Genetic and Epigenetic Effects of Environmental Mutagens and Carcinogens

BioMed Research International

Guest Editors: Alessandra Pulliero, Jia Cao, Luciana dos Reis Vasques, and Francesca Pacchierotti Genetic and Epigenetic Effects of Environmental Mutagens and Carcinogens BioMed Research International Genetic and Epigenetic Effects of Environmental Mutagens and Carcinogens

This is a special issue published in “BioMed Research International.” All articles are open access articles distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Contents

Genetic and Epigenetic Effects of Environmental Mutagens and Carcinogens, Alessandra Pulliero, Jia Cao, Luciana dos Reis Vasques, and Francesca Pacchierotti Volume 2015, Article ID 608054, 3 pages

Effect of Environmental Chemical Stress on Nuclear Noncoding RNA Involved in Epigenetic Control, Patrizio Arrigo and Alessandra Pulliero Volume 2015, Article ID 761703, 10 pages

Association between Genetic Polymorphisms of DNA Repair Genes and Chromosomal Damage for 1,3-Butadiene-Exposed Workers in a Matched Study in China, Menglong Xiang, Lei Sun, Xiaomei Dong, Huan Yang, Wen-bin Liu, Niya Zhou, Xue Han, Ziyuan Zhou, Zhihong Cui, Jing-yi Liu, Jia Cao, and Lin Ao Volume 2015, Article ID 234675, 7 pages

Applying a Weight-of-Evidence Approach to Evaluate Relevance of Molecular Landscapes in the Exposure-Disease Paradigm,SherilynA.GrossandKristenM.Fedak Volume 2015, Article ID 515798, 11 pages

Heterogeneity of Genetic Damage in Cervical Nuclei and Lymphocytes in Women with Different Levels of Dysplasia and Cancer-Associated Risk Factors, Carlos Alvarez-Moya, Monica´ Reynoso-Silva, Alejandro A. Canales-Aguirre, JoseO.Chavez-Chavez,HugoCasta´ neda-V˜ azquez,´ and Alfredo I. Feria-Velasco Volume 2015, Article ID 293408, 6 pages

Oxidative Stress Alters miRNA and Gene Expression Profiles in Villous First Trimester Trophoblasts, CourtneyE.Cross,MaiF.Tolba,CatherineM.Rondelli,MeixiangXu,andSherifZ.Abdel-Rahman Volume 2015, Article ID 257090, 11 pages

Bisphenol A Effects on Mammalian Oogenesis and Epigenetic Integrity of Oocytes: A Case Study Exploring Risks of Endocrine Disrupting Chemicals, Ursula Eichenlaub-Ritter and Francesca Pacchierotti Volume 2015, Article ID 698795, 11 pages

Environmental Impact on DNA Methylation in the Germline: State of the Art and Gaps of Knowledge, Francesca Pacchierotti and Marcello Spano` Volume 2015, Article ID 123484, 23 pages

Putative Epimutagens in Maternal Peripheral and Cord Blood Samples Identified Using Human Induced Pluripotent Stem Cells, Yoshikazu Arai, Koji Hayakawa, Daisuke Arai, Rie Ito, Yusuke Iwasaki, Koichi Saito, Kazuhiko Akutsu, Satoshi Takatori, Rie Ishii, Rumiko Hayashi, Shun-Ichiro Izumi, Norihiro Sugino, Fumio Kondo, Masakazu Horie, Hiroyuki Nakazawa, Tsunehisa Makino, Mitsuko Hirosawa, Kunio Shiota, and Jun Ohgane Volume 2015, Article ID 876047, 13 pages

Noncoding RNAs: Possible Players in the Development of Fluorosis,AtulP.Daiwile, Saravanadevi Sivanesan, Alberto Izzotti, Amit Bafana, Pravin K. Naoghare, Patrizio Arrigo, Hemant J. Purohit, Devendra Parmar, and Krishnamurthi Kannan Volume 2015, Article ID 274852, 10 pages Genotoxic Effect in Autoimmune Diseases Evaluated by the Micronucleus Test Assay: Our Experience and Literature Review, Olivia Torres-Bugar´ın, Nicole Macriz Romero, Mar´ıa Luisa Ramos Ibarra, Aurelio Flores-Garc´ıa, Penelope´ Valdez Aburto, and Mar´ıa Guadalupe Zavala-Cerna Volume 2015, Article ID 194031, 11 pages

Effect of 50 Hz Extremely Low-Frequency Electromagnetic Fields on the DNA Methylation and DNA Methyltransferases in Mouse Spermatocyte-Derived Cell Line GC-2, Yong Liu, Wen-bin Liu, Kai-jun Liu, Lin Ao, Julia Li Zhong, Jia Cao, and Jin-yi Liu Volume 2015, Article ID 237183, 10 pages

Environmental Epigenetics: Crossroad between Public Health, Lifestyle, and Cancer Prevention, Massimo Romani, Maria Pia Pistillo, and Barbara Banelli Volume 2015, Article ID 587983, 13 pages

Impact of Cadmium on Intracellular Zinc Levels in HepG2 Cells: Quantitative Evaluations and Molecular Effects, Chiara Urani, Pasquale Melchioretto, Maurizio Bruschi, Marco Fabbri, Maria Grazia Sacco, and Laura Gribaldo Volume 2015, Article ID 949514, 10 pages Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 608054, 3 pages http://dx.doi.org/10.1155/2015/608054

Editorial Genetic and Epigenetic Effects of Environmental Mutagens and Carcinogens

Alessandra Pulliero,1 Jia Cao,2 Luciana dos Reis Vasques,3 and Francesca Pacchierotti4

1 Department of Health Sciences, University of Genoa, Via Antonio Pastore 1, 16132 Genoa, Italy 2Toxicology Institute, Preventive Medical College, Third Military Medical University, Chongqing 400038, China 3Department of Genetics and Evolutionary Biology, Institute of Bioscience, University of Sao˜ Paulo, 05508-090 Sao˜ Paulo, SP, Brazil 4Laboratory of Toxicology UT-BIORAD, ENEA, CR Casaccia, Via Anguillarese 301, 00123 Rome, Italy

Correspondence should be addressed to Alessandra Pulliero; [email protected]

Received 21 June 2015; Accepted 22 June 2015

Copyright © 2015 Alessandra Pulliero et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Identifying the effects of environmental exposures on human is a general mechanism playing an important pathogenic health is a major objective of life sciences and biomedi- role in linking exposure to environmental toxic agents with cal research. In environmental health, the recognition that their pathological consequences. The underlying molecular exposures could produce DNA mutations represents a major mechanism involves both p53-microRNA interconnection landmark for risk assessment and prevention [1]. Conse- and alterations of Dicer function [5]. MicroRNAs have been quently, chemical agents have been categorized according to also implicated in transgenerational epigenetic inheritance their ability to alter the DNA sequence. Such information via the gametes. Other epigenetic mechanisms, like DNA has been fundamental to determine environmental risks methylation, are susceptible to environmental influence in and shape current regulatory efforts for exposure reduction. both somatic and germ cells modulating the mechanisms and Recent evidences suggest that the molecular influence of the risk of chemically induced genetic and epigenetic effects. the environment may extend beyond the interaction with The present special issue provides a comprehensive overview the DNA sequence [2]. Epigenetics is the study of heritable of the current status of this interesting field of research. changes in gene expression that occur without changes in It comprises manuscripts reporting novel data as well as DNA sequence [3]. This is particularly fascinating because state-of-the-art reviews. We have composed a balanced issue epigenetics involves factors that cause chemical changes to combining in vitro studies, as well as studies performed in occur in our genomes. Epigenetic mechanisms, with particu- humans.Theissueisstartedbyacomprehensivereviewby lar reference to microRNA alterations, DNA methylation, and M. Romani et al., describing the mechanisms of epigenetic histone modifications, can change genome function under control and their role in disease development, the influence exogenous influence. MicroRNAs are posttranscriptional of the environment, lifestyle, and nutritional habits influence regulators of gene expression involved in carcinogenesis, on epigenetic markers, and how these factors are related metastasis, and invasion. Thus, microRNAs are potentially to the development of cancer and other noncommunicable useful as diagnostic and prognostic biomarkers as well as diseases. Diet is a major exposure daily route to many anticancer therapeutic targets. Many microRNAs undergo toxic substances including endocrine disruptors such as altered expression as consequence of exposure to carcinogens bisphenol A. The epigenome of mammalian female germ and these changes may be useful in cancer prevention [4]. The cells and oocyte development is modified by low bisphenol relationship between exposure to environmental carcinogens A concentrations with functional consequences on gene and epigenetics revealed that toxicants modify epigenetic expression, chromosome dynamics in meiosis, and oocyte states. Accordingly, the alteration of microRNA expression development. These important aspects are described by U. 2 BioMed Research International

Eichenlaub-Ritter and F. Pacchierotti underlying that there the presence of any autoimmune disease condition. The are specific time windows, during which profound chromatin authors propose that micronuclei could be used as an early remodeling occurs and maternal imprints are established or biomarker of disease progression and/or response to therapy. protected that appear particularly vulnerable to epigenetic The effect of toxicological agents on genes is an area that deregulation by bisphenol A. Indeed, the BPA exposure warrants further research as a major public health issue. In has long lasting consequences on the female reproductive particular, pollution in Asia is a growing concern also in health, reducing fertility and potentially leading to offspring occupational places. Consequently, human molecular epi- defects. At the same time, certain components of the diet demiological studies have explored the association between can modify the epigenetic pattern through natural bioactive polymorphisms of DNA repair genes and chromosomal components that can act on DNA methylation or histone damage by 1,3-butadiene in current exposure workers. The modification or, as in the case of exogenous miRNA that can paper by M. Xiang et al. focuses on genetic polymorphisms be direct epigenetic actors. Although much effort has focused of DNA repair genes and chromosomal damage. In their on the gene-environment axis, there is a growing body paper, they provide evidence for DNA repair genes (XRCC1, of information suggesting that environmental influences MGMT, APE1,andADPRT)thatmodifythegeneticinstabil- extend beyond the DNA sequences of our genes. Epigenetics ity induced by 1,3-butadiene, group 1 carcinogen, and widely includes the study of the protein constituents of chromatin, used as an industrial chemical and also present in autoe- the interaction of microRNAs with the genome, and the mission. Newly developed biomarkers (MNi: micronuclei; protein and DNA modifications that appear to define biologic NPBs: nucleoplasmic bridges; NBUDs: nuclear buds) in the states in local regions of chromosomes. However, only more cytokinesis-blocked micronucleus cytome assay were used recently the role of noncoding RNAs (ncRNAs) in epigenetic to detect chromosomal damage, showing that 1,3-butadiene processes has been highlighted. This topic is discussed in exposed workers have increased frequencies of MNi and P. Arrigo and A. Pulliero’s paper dealing with the effect of NPBs compared to control subjects. Actually, the effect on environmental chemical stress on nuclear noncoding RNA cell transformation of many environmental factors, which are involved in epigenetic control. This paper summarizes how reported in epidemiological studies to be associated with a a chemical mutagen can interfere with the conformation certain risk for a given cancer type, might well be mediated by of specific regulatory domains in ncRNAs. The majority of epimutations. The paper by S. A. Gross and K. M. Fedak nicely human cellular RNA consists of rRNA (∼90% of total RNA). summarizes the role of such epimutations applying a weight The total number of RNA molecules is estimated tobe of evidence approach to evaluate molecular characteristics 7 about 10 per cell, and ncRNAs include snRNA, snoRNA, for environmental carcinogens as well as associated disease and miRNA. The overexpression of lncRNAs can potentially outcomes. The authors consider the example of benzene as measure the cytotoxicity signals of various environmental an environmental carcinogen and myelodysplastic syndrome stresses [6]. Knowing the location of such ncRNAs could help as a causative disease endpoint. Using the weight of evi- in selecting the best candidates for starting the ncRNA-based dence method, they find overlapping polymorphisms in the gene therapy trials. For example, the fact that miRNA altered metabolic enzymes GST and NQO1. Additional mutations expression in cancer cells has a pathogenic effect provides the including RUNX1 and glycophorin A, epigenetic events rationale for using miRNAs as potential therapeutic targets in associated with DNMT3a, and polymorphism in interleukin cancer. It is interesting to note how the chemical entity could cytokines induce dysregulation of the hematopoietic sys- induce a slight change in the conformation that is critical tem in benzene-exposed workers. The weight of evidence formolecularrecognitionintheprocessofRNAediting. approachillustratedbyS.A.GrossandK.M.Fedakisauseful Environmental exposures to potentially aneugenic agents, tool for sorting, categorizing, and prioritizing the most mean- including therapeutic exposures and exposures associated ingful information. Using such a framework, researchers can with lifestyle habits, influence the epigenetic alterations. In determine ways to take advantage of best practices in iden- India, the health of approximately 62 million people is at risk tifying exposure scenarios and defining biomarkers relevant due to high concentration of fluoride in drinking water/diet. to the exposure-disease paradigm. In addition, certain toxic The paper of A. P. Daiwile et al. describes the expression metals (i.e., cadmium, nickel, arsenic, and chromium) have profiling of short noncoding RNAs, in particular miRNAs been identified as persistent environmental pollutants due to and snoRNAs, carried out in sodium fluoride (NaF) treated their indestructible chemical and physical properties. While human osteosarcoma cells to understand their possible role the ability of some metal compounds to cause cancer in in the development of fluorosis. They found that epigenetic exposed workers has been known for a long time, epigenetic- modifications in miRNAs and C/D box snoRNAs that can based research has recently shed light upon mechanisms regulate the key genes are involved in the development of involved in metal-induced carcinogenesis [7]. For example, fluorosis, suggesting the possible involvement of miR-124 and recent research has shown that cadmium exposure stim- miR-155 in the development of fluorosis. The special issue ulates cellular proliferation, increases oxidative stress and is continued by three contributions describing mutations DNA damage, and induces global DNA hypomethylation and epigenetic events important in our understanding of [8]. C. Urani et al., in the experimental study, provide new molecular mechanisms associated with exposure-disease out- information on the novel mechanisms for cadmium toxicity comes. The review by O. Torres-Bugarin et al. provides strong and biological effects and, along with those already reported information that sustains the observation that increased in the literature, open new intriguing interpretations on micronuclei frequencies are observed in association with cadmium-induced carcinogenicity. The authors show that BioMed Research International 3

