- To this end, in this study we aimed to: i) per­ processing of DNA methylation data. In the sample form a genome-wide association study of DNA quality control, we removed the sex detection mis­ methylation with tobacco, alcohol, and marijuana matches and the sex detection outliers (based on the consumption; ii) identify the physiological path­ difference between median chromosome Y and chro­ ways whose methylation is affected by those mosome X probe intensities). We also discarded those recreational drugs; iii) evaluate the mediation samples whose predicted median methylated signal between substance consumption and hypertension was more than 3 standard deviations from the by methylation changes at the CpG site level. expected. We excluded the outliers based on devia­ tions from mean values for control probes (dye bias, bisulphite 1, and bisulphite 2). Finally, we removed Materials and methods those samples with more than 5% of undetected probes (detection P-value larger than 0.01) or with The study cohort a low number of beads (less than 3). This quality Our study sample included 3,590 individuals from the control resulted in 3,424 individuals, indicating that TruDiagnostic DNA biobank recruited between 90,3% of the samples met our QC standards. In the October 2020 and February 2022. Those individuals feature quality control, we removed those probes have taken the commercial TruDiagnostic TruAge test undetected or with low bead numbers in more than and methylation data was generated from them. This 5% of the samples. We used InfiniumAnnotation [49] is an EEUU population-based cohort aged between 13 to filter probes where the 30bp 3’-subsequence of the and 97 years old. Among them, 58.7% are male. probe is non-unique, probes with INDELs, probes Demographic and substance use characteristics of with extension base inconsistent with specified colour the samples that met the QC requirements (N = channel (type-I) or CpG (type-II) based on mapping, 3,424) are displayed in Table 1. As this testing is priced probes with a SNP in the extension base that causes to the consumer at approximately $500, this study a colour channel switch from the official annotation, cohort is relatively more affluent than random sam­ and probes where 5bp 3’-subsequence overlap with pling or traditional banked cohorts. Additionally, any of the SNPs with global population frequency these individuals may experience a healthy donor higher than 1%. The functional normalization method effect whereby they seek preventative medicine and was further applied based on the first 10 principal have fewer comorbidities than normal patient popula­ components of the control probes. Consequently, the tions, as is the case with blood donors [46,47]. The number of CpG probes kept was 745,150, which repre­ study involving human participants was reviewed and sents 86% of the total EPIC array manifest. CpG sites 4 N. CARRERAS-GALLO ET AL. Table 1. Characteristics of participants in the TruDiagnostic DNA Biobank. N = 3424 Sex, male 2010 (58.7%) Age in years, mean (range) 52.9 (13.3–97.8) Ethnicity European 2584 (75.5%) African American or Black 70 (2.0%) Asian 41 (1.2%) Latino or Hispanic 276 (8.1%) Middle Eastern or North African 76 (2.2%) Native American or Alaska Native 26 (0.8%) Pacific Islander or Oceanian 23 (0.7%) Sub-Saharan African 7 (0.2%) Other 321 (9.4%) BMI (kg/m2), median (range) 25.4 (10.1–58.2) Tobacco consumption None 3275 (95.6%) Less than 1 cigarette per week 48 (1.4%) Less than 1 cigarette per day 25 (0.7%) 1–5 cigarettes per day 27 (0.8%) 6–10 cigarettes per day 21 (0.6%) 11–20 cigarettes per day 20 (0.6%) More than 20 cigarettes per day 8 (0.2%) Alcohol consumption Never 634 (18.5%) On special occasions 976 (28.5%) Once per week 578 (16.9%) 3–5 times per week 794 (23.2%) Regularly 442 (12.9%) Marijuana consumption Missing 149 Never 2908 (88.8%) On special occasions 180 (5.5%) Once per week 46 (1.4%) 3–5 times per week 73 (2.2%) Regularly 68 (2.1%) Note: All the continuous variables are shown as mean (range) and the categorical variables as n (%). BMI: body mass index. were annotated to genes using EPIC Illumina annota­ level of smoking according to 7 possible answers tion ilm10b4.hg19. Blood cell type proportions were (‘none’ to ‘more than 20 cigarettes per day’). estimated using the blood gse35069 complete cell type Regarding the medical history, the survey covered methylation profile references from the meffil package. information about the blood type, medications and We then performed a surrogate variable analysis supplements, and diagnosis of any type of disease (SVA) to remove the batch effects using the (cardiovascular, respiratory, skin and hair, endocrine, SmartSVA package [50]. We estimated the number gastrointestinal, genitourinary, musculoskeletal, neu­ of surrogate variables (SVs) using the isva package ropsychological, reproductive, immune, and cancer). [51]. Methylation levels were expressed as residual The main clinical outcome of this study was hyperten­ values after adjusting beta values for the first 60 SVs. sion, assessed as a dichotomic variable (affected and unaffected. Exposure and clinical history assessment Statistical analyses and reproducibility During the recruitment of participants, they were Epigenome-wide association analysis asked to complete a survey that included questions The epigenome-wide association study (EWAS) about personal information, medical history, social was performed using the MEAL Bioconductor history, lifestyle, and family history. Alcohol and mar­ package [52]. We performed a differential mean ijuana consumption was assessed on a 5-point scale analysis on different substance consumption (‘never’ to ‘regularly’). Participants also reported their (tobacco, alcohol, and marijuana) using the EPIGENETICS 5 function runPipeline that calls limma [53]. Based containing information on the associations on previous analyses [32], we adjusted all the between DNA methylation and tobacco, alcohol, regression models by sex, age, ethnicity, body and marijuana consumption. The dataset for mass index (BMI), level of education, alcohol and tobacco smoking included 30 publications, while tobacco consumption (except when they were the the dataset for alcohol consumption had 6 publi­ variable of interest), slide, cell type, and surrogate cations, and the dataset for marijuana consump­ variables. For each substance, we fitted models tion contained 2 publications. After excluding associations that lacked information on the beta Ej¼ αj þβj SþΣr γr Cr þεj (1) value, we obtained 26 publications for smoking, no where Ej denotes the methylation level vector publications for alcohol consumption, and 1 pub­ across individuals at probe j (j = 1, . . . 745150), lication for marijuana consumption. For CpG sites S is the individuals’ consumption (separated mod­ that were reported in more than one study, we els for alcohol, smoking, and marijuana where performed a metanalysis using the meta package fitted) with its associated effect, βj, Cr is the [59]. However, due to the high heterogeneity r adjusting covariate and its effect γr, and εj is between studies, we reported the proportion of the noise that follows the distribution of methyla­ studies that showed the same direction of effect tion levels with mean 0. Adjusted P-values were as our study. calculated using Bonferroni’s correction for con­ sidering multiple comparisons. The inflation or Estimation of epigenetic clocks deflation of P-values across the methylome was assessed with Q-Q plots and lambda values [54]. We further evaluated whether the three behaviours Enrichment analysis had an impact on epigenetic age. To calculate the We performed an enrichment in Kyoto principal component-based epigenetic clock for Encyclopedia of Genes and Genomes (KEGG) the Horvath multi-tissue clock, Hannum clock, and Gene Ontology (GO) pathways [55,56] using PhenoAge clock, GrimAge clock, and telomere the enrichKEGG and enrichGO functions from the length we used the custom R script available via clusterProfiler Bioconductor package [57], respec­ GitHub (https://github.com/MorganLevineLab/ tively. Using the same package, we also evaluated PC-Clocks). Non-principal component-based the over-representation of diseases using the (non-PC) Horvath and Hannum epigenetic esti­ DisGeNET platform which contains 1,134,942 mates were calculated using the agep function gene-disease associations [58]. Associations were available in the wateRmelon R package [60], and corrected for multiple comparisons using non-PC PhenoAge was calculated using the Bonferroni adjustment, as computed by methyAge function in the ENMix R package [61]. clusterProfiler. Finally, the pace of ageing clock, DunedinPACE, Comparison of the results with previous studies was calculated using the PACEProjector package For the validation of our results, we used pub­ available via GitHub (https://github.com/dan licly available data from the EWAS catalogue. This belsky/DunedinPACE). catalogue includes published epigenome-wide association studies from PubMed using the search Mediation analysis between recreational drug terms ‘epigenome-wide’ OR ‘epigenome wide’ OR use and hypertension ‘EWAS’ OR ‘genome-wide AND methylation’ OR ‘genome wide AND methylation.’ Studies are To investigate the relationship between tobacco, selected according to the year of publication alcohol, and marijuana consumption and hyper­ (>2010), the use of human samples, and the num­ tension, we conducted logistic regression analyses. ber of genome-wide CpG sites (>100,000) and Next, we performed a mediation analysis to individuals (>100). Reported CpG sites have nom­ explore whether methylation played a role in the inal P-values lower than 1·10−4. The data for the relationship between recreational drug use and EWAS Catalogue is then manually curated. We hypertension. We first identified CpG sites that obtained separate datasets for each drug of abuse were significantly associated with those behaviours 6 N. CARRERAS-GALLO ET AL. and hypertension after correcting for multiple overlap was observed between alcohol and mar­ comparisons using Bonferroni. Then, we con­ ijuana, with 299 individuals indicating con­ ducted a univariate mediation analysis for each sumption of both substances. Notably, 98 significant CpG site using the mediate function participants reported consuming both tobacco from the mediation package [62]. This analysis and alcohol, highlighting the co-occurring use allowed us to estimate the total effect, the effect of different substances (Additional File 2: of substance consumption on CpG methylation, Fig. S1). and the effect of the mediator and substance con­ sumption on hypertension. Finally, we performed Genome-wide effect of tobacco smoking on a causal mediation analysis and estimated the aver­ DNA methylation age causal mediation effects, average direct effects, total effect of the independent variable on the We tested the association between the level of dependent variable, and the proportion of the smoking (codified as a 7-point scale from ‘none’ effect of the independent variable on the depen­ to ‘more than 20 cigarettes per day;’ see Table 1) dent variable that goes through the mediator. We with each CpG site using linear regression models adjusted all models for the same covariates as in run in the MEAL R package [52]. We found 528 the EWAS. CpG sites associated with smoking levels after correcting for multiple comparisons and genomic inflation was not observed (λ = 1.031). Table 2 lists Results the top 15 CpG sites for tobacco smoking (see We analysed 3,424 individuals from the Additional File 1: Table S1 for all the significant TruDiagnostic DNA Biobank recruited from CpG sites). Figure 1a shows how the CpG sites are the general population in EEUU. Table 1 pre­ distributed in the genome using a Manhattan plot. sents demographic and substance use character­ Among tobacco-related methylation sites, 68.2% istics. The mean age was 52.9 years (range: were hypomethylated (that is, lower DNA methy­ 13.3–97.8) and 58.7% were male. The partici­ lation associated with higher tobacco consump­ pants were classified according to 7 ethnic tion). From the 528 probes differentially groups and ‘other’ for those who had a mixed methylated, 374 CpG sites were mapped to 344 ethnicity. Most participants were Europeans unique genes. AHRR, GFI1, PRSS23, and (75.5%), and Latino American was the second IMMP2L had 10, 6, 4, and 4 probes differentially most common ethnicity (8.1%). There were 149 methylated, respectively. Moreover, these 344 current tobacco smokers, classified into seven genes were enriched in morphine addiction groups according to the number of cigarettes (P-value = 4.3 · 10−6), dopaminergic synapse smoked, and 3,275 non-smokers. Regarding (P-value = 1.1 · 10−4), and cholinergic synapse alcohol consumption, there were 2,790 drinkers (P-value = 1.7 · 10−4) KEGG pathways grouped by consumption frequency and 634 non-drinkers. Marijuana consumption was also (Additional File 2: Fig. S2). Consistent with pre­ classified according to consumption frequency. vious studies, cg05575921 was the top-ranked CpG In total, 2,908 did not smoke marijuana, and with a P-value = 1.3·10−226. We further demon­ 367 smoke marijuana at least on special occa­ strated that the effect of tobacco in this CpG site sions. It is worth noting that out of the parti­ was proportional to the number of cigarettes cipants, 465 individuals reported consuming smoked (Figure 2a). In addition, we cross- more than one drug, at least on special occa­ referenced our findings with those previously sions. On the other hand, 584 participants documented in the EWAS catalogue. Out of the reported not consuming any of the drugs 528 CpG sites that displayed differential methyla­ included in this study. Among those who tion in our investigation, 183 had already been reported using recreational drugs, the largest documented in the EWAS catalogue, with 181 Table 2. Top 15 differentially methylated CpG sites by tobacco consumption. Less than EWAS catalogue Gene Gene 1 per Less than More than Adjusted (same direction/total CpG chr position Symbol Group week 1 per day 1–5 per day 6–10 per day 11–20 per day 20 per day F P-Value P-Value studies) cg05575921 chr5 373378 AHRR Body −0.028 −0.054 −0.097 −0.216 −0.183 −0.217 208.2 1.3E–226 9.7E–221 26/26 cg21566642 chr2 233284661 −0.030 −0.058 −0.066 −0.109 −0.111 −0.160 59.7 1.6E–70 5.9E–65 22/22 cg01940273 chr2 233284934 −0.016 −0.034 −0.045 −0.066 −0.069 −0.109 53.6 2.4E–63 5.9E–58 22/22 cg03636183 chr19 17000585 F2RL3 Body −0.007 −0.017 −0.029 −0.064 −0.053 −0.094 50.9 2.8E–60 5.2E–55 21/21 ch.6.2768623F chr6 144037285 PHACTR2 Body −0.002 −0.001 0.000 0.009 −0.005 0.350 49.1 4.7E–58 7.0E–53 Not reported cg21161138 chr5 399360 AHRR Body −0.008 −0.016 −0.015 −0.071 −0.043 −0.069 40.6 5.7E–48 7.1E–43 16/16 cg17739917 chr17 38477572 RARA 5‘UTR −0.008 −0.053 −0.034 −0.069 −0.082 −0.116 38.2 4.5E–45 4.8E–40 7/7 cg15928392 chr2 81694446 0.000 0.000 0.001 −0.001 −0.002 −0.106 34.1 3.4E–40 3.2E–35 Not reported cg16774290 chr7 4824628 AP5Z1 Body 0.000 0.000 0.000 −0.001 0.000 −0.100 33.2 4.1E–39 3.4E–34 Not reported cg11173636 chr10 65632259 −0.008 0.017 0.024 0.018 0.339 −0.020 30.5 8.5E–36 6.3E–31 Not reported cg09935388 chr1 92947588 GFI1 Body −0.018 −0.030 −0.029 −0.100 −0.069 −0.134 30.0 3.5E–35 2.4E–30 20/20 cg26703534 chr5 377358 AHRR Body −0.004 −0.013 −0.016 −0.052 −0.034 −0.049 29.9 4.0E–35 2.4E–30 14/14 cg14990808 chr6 28493651 GPX5 TSS200 −0.002 −0.022 0.005 0.002 0.001 0.246 26.3 1.1E–30 6.2E–26 Not reported cg25648203 chr5 395444 AHRR Body −0.006 −0.005 −0.011 −0.047 −0.037 −0.057 24.8 6.2E–29 3.3E–24 18/18 cg04176674 chr14 21121564 −0.004 −0.002 0.001 0.005 0.000 0.343 24.7 7.8E–29 3.9E–24 Not reported Note: The CpG sites are annotated based on the chromosome (chr), the position (pos), the gene symbol from HGNC, and the gene group (based on the position of the CpG regarding the nearest gene). Each CpG site has a beta value for each consumer group vs non-consumers, a F-statistic (F), a nominal p-value, and an adjusted p-value by Bonferroni. The last column represents the number of studies in the EWAS catalogue