HepG2 cells, exposed to 10 𝜇Mcadmiumfor24hrs,increase and epigenetic effects in the diseases risk, and we are grateful microarray expression profiling of labile zinc after cadmium to them for their contributions. exposure. In addition, the downregulation of miR-34a and miR-200a, both implicated in the epithelial-mesenchymal Alessandra Pulliero transition, supports the role played by zinc in affecting Jia Cao gene expression at the posttranscriptional level. Research on Luciana dos Reis Vasques epimutations is also a rapidly growing area. These changes Francesca Pacchierotti may involve modification of DNA bases and histones, such as methylation, ubiquitination, or acetylation. The regulation References of transcription and genome stability by epigenetic systems is crucial for the proper development of mammalian embryos. [1] G. N. Wogan, “Molecular epidemiology in cancer risk assess- Y. Arai et al. confirm the hypothesis that combined exposure ment and prevention: recent progress and avenues for future to chemicals (diethyl phosphate, mercury, cotinine, selenium, research,” Environmental Health Perspectives, vol. 98, pp. 167– and octachlorodipropyl ether) detected at low concentrations 178, 1992. in maternal peripheral and cord blood samples affect embry- [2] R. L. Jirtle and M. K. Skinner, “Environmental epigenomics and oid body formation and neural differentiation from human disease susceptibility,” Nature Reviews Genetics,vol.8,no.4,pp. 253–262, 2007. induced pluripotent stem cells (hiPSCs). C. E. Cross et al. add another detail reporting that short-term exposure of [3] A.P.WolffeandM.A.Matzke,“Epigenetics:regulationthrough repression,” Science,vol.286,no.5439,pp.481–486,1999. villous first trimester trophoblasts to low concentrations of [4] A. Izzotti, “Molecular medicine and the development of cancer H2O2 significantly alters miRNA profile and expression of chemopreventive agents,” Annals of the New York Academy of genesimplicatedinplacentaldevelopment.Thelastsection Sciences,vol.1259,no.1,pp.26–32,2012. of this special issue comprises two papers describing the features of DNA methylation and the exposure to a variety [5] A. Izzotti and A. Pulliero, “The effects of environmental chemical carcinogens on the microRNA machinery,” International of chemical stressors, occurring during the prenatal or the Journal of Hygiene and Environmental Health,vol.217,no.6,pp. adult life. The first paper by F. Pacchierotti and M. Spano` 601–627, 2014. reviews the literature on the impact of chemicals and lifestyle [6]F.Doolittle,T.D.P.Brunet,S.Linquist,andT.R.Gregory, factors on DNA methylation in germ cells. Studies on rodents “Distinguishing between ‘Function’ and ‘Effect’ in genome show that DNA methylation can be altered by many different biology,” Genome Biology and Evolution,vol.6,no.5,pp.1234– kinds of exposures during the fetal as well as the adult 1237, 2014. life. Still these studies suffer from some limitations: few [7] J. E. Klaunig, L. M. Kamendulis, and Y. Xu, “Epigenetic mech- experiments were conducted on the female germline; the anisms of chemical carcinogenesis,” Human and Experimental analyses were rarely extended to a genomewide scale; only Toxicology,vol.19,no.10,pp.543–555,2000. in a few cases the functional impact of epigenetic changes [8]D.Huang,Y.Zhang,Y.Qi,C.Chen,andW.Ji,“Global was investigated; dose-response relationships were scarcely DNA hypomethylation, rather than reactive oxygen species considered. Nevertheless, their results are very important (ROS), a potential facilitator of cadmium-stimulated K562 cell because they establish proof of principle demonstration that proliferation,” Toxicology Letters,vol.179,no.1,pp.43–47,2008. a variety of exogenous stressors may alter DNA methylation at developmentally important imprinted or metabolic genes and warrant future investigations on human mother-child cohorts and biomonitoring epigenetic analyses on semen collected from human exposed cohorts. In a final paper, Y. Liu et al. demonstrated for the first time that 50 Hz ELF-EMF exposure can induce the alterations of genomewide methylation and the expression of DNMTs in spermatocyte-derived GC-2 cells. Indeed, the epigenetics plays an important role in the biological effects of 50 Hz ELF-EMF exposure. Understanding the mechanistic basis of how epigenetic regulation is achieved is fundamental to placing this level of biologic regulation in the toolbox of environmental health science and medicine. Based on these contributions, it may be stated that the genetic and epigenetic effects of environmental mutagens and carcinogens are subject of innovative research.

Acknowledgment The authors of the papers included in this special issue have provided new inputs to establishing a synergistic interaction between environmental exposure and genetic predisposition Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 761703, 10 pages http://dx.doi.org/10.1155/2015/761703

Review Article Effect of Environmental Chemical Stress on Nuclear Noncoding RNA Involved in Epigenetic Control

Patrizio Arrigo1 and Alessandra Pulliero2

1 National Research Council (CNR), Institute for Macromolecular Studies (ISMAC), Via De Marini 6, 16149 Genoa, Italy 2Department of Health Sciences, University of Genoa, Via A. Pastore 1, 16132 Genoa, Italy

Correspondence should be addressed to Patrizio Arrigo; [email protected]

Received 23 March 2015; Accepted 11 May 2015

Academic Editor: Diana Bahia

Copyright © 2015 P. Arrigo and A. Pulliero. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

In the last decade the role of noncoding RNAs (ncRNAs) emerges not only as key elements of posttranscriptional gene silencing, but also as important players of epigenetic regulation. New kind and new functions of ncRNAs are continuously discovered and one of theirmostimportantrolesisthemediationofenvironmentalsignals, both physical and chemical. The activity of cytoplasmic short ncRNA is extensively studied, in spite of the fact that their function and role in the nuclear compartment are not yet completely unraveled. Cellular nucleus contains a multiplicity of long and short ncRNAs controlling at different levels transcriptional and epigenetic processes. In addition, some ncRNAs are involved in RNA editing and quality control. In this paper we review the existing knowledge dealing with how chemical stressors can influence the functionality of short nuclear ncRNAs. Furthermore, we perform bioinformatics analyses indicating that chemical environmental stressors not only induce DNA damage but also influence the mechanism of ncRNAs production and control.