that reported the same effect direction compared to the total studies that reported the beta value for the specific CpG site. EPIGENETICS 7 8 N. CARRERAS-GALLO ET AL. Figure 1. Manhattan plots of the epigenome-wide association study (EWAS) of tobacco (a), alcohol (b), and marijuana (c) consumption. The Y-axis represents the -log10(p) values and the X-axis the position of the CpG sites within the chromosome. The blue line is the suggestive nominal p-value threshold (0.0001) and the red line is the p-value adjusted threshold lower than 0.05. showing the same direction of change and 2 Genome-wide effect of alcohol consumption exhibiting an opposing direction. Among the on DNA methylation 345 not-reported, 6 of them were in the top 15 We identified 2,569 CpG sites differentially methy­ CpG sites in our data, evidencing that tobacco lated according to alcohol consumption frequency (5 may have important effects on them (Table 1). levels from ‘never’ to ‘regularly;’ see Table 1). Model EPIGENETICS 9 Figure 2. Boxplots showing the association between CpG methylation and substance consumption. (a) Association between cg05575921 methylation (AHRR) and tobacco consumption. (b) Association between cg06690548 methylation (SLC7A11) and alcohol consumption. The Y-axis represents the residuals for beta values after adjusting by covariates. The X-axis represents the number of cigarettes smoked and the frequency of drinking, respectively. Methylation means for each tobacco consumption level are represented with their 95% confidence intervals. fitting showed no indication of genomic inflation (λ 36.9% were hypomethylated for regular consumers = 1.044). The top 15 CpG sites are shown in Table 3 compared with non-consumers. However, the per­ and all the epigenome-wide significant CpG sites are centage of hypomethylated probes was increased for listed in Additional File 1: Table S3 and represented the most significant probes. Among the 2,569 alco­ as a Manhattan Plot in Figure 1b. Among them, hol-related methylation sites, 609 were intergenic 10 N. CARRERAS-GALLO ET AL. Table 3. Top 15 differentially methylated CpG sites by alcohol consumption. 3–5 On Once times Lohof Gene Gene special per per Adjusted et al CpG chr position Symbol Group occasions week week Regularly F P-Value P-Value (direction) cg06690548 chr4 139162808 SLC7A11 Body −0.003 −0.005 −0.015 −0.031 103.1 5.9E–83 4.4E–77 – cg11376147 chr11 57261198 SLC43A1 Body −0.001 −0.002 −0.006 −0.010 49.3 2.4E–40 8.9E–35 – cg26457483 chr1 120256112 PHGDH Body −0.001 −0.004 −0.017 −0.028 39.7 1.5E–32 3.8E–27 – cg14476101 chr1 120255992 PHGDH Body −0.002 −0.005 −0.016 −0.031 36.2 1.1E–29 2.0E–24 – cg18120259 chr6 43894639 LINC01512 Body −0.002 −0.004 −0.009 −0.019 35.0 1.1E–28 1.6E–23 – cg06088069 chr14 75895604 JDP2 5‘UTR −0.002 −0.003 −0.007 −0.013 34.2 5.0E–28 6.1E–23 – cg21998542 chr10 33605101 NRP1 Body −0.002 −0.005 −0.009 −0.019 29.0 9.2E–24 9.8E–19 – cg15837522 chr8 117892654 −0.003 −0.011 −0.022 −0.031 28.7 1.7E–23 1.6E–18 – cg06644515 chr1 173834831 SNORD47; TSS1500; −0.001 −0.001 −0.007 −0.013 27.6 1.3E–22 1.1E–17 – GAS5; Body; SNORD81; TSS1500; SNORD80; TSS1500; SNORD78; TSS200; SNORD79 TSS1500 cg12116137 chr17 1576449 PRPF8 Body 0.001 0.001 0.008 0.018 25.5 7.5E–21 5.5E–16 ++ cg19693031 chr1 145441552 TXNIP 3‘UTR −0.004 −0.003 −0.011 −0.026 25.1 1.8E–20 1.2E–15 – cg12825509 chr3 185648568 TRA2B Body 0.000 −0.004 −0.010 −0.016 24.6 4.5E–20 2.8E–15 – cg14346162 chr21 44490229 CBS Body 0.001 0.004 0.009 0.014 23.5 3.1E–19 1.8E–14 Not reported cg12973487 chr19 1623075 TCF3 Body 0.004 0.005 0.007 0.014 22.6 2.0E–18 1.0E–13 ++ cg05713943 chr13 97912352 MBNL2 5‘UTR 0.001 0.000 −0.005 −0.008 21.8 8.7E–18 4.2E–13 – Note: The CpG sites are annotated based on the chromosome (chr), the position (pos), the gene symbol from HGNC, and the gene group (based on the position of the CpG regarding the nearest gene). Each CpG site has a beta value for each consumer group vs non-consumers, a F-statistic (F), a nominal p-value, and an adjusted p-value by Bonferroni. We also included one column with the direction of the effect in a previous study (Lohoff et al (31)) for those that were significant in their analysis. and 1,960 were annotated to 1,670 unique genes. information in the studies included in the EWAS Five genes had seven or more significant probes catalogue, we compared our results with the largest mapping to their locus, including RPTOR (11 single-cohort EWAS of alcohol consumption per­ probes), JARID2 [8], and ABCG1 [8]. The enrich­ formed by Lohoff et al [31]. In this study, they ment revealed an over-representation of autistic dis­ found 2,504 CpG sites and 909 were overlapping order (P-value = 4.2 · 10−7), acquired scoliosis and with our study. All the CpG sites overlapping had curvature of the spine (P-value = 2.7 · 10−4), small the same direction of the effect. nose (P-value = 2.5 · 10−4), small midface (P-value = 6.1 · 10−5), and self-injurious behaviour (P-value = 1.5 · 10−4) (Additional File 2: Fig. S3A). These genes Genome-wide effect of marijuana were also enriched in the PI3k-Akt signalling path­ consumption on DNA methylation way (P-value = 4.0 · 10−4), which is involved in the In the EWAS for the frequency of marijuana use cell cycle, cholinergic synapse (P-value = 2.7 · 10−5), (codified as a 5-point scale from ‘never’ to ‘reg­ and longevity regulating pathway (P-value = 6.7 · ularly;’ see Table 1), we did not find any CpG 10−5) (Additional File 2: Fig. S3B). The GO enrich­ site with a P-value adjusted lower than 0.05 ment revealed regulation of Wnt signalling pathway (Figure 1c). However, we used the P-value (P-value = 3.2 · 10−8) and heart growth (P-value = 2.6 threshold from the EWAS catalogue (P-value < 1·10−4) and we found 195 CpG sites at · 10−5) (Additional File 2: Fig. S3C). The top CpG site a suggestive significant level (Additional File 1: (cg06690548, P-value = 5.9·10−83) mapped to the Table S3). From them, almost 50%were hypo­ SLC7A11 gene and its methylation was significantly methylated for regular consumption compared reduced proportionally to the alcohol consumption to no consumption (Table 4). Gene symbols for (Figure 2b). Due to the absence of beta value the 195 CpG sites were tested for enrichment in EPIGENETICS 11 KEGG pathways and Gene Ontology (GO) and were implicated in various biological processes, we found an enrichment of myelin assembly including positive regulation of heart rate (P-value = 8.0·10−6). We did not find CpG sites (P-value = 5.3·10−5), inositol lipid-mediated signal overlapping between our results and the ones (P-value = 1.7·10−4), and positive regulation of reported in the EWAS catalogue. This may be blood circulation (P-value = 1.9·10−4), among in part due to the differences in the variable of others (Additional File 2: Fig. S4). interest and the study population, such as the Then, we evaluated whether recreational drug evaluation of the effect of cannabis use on non- consumption was associated with multiple epige­ Hispanic white women and the risk of breast netic clocks. For tobacco smoking, we found cancer [36]. a significant association with telomere length (β = −0.01, P-value = 1.2·10−3), Grim Age Principal Component (β = 0.72, P-value = 1.6·10−17), and Comparison between recreational drugs DunedinPACE (β = 0.01, P-value = 9.4·10−5). Alcohol consumption was associated also with We compared the differentially methylated CpG a lower telomere length (β = −0.004, P-value = sites for the different drugs of abuse. Since no CpG 7.5·10−3) and Grim Age Principal Component (β = sites were significant for marijuana consumption 0.13, P-value = 7.7·10−4). Finally, marijuana con­ after correcting by multiple comparisons, we sumption was not associated with any epigenetic focused on comparing the results between tobacco clock. and alcohol. We found that 12 CpG sites were overlapping between tobacco and alcohol, and nine of them had the same direction effect in both analyses (Figure 3a). Next, we extracted the Mediation between substance consumption and hypertension by CpG methylation genes where all the differentially methylated CpG sites were annotated and compared the alcohol We evaluated whether the changes at the CpG and tobacco related genes. We observed a larger methylation level mediated the effect of recreational overlap and found that 61 genes were differentially drugs on hypertension. We first tested the associa­ methylated by both substances (Figure 3b). To tion between smoking and hypertension. We con­ determine whether these genes were involved in sidered that the group with the highest levels of similar pathways, we performed an enrichment smoking were those who smoked more than 11 analysis, and the results revealed that these genes cigarettes per day, joining the categories 11–20 Table 4. Top 15 differentially methylated CpG sites by marijuana consumption. Gene Gene On special Once per 3–5 times Adjusted CpG chr position Symbol Group occasions week per week Regularly F P-Value P-Value cg05107281 chr11 47072710 CSTPP1 Body −0.004 −0.010 −0.014 0.001 9.4 1.6E–07 0.12 cg06499565 chr20 35374747 NDRG3 TSS1500 0.001 0.006 0.000 −0.001 8.7 5.7E–07 0.17 cg10054857 chr18 61816543 LINC00305 TSS1500 −0.004 −0.007 −0.002 −0.002 8.4 1.0E–06 0.17 cg24344693 chr10 133273964 −0.004 0.003 −0.001 −0.002 8.2 1.5E–06 0.17 cg20226924 chr5 133985272 SEC24A Body 0.001 0.003 0.003 0.001 8.2 1.5E–06 0.17 cg05575921 chr5 373378 AHRR Body −0.003 −0.011 −0.012 −0.020 8.1 1.7E–06 0.17 cg21161138 chr5 399360 AHRR Body −0.004 −0.011 −0.009 −0.012 8.1 1.7E–06 0.17 cg19730404 chr9 139361517 SEC16A; ExonBnd; −0.003 −0.001 0.001 −0.004 7.8 2.8E–06 0.25 SEC16A Body cg19308363 chr12 116290566 0.006 0.005 −0.005 −0.015 7.7 3.4E–06 0.25 cg08258765 chr5 24841586 LOC340107 TSS1500 0.005 0.002 0.004 −0.002 7.7 3.8E–06 0.25 cg03838168 chrX 48433876 RBM3 Body −0.003 −0.004 0.015 0.005 7.7 3.9E–06 0.25 cg16822035 chr13 113633379 MCF2L; Body; 0.001 0.004 0.009 0.012 7.6 4.0E–06 0.25 MCF2L TSS1500 cg11756734 chr12 9028945 A2ML1 3‘UTR −0.005 0.002 −0.007 0.003 7.5 4.8E–06 0.26 cg08415592 chr22 36648973 APOL1 TSS200 0.008 0.016 −0.004 0.007 7.4 5.8E–06 0.26 cg17325792 chr1 77042560 ST6GALNAC3; 3‘UTR; 0.003 0.001 0.003 −0.011 7.4 6.6E–06 0.26 ST6GALNAC3 Body Note: The CpG sites are annotated based on the chromosome (chr), the position (pos), the gene symbol from HGNC, and the gene group (based on the position of the CpG regarding the nearest gene). Each CpG site has a beta value for each consumer group vs non-consumers, a F-statistic (F), a nominal p-value, and an adjusted p-value by Bonferroni. 12 N. CARRERAS-GALLO ET AL. Figure 3. Venn diagrams comparing CpG sites and genes differentially methylated for tobacco and alcohol use. (a) Comparison between differentially methylated CpG sites with a P-value adjusted lower than 0.05. (b) Comparison between genes where the CpG sites differentially methylated are annotated. and >20 cigarettes due to their low number of comparing the 4 levels of consumption (from ‘on participants in each. We evaluated the association special occasions’ to ‘regularly’) with non- between smoking codified as numeric and high consumers (Additional File 2: Fig. S6). blood pressure as a binary variable and we did not As for alcohol use, we found a significant asso­ detect a significant tendency (P-value = 0.26). The ciation between alcohol and higher blood pressure forest plot shows that the risk of hypertension (P-value = 0.001, OR = 1.13). In addition, the forest increases with a higher number of cigarettes plot in Figure 4b revealed a significant association smoked except for the last group (more than 11 between daily consumption with high blood pres­ cigarettes per day) (Additional File 2: Fig. S5). sure (P-value = 0.014, OR = 1.39) and a non- Although we expected this group to be the one at significant association between light to moderate the highest risk, we also observed that those parti­ consumption with the phenotype, as expected. cipants were also the youngest (average of 5.75 Thus, we evaluated whether DNA methylation years less, P-value = 0.005), suggesting was mediating this association. We first performed a particularly strong healthy donor effect for this logistic regression between the alcohol-related group. We tested the association after removing this CpG sites and hypertension and found 76 signifi­ group and we found a significant association cant CpG sites after correcting by multiple com­ between tobacco smoking and hypertension parisons. To see the effect of each CpG site, we (P-value = 0.009, OR = 1.28). The forest plot in performed a univariate mediation analysis for each Figure 4a revealed a clear dose-response relation­ of these probes. Among them, 66 CpG sites sig­ ship where individuals who consume 6 to 10 cigar­ nificantly mediated the association between alco­ ettes per day have 3.19 times of high blood pressure hol and hypertension with estimated proportions risk compared with non-smokers (P-value = 0.023). ranged between 3 and 71%. The most differentially For the mediation analysis using CpG sites as med­ methylated CpG site by alcohol consumption, iators, we selected those CpG sites that were sig­ cg06690548, was also the most significant media­ nificantly associated with hypertension and tobacco tor between alcohol consumption and high blood smoking. The analysis revealed 10 CpG sites that pressure after adjusting by covariates. This CpG were potential mediators. However, the univariate site mediated 71% of the total effect of alcohol on mediation analyses showed that none of those CpG the phenotype (FDR = 0.019). Furthermore, we sites was significantly mediating the effect of observed that 12.8% of the hypertension variance tobacco on hypertension with a FDR lower than was explained by alcohol consumption after 0.05 (Additional File 1: Table S4). adjusting by covariates. The variance explained We also tested the association between mari­ was increased to 13.6% after including the juana consumption and high blood pressure. We cg06690548 methylation in the model, and to did not find any significant association when 20.5% after including the 66 significant CpG EPIGENETICS 13 Figure 4. Forest plot of the association between tobacco (a) and alcohol (b) consumption with hypertension. OR: Odds Ratio. sites. In these models, the association between to tobacco smoking, 2,569 according to alcohol hypertension and alcohol was no longer signifi­ consumption, and no significant associations for cant, further suggesting that methylation has an marijuana consumption. Second, we detected important mediation role in hypertension risk. a large overlapping between the differentially methylated genes by alcohol and tobacco con­ sumption. Third, we found a significant media­ Discussion tion between alcohol consumption and The current study evaluated the effect of tobacco, hypertension by many alcohol-related methyla­ alcohol, and marijuana consumption on genome- tion sites. wide DNA methylation in 3,424 individuals from A considerable amount of literature has been an EEUU population-based cohort. We identified published on DNA methylation changes due to 528 CpG sites differentially methylated according smoking. The first studies evaluating these 14 N. CARRERAS-GALLO ET AL. changes were carried out in single genes or small Alcohol is known to affect DNA methylation. panels of genes [63–65]. It was not until 2012 that To date, several EWAS have detected CpG sites the first epigenome-wide association study on associated with alcohol consumption [29,31– tobacco was reported by Wan et al. [66]. Since 33,72,73]. Here, we identified 2569 probes asso­ that time, several studies have demonstrated the ciated with drinking. As we found, many studies huge impact of tobacco on DNA methylation detected cg06690548 mapped at the SLC7A11 pro­ across the human genome, even in newborns moter as the most alcohol-related methylation site when the tobacco exposure was during pregnancy [29,31,32]. Furthermore, Lohoff et al. demon­ [22,23,25–27,67,68]. Our findings were in line strated that various liver biomarkers were robustly with previous research, demonstrating a large associated with SLC7A11 methylation