1. Introduction The next-generation sequencing and other high throughput methods have underlined the relevance of epigenetic The term “epigenetic” was originally used in the field of process in cellular homeostasis. A large number of exper- evolution and development [1]. The advancement of knowl- iments have demonstrated epigenetic alteration in several edge about molecular mechanisms that control the gene pathological processes [3]. It is important to note the exis- expression has profoundly changed the meaning of the term, tence of a feedback between metabolic network and epige- which can assume different facets that are based on the field netic activity [4, 5]. The knowledge about the mechanisms of interest of a researcher. Recent studies shed light on the connecting metabolic network and epigenetic is even now role of environment in epigenetic processes and, as a con- to be deeply investigated. The goal of epigenetic is the sequence, a new definition of epigenetic has been proposed transcriptional control by several different mechanisms. The [2]. This interpretation considers the epigenetic as structural fundamental elements are (a) the DNA methylation; (b) the changes, at chromosome level, that allow sensing, storing, histone modification; (c) nucleosomal positioning; and (d) and transmitting modifications of biological activity. This noncoding RNAs. DNA methylation, histone modifications, proposed definition considers the modification of chromoso- and nucleosomal positioning are the mechanism studied mal marks associated with genomic and cellular changes. The for a long time; however only more recently the role of proposed definition does not take into account other events, noncoding RNAs (ncRNAs) in epigenetic processes has been both endogenous (structural) and exogenous, except when highlighted. The results of deep sequencing project, such they affect chromosome functionality. Probably epigenetic asENCODE,havepointedoutagreaterthanexpected mechanisms operate as a buffer for genetic variations that transcriptional activity of genome. These high throughput are waiting to change the expression of genes leading to a experiments have also screened an amount of expressed different molecular phenotype. ncRNAs larger than protein-coding genes [6]. Systematic 2 BioMed Research International bioinformatics analysis has supported the investigation of a new more refined categorization has been proposed [15]. relation between ncRNAs and epigenetic mechanisms [7]. Volders et al. describe a novel database for human long The advancement of experimental activity has unrevealed noncoding RNA, constituted by a large and diverse class the complexity of ncRNA world. Nowadays it is known that of noncoding RNA genes. LNCipedia offers 21 488 human ncRNAsareinvolvedinprocessessuchastranscription, lncRNA transcripts obtained from different sources. They translation, RNA editing, and protection against exogenous underline that, much like microRNAs, many lncRNAs have nucleicacids.TheroleofncRNAsintheresponseagainstthe a significant secondary structure, in line with their presumed environmental stress has initially been identified in bacteria association with proteins or protein complexes [16]. [8], subsequently in plant [9],andtheninanimals.Thelength According to this rule we can classify them into “nuclear” of ncRNAs is a discriminating feature and, even if the initial and “cytoplasmic” noncoding RNAs. It is important to clarify focus of research has been the class of short ncRNAs, the long that even if the ncRNA is produced in the cytoplasm but acts in the nucleus it is considered nuclear. The proposed noncoding RNAs have acquired a considerable significance in approach allows better identifying those classes that can the exploration of epigenetic mechanisms. influence epigenetic mechanisms and RNA processing and It is difficult to investigate how each type of ncRNAs also having more valuable information about the differen- cooperates in the epigenetic control and to examine the tial effect of environmental stressors on the two cellular hierarchy and timing of their control activity. The aim of this compartments. Knowing the location of such ncRNAs could paper is to give a general framework of the overall epigenetic help in selecting the best candidates for starting the ncRNA- control regulated by noncoding RNA taking their structural based gene therapy trials. For example, the fact that miRNA characteristics into account. In particular, we would like to overexpression in cancer cells has a pathogenic effect provides highlight those elements that are predominantly involved in the rationale for using miRNAs as potential therapeutic the response against environmental stress. targets in cancer. Cytoplasmic compartment could also be influenced by 2. The Noncoding RNA World noncoding RNA originated far from one tissue (circulating ncRNAs) or by exogenous ncRNAs (viral miRNAs). In an It has been commonly accepted that ncRNAs were not epigenetic perspective the role of ncRNA in the nucleus bears functional and were therefore labeled as “junk RNAs.” At the particular interest. Noncoding RNAs act in nucleus as long same time, the function of the noncoding transcripts and ncRNAsandinnucleolusasshortncRNAs. their relevance to diseases have remained undefined. The long ncRNA (lncRNA) action affects directly the Many of the putative functional ncRNAs are present at transcriptional process. The short ncRNA acting in the very low levels and thus unlikely to be of any importance nucleolusisprincipallyentailedinRNAqualitycontrol.It withrespecttocellororganismphysiology.Additionally, is crucial to recall that ncRNAs are originated by a tran- the abundance of an ncRNA species shortly correlates with scriptional process comparable with those regulating protein its level of conservation [10], which is a good agent for transcription. The localization of genomic regions encoding function [11, 12], thus determining the relative abundance ncRNA is not yet completely clarified. In terms of ncRNA of a given ncRNA in the relevant cell type. The majority transcription, it is known they can be originated near the of human cellular RNA consists of rRNA (∼90% of total locus that they regulate (cis-regulation) or distally to their tar- RNA). Total number of RNA molecules is estimated to be get (trans-regulation). In the paragraph below we summarize 7 about 10 per cell, and ncRNAs include snRNA, snoRNA, the functional differences between strictly nuclear ncRNAs and miRNA. Although there is less tRNA by mass, ncRNAs (e.g., lncRNAs, short ncRNAs such as piRNAs, paRNAs, and rasiRNAs) and nucleolar RNAs. In order to better correlate small size results in their molar level being higher than their structural and functional characteristics, we separately rRNA [13]. Other abundant RNAs, such as mRNAs, snRNAs, analyzed the main groups of nuclear and nucleolar ncRNAs. and snoRNAs, are present in aggregate at levels that are This approach is valuable to understand ncRNA cooperative about 1-2 orders of magnitude lower than rRNA and tRNA. effects on epigenetic regulation. In particular our interest is Certain small RNAs, such as miRNAs and piRNAs, can to describe the functional role of those ncRNAs operating be present at very high levels; however, this appears to be in nuclear environment because, leaving out the important cell type dependent [11]. Moreover, the overexpression of exception of mitochondria, the epigenetic control occurs lncRNAs can potentially measure the cytotoxicity signals in the nucleus. This survey also attempts to find common of various environmental stresses. Indeed, some researchers elements that help to analyze the DNA damage and repair by a have developed human cells transfected with lncRNAs used different integrative perspective. We guess that clarifying the as a sensor of cytotoxicity for a broad range of environmental functional feedback between different noncoding regulatory stresses. They identified three lncRNAs (GAS5, IDI2-AS1, and RNAs helps to decipher those steps that are more sensitive to SNHG15)thatrespondtoseveralenvironmentalstressesand induced damage by environmental factors in the perspective found overexpression of these lncRNAs sensitized human toassessthelongtermriskforhumanhealth. cells to cell death by the stresses. These sensor cells with overexpressed lncRNAs can potentially report cytotoxicity 3. The Nuclear Small ncRNAs (snRNAs) signals of various environmental stresses [14]. This explosion can be related with the increased number The nuclear activity regulation by RNA is a dynamic process of biological functionalities operated by these molecules. In in which are involved ncRNAs originated from nuclear com- spite of the simple initial classification based on their length, partmentandsomeofthemareprocessedinthecytoplasm. BioMed Research International 3

The regulatory ncRNAs, according to their length, canbe there can be as many as 1000 copies of piRNAs in a cluster. divided into small ncRNAs (sncRNAs, smaller than 200 piRNAs are processed from long precursors transcripts but nucleotides (nts)) and long ncRNAs (lncRNAs, longer than little is known of the biogenesis of piRNAs. Some piRNA 200 nts). There are also ncRNAs with the length of 60– clustersconsistoftransposonsequences.Amajorroleofthe 300nts, called small nucleolar RNAs (snoRNAs). The small piRNA/Piwi protein complex in germ line cells is to protect ncRNAs include Piwi-associated RNAs (piRNAs), microR- cells from invading transposons. When the cell encounters NAs (miRNAs), small interfering RNAs (siRNAs), transcrip- atransposonthatithasnotbeenexposedtobefore,the tion initiation RNAs (tiRNAs), and other small ncRNAs [17]. transposable elements (TE) by chance may incorporate into The short nuclear RNAs (snRNAs) are classified in two the DNA in a piRNA-encoding cluster and thus its sequence different groups according to the proteins they interact with can become part of the piRNA cluster. This is a type of [18]. The first group is identified as “Sm-class” because they “genetic immune system” that is found in both eukaryotes and interact with Sm proteins. The second group is called “Lsm- prokaryotes. For example, the CRISPR complex in bacteria class” according to the name of their interacting proteins. hasamechanismtoprotectcellsfrominvasionbyplasmids These proteins and snRNA cooperate in spliceosome for- and viruses [25]. The piRNA/Piwi complex is also essential in mation [19]. Conversely, it is important to underline the genetic imprinting in the case involving DNA methylation of transcriptional differences between the two classes. The “Sm- the imprinted locus Ras protein-specific guanine nucleotide- class” is transcribed by Pol II while the “Lsm-class” is releasing factor 1 locus in mouse germ line cells [26]. In transcribed by Pol III. nematodes, piRNAs detect a TE sequence via imperfect base- Another important discriminative element between these pairing and then induce another small RNA class, termed two types of snRNAs is their biogenetic pathway. The “Sm- 22G-RNAs to silence a transposon [27]. Some processes class” requires a cytoplasmic phase, in spite of the fact involve epigenetic mechanisms. For example, in Drosophila, that “Lsm-class” biogenesis is exclusively completed in the nuclear piRNAs can target a transposon and direct Piwi nucleus. The two groups show structural differences as proteins repress chromatin and transcription of the TE [28]. highlighted in Figure 1. From the structural point of view, 󸀠 Additionally, piRNAs may also induce the methylation of TE the 5 end is discriminative between Sm-class and Lsm-class. LINE-1 DNA in humans. In the first group there is a trimethylguanosine, while the The primary task of piRNAs is to preserve the genome second group of ncRNAs only have one methyl group in the 󸀠 integrity by repression TE. The transposon repositioning can 5 cap. The protein selectivity is driven by the presence of induce some genomic damage in germ line. It is interesting to specific nucleotide motifs. In the “Sm-class” the recognition underline that, at variance with the precursors of other short motif is located between two regular stem loops and it has the following composition: AUUUGUG (Sm site) and ncRNAs such as miRNA or snRNA, the precursors of piRNA GAAGCUG (Lsm site) [20]. are vesiculated in cytoplasm without a shuttle complex like Among these small snRNAs the U7 seems to have a role RanGTp-exportin 5 required by pre-miRNA. piRNA matin epigenetic control by repression of histone genes [21]; uration and Piwi-piRNA complex (Piwi-piRISC) formation its structure is shown in Figure 2. It differs from other Sm occur in the cytoplasm. Piwi-piRISCs are then imported into snRNAs because it recruits a different set of Sm proteins. the nucleus where they repress TEs at transcriptional level In addition to this specific nuclear small noncoding RNA by directing specific histone modifications to TE loci. The other ribonucleotide oligomers have function in epigenetic piRNAs originate from intergenic repetitive elements that control. In particular we consider the Piwi-interacting RNAs are in many cases clustered. Currently, available knowledge (piRNAs), the promoter-associated RNAs (paRNAs), and the suggests a possible involvement of piRNA in “de novo” repeat associated RNAs (rasiRNAs). methylation of DNA in transposable elements. The piRNAs Piwi-interacting RNAs (piRNAs) are a novel class of are indicated as pivotal element in the process of genomic sncRNAs, with a length of 26–31 nts, which specifically imprinting. It is important to underline that also these small interact with P-element-induced wimpy testis (Piwi) protein ncRNAs are prone to be methylated [29]. Their stability is in the Argonaute group of proteins. Piwi proteins in humans affected by methylation because it protects the piRNAs, and have four homologs: PiwiL1/Hiwi, PiwiL2/Hili, PiwiL3, and other small ncRNAs, from other kinds of modifications, such PiwiL4/Hiwi2 [22]. Studies have reported that Hiwi may take as uridylation, that determine the speed of RNA degradation. part in germ cell proliferation and carcinogenesis process The enzyme responsible for piRNA methylation is the [23]. HEN1, that is, a methyltransferase that methylates miRNAs The piRNAs are short noncoding RNAs little bit longer andsiRNAsontheriboseofthelastnucleotide.Thisenzyme 󸀠 󸀠 than miRNAs; they do not show a significant sequence introduces a 2 -O-methyl group at 3 end of a small RNA, conservation. They, at variance with miRNAs and siRNAs, including piRNAs. This stabilization process is a critical step are functionalized by a DICER independent pathway. They to protect the genome integrity as demonstrated by exper- interact with the AGO3, a member of Argonaute family that iments carried out on protozoa [30]. Docking experiments is peculiar of germ line cells [24]. with HEN1 and synthetic short RNA oligomers have sug- piRNAs are found in germ line cells, especially in mam- gested, as possible recognition site, a tetramer such as NCGN; mals; for example, several million piRNAs are found in mam- these results confirm the well established knowledge about malian testes. Genetic regions that encode piRNAs consist of thecentralityofCGdinucleotideinmethylationprocess[31]. clusters. These clusters have repeats of piRNA sequences and In Figure 3 piRNA 2D structure and folding are shown. 4 BioMed Research International

G U U G U A A G A C U U A G A U U U C U A U A U A A C G C A A U G U U G U U G A A G A G U G AU A A C U C G G U C A A U U U G C A C C G A A G U C A G G A A A A U G U A C A C G G A C C G A A C A A G C U A C G U C A

0 1

(a) A U U U A G C A C G U U U U G G U U G U A U A U GU AC U G C A U G A G U G A C G A C A A A A U C U U G C AC G A C U G A U G A A G A G U G C C U A C G G U U C A G A UAAU C G A A C G C U C U G A C A A G C C G A U G U U A UU CU A C G C G A U G C G C C G C A U A A G G G U C C U G U U A A U U A C G A C G GA U U CA U AU C AUCC C G A G C U U C

0 1

(b)

Figure 1: (a) 2D structure of U6 snRNA (Lsm-class). (b) 2D structure of U1 snRNA (Sm-class). BioMed Research International 5

U U features contribute to specifying the regulatory function of A G these snoRNAs. The two conformations underline the small C U A G G but significant differences between the piRNA located on C C A U C thestrandplusandthepiRNAlocatedonthestrandminus. U C A U C U C C A A G G U C It is interesting to note that only the piRNA located on the U G U U 󸀠 U A G G U U C C A U U UU strand plus shows a C in the 3 end. This suggests that this U G C oligomer can be prone to stabilization by methylation. The A A main distinctive feature is the presence of a small symmetric G U A C bulgeinthepiRNAlocatedontheminusstrand. A U U U Another ncRNA class containing both types of recogni- U U tion sites are scaRNA, which are longer than the previously described types. They are associated with the Cajal body, 0 1 another organized nuclear structure in which also RNPs are processed. Cajal bodies have role in epigenetic control because in these structure the mRNAs encoding histones are Figure 2: Conformation of Human U7 cDNA (ENST00000459276). processed by snRNAs [33]. The scaRNAs unveil a complex conformational structure that combines both nucleotide feature of C/D and H/ACA box RNAs. Their complexity is The repeat associated siRNAs (rasiRNAs) can be con- exemplified in Figure 5. sidered as subclass of piRNAs. They look like shorter than The scaRNAs include a specific nucleotide motif UGAG, piRNAs, although their length depends on specific organism. the Cajal body box (CAB), in the H/ACA domain of the They are originated from an antisense transcript, in spite of ncRNA [34]. the fact that endo-siRNA and miRNA and their fictionaliza- tion appear to be DICER1 independent [32]. At the moment 5. Long Noncoding RNAs the rasiRNAs were identified in low eukaryotes. In order to focus our survey on human we mention the class of small Long noncoding RNAs (lncRNAs) are transcripts greater nuclear ncRNA but we do not have enough evidences about than 200 nucleotides in length with little or no protein-coding their role in human epigenetic control. capability [35]. This arbitrary size threshold distinguishes lncRNAs from other distinct classes of small RNAs such as 4. The Nucleolar Small Noncoding RNA microRNAs, tRNAs, and snoRNAs. A recent definition of long ncRNA has established a new minimal length of 1000 bp Thenucleolusisthemajorcompartmentofthenucleus [36]. and it is the place where rRNAs are synthesized. A nuclear In the last few years the functional explorations of compartment can contain more than one nucleolus, that is, individual lncRNAs have seen rapid growth. Concomitant surroundedbycondensedchromatinlayer.Inthenucleus with this increased growth of characterized lncRNAs is an there are other organized elements such as Cajal bodies. The increasing understanding toward biological mechanisms, as regulation of nucleolus assembly and function is primarily well as a growing awareness and recognition of the impor- committed to specific ncRNAs called small nucleolar RNAs tance of lncRNAs in virtually every cellular and regulatory (snoRNAs). Their primary function is to modify other RNAs, process [37]. not only rRNA but also tRNA and other ncRNAs. These Short ncRNAs are mainly devoted to inducing rela- short ribonucleic acids, the most abundant group of intron- tive small changes. Conversely, long ncRNAs are capable encoded ncRNAs, are classified, according to their specific of directly affecting the transcriptional process mainly by conformational properties, into two main groups: the C/D chromatin remodeling [38]. The long ncRNAs are confined box and the H/ACA box RNAs. Each group executes a specific in the nucleus and the experimental findings have con- modification reaction: the C/D box RNAs operate a RNA firmed their role in transcriptional control by recruitment methylation reaction, whereas H/ACA box RNAs trigger of proteins responsible for chromatin remodeling. The long the change of Uridine into pseudouridine. Figure 4 shows ncRNAs interact with promoter of silenced genes but they the architectural differences between C/D box and H/ACA can also target other distal transcriptional regulatory regions box RNAs. The C/D box is characterized by two nucleotide (enhancer or LCR). Indeed, lncRNA-p21 is induced by DNA motifs: the “C box” (RUGAUGA) and the “D box” (CUGA), damagecausedbydoxorubicinandplaysakeyroleinthe 󸀠 󸀠 respectively, located in the 5 and 3 ends of the sequence. p53 transcriptional response [39]. This lncRNA represses p53- The H/ACA snoRNAs have the following canonical regulated genes through binding to heterogeneous nuclear motifs: the “H motif” (AgAnnA) located between two con- ribonucleoprotein K and modulating its localization, which is 󸀠 served stem loops and the box “ACA” near to the 3 end. necessary for the p53-dependent apoptotic response to DNA Figure 4 underlines the conformational differences damage. Moreover, cycloheximide and hydrogen peroxide between the two classes of snoRNAs, in particular in the dramatically induced these lncRNAs and respond to cellular terminal loop that in the C/D box is smaller than in the stresses. Distinct sets of lncRNAs play roles in cellular defense H/ACA box. The differences in the two ends and inthe mechanisms against specific stresses, and particular lncRNAs symmetry of internal bulges are also noticeable. All these havethepotentialtobesurrogateindicatorsforcellular 6 BioMed Research International

A A U A A A C C A C G C C U C G A A A G G A G A C G A C G A U C G G U U G C G G A U U A G C G U U U U U

0101

(a) (b)

Figure 3: (a) Conformational analysis of a piRNA (NCBI code DQ569927 coding strand minus chromosome 11), (b) conformational analysis of a piRNA (NCBI code DQ569913 coding strand plus chromosome 21).

A C C C C C C C C U U G U G C C G A U G C C G C G C G C G C U A A C G U U G C A U A G G A U U U U C G U A U A C G U A U A UC U C C C GC U A G G A U G C UG G C A G CAA C C A C A C U UUAA A A C UG U C CU C G UG UG G C C G A C A A G G C G A C U C G G C C U G G U U C UG U C U U U UUCAG G G A G C G G A C C C U A C G G C G G A U U UG U C U C G G A A G C C U C C U A G U U A GACAGA GCC G G A AU AC C C C A C G AG A A GC C A U U A G 0 1 01

(a) (b)

Figure 4: (a) C/D box RNA (SNORD125 NCBI code AM13037.1), (b) H/ACA box RNA (ACA10 NCBI code AJ609432.1). BioMed Research International 7

A A U A A A A U G C U U G C G A AA UA G C A G U U U U C G U G C C G A U U A U A G C G GU A AAAU A A U G C A U G C A U A U A A A U A U A G C U G U G C U A U U U G A U A A A U U U G C U U U A U U U A A A A G C U A G GU A U G G AU U G U A GU G U A C UAA C A U UA G U G UC A U GGG C A U C G G G G G A U U A C A UU A A U U GA C C A G UG G C A U U A C U GU C GA A G UC G CU AUG G U C A A A A G G U C G AUAAU C U G AG U G A AA AG G U A UU C G U G C C UG GAG AAUA UG G U U U C U U A G U G C A AG U A U U U A G G C C U U C C A G G A GG GUU U U C AC G G UG A U U AC UG A U UUA AG G UU A AA G U UU AA G G C G A U

0 1

Figure 5: Example of scaRNA conformation (SCANRNA7 NCBI code AY077740.1). stress responses in human-induced pluripotent stem cells a prolonged environmental stress, this second analytical [40]. Another study identifies six long ncRNAs (MIR22HG, perspective being valuable in terms of risk assessment and GABPB-AS1, LINC00152, IDI2-AS1, SNHG15, and FLJ33630) regulatory activity. This survey describes the effect of envi- that responded to chemical stressors (cisplatin, cyclohex- ronmental stress on short snRNAs taking their characteristic imide, and mercury (II) oxide) in HeLa Tet-off cells. The conformational features into account. It is important to results indicate that long ncRNAs respond to general and underline the functional feedback between the cytoplasmic specific chemical stressors. The expression levels of the long phase of posttranscriptional gene regulation and nuclear epi- ncRNAs were elevated because of prolonged decay rates in genetic control. We guess that, in an epigenetic point of view, response to chemical stressors [41]. nuclear compartment seems to be more interesting because also the cytoplasmic ncRNAs such as miRNAs are transcribed 6. Environmental Stress and Small Nuclear inthenucleusandconsequentlytheyarealsosubjectto ncRNA and In Silico Test modification induced by snRNAs. In a previous paper [43] we have investigated, by an “in silico” analysis of interaction The above reported sections have summarized the nuclear between a selected sample of chemical environmental pol- ncRNA characteristics, functions, and their role in epige- lutant and DICER, a pivotal enzyme of RISC complex. In netic control. The knowledge about the damage induced by the present survey, in order to exemplify the possible effects environmental agents at epigenetic level is well established of environmental chemical stress on epigenetic regulation, principally at DNA level, in spite of the fact that the role we have estimated the interaction between three chemical of noncoding RNAs has become visible more recently. The mutagens and some available structures of ncRNAs. These initial evidence of the ncRNA’s role in stress response was structures represent specific domains of noncoding RNA obtained by experiments in plants [42] and then in animal becausetoobtainacompletestructureisverydifficult.In models in higher vertebrate and in humans remains to be addition some of them are obtained from yeast but, in this done. It is necessary to distinguish between an acute and case, the evolutionary conservation ensures a quite reliable 8 BioMed Research International

Table 1: Interaction between three different chemical mutagens and some available structures of ncRNAs.

Benzo(a)pyrene 2-Nitrofluorene 4-Nitrosomorpholine PDB code Type of RNA CID 2336 CID 11 831 CID60 46 ACE Contacts ACE Contacts ACE Contacts 2PCW rRNA bound with H/ACA 𝜓 domain −274.04 57 −231.89 61 −110.60 24 U65 H/ACA 󸀠 2EUY 3 hairpin −281.06 70 −237.26 55 −113.09 37 loop U64 H/ACA − − − 2QH3 󸀠 279.60 90 219.53 60 109.68 31 3 hairpin loop U4 snRNA − − − 1MFJ 󸀠 266.35 76 218.45 56 109.02 35 3 stem loop 2LK3 U2/U6 snRNA complex −224.25 48 −235.16 48 −108.79 21 2LX1 Internal loop −253.01 79 −189.51 55 −98.88 34 2O33 U2 stem I −283.93 58 −223.18 78 −117.37 30 1LC6 U6 stem loop −273.62 65 −226.94 69 −129.48 19 1KKS Histone mRNA hairpin −250.27 162 −230.73 165 −133.11 51

(A) (B)

(a)

(A) (B)

(b)

Figure 6: (a) Comparison of complexes between Benzo(a)pyrene and RNA. (A) Complex with a stem loop of human histone mRNA (PDB 1KK);(B)complexwithstemloopIofS. cerevisiae stem loop I of U2 snRNA (PDB 2O33). Each nucleotide is identified by a different color: A 󸀠 = yellow; C = light blue; G = green; U = red. (b) Comparison of complexes between Benzo(a)pyrene and RNA. (A) Complex with 3 hairpin 󸀠 of human U65 snRNAs (PDB 2EUY); (B) complex with 3 stem loop of human U4 snRNA. Each nucleotide is identified by a different color: A=yellow;C=lightblue;G=green;U=red.

estimation of interaction. Table 1 summarizes these results. not allow estimating the binding energy but only the atomic The 3D structures were retrieved from Nucleic Acid Database contact energy and the number of atomic contacts. (http://ndbserver.rutgers.edu/)andthechemicalcompound Table 1 summarizes demonstrative test about the effects for Pubchem (https://pubchem.ncbi.nlm.nih.gov/). We have of chemical contaminants on small nuclear noncoding RNAs. usedthesamedockingsystemPatchDock[44]. The table It is important to underline some interesting considerations: underlines the type of RNA for each structure. We have used first of all the complex between two small nuclear ncRNAs a similar approach using solved 3D structure of snRNA and (U6 and U2) shows the minimal value of estimated contact some synthetic oligomers rather than embedding nucleotide energy if compared with the other structures. The second motifs mimicking the natural functional ones such as C box interesting finding is the variation of estimated parameter in 󸀠 or H box characterizing the snRNAs. Table 1 summarizes the the 3 hairpin loop of two different H/ACA snoRNAs and a set of ncRNA structures, the site of interaction, in spite of similarbehaviourwaspointedoutinthetwostemloopsofU2 the protein-nucleic acid interaction; this kind of analysis does (stem loop I) and U6 snRNA. Both these local conformations BioMed Research International 9