status [31], number of CpG sites differentially methylated suggesting an implication of this gene in the dis­ along the genome. We were able to compare our turbance of the gastrointestinal system when con­ results with the EWAS catalogue, and this suming alcohol. In addition, we compared our revealed a high overlap in the CpG sites differen­ results with those reported in study where they tially methylated. Moreover, 98.9% of these CpG also performed an EWAS for alcohol consumption sites had the same direction of the effect. Among in 8,161 individuals. We found that 909 probes the probes that were not previously reported in overlapped, all showing the same direction of the EWAS catalogue, it is remarkable the CpG site effect. We observed that certain probes did not at PHACTR2 gene, because it is involved in actin overlap, which could be attributed to differences cytoskeleton organization and implicated in in study design. Specifically, Lohoff et al used Parkinson’s disease [69], and the CpG site at information on alcohol consumption during the GPX5 gene, since it protects cells and enzymes previous week, while our study assessed general from oxidative damage. Additionally, our results alcohol consumption habits. As a result, these dis­ confirm previous observations where the crepancies suggest that the timing and specificity cg05575921 mapped to AHRR (P-value = of the alcohol consumption measure used in dif­ 9.7·10−221) and the cg21566642 in the 2q37.1 ferent studies can influence epigenetic changes. region (P-value = 5.9·10−65) were the most signifi­ Besides, the genes differentially methylated in our cantly associated CpG sites to tobacco consump­ study were highly enriched in autistic disorder, tion [26–28]. Besides, the enrichment of the CpG acquired scoliosis, curvature of the spine, small sites differentially methylated revealed remarkable nose, and small midface. Engagingly, all these fea­ findings. Dopaminergic and cholinergic synapses tures are symptoms of the foetal alcohol spectrum are two systems that are affected by nicotine, the disorder (FASD), which encompasses the range of psychoactive substance found in tobacco. On one adverse effects associated with alcohol exposure hand, nicotine activates the release of dopamine, during pregnancy [74–77]. In addition, previous inducing feelings of pleasure and reward [70]. studies have already investigated the epigenetic This activation is one of the main reasons for mechanism linking autism and FASD [78]. Our which nicotine is so addictive. On the other results suggest that DNA methylation changes are hand, nicotine binds to receptors in the choliner­ important contributors to the relationship between gic synapse, enhancing the release of acetylcholine alcohol consumption and FASD. Future studies [71]. This neurotransmitter can improve cogni­ should examine whether pregnant women with tion, attention, and memory. However, chronic alcohol disorders could be monitored using epige­ exposure to nicotine can lead to a decrease of netic changes to prevent disorders in children. the numbers of receptors, which can produce The studies that have investigated DNA methyla­ cognitive impairment and other negative effects tion modifications after marijuana consumption are in the brain. Thus, the CpG sites differentially limited. In 2015, Watson et al. evaluated in rats the methylated by tobacco may be used as biomarkers effect of cannabis parental exposure on the epigen­ to monitor addiction and neurotoxic effects on ome of the nucleus accumbens [79] and they identi­ tobacco smokers. fied 1027 differentially methylated regions. Five EPIGENETICS 15 years later, Osborne et al. carried out the first EWAS epigenetic changes may explain the shared con­ on heavy cannabis consumption with and without sequences of drug abuse. Targeting these specific tobacco comparing 48 consumers with 48 controls pathways could be a strategy for preventing [35]. They found five differentially methylated sites addiction, as well as neurological and cardiovas­ in cannabis and tobacco users that replicated pre­ cular disorders. In line with this, we found that vious studies on the effects of tobacco. However, tobacco and alcohol consumption were associated cannabis-only users had no evidence of significant with shorter telomere lengths, as reported pre­ differential methylation in any gene. Markunas et al. viously [82–84]. performed another EWAS with a larger sample size In our data, hypertension was partially asso­ (1,247 ever users) consisting of women at risk of ciated with tobacco and highly associated with developing breast cancer [36]. They identified alcohol consumption, as demonstrated previously a unique significant CpG mapped to CEMIP 5’ [39–42]. Marijuana was not associated with hyper­ region. However, they designed a biomarker for life­ tension, in line with previous studies that have time cannabis use based on the top 50 EWAS CpG revealed ambiguous results [43–45]. In the case sites. In our study, 367 individuals smoked mari­ of tobacco exposure, we found unexpected results juana from occasionally to daily. The EWAS did because the individuals who smoked the most not reveal significant CpG sites at the Bonferroni were the ones who had less risk to develop hyper­ adjustment. Nonetheless, the genes where the 195 tension. These results may be explained due to the CpG sites with a P-value lower than 1·10−4 were lower age of the individuals in that group and also annotated were enriched in myelin assembly, essen­ by the healthy donor effect of the data. This means tial for the proper functioning of the nervous system, that participants are volunteers who have paid for as it enables the rapid and efficient transmission of the TruAge test and may have healthy habits that nerve impulses between neurons. This suggests protect them against hypertension although they a possible implication of DNA methylation changes are heavy smokers. on the effects of THC, the active substance in mar­ Our data replicated prior studies where light to ijuana, in cognitive and behavioural impairments. In moderate drinking was not associated with high addition, we detected cg05575921 (AHRR), the most blood pressure and heavy drinking increased the significant tobacco-associated CpG site, differen­ risk of the disease [8,40]. Another important find­ tially methylated according to marijuana use with ing was that 66 CpG sites significantly mediated a nominal P-value equal to 1.7·10−6. Allen et al. the effect of alcohol consumption on hypertension. already found that the link between marijuana use Importantly, the methylation levels of these probes and epigenetic age acceleration was statistically increased the variance explained in hypertension mediated via hypomethylation at site cg05575921 by alcohol from 13.6% to 20.5%. More interest­ [80]. This is consistent with the association of the ingly, lower methylation levels of cg06690548 at AHRR gene with exposure to tobacco and fine parti­ SLC7A11, cg18120259 at LINC01512, and culate matter (PM2.5) which suggests that marijuana cg19693031 at TXNIP have been seen previously inhalation can produce similar effects [80,81]. associated with higher systolic and diastolic blood In the current study, we aimed to investigate pressure [85–87]. Additionally, hypomethylation whether the adverse health effects commonly of cg06690548 was associated with higher expres­ associated with the use of multiple substances, sion of SLC7A11 [87]. SLC7A11 enhances antiox­ such as cardiovascular disease, could be attribu­ idant defence and protects against endothelial ted to epigenetic alterations. Our findings dysfunction and vascular inflammation. This revealed a significant overlap of 12 CpG sites increases vascular tone and rigidity, and conse­ that displayed differential methylation levels in quently blood pressure. Also, Richard et al. evi­ response to both tobacco and alcohol. Moreover, denced triangular associations between 61 genes overlapped between these two sub­ methylation, gene expression, and blood pressure stances and were enriched in biological