󸀠 are located near to 5 terminal. The last interesting result is [6] F. F. Costa, “Non-coding RNAs: meet thy masters,” BioEssays, the clearly different estimated ACE and contact number in vol. 32, no. 7, pp. 599–608, 2010. the histone mRNA hairpin. Figures 6(a) and 6(b) illustrate [7]M.G.Dozmorov,C.B.Giles,K.A.Koelsch,andJ.D.Wren, how a chemical contaminant could influence the snRNA “Systematic classification of non-coding RNAs by epigenomic functionality by interaction with its conformation. similarity,” BMC Bioinformatics,vol.14,supplement14,article S2, 2013. [8]S.Gottesman,C.A.McCullen,M.Guillieretal.,“SmallRNA 7. Conclusions regulators and the bacterial response to stress,” Cold Spring Harbor Symposia on Quantitative Biology,vol.71,pp.1–11,2006. Even if the knowledge about epigenetic mechanisms is [9] J. R. Phillips, T. Dalmay, and D. Bartels, “The role of small RNAs well established, the new high throughput techniques have in abiotic stress,” FEBS Letters, vol. 581, no. 19, pp. 3592–3597, disclosed new research perspective. The complexity of RNA 2007. world does not yet permit unrevealing the intricate network [10] D. Managadze, I. B. Rogozin, D. Chernikova, S. A. Shabalina, of feedbacks among different kinds of ncRNAs. The nuclear and E. V. Koonin, “Negative correlation between expression transcriptional control by nCRNA is pivotal not only for level and evolutionary rate of long intergenic noncoding RNAs,” protein expression but also for the production of other Genome Biology and Evolution,vol.3,no.1,pp.1390–1404,2011. ncRNAs. The epigenetic damage, induced by chemical and [11] W. F. Doolittle, T. D. P. Brunet, S. Linquist, and T. R. Gregory, physical agents, on DNA is well established, in spite of the “Distinguishing between ‘function’ and ‘effect’ in genome biol- fact that their effects on the different processes in which ogy,” Genome Biology and Evolution,vol.6,pp.1234–1237,2014. ncRNAsareinvolvedarenotyetcompletelyinvestigated. [12] T. A. Elliott, S. Linquist, and T. R. Gregory, “Conceptual and Indeed, the experimental studies are principally focused on empirical challenges of ascribing functions to transposable the fold change of ncRNAs but precise details of ncRNAs elements,” The American Naturalist,vol.184,no.1,pp.14–24, related to structural damage by chemicals are unknown. 2014. In this paper we have attempted to demonstrate how a [13] S.Djebali,C.A.Davis,A.Merkeletal.,“Landscapeoftranscrip- tion in human cells,” Nature,vol.489,pp.101–108,2012. chemical mutagen can trigger damage not only at DNA but also at RNA level. Despite the lack of enough structural [14] H. Tani and M. Torimura, “Development of cytotoxicity- sensitive human cells using overexpression of long non-coding information about snoRNAs and the limited set of tested RNAs,” Journal of Bioscience and Bioengineering,vol.119,no.5, compounds, it is interesting to underline how a chemical pp.604–608,2015. mutagen can interfere with the conformation of specific [15] P.-J. Volders, K. Helsens, X. Wang et al., “LNCipedia: a database regulatory domains in ncRNAs. It is interesting to note how for annotated human IncRNA transcript sequences and struc- the chemical entity could induce a slight change in the tures,” Nucleic Acids Research,vol.41,no.1,pp.D246–D251, conformation that could be critical for molecular recognition 2013. mainly in the process of RNA editing. Our demonstrative [16]J.L.RinnandH.Y.Chang,“Genomeregulationbylong analysis has also pointed out the chemical sensitivity of a noncoding RNAs,” Annual Review of Biochemistry,vol.81,no. hairpin placed into a histone mRNA. It was not possible, for 1,pp.145–166,2012. lack of structural information, to analyze the effect of snRNA [17] T. R. Cech and J. A. Steitz, “The noncoding RNA revolution— directly involved in transcriptional process such as promoter- trashing old rules to forge new ones,” Cell,vol.157,no.1,pp. associated, Piwi-interacting, and long noncoding RNAs. We 77–94, 2014. guess that environmental contaminants have the potentiality [18] A. G. Matera, R. M. Terns, and M. P.Terns, “Non-coding RNAs: to modify both levels of genetic information processes. lessons from the small nuclear and small nucleolar RNAs,” Nature Reviews Molecular Cell Biology,vol.8,no.3,pp.209– 220, 2007. Conflict of Interests [19] B. Seraphin, “Sm and Sm-like proteins belong to a large family: identification of proteins of the U6 as well as the U1, U2, U4 and The authors declare that there is no conflict of interests U5 snRNPs,” The EMBO Journal,vol.14,no.9,pp.2089–2098, regarding the publication of this paper. 1995. [20] C. van der Feltz, K. Anthony, A. Brilot, and D. A. Pomeranz References Krummel, “Architecture of the spliceosome,” Biochemistry,vol. 51, no. 16, pp. 3321–3333, 2012. [1] G. Felsenfeld, “A brief history of epigenetics,” Cold Spring [21] T. Ideue, S. Adachi, T. Naganuma, A. Tanigawa, T. Natsume, Harbor Perspectives in Biology,2014. and T. Hirose, “U7 small nuclear ribonucleoprotein represses [2] A. Bird, “Introduction perceptions of epigenetics,” Nature,vol. histone gene transcription in cell cycle-arrested cells,” Proceed- 447, no. 7143, pp. 396–398, 2007. ings of the National Academy of Sciences of the United States of America,vol.109,no.15,pp.5693–5698,2012. [3] F. J. Carmona and M. Esteller, “Human cancer epigenetics,” in Environmental Epigenomics and Diseases,R.L.JirltleandF.L. [22] T. Sasaki, A. Shiohama, S. Minoshima, and N. Shimizu, “Identi- Tyson, Eds., Springer, Berlin, Germany, 2013. fication of eight members of the Argonaute family in the human genome,” Genomics,vol.82,no.3,pp.323–330,2003. [4] C. Lu and C. B. Thompson, “Metabolic regulation of epigenet- [23] D. Qiao, A.-M. Zeeman, W. Deng, L. H. J. Looijenga, and H. ics,” Cell Metabolism,vol.16,no.1,pp.9–17,2012. Lin, “Molecular characterization of hiwi, a human member [5] S. Katada, A. Imhof, and P. Sassone-Corsi, “Epigentics and of the piwi gene family whose overexpression is correlated to metabolism,” Cell,vol.148,pp.20–28,2012. seminomas,” Oncogene, vol. 21, no. 25, pp. 3988–3999, 2002. 10 BioMed Research International

[24] K. Sato and M. C. Siomi, “Piwi-interacting RNAs: Biological [42] A. Matsui, A. H. Nguyen, K. Nakaminami, and M. Seki, functions and biogenesis,” Essays in Biochemistry,vol.54,no. “Arabidopsis non-coding RNA regulation in abiotic stress 1,pp.39–52,2013. responses,” International Journal of Molecular Sciences,vol.14, [25] B. Wiedenheft, E. van Duijn, J. Bultema et al., “RNA-guided no. 11, pp. 22642–22654, 2013. complex from a bacterial immune system enhances target [43] M. Ligorio, A. Izzotti, A. Pulliero, and P. Arrigo, “Mutagens recognition through seed sequence interactions,” Proceedings of interfere with microRNA maturation by inhibiting DICER. An the National Academy of Sciences of the United States of America, in silico biology analysis,” Mutation Research—Fundamental vol. 108, no. 25, pp. 10092–10097, 2011. and Molecular Mechanisms of Mutagenesis,vol.717,no.1-2,pp. [26] T. Watanabe, S.-I. Tomizawa, K. Mitsuya et al., “Role for piRNAs 116–128, 2011. and noncoding RNA in de novo DNA methylation of the [44]D.Schneidman-Duhovny,Y.Inbar,R.Nussinov,andH.J. imprinted mouse Rasgrf1 locus,” Science,vol.332,no.6031,pp. Wolfson, “PatchDock and SymmDock: servers for rigid and 848–852, 2011. symmetric docking,” Nucleic Acids Research,vol.33,no.2,pp. [27] M. P. Bagijn, L. D. Goldstein, A. Sapetschnig et al., “Function, W363–W367, 2005. targets, and evolution of Caenorhabditis elegans piRNAs,” Science,vol.337,no.6094,pp.574–578,2012. [28] A. Le Thomas, A. K. Rogers, A. Webster et al., “Piwi induces piRNA-guided transcriptional silencing and establishment of a repressive chromatin state,” Genes & Development,vol.27,no.4, pp. 390–399, 2013. [29] L. Ji and X. Chen, “Regulation of small RNA stability: methylation and beyond,” Cell Research,vol.22,no.4,pp.624–636, 2012. 󸀠 [30] H. M. Kurth and K. Mochizuki, “2 -O-methylation stabilizes Piwi-associated small RNAs and ensures DNA elimination in Tetrahymena,” RNA, vol. 15, no. 4, pp. 675–685, 2009. [31] M. C. Chio, C. Zhou, J. S. Brunzelle, and R. H. Huang, 󸀠 “Structural and biochemical insights into 2 -O-methylation at 󸀠 the 3 -terminal nucleotide of RNA by Hen1,” Proceedings of the National Academy of Sciences of the United States of America, vol.106,no.42,pp.17699–17704,2009. [32] V. V. Vagin, A. Sigova, C. Li, H. Seitz, V. Gvozdev, and P. D. Zamore, “A distinct small RNA pathway silences selfish genetic elements in the germline,” Science,vol.313,no.5785,pp.320– 324, 2006. [33] G. E. Morris, “The Cajal body,” Biochimica et Biophysica Acta— Molecular Cell Research, vol. 1783, no. 11, pp. 2108–2115, 2008. [34]T.Kiss,E.Fayet,B.E.Jady,´ P. Richard, and M. Weber, “Biogenesis and intranuclear trafficking of human box C/D and H/ACA RNPs,” Cold Spring Harbor Symposia on Quantitative Biology,vol.71,pp.407–417,2006. [35]K.C.WangandH.Y.Chang,“Molecularmechanismsoflong noncoding RNAs,” Molecular Cell,vol.43,no.6,pp.904–914, 2011. [36]J.H.Yoon,J.Kim,andM.Gorospe,“LongnoncodingRNA turnover,” Biochimie,2015. [37] P. P. Amaral, M. E. Dinger, T. R. Mercer, and J. S. Mattick, “The eukaryotic genome as an RNA machine,” Science,vol.319,no. 5871, pp. 1787–1789, 2008. [38]I.V.Novikova,S.P.Hennelly,andK.Y.Sanbonmatsu,“Sizing up long non-coding RNAs: do lncRNAs have secondary and tertiary structure?” BioArchitecture,vol.2,no.6,pp.189–199, 2014. [39] A. Saxena and P. Carninci, “Long non-coding RNA modifies chromatin: epigenetic silencing by long non-coding RNAs,” BioEssays, vol. 33, no. 11, pp. 830–839, 2011. [40]M.Huarte,M.Guttman,D.Feldseretal.,“Alargeintergenic noncoding RNA induced by p53 mediates global gene repression in the p53 response,” Cell,vol.142,no.3,pp.409–419,2010. [41]H.Tani,Y.Onuma,Y.Ito,M.Torimura,andA.Asakura, “Long non-coding RNAs as surrogate indicators for chemical stress responses in human-induced pluripotent stem cells,” PLoS ONE,vol.9,no.8,ArticleIDe106282,2014. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 234675, 7 pages http://dx.doi.org/10.1155/2015/234675

Research Article Association between Genetic Polymorphisms of DNA Repair Genes and Chromosomal Damage for 1,3-Butadiene-Exposed Workers in a Matched Study in China

Menglong Xiang,1 Lei Sun,2 Xiaomei Dong,2 Huan Yang,2 Wen-bin Liu,2 Niya Zhou,2 Xue Han,2 Ziyuan Zhou,1 Zhihong Cui,2 Jing-yi Liu,2 Jia Cao,2 and Lin Ao2

1 DepartmentofEnvironmentalHygiene,CollegeofPreventiveMedicine,ThirdMilitaryMedicalUniversity,Chongqing400038,China 2InstituteofToxicology,CollegeofPreventiveMedicine,ThirdMilitaryMedicalUniversity,Chongqing400038,China

Correspondence should be addressed to Lin Ao; [email protected]

Received 1 December 2014; Accepted 2 February 2015

Academic Editor: Alessandra Pulliero

Copyright © 2015 Menglong Xiang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The aim of the study was to examine the association between polymorphisms of DNA repair genes and chromosomal damageof 1,3-butadiene- (BD-) exposed workers. The study was conducted in 45 pairs of occupationally exposed workers in a BD product workshop and matched control workers in an administrative office and a circulatory water workshop in China. Newly developed biomarkers (micronuclei, MNi; nucleoplasmic bridges, NPBs; nuclear buds, NBUDs) in the cytokinesis-blocked micronucleus (CBMN) cytome assay were adopted to detect chromosomal damage. PCR and PCR-restriction fragment length polymorphism (RFLP) are adopted to analyze polymorphisms of DNA repair genes, such as X-ray repair cross-complementing Group 1 (XRCC1), O6-methylguanine-DNA methyltransferase (MGMT), poly (adenosine diphosphate-ribose) polymerases (ADPRT), and apurinic/apyrimidinic endonucleases (APE1). The BD-exposed workers exhibited increased frequencies of MNi and NPBs when compared to subjects in the control group. The results also show that the BD-exposed workers carrying XRCC1 diplotypes TCGA- CCGG (4.25 ± 2.06‰) (FR = 2.10, 95% CI: 1.03–4.28) and TCGG-TCGA (5.80 ± 3.56‰) (FR = 2.75, 95% CI: 0.76–2.65) had statistically higher NBUD frequencies than those who carried diplotype TCGG-TCGG (1.89 ± 1.27‰). Our study suggests that polymorphisms of XRCC1 gene may influence chromosomal damage in BD-exposed workers.

1. Introduction However, the results of human molecular epidemiological studies on BD have been mixed. In terms of common 1,3-Butadiene (BD), a Group 1 carcinogen as classified by genotoxic endpoints, only a few studies have yielded positive IARC in 2008 [1], is widely used as an industrial chemical results. For example, one group studied the population in and is also present in autoemission and tobacco smoke Texas in the US and reported significantly elevated frequen- [2]. The carcinogenicity of BD toward rodent animals was cies of hypoxanthine-guanine phosphoribosyltransferase realized early [3]. Meanwhile, a series of epidemiological studies concerning North American BD-exposed workers (HPRT) gene mutations in the peripheral blood lymphocytes found associations with leukemia [4]. Hence, there is a (PBLs) of BD-exposed workers [5],whileothersfailedtofind critical need to identify the early events and factors that increases in gene mutations [6–9]. For chromosome-level are a potential for predicting health effects of BD exposure. damage, a study in China [10] indicated positive induction Since the major metabolites of BD have been proved to be of micronucleus (MN) in the PBLs of heavily exposed mutagenic carcinogens [3], the research on the mutagenicity BD workers, but similar cytogenetic effects have not been of BD provided by molecular epidemiological studies may indicated by several studies conducted on BD-exposed offer useful insights. workers in Italy and the Czech Republic [11, 12]. Additional 2 BioMed Research International reports describe studies in Texasrelating metabolic genotypes of XRCC1 gene and associated DNA repair genes are worthy to HPRT mutation frequencies (MF), and some positive of further research to clarify their roles in BD-related geno- 6 associations were found for mEH genotypes/phenotypes toxicity. The DNA-repair enzyme O -methylguanine-DNA and HPRT MF [13, 14]. Significantly, further analysis on methyltransferase (MGMT) is a key factor in the resistance theoriginaldataofCzechworkershasshownnoBD to alkylating agents. The MGMT protein can rapidly reverse 6 effect revealed in the increased chromosomal aberrations alkylation at the O position of guanine, thereby averting the among workers lacking the glutathione S-transferases T1 formation of lethal cross-links [20]. (GSTT1) gene compared to BD workers with the gene [11]. Our group has conducted two studies concerning BD- These results indicate the possibility of a different genetic exposed workers in 2002 and 2009, respectively. The 2002 background such as single nucleotide polymorphism (SNPs) study found that BD exposure did not statistically increase that may play a critical role in BD’s genotoxicity. HPRT genes [6], while the following study conducted in In recent years, BD epidemiological research has focused 2009 showed significant chromosomal damage and positive on clarifying the relationship between gene polymorphisms associations with some metabolic genotypes in BD-exposed and the risk of mutagenicity and carcinogenicity. Two workers [17]. The aims of the present study were to determine groups of environmental interactive genes: metabolic enzyme whether the DNA repair genes (XRCC1, MGMT, APE1,and genes and DNA repair genes are the mostly studied ones. ADPRT) can modify the genetic instability induced by BD Metabolism is a focal point when evaluating the genotoxicity exposure in current BD workers. of BD in humans, because the epoxide metabolites of BD conduct the most DNA damage by bioalkylating DNA and 2. Materials and Methods forming adducts. Thus, the polymorphisms of metabolic genes (CYP2E1, GSTs, and mEH) involved in BD metabolism 2.1. Study Population. As described earlier [17], we conducted were included in several studies to understand their rela- a1:1matchedpairstudyatapetrochemicalproductcompany tionship with BD genotoxicity. Epidemiological studies on in the Nanjing area of China. Forty-five BD-exposed workers BD-exposed workers indicated that many polymorphic loci were paired with an appropriate control from the same of metabolic genes can impact the chromosomal damage plant and were engaged for the present study and matched induced by BD exposure [13–16]. In 2009, using cytokinesis- by gender, smoking habits, and close age (±3years).The blocked micronucleus (CBMN) cytome assay, we found controlsubjectswereselectedfromemployeesworkingin that BD-exposed workers exhibited increased frequencies of the administrative office or circulating water workshop and micronuclei (MNi) and nucleoplasmic bridges (NPBs) when showed no evidence of exposure to known genotoxic agents. compared to subjects in the control group. Further poly- Questionnaires for all the subjects were accompanied by reg- morphism analysis indicated that the BD-exposed workers ular physical examinations at the Yangzi Employee Hospital. carrying CYP2E1 c1c2/c2c2 or mEH intermediate (I)/high (H) Meanwhile, blood samples and urine samples were collected grouphadasignificantlyhigherNPBfrequencythanthose for further study. An informed consent was obtained from carrying CYP2E1 c1c1 or the mEH low (S) group, respectively eachsubjectatthestartofthisstudy. [17]. DNA repair is a universal process occurring in living 2.2. Exposure Measurement. The sampling methods were cells. This process is responsible for the maintenance of the described previously [17]. In brief, two air sampling ways structural integrity of DNA in the face of damage arising were conducted for exposure assessment in the present study, from environmental insults, as well as from the normal namely, personal sampling and stationary sampling. Both metabolic processes. A study on the BD-exposed workers of sampling ways were performed by active air samplers, with Ningbo, China, examined the polymorphic variants in DNA Gilian-LFS3 (Sensidyne, Inc., USA) carried by workers for repair genes, assuming that the ability of DNA repair that is personal sampling and GilAir-5 (Sensidyne, Inc., USA) for different between individuals can modify the genotoxic effect stationary sampling. Nine BD workers and 4 administrative of BD exposure [10]. The results showed that some SNP loci officers (as controls) carried active air samplers at a flow rate of XRCC1 did impact the MNi frequencies of BD-exposed of 50 mL/min for 8 hours during the consecutive 3 sampling workers. XRCC1 is a protein essential to the repair of single days for personal sampling, while 11 locations in the BD strand breaks (SSBs) and base excision repair (BER) pathway production workshop and 2 locations in the workplace of [18]. XRCC1 acts as a scaffold protein and interacts with mul- control group were chosen for stationary sampling, with 3 tiple DNA repair enzymes like poly (adenosine diphosphate- samples being collected at a pumping rate of 200 mL/min ribose) polymerases (ADPRT) and apurinic/apyrimidinic for 15 minutes in 3 consecutive days at each location. After endonucleases (APE1). However, the research conducted on sampling,eachcharcoaltubesamplewassentforquantitative workers employed in tire plants of the Czech Republic did analysis, conducted according to the standard method (GBZ not find any significant association between genetic poly- 160.39–2007). morphism of XRCC1 exon 10 (Arg399Gln)andDNAdamage biomarkers including chromosome aberrations and single 2.3. Urinary Metabolite. Inhaled BD in the human body is strand breaks, where these workers were exposed to a variety metabolized through cytochrome P450-catalyzed oxidation of xenobiotics, the most prominent being BD and soot con- processes to highly reactive epoxides. The epoxides can taining polycyclic aromatic hydrocarbons (PAHs) [19]. Thus, behydrolyzedandconjugatedwithglutathione,leading these inconsistent results indicated that the polymorphisms to mercapturic acids which are excreted in urine. One of BioMed Research International 3

Table 1: PCR primers and restricted endonucleases for each of DNA repair genes in genotyping process.

Primers PCR Restricted Gene Forward Reverse method endonucleases 󸀠 󸀠 󸀠 󸀠 XRCC1 194 5 -GCCCCGTCCCAGGTA-3 5 -AGCCCCAAGACCCTTTCACT-3 RFLP MspI 󸀠 󸀠 󸀠 󸀠 XRCC1 280 5 -TGGGGCCTGGATTGCTGGGTCTG-3 5 -CAGCACCACTACCACACCCTGAAGG-3 RFLP RsaI 󸀠 󸀠 󸀠 󸀠 XRCC1 399 5 -TTGTGCTTTCTCTGTGTCCA-3 5 -TCCTCCAGCCTTTTCTGATA-3 RFLP MspI 󸀠 󸀠 󸀠 󸀠 XRCC1 -77 5 -GAGGAAACGCTCGTTGCTAAG-3 5 -TCCTCATTAATTCCC TCACGTC-3 RFLP BsrBI 󸀠 󸀠 󸀠 󸀠 ADPRT 762 5 -TTTTGCTCCTCCAGGCCAAcG-3 5 -CCTGACCCTGTTACCTTAATGTCAGTTTT-3 RFLP Hinf1 󸀠 󸀠 󸀠 󸀠 MGMT 84 5 -AAGAGTTCCCCGTGCCGAC-3 5 -GCCAAACGCTGCCTCTGT-3 RFLP HinfI 󸀠 󸀠 󸀠 󸀠 APE1 148 5 -CTGTTTCATTTCTATAGGCTA-3 5 -AGGAACTTGCGAAA GGCTTC-3 RFLP Bfa1