processes [86]. The univariate mediation based on involved in the proper functioning of the nervous cg06690548 methylation revealed that 70.5% of and cardiovascular systems. Therefore, similar the effect of alcohol on high blood pressure was 16 N. CARRERAS-GALLO ET AL. mediated by the CpG methylation level (P-value = suggests that marijuana may have a lower effect on 0.006). In essence, we have demonstrated that the blood DNA methylation compared to other sub­ effect of heavy drinking on high blood pressure is stances. However, further studies with better con­ partially mediated by the methylation of CpG sites sumption assessment are needed to confirm this that are significantly associated with the disease. observation. This finding provides new insights on targets to prevent and manage hypertension in individuals Conclusions with regular alcohol consumption. The generalizability of these results is subject To the best of our knowledge, this is the first to certain limitations. First, DNA methylation study to assess simultaneously the effect of was obtained from blood samples, thus, further tobacco, alcohol, and marijuana on DNA methy­ research is required to understand the implication lation. We have shown that tobacco and alcohol of the identified markers in each tissue. Second, have large effects on genome-wide DNA methyla­ genetics has an important role in substance use tion, while marijuana consumption has small predisposition. In our analysis, we were not able effects. Most importantly, many genes differen­ to remove the genetic factor because of the lack of tially methylated by smoking are also affected by data. Some of the differentially methylated probes alcohol consumption, suggesting a similar epige­ may be a consequence of the genetic differences netic impact after the consumption of recreational and not the exposure itself. Notwithstanding this drugs. The results of this research also have sig­ limitation, we filtered all the probes with a SNP in nificant implications for the understanding of the extension base and all probes where 5bp 3’- how alcohol consumption increases hypertension. subsequence overlapped with any of the SNPs We demonstrated that 66 CpG sites were partially with a global population frequency higher than mediating the association between heavy drinking 1%. Third, the consumption assessment was self- and hypertension and the most alcohol-related reported and not specific for a time period, limit­ CpG site mediated 70.5% of this association. ing credibility and enhancing misclassification. In Finally, the current data highlight the importance addition, we did not have information on of using blood methylation biomarkers in clinical whether marijuana was smoked mixed or not practice to detect and monitor the adverse effects, with tobacco. This information could benefit such as addiction, derived from substance future studies on removing the tobacco effect. consumption. Fourth, we have compared our results with the EWAS catalogue for smoking and with a unique paper for alcohol consumption. It would be useful Acknowledgments to compare our results with other published We are grateful to all participants and researchers who took papers. part in this study. Our study also had notable strengths, including a large number of drinkers and a high variability Disclosure statement in drinking frequency. This allowed us to test the mediation analysis between alcohol consumption No potential conflict of interest was reported by the authors. and hypertension. Moreover, most studies are focused on evaluating the effects of one substance Funding in drug-specific cohorts. Our data provided infor­ mation on tobacco, alcohol, and marijuana con­ This research has received funding from the Spanish Ministry sumption in the same individuals, along with of Science and Innovation through the “Centro de Excelencia clinical data. While marijuana consumption may Severo Ochoa 2019-2023 (CEX2018-000806-S) program, and support from the Generalitat de Catalunya through the be reported or suffer from healthy donor effect, we CERCA Program. NC is supported by the Spanish regional observed that fewer individuals smoked tobacco program PERIS (Ref.: SLT017/20/000061), granted by compared with marijuana, yet the effect of tobacco Departament de Salut de la Generalitat de Catalunya. and alcohol on DNA methylation was strong. This TruDiagnostics also provided funding for data analysis. EPIGENETICS 17 Data availability statement [10] Zakhari S. Director is. Overview: how is alcohol metabo­ lized by the body? Alcohol Res Heal. 2006 [cited 2023 Feb The data that support the findings of this study are available 14];29(4):245. https://pmc/articles/PMC6527027/ upon reasonable request due to ensure the privacy of the [11] Gardner JD, Mouton AJ. Alcohol effects on cardiac participants. Please email varun@trudiagnostic.com for data function. Compr Physiol. 2015 Apr 1 [cited 2023 Feb requests. Any custom code or software used in our analysis is 14];5(2):791–802. https://pubmed.ncbi.nlm.nih.gov/ available at DOI: 10.5281/zenodo.6417926 (URL: https:// 25880513/ zenodo.org/badge/latestdoi/296552532). [12] La Monte SM D, Kril JJ. Human alcohol-related neuropathology. Acta Neuropathol. 2014 Jan-Mar [[cited 2023 6]];127(1):71–90. https://pubmed.ncbi. nlm.nih.gov/24370929/ ORCID [13] Atakan Z. Cannabis, a complex plant: different com­ Natàlia Carreras-Gallo http://orcid.org/0000-0001-8559- pounds and different effects on individuals. Ther Adv 2313 Psychopharmacol. [cited 2023 Feb 14] 2012;2:241. Alejandro Cáceres http://orcid.org/0000-0001-8551-6695 https://pmc/articles/PMC3736954/ Juan R Gonzalez http://orcid.org/0000-0003-3267-2146 [14] Egger G, Liang G, Aparicio A, et al. Epigenetics in human disease and prospects for epigenetic therapy. Nature. 2004 May 27 [cited 2022 May 3];429 (6990):457–463. https://pubmed.ncbi.nlm.nih.gov/ References 15164071/. [1] Substance Abuse and Mental Health Services [15] Szutorisz H, Hurd YL. Epigenetic Effects of Cannabis Administration (SAMHSA). Results from the 2015 Exposure. Biol Psychiatry. [cited 2022 May 4] 2016;79 national survey on drug use and health: Summary of (7):586. https://pmc/articles/PMC4789113/ National Findings [Internet]. 2017 [cited 2022 Jun 27]. [16] Cecil CAM, Walton E, Methylation VED. Substance Use Available from: https://www.samhsa.gov/ and Addiction: a Systematic Review of Recent Animal [2] World Health Organization. Management of substance and Human Research from a Developmental Perspective. abuse team. The health and social effects of nonmedical Curr Addict Reports. 2015 Dec 1 [cited 2022 Jul 8];2 cannabis use. World Health Organization; 2016. (4):331–346. DOI:10.1007/s40429-015-0072-9. [3] Substance Abuse and Mental Health Services [17] Moosavi A, Ardekani AM. Role of Epigenetics in Biology Administration. Key substance use and mental health and Human Diseases. Iran Biomed J. 2016 Nov 1 [cited indicators in the United States: results from the 2017 2022 Jul 13];20(5):246. https://pmc/articles/PMC5075137/ national survey on drug use and health. Rockville, MD: [18] Leenen FAD, Muller CP, Turner JD. DNA methylation: Center for Behavioral Health Statistics and Quality, conducting the orchestra from exposure to phenotype? Substance Abuse and Mental HealthServices Clin Epigenetics. 2016 Sep 6 [cited 2022 May 2]. 8(1). Administration; 2018. https://www.samhsa.gov/data/ https://pubmed.ncbi.nlm.nih.gov/27602172/ [4] Geneva: World Health Organization. WHO global [19] Carreras-Gallo N, Cáceres A, Balagué-Dobón L, et al. report on trends in prevalence of tobacco use The early-life exposome modulates the effect of poly­ 2000-2025, fourth edition. 