∘ the major metabolites excreted in urine is N-acetyl-S-(3,4- each locus was as follows: XRCC1 Arg194Trp (58 C),XRCC1 ∘ ∘ dihydroxybutyl)-L-cysteine (DHBMA or M1). DHBMA was Arg280His (69.5 C), XRCC1 Arg399Gln (56 C), XRCC1 T-77C ∘ ∘ ∘ chosen as an internal exposure biomarker in the present (58 C), ADPRT Val762Ala (58 C), MGMT Leu84Phe (58 C), ∘ study. A liquid chromatography tandem mass spectrometry and APE1 Asp148Glu (53 C). PCR products were digested (LC-MS/MS) was adopted to identify the urinary concen- with specific restriction enzymes that were recognized and trations of DHBMA in both groups. Urine samples were cuteitheratthewild-typeorvariantsequencesite.Primers collected after the work shift. After excluding unqualified andrestrictedendonucleaseswereshowninTable1.The urine samples according to the standard sampling and storage genotype results were regularly confirmed via direct DNA procedures (GBZ159-2004), 23 pairs of subjects were selected sequencing of the amplified fragments. within the 45 pairs of included subjects to identify the concentrations of DHBMA. Briefly, urinary samples were thawed 2.6. Statistical Methods. Normality tests showed that both − ∘ from 80 C to room temperature initially. Then, analytes exposure measurements and chromosomal damage data were extracted from urine using solid phase extraction with + did not distribute normally; so Wilcoxon’s rank sum test a SAX column (Isolute ENV )andquantifiedbyLC-MS/MS and Wilcoxon’s signed rank test were applied to assess the analysis performed with an ion trap spectrometer. Ionization differences between these two groups. Haplotype analysis was of the analytes was obtained by electrospray in negative performed by PHASE 2.1 software. Poisson regression models mode and acquisition was performed in multiple-reaction as described by Wang et al. [10]wereproducedtoquantify 250.1 → monitoring mode, following the reaction m/z: the relationship of chromosomal damage and the genotypes 121 257.1 → 128.1 𝛽 and for DHMBA and DHBMA-D7. or diplotypes, estimated by the frequency ratio (FR) (FR =𝑒 , DHBMA and DHBMA-D7 standard materials were obtained 𝑒 = 2.71828, 𝛽: regression coefficient) with 95% confidence from Toronto Research Chemicals (TRC), Ontario, Canada. 𝜇 intervals. FR was adjusted for age, sex, smoking status, and The detection limit for the test substance was 10 g/L. alcoholdrinkinginamultivariatePoissonregressionanalysis. For categorical variables, the FR indicated a proportional 2.4. CB-MN Assay. The CBMN assay was performed accord- increase/decrease of the MN/NPB/NBUD frequency in a ing to standard methods described by Fenech [21]. This comparison group relative to the reference. Statistical anal- methodologywaspublishedpreviously[17]. In the present yses were performed using SAS 9.0 (SAS Institute Inc., USA). study, 0.5 mL of fresh blood was used to set up cultures for measuring. One thousand binucleated lymphocytes per subjectwerescoredblindlybyasingleinvestigatorforthe 3. Results presence of MNi, NPBs, and NBUDs. The MNi, NPBs, and 3.1. Baseline Information. Thematchupprocessresultedin NBUDs frequencies were the number of MNi, NPBs, and 45 pairs of subjects. As described earlier [17], we found that NBUDs observed per 1000 lymphocytes, expressed as a count the pairs were well matched for baseline information, such as per thousand (‰). The numbers of mono-, bi-, tri-, and gender (34 males and 11 females), age, and smoking habits (26 tetranucleated cell in 500 lymphocytes were also scored for ex- or present smokers and 19 nonsmokers), with a mean age NDI calculation. of 40.6 in both the BD-exposed group and the control group. 2.5. DNA Extraction and Genotyping. Genomic DNA was AllsubjectsareHanraceChinese. directly extracted from EDTA-anticoagulated whole blood using a wizard genomic DNA purification kit (Promega 3.2. Exposure Measurement. Environmental exposure data Corp., Madison, WI, USA) according to the manufacturer’s have been published [17]. Briefly, for personal sampling, each instructions. PCR-RFLP was the main genotyping method measurement was recorded as the 8 h time-weighted average employed. PCR-RFLP for XRCC1, ADPRT, MGMT,and APE1 (TWA), and, for the subjects’ workshift, the average BD mea- SNP loci were performed under the following conditions: surement for the exposed group (0.34 ± 0.61 p.p.m. or 0.75 ± ∘ ∘ 3 95 C for 5 min, followed by 30 cycles of 94 Cfor40s, 1.35 mg/m )wassignificantlyhigher(𝑃 < 0.01) than that for ∘ 3 annealing for 20 s and 72 C for 35 s, and a final elongation step the control group (0.04 ± 0.01 p.p.m. or 0.09 ± 0.02 mg/m ). ∘ at 72 Cfor10min;therespectiveannealingtemperaturefor For stationary sampling, the BD production plant had a mean 4 BioMed Research International

Table2:UrinarymetabolitesofBD-exposedworkersandcontrols(23pairs).

Exposure (p.p.b.) Control (p.p.b.) Selected percentile Selected percentile Range Range 25th 50th (Median) 75th 25th 50th (Median) 75th ∗ DHBMA <10∼1080.10 56.10 122.54 199.23 <10∼214.76 <10 10.70 61.09 ∗ Compared between exposure group and matched controls, P < 0.01; p.p.b.: 𝜇g/L.

Table 3: Distribution of genotypes and allele frequencies among 1,3-butadiene- (BD-) exposed workers.

SNPs Genotypes 𝑁 Rate% Frequency TT 30 66.7 T: 0.83 XRCC1 -77 CT 15 33.3 C: 0.27 CC 0 CC 37 82.2 C: 0.91 XRCC1 194 CT 8 17.8 T: 0.09 TT 0 GG 36 80.0 G: 0.88 XRCC1 280 GA 7 15.6 A: 0.12 AA 2 4.4 GG 27 60.0 G: 0.76 XRCC1 399 GA 14 31.1 A: 0.24 AA 4 8.9 TT 15 33.3 T: 0.56 ADPRT 762 CT 20 44.4 C: 0.44 CC 10 22.3 CC 32 71.1 C: 0.84 MGMT 84 CT 12 26.7 T: 0.16 TT 1 2.2 TT 11 24.4 T: 0.54 APE1 148 GT 27 60.0 G: 0.46 GG 7 15.6

3 concentration of 2.27±3.33 p.p.m. or 5.02±7.36 mg/m .Inthe 0.053)higherMNifrequency(9.45 ± 3.21‰) than the male control administration office and the circulatory water plant, workers (7.53 ± 3.87‰) in the BD-exposed group [17]. six measurements showed a low-level mean concentration 3 of 0.84 ± 0.20 p.p.m. or 1.86 ± 0.44 mg/m ,whichwas 3.5. Distribution of Genotypes and Allele Frequencies. The significantly lower𝑃 ( < 0.01) than that for the BD production allele frequencies for each single nucleotide polymorphism plant. site are shown in Table 3.Thegenotypedistributionsateach locus were consistent with the Hardy-Weinberg equilibrium. 3.3. Urinary Metabolite. Using LS-MS/MS methods, we identified urinary metabolites (DHBMA) concentrations of 23 pairs of subjects and found that the DHBMA median concen- 3.6. Polymorphism Analysis of DNA Repair Genes. Amul- trations (as shown in Table 2) of BD-exposed workers were tivariate analysis using a backward stepwise selection of statistically higher than those for the controls. variables did not find that any SNP locus of XRCC1, MGMT, ADPRT, or APE1 could impact chromosomal instability 3.4. Chromosomal Damage. The data on the frequencies of induced by BD exposure (data not shown). MNi, NPBs, NBUDs, and NDI have been published [17]. To elucidate the relevance of XRCC1 variants with chro- Briefly, the numbers of MNi (8.00 ± 3.78‰versus5.62 ± mosomal damage further, haplotypes among the four XRCC1 2.41‰) and NPBs frequency (2.58 ± 2.79‰versus1.13 ± polymorphisms (XRCC1 -77 C/T, Arg194Trp, Arg280His, and 1.34‰) in the BD-exposed workers were significantly higher Arg399Gln) were reconstructed. For BD-exposed workers, (𝑃 < 0.01)andNDI(2.20 ± 0.14 versus 2.35 ± 0.27)was 14 XRCC1 (-77)-(194)-(280)-(399) diplotypes were identified significantly lower𝑃 ( < 0.01) in BD-exposed workers than in the analysis. The diplotype TCGG-TCGG which consists in the control subjects, respectively. When age, gender, and of the wild-type sequence in all loci was selected as the smoking status factors were taken into consideration, the reference. Among these haplotype pairs, the rare diplotypes results showed that female workers had a borderline (𝑃= (<5% frequency) were analyzed as a group. BioMed Research International 5

Table 4: Chromosomal damage between diplotypes of XRCC1 in BD-exposed workers.

Diplotypes 𝑁 (%) MN (‰) NPB (‰) NBUD (‰) NDI TCGG-TCGG 9 (20.0) 8.89 ± 4.40 2.44 ± 2.01 1.89 ± 1.27 2.21 ± 0.17 TCGG-CCGG 7 (15.6) 6.14 ± 3.13 2.00 ± 2.38 3.43 ± 1.98 2.11 ± 0.09 TCGG-TCAG 5 (11.1) 7.00 ± 2.24 3.60 ± 4.04 2.60 ± 1.67 2.22 ± 0.14 ∗ TCGA-CCGG 5 (11.1) 8.50 ± 2.89 4.25 ± 2.87 4.25 ± 2.06 2.07 ± 0.11 ∗ TCGG-TCGA 4 (8.9) 11.60 ± 4.61 2.40 ± 0.55 5.80 ± 3.56 2.25 ± 0.05 Others 15 (33.3) 7.33 ± 3.54 2.20 ± 3.43 2.73 ± 2.19 2.25 ± 0.13 The diplotype is defined as the allele present at positions −77 (C/T), 194 (C/T), 280 (G/A),and399 (G/A),respectively. ∗ As seen in Table 5, P < 0.05.

Table 5: Association between diplotypes of XRCC1 and the frequency of nucleus buds (NBUDs).

a 95% CI 2 Adjusted FR Name 𝛽 𝑋 𝑃 Low Upper (95% CI) Intercept 0.6047 −0.3894 1.5988 1.42 0.2332 Gender (female) 0.3118 −0.2169 0.8405 1.34 0.2478 1.37 (0.81–2.32) Age (⩽40) 0.1287 −0.4292 0.6865 0.20 0.6512 1.14 (0.65–1.99) OL −0.1447 −0.6655 0.3760 0.30 0.5859 0.87 (0.51–1.46) Smoking (no) −0.0894 −0.5808 0.4021 0.13 0.7215 0.91 (0.56–1.49) Drinking (no) −0.3682 −0.8256 0.0891 2.49 0.1145 0.69 (0.44–1.09) TCGG-TCGG ————— 1 TCGG-CCGG 0.5219 −0.1652 1.2091 2.22 0.1366 1.69 (0.85–3.35) TCGG-TCAG 0.2441 −0.4885 0.9767 0.43 0.5137 1.28 (0.61–2.66) ∗ TCGA-CCGG 0.7404 0.0274 1.4534 4.14 0.0418 2.10 (1.03–4.28) ∗ TCGG-TCGA 1.0101 0.3804 1.6399 9.88 0.0017 2.75 (1.46–5.15) Others 0.3530 −0.2696 0.9757 1.23 0.2665 1.42 (0.76–2.65) The diplotype is defined as the allele present at positions −77 (C/T), 194 (C/T), 280 (G/A),and399 (G/A),respectively. Others: grouping of all diplotypes with <5% frequency. aMultiple Poisson regression: FR adjusted by age, gender, smoking, and occupational longevity (OL). ∗ P < 0.05.

When XRCC1 diplotypes were taken as categorical vari- airborne environmental exposure samplings, the levels of ables, after being adjusted by age, gender, occupational chemicalmetabolitesinthehumanbodymayreflectcumu- length, smoking, and drinking status in the multivariate lative exposure status more accurately. Therefore, in a further Poissonregressionmodel,wefoundthattheworkerscarrying study, we tested the internal exposure levels of BD in subjects diplotypes TCGA-CCGG and TCGG-TCGA had statistically as reflected by DHBMA. DHBMA and MHBMA are the higher NBUDs frequencies than those who carried diplotype main urinary metabolites of BD. It is claimed that DHBMA TCGG-TCGG among the BD-exposed workers (Table 4). The assay has the sensitivity to measure average BD exposure FRs associated with various diplotypes in all study subjects of 3 to 4 p.p.m. during the workday [22]. A study in Czech are presented in Table 5. found that urinary DHBMA concentrations had increased in BD-exposed workers (median = 508 𝜇g/L) compared 4. Discussion to controls (median = 355 𝜇g/L). And when BD-exposed workers were divided into >0.7 p.p.m. group (median = There is sufficient evidence of potential carcinogenicity ofBD 2719 𝜇g/L) and <0.7 p.p.m. group (median = 669 𝜇g/L), in humans; thus, relevant occupational health organizations the difference of urinary DHBMA concentrations were more have reduced occupational exposure limits of BD. For exam- significant22 [ ]. In a study concerning butadiene-polymer ple, in 1996, OSHA decided to lower permissible exposure workers in China [23], urinary DHBMA (M1) levels were limit (PEL) for BD in occupational-exposed workers from substantially elevated in BD workers (median = 1.3 g M1/mg 1000p.p.m. to 1p.p.m. The data from two studies conducted creatinine) too. A recent study in China (Ningbo) found by our group indicate that stationary sampling results in that the average DHBMA levels of the BD-exposed workers the present study decreased rapidly compared with that of were up to 617.82 ng/mL (or 𝜇g/L) [24]. In the present 2002 study, and the personal sampling results were also study, urine DHBMA concentrations were also significantly below the OSHA limits. The decrease of the stationary higher in the BD-exposed workers (median = 122.54 𝜇g/L, sampling results in our studies, we believed, was due to range = 1080.10 𝜇g/L).TheDHBMAlevelsofthepresent the occupational protection effort made by the company study were much lower than those of the Ningbo study, administrators. However, in a comparison with short-time which was consistent with differences of airborne BD levels 6 BioMed Research International determined by TWAmeasurement between these two studies haplotypes used as a susceptibility biomarker to monitor 3 3 (0.75 ± 1.35 mg/m versus 2.40 ± 2.93 mg/m ). In conclusion, BD occupational exposure in the future; however, extensive the present study demonstrated that the DHBMA urine research is still needed on a larger sample size of population, metabolite concentrations could be a suitable biomarker for especially outside China. BD exposure. In our previous study, we found that MNi and NPBs In recent years, cytokinesis-block micronucleus cytome frequencies differed significantly between BD-exposed work- (CBMN Cyt) assay has evolved into a comprehensive ers and the controls, but no such response was observed method for measuring chromosome breakage, DNA mis- for NBUDs frequencies [17]. When genetic susceptibility is repair, chromosome loss, nondisjunction, necrosis, apop- taken into account in the present study, the polymorphisms tosis, and cytostasis [21]. Using CBMN cytome assay, our of DNA repair genes (XRCC1)werefoundtobeassociated study shows that BD exposure could induce chromosomal with NBUD frequencies in the BD-exposed workers. NBUD damage (MNi and NPBs) in BD-exposed workers, which is regarded as a biomarker for the DNA repair process was consistent with two other studies in China [10, 16]. and is associated with molecular events, such as nuclear However, several studies have yielded negative results [6– elimination of amplified DNA and/or DNA repair DNA- 9]. Individual differences especially genetic polymorphisms protein complexes [21]. Nuclear budding and micronucle- may contribute such inconsistent results. More and more ation have often been seen as part of a p53-dependent DNA attention has been paid to the polymorphisms of DNA repair repair mechanism for the removal of promiscuous DNA [28]. genes when considering susceptibility factors involved in BD- Dutra et al. reported NBUDs highly induced in cultured induced genotoxicity in occupationally exposed workers. A fibroblasts taken from transgenic mice with a DNA repair recent study found that several polymorphic genes including deficiency and a Trp53 deficiency [29]. The present results XRCC1 were associated with higher MN frequencies among providesomeevidencethattheinductionofNBUDsisrelated BD-exposed workers [10]. Thus, it would seem likely that to DNA repair genes in human epidemiological studies. susceptible XRCC1 might play a critical role in BD-related However, whether NBUDs are a mechanism to eliminate genotoxicity. excess chromosomes in a hypothesised process known as aneuploidy rescue remains unclear [30]. And the mechanism Eleven loci among 4 DNA repair genes were included for in detail for how the XRCC1 gene was involved in this process polymorphism analysis in the present study, but no single still needs more studies. The polymorphisms of the metabolic genotype of DNA repair genes could impact chromosomal genes and DNA repair genes associated with NPBs or NBUDs damage in BD-exposed workers. However, in the further frequencies in BD-exposed workers highlight the importance haplotype analysis, we found that those BD-exposed work- ofthenewlydevelopedbiomarkersfortheCBMNcytome ers carrying diplotype TCGA-CCGG (XRCC1 -77-194-280- assay in occupational exposure biomonitoring. 399)anddiplotypeTCGG-TCGA exhibited more serious In conclusion, our study indicates that occupational chromosomal damage (as reflected by NBUDs, compared to exposure to BD may impact chromosomal instability (MNi wild-type diplotype as a reference). We noticed that both of and NPBs) of those workers. The results of polymorphism these two diplotypes shared one haplotype TCGA.Similarly, analysis and haplotype analysis suggest that the diplotype further haplotype analysis in a population of Ningbo area of of XRCC1 (TCGA-CCGG and TCGG-TCGA)elevatedthe China study found that BD-exposed workers with diplotype NBUD frequencies among BD-exposed workers. However, TCGA-TCGA had a higher MN frequency than the others additional studies are recommended to validate these find- [10].Thusfar,wehaveassumedthatTCGA of XRCC1 can be ings in larger population in the future. regarded as a risky haplotype for BD induced genetic damage. The XRCC1 protein was found to be responsible for the BER pathway. Evidence indicates that the VCM (vinyl chloride Conflict of Interests monomer) derivatives etheno-DNA adducts can be repaired The authors declare that there is no conflict of interests through the BER pathway [25]. Meanwhile, the formation regarding the publication of this paper. of DNA adducts by various BD metabolites is central to initiating the mutagenic process for BD exposure. XRCC1 protein may also play a role in the repair of BD-induced Acknowledgment chromosomal damage through the BER pathway because of similar DNA adducts formation. Besides, Chinese hamster The work was supported by the National Natural Science ovary cell lines (EM9 and EM-C11) with XRCC1 mutant have Foundation of China (nos. 30872138 and 81273105). revealed an unusually high frequency of sister chromatid exchange induced by alkylating agents, such cells reverting subsequently to the transfection of human XRCC1 [26]. Other References studies indicated that XRCC1 haplotypes may be suitable to [1] IARC Working Group on the Evaluation of Carcinogenic Risks study the association of environmental factors and diseases. to Humans, “IARC monographs on the evaluation of car- For example, Leng et al. reported that XRCC1 haplotypes cinogenic risks to humans. Volume 97. 1,3-butadiene, ethylene could impact chromosomal damage in Chinese coke oven oxide and vinyl halides (vinyl fluoride, vinyl chloride and vinyl workers [27]. The present study as well as the study in the bromide),” IARC Monographs on the Evaluation of Carcinogenic Ningbo area of China indicates the possibility of XRCC1 Risks to Humans,vol.97,pp.3–471,2008. BioMed Research International 7

[2]N.L.Morrow,“Theindustrialproductionanduseof1,3- [18]R.D.Wood,M.Mitchell,J.Sgouros,andT.Lindahl,“Human butadiene,” Environmental Health Perspectives,vol.86,pp.7–8, DNA repair genes,” Science, vol. 291, no. 5507, pp. 1284–1289, 1990. 2001. [3]R.J.Albertini,M.L.Carson,C.R.Kirman,andM.L.Gargas, [19] P. Vodicka, R. Kumar, R. Stetina et al., “Markers of individ- “1,3-Butadiene: II. Genotoxicity profile,” vol. 40, supplement 1, ual susceptibility and DNA repair rate in workers exposed pp.12–73,2010. to xenobiotics in a tire plant,” Environmental and Molecular [4] H.Cheng,N.Sathiakumar,J.Graff,R.Matthews,andE.Delzell, Mutagenesis, vol. 44, no. 4, pp. 283–292, 2004. “1,3-Butadiene and leukemia among synthetic rubber industry [20] M. Esteller, J. Garcia-Foncillas, E. Andion et al., “Inactivation workers: exposure-response relationships,” Chemico-Biological of the DNA-repair gene MGMT and the clinical response Interactions,vol.166,no.1–3,pp.15–24,2007. of gliomas to alkylating agents,” The New England Journal of [5]J.B.WardJr.,S.Z.Abdel-Rahman,R.F.Hendersonetal., Medicine,vol.343,no.19,pp.1350–1354,2000. “Assessment of butadiene exposure in synthetic rubber man- [21] M. Fenech, “Cytokinesis-block micronucleus cytome assay,” ufacturing workers in Texas using frequencies of hprt mutant Nature Protocols,vol.2,no.5,pp.1084–1104,2007. lymphocytes as a biomarker,” Chemico-Biological Interactions, [22]N.J.vanSittert,H.J.J.J.Megens,W.P.Watson,andP.J. vol. 135-136, pp. 465–483, 2001. Boogaard, “Biomarkers of exposure to 1,3-butadiene as a basis [6]S.X.Liu,L.Ao,B.Duetal.,“HPRTmutationsinlymphocytes for cancer risk assessment,” Toxicological Sciences,vol.56,no.1, from 1,3-butadiene-exposed workers in China,” Environmental pp.189–202,2000. Health Perspectives,vol.116,no.2,pp.203–208,2008. [23] R. B. Hayes, L. Zhang, J. A. Swenberg et al., “Markers for [7] A.D.Tates,F.J.vanDam,F.A.DeZwartetal.,“Biologicaleffect carcinogenicity among butadiene-polymer workers in China,” monitoring in industrial workers from the Czech Republic Chemico-Biological Interactions, vol. 135-136, pp. 455–464, 2001. exposed to low levels of butadiene,” Toxicology, vol. 113, no. 1– [24] X. Cheng, T. Zhang, J. Zhao et al., “The association between 3, pp. 91–99, 1996. genetic damage in peripheral blood lymphocytes and polymor- [8]R.J.Albertini,R.J.Sram,P.M.Vaceketal.,“Biomarkers´ phisms of three glutathione S-transferases in Chinese workers in Czech workers exposed to 1,3-butadiene: a transitional exposed to 1,3-butadiene,” Mutation Research: Genetic Toxicol- epidemiologic study,” Research Report (Health Effects Institute), ogy and Environmental Mutagenesis,vol.750,no.1-2,pp.139– no. 116, pp. 1–162, 2003. 146, 2013. [9]R.J.Albertini,R.J.Sram,P.M.Vaceketal.,“Biomarkersfor [25] S.-M. Zhu, Z.-L. Xia, A.-H. Wang et al., “Polymorphisms and assessing occupational exposures to 1, 3-butadiene,” Chemico- haplotypes of DNA repair and xenobiotic metabolism genes Biological Interactions, vol. 135-136, pp. 429–453, 2001. and risk of DNA damage in Chinese vinyl chloride monomer [10] Q. Wang, A.-H. Wang, H.-S. Tan et al., “Genetic polymorphisms