2021. morphic inversions on DNA methylation. Commun [5] West R. Tobacco smoking: health impact, prevalence, Biol. 2022 DecJul. [cited 2022 20];5(1). https:// correlates and interventions. Psychol Health Internet. pubmed.ncbi.nlm.nih.gov/35550596/ 2017 Aug 3[cited 2022 Jun 20];32(8):1018–1036. [20] Li E, Zhang Y. DNA Methylation in Mammals. Cold https://pubmed.ncbi.nlm.nih.gov/28553727/. Spring Harb Perspect Biol. [cited 2022 Jul 19] 2014;6: [6] (US) C for DC and P, (US) NC for CDP and HP, (US) a019133. https://pmc/articles/PMC3996472/ O on S and H. Chemistry and Toxicology of Cigarette [21] Laurent L, Wong E, Li G, et al. Dynamic changes in the Smoke and Biomarkers of Exposure and Harm. 2010 human methylome during differentiation. Genome [cited 2023 Feb 14]; Available from: https://www.ncbi. Res. 2010 MarJul [cited 2022 20];20(3):320–331. nlm.nih.gov/books/NBK53014/ https://pubmed.ncbi.nlm.nih.gov/20133333/. [7] Levin ED. Neurotoxicology of nicotine and tobacco. [22] Dugué PA, Jung CH, Joo JE, et al. Smoking and blood Adv Neurotoxicology. 2022 Jan 1;8:93–103. DNA methylation: an epigenome-wide association [8] O’Keefe JH, Bhatti SK, Bajwa A, et al. Alcohol and study and assessment of reversibility. Epigenetics. cardiovascular health: the dose makes the poison . . . 2020 Apr 2 [cited 2022 Jun 2];15(4):358–368. https:// or the remedy. Mayo Clin Proc. [cited 2022 Jun 27] pubmed.ncbi.nlm.nih.gov/31552803/ 2014;89:382–393. https://pubmed.ncbi.nlm.nih.gov/ [23] Zeilinger S, Kühnel B, Klopp N, et al.Tobacco smoking 24582196/ leads to extensive genome-wide changes in DNA [9] Geneva: World Health Organization. Global status methylationPLoS ONE. 2013 May 17 [cited 2022 Jun report on alcohol and health 2018. 2018. 2];8(5). https://pubmed.ncbi.nlm.nih.gov/23691101/ 18 N. CARRERAS-GALLO ET AL. [24] Dogan MV, Beach SRH, Philibert RA. Genetically con­ [33] Wilson LE, Xu Z, Harlid S, et al. Alcohol and DNA textual effects of smoking on genome wide DNA Methylation: an Epigenome-Wide Association Study in methylation. Am J Med Genet B Neuropsychiatr Blood and Normal Breast Tissue. Am J Epidemiol. Genet. 2017 Sep 1 [cited 2022 Jun 2];174(6):595–607. 2019 Jun 1 [cited 2022 May 12];188(6):1055–1065. https://pubmed.ncbi.nlm.nih.gov/28686328/. https://pubmed.ncbi.nlm.nih.gov/30938765/ [25] Christiansen C, Castillo-Fernandez JE, Domingo- [34] Stephenson M, Bollepalli S, Cazaly E, et al. Associations Relloso A, et al.Novel DNA methylation signatures of of Alcohol Consumption with Epigenome-Wide DNA tobacco smoking with trans-ethnic effectsClin Methylation and Epigenetic Age Acceleration: Epigenetics. 2021 Dec 1 [cited 2022 Jun 2];13(1). individual-Level and Co-twin Comparison Analyses. https://pubmed.ncbi.nlm.nih.gov/33593402/ Alcohol Clin Exp Res. 2021 Feb 1 [cited 2022 May [26] Sikdar S, Joehanes R, Joubert BR, et al. Comparison of 12];45(2):318–328. https://pubmed.ncbi.nlm.nih.gov/ smoking-related DNA methylation between newborns 33277923/ from prenatal exposure and adults from personal [35] Osborne AJ, Pearson JF, Noble AJ, et al. Genome-wide smoking. Epigenomics. 2019 Oct 1 [cited 2022 Jun DNA methylation analysis of heavy cannabis exposure 2];11(13):1487–1500. https://pubmed.ncbi.nlm.nih. in a New Zealand longitudinal cohort. Transl gov/31536415/ Psychiatry. 2020 Apr 22 [cited 2022 Apr 5];10 [27] Joubert BR, Håberg SE, Nilsen RM, et al. 450K (1):1–10. https://www.nature.com/articles/s41398-020- epigenome-wide scan identifies differential DNA 0800-3 methylation in newborns related to maternal smoking [36] Markunas CA, Hancock DB, Xu Z, et al. Epigenome- during pregnancy. Environ Health Perspect. 2012 wide analysis uncovers a blood-based DNA methyla­ Oct;120(10):1425–1431. [cited 2022 Jul 7] https:// tion biomarker of lifetime cannabis use. Am J Med pubmed.ncbi.nlm.nih.gov/22851337/. Genet B Neuropsychiatr Genet. 2021 Apr 1 [cited [28] Richmond RC, Simpkin AJ, Woodward G, et al. 2022 May 5];186(3):173–182. https://pubmed.ncbi. Prenatal exposure to maternal smoking and offspring nlm.nih.gov/32803843/ DNA methylation across the life course: findings from [37] Kumar J. Epidemiology of hypertension. Clin Queries the Avon Longitudinal Study of Parents and Children Nephrol. 2013 Apr 1 [cited 2022 Jul 25];2(2):56–61. (ALSPAC). Hum Mol Genet. 2015 Apr 15 [cited 2022 https://linkinghub.elsevier.com/retrieve/pii/ Jul 7];24(8):2201–2217. https://pubmed.ncbi.nlm.nih. S2211947713000162. gov/25552657/ [38] Singh S, Shankar R, Singh GP. Prevalence and [29] Xu K, Montalvo-Ortiz JL, Zhang X, et al. Epigenome- Associated Risk Factors of Hypertension: a Wide DNA Methylation Association Analysis Cross-Sectional Study in Urban Varanasi. Identified Novel Loci in Peripheral Cells for Alcohol Int J Hypertens. 2017 [cited 2022 Jul 25];2017:1–10. Consumption Among European American Male https://pubmed.ncbi.nlm.nih.gov/29348933/ Veterans. Alcohol Clin Exp Res. 2019 Oct 1 [cited [39] Minzer S, Losno RA, Casas R. The Effect of Alcohol on 2022 May 12];43(10):2111–2121. https://pubmed.ncbi. Cardiovascular Risk Factors: is There New nlm.nih.gov/31386212/ Information? Nutrients. 2020 Apr 1 [cited 2022 Jun [30] Toinét Cronjé H, Elliott HR, Nienaber-Rousseau C, 20];12(4):1–22. https://pmc/articles/PMC7230699/. et al. Replication and expansion of epigenome-wide [40] Tasnim S, Tang C, Musini VM, et al. Effect of alcohol association literature in a black South African on blood pressure. Cochrane Database Syst Rev. 2020 population. Clin Epigenetics. 2020 Jan 7 [cited 2022 Jul 1 [cited 2022 Jul 18]. 2020(7). https://pmc/articles/ May 12];12(1). https://pubmed.ncbi.nlm.nih.gov/ PMC8130994/ 31910897/. [41] Virdis A, Giannarelli C, Fritsch Neves M, et al. [31] Lohoff FW, Clarke TK, Kaminsky ZA, et al. Cigarette smoking and hypertension. Curr Pharm Epigenome-wide association study of alcohol con­ Des. 2010 Aug 6 [cited 2022 Jun 20];16 sumption in N = 8161 individuals and relevance to (23):2518–2525. https://pubmed.ncbi.nlm.nih.gov/ alcohol use disorder pathophysiology: identification of 20550499/. the cystine/glutamate transporter SLC7A11 as a top [42] Dikalov S, Itani H, Richmond B, et al. Tobacco smok­ target. Mol Psychiatry. [cited 2022 May 18]. ing induces cardiovascular mitochondrial oxidative 2022;27:1754–1764. https://pubmed.ncbi.nlm.nih.gov/ stress, promotes endothelial dysfunction, and enhances 34857913/ hypertension. Am J Physiol Heart Circ Physiol. 2019 [32] Dugué PA, Wilson R, Lehne B, et al.Alcohol con­ Mar 1 [cited 2022 Jun 20];316(3):H639–46. https:// sumption is associated with widespread changes in pubmed.ncbi.nlm.nih.gov/30608177/ blood DNA methylation: analysis of cross-sectional [43] Abuhasira R, Haviv YS, Leiba M, et al., Cannabis is and longitudinal dataAddict Biol. 2021 Jan 1 [cited associated with blood pressure reduction in older 2022 May 5];26(1). https://pubmed.ncbi.nlm.nih.gov/ adults - a 24-hours ambulatory blood pressure mon­ 31789449/ itoring study. Eur J Intern Med. 2021 Apr 1 [cited 2022 EPIGENETICS 19 Jul 25];86:79–85. https://pubmed.ncbi.nlm.nih.gov/ [56] Ashburner M, Ball CA, Blake JA, et al. Gene ontology: 33483174/ tool for the unification of biology. Nature genet. [44] Alshaarawy O, Elbaz HA. Cannabis use and blood 2000;25(1):25–29. pressure levels: United States National Health and [57] Yu G, Wang L-G, Han Y, et al. clusterProfiler: an Nutrition Examination Survey, 2005-2012. R Package for Comparing Biological Themes Among J Hypertens. 2016 Aug 1 [cited 2022 Jul 25];34 Gene Clusters. Omi A J Integr Biol. 2012;16 (8):1507–1512. https://pubmed.ncbi.nlm.nih.gov/ (5):284–287. DOI: 10.1089/omi.2011.0118. 27270185/. [58] Piñero J, Piñero P, Bravo A, et al. DisGeNET: [45] Jones RT. Cardiovascular system effects of marijuana. a comprehensive platform integrating information on J Clin Pharmacol. 2002 Nov 1 [cited 2022 Jul 25];42 human disease-associated genes and variants. Nucleic (S1):58S–63S. https://pubmed.ncbi.nlm.nih.gov/ Acids Res. 2016 [cited 2020 Apr 9];45:833–839. http:// 12412837/. www.disgenet.org [46] Atsma F, De Vegt F. The healthy donor effect: a matter [59] Schwarzer G, Carpenter JR, Rücker G. Meta-Analysis of selection bias and confounding. Transfusion. 2011 with R. 2015 [cited 2023 Mar 6]. DOI:10.1007/978- Sep [cited 2023 6];51(9):1883–1885. https://pubmed. 3-319-21416-0. ncbi.nlm.nih.gov/21790637/ [60] Pidsley R, Wong CC Y, Volta M, et al. A data-driven approach to preprocessing Illumina 450K methylation [47] Brodersen T, Rostgaard K, Lau CJ, et al.The healthy array data. BMC Genomics. 2013;14(1):293. DOI: 10. donor effect and survey participation, becoming 1186/1471-2164-14-293. a donor and donor careerTransfusion. 2023 Dec 8 [61] Xu Z, Niu L, Li L, et al. Enmix: a novel background [cited 2023 Feb 6];631https://pubmed.ncbi.nlm.nih. correction method for Illumina HumanMethylation450 gov/36479702/ BeadChip. Nucleic Acids Res. 2016 Feb 18;44(3):e20. [48] Min JL, Hemani G, Smith GD, et al. Meffil: efficient DOI:10.1093/nar/gkv907. normalization and analysis of very large DNA [62] Tingley D, Yamamoto T, Hirose K, et al. Mediation: methylation datasets. Bioinformatics. 2018 Dec 1 r Package for Causal Mediation Analysis. J Stat Softw. [cited 2022 Jun 23];34(23):3983–3989. https:// 2014;59(5):1–38. http://www.jstatsoft.org/v59/i05/ pubmed.ncbi.nlm.nih.gov/29931280/. [63] Enokida H, Shiina H, Urakami S, et al. Smoking influ­ [49] Zhou W Infinium Annotation. https://zwdzwd. ences aberrant CpG hypermethylation of multiple github.io/InfiniumAnnotation. Accessed 14 June genes in human prostate carcinoma. Cancer. 2006 Jan 2022. 1 [cited 2022 Jul 8];106(1):79–86. https://pubmed.ncbi. [50] Chen J, Behnam E, Huang J, et al. Fast and robust adjust­ nlm.nih.gov/16323173/ ment of cell mixtures in epigenome-wide association stu­ [64] Ostrow KL, Michalidi C, Guerrero-Preston R, et al. dies with SmartSVA. BMC Genomics. 2017 May 26 [cited Cigarette smoke induces methylation of the tumor 2021 Mar 9];18(1):413. DOI:10.1186/s12864-017-3808-1. suppressor gene NISCH. Epigenetics. 2013 Apr 4 [51] Teschendorff AE, Zhuang J, Widschwendter M. [cited 2023 Feb 9];8(4):383. https://pmc/articles/ Independent surrogate variable analysis to deconvolve con­ PMC3674047/ founding factors in large-scale microarray profiling studies. [65] Breitling LP, Salzmann K, Rothenbacher D, et al. Bioinformatics. 2011 Mar-Jun [cited 2021 9];27 F2RL3 methylation, and prognosis in stable coronary (11):1496–1505. https://pubmed.ncbi.nlm.nih.gov/ heart disease. Eur Heart J. 2012 Feb-Nov [cited 2023 21471010/ 9];33(22):2841–2848. https://pubmed.ncbi.nlm.nih.gov/ [52] Ruiz-Arenas C, JR G. MEAL: perform methylation 22511653/ analysis. R package version 1.22.0. 2019 [cited 2020 [66] Wan ES, Qiu W, Baccarelli A, et al. Cigarette smoking Apr 9]; Available from: https://bioconductor.org/ behaviors and time since quitting are associated with dif­ packages/release/bioc/html/MEAL.html ferential DNA methylation across the human genome. Hum Mol Genet. 2012 Jul [cited 2022 8];21 [53] Ritchie ME, Phipson B, Wu D, et al. Limma powers (13):3073–3082. https://pubmed.ncbi.nlm.nih.gov/ differential expression analyses for {RNA}-sequencing 22492999/ and microarray studies. Nucleic Acids Res. 2015;43(7): [67] Shenker NS, Polidoro S, van Veldhoven K, et al. e47. DOI:10.1093/nar/gkv007 Epigenome-wide association study in the European [54] Guintivano J, Shabalin AA, Chan RF, et al. Test- Prospective Investigation into Cancer and Nutrition statistic inflation in methylome-wide association (EPIC-Turin) identifies novel genetic loci associated studies. Epigenetics. 2020 Nov 1 [cited 2022 Jul 15];15 with smoking. Hum Mol Genet. 2013 Mar-Jul;22 (11):1163–1166. DOI:10.1080/15592294.2020.1758382. (5):843–851. [cited 2022 8] https://pubmed.ncbi.nlm. [55] Ogata H, Goto S, Sato K, et al. Kyoto Encyclopedia of nih.gov/23175441/ Genes and Genomes [Internet]. Nucleic Acids Res. [68] Monick MM, Beach SRH, Plume J, et al. Coordinated 1999 [cited 2020 Apr 9];27(1):29–34. https://academic. Changes in AHRR Methylation in Lymphoblasts and oup.com/nar/article-abstract/27/1/29/1238108 Pulmonary Macrophages from Smokers. Am J Med 20 N. CARRERAS-GALLO ET AL. Genet. [cited 2022 Jul 8] 2012;159B:141. https://pmc/ [79] Watson CT, Szutorisz H, Garg P, et al. Genome-Wide articles/PMC3318996/ DNA Methylation Profiling Reveals Epigenetic [69] Wider C, Lincoln SJ, Heckman MG, et al. Phactr2 and Changes in the Rat Nucleus Accumbens Associated Parkinson’s disease. Neurosci Lett. 2009 Mar 3 [cited with Cross-Generational Effects of Adolescent THC 2022 Jul 26];453(1):9. https://pmc/articles/ Exposure. Neuropsycho pharmacology. 2015 Dec 1 PMC2684848/ [cited 2022 May 13];40(13):2993–3005. https:// [70] Wise RA, Robble MA. Dopamine and Addiction. Annu pubmed.ncbi.nlm.nih.gov/26044905/ Rev Psychol. 2020 Jan 4 [cited 2023 Mar 9];71(1):79–106. [80] Allen JP, Danoff JS, Costello MA, et al.Lifetime mar­ https://pubmed.ncbi.nlm.nih.gov/31905114/. ijuana use and epigenetic age acceleration: a 17-year [71] Subramaniyan M, Dani JA. Dopaminergic and cho­ prospective examination. Drug Alcohol Depend. [cited linergic learning mechanisms in nicotine addiction. 2022 Apr 5];233109363https://pubmed.ncbi.nlm.nih. Ann N Y Acad Sci. 2015 Sep 1 [cited 2023 Mar gov/35231715/ 9];1349(1):46–63. https://pubmed.ncbi.nlm.nih.gov/ [81] Tantoh DM, Lee KJ, Nfor ON, et al.Methylation at 26301866/. cg05575921 of a smoking-related gene (AHRR) in [72] Liang X, Justice AC, So-Armah K, et al. DNA methyla­ non-smoking Taiwanese adults residing in areas with tion signature on phosphatidylethanol, not on different PM 2.5 concentrations. Clin Epigenetics. 2019 self-reported alcohol consumption, predicts hazardous May 6 [cited 2022 Jul 8];111. https://pubmed.ncbi.nlm. alcohol consumption in two distinct populations. Mol nih.gov/31060609/ Psychiatry. 2021 Jun 1 [cited 2022 May 12];26 [82] Astuti Y, Wardhana A, Watkins J, et al. Cigarette smoking and telomere length: a systematic review of 84 studies and (6):2238–2253. https://pubmed.ncbi.nlm.nih.gov/ meta-analysis. Environ Res. 2017 [cited 2023 Mar 32034291/. 9];158:480–489. https://pubmed.ncbi.nlm.nih.gov/ [73] Liu C, Marioni RE, Hedman AK, et al. A DNA methylation 28704792/ biomarker of alcohol consumption. Mol Psychiatry. 2016 [83] Topiwala A, Taschler B, Ebmeier KP, et al. Alcohol Nov 15 [cited 2022 May 13] 2018;23:422–433. https://www. consumption and telomere length: Mendelian ran­ nature.com/articles/mp2016192 domization clarifies alcohol’s effects. Mol [74] Aragona J, Lee CK. Scoliosis in fetal alcohol syndrome: Psychiatry. 2022 Jul 26 [cited 2023 Mar a case report. Orthopedics. 1981;4(10):1141–1143. 9];27:4001–4008. https://www.nature.com/articles/ [75] Gallagher C, McCarthy FP, Ryan RM, et al. Maternal s41380-022-01690-9 Alcohol Consumption During Pregnancy and the [84] Maugeri A, Barchitta M, Lio RMS, et al. The effect of Risk of Autism Spectrum Disorders in Offspring: alcohol on telomere length: a systematic review of a Retrospective Analysis of the Millennium Cohort epidemiological evidence and a pilot study during Study. J Autism Dev Disord. 2018 Nov 1 [cited 2022 pregnancy. Int J Environ Res Public Health. 2021 Jul 26];48(11):3773. https://pmc/articles/ May 1 [cited 2023 Mar 9];18(9):5038. https://pmc/arti PMC6182718/. cles/PMC8126216/ [76] Singer AB, Aylsworth AS, Cordero C, et al. Prenatal [85] Gonzalez-Jaramillo V, Portilla-Fernandez E, Glisic M, Alcohol Exposure in Relation to Autism Spectrum et al. The role of DNA methylation and histone mod­ Disorder: findings from the Study to Explore Early ifications in blood pressure: a systematic review. J Hum Development (SEED). Paediatr Perinat Epidemiol. Hypertens. 2019 Oct 1 [cited 2022 Jul 11];33 2017 Nov 1 [cited 2022 Jul 26];31(6):573. https://pmc/ (10):703–715. https://pubmed.ncbi.nlm.nih.gov/ articles/PMC5690833/ 31346255/ [77] Wilhoit LF, Scott DA, Simecka BA. Fetal Alcohol [86] Richard MA, Huan T, Ligthart S, et al. DNA Spectrum Disorders: characteristics, Complications, Methylation Analysis Identifies Loci for Blood and Treatment. Community Ment Health J. 2017 Aug Pressure Regulation. Am J Hum Genet. 2017 Dec 7 1 [cited 2023 Feb 17];53(6):711–718. https://pubmed. [cited 2022 Jul 11];101(6):888–902. https://pubmed. ncbi.nlm.nih.gov/28168434/. ncbi.nlm.nih.gov/29198723/ [78] Varadinova M, Boyadjieva N. Epigenetic mechanisms: [87] Syme C, Shin J, Richer L, et al. Epigenetic Loci of Blood a possible link between autism spectrum disorders and Pressure. Circ Genomic Precis Med. 2019 Jan 1 [cited fetal alcohol spectrum disorders. Pharmacol Res. 2015 2022 Jul 11];12(1):E002341. https://pubmed.ncbi.nlm. Dec 1;102:71–80. nih.gov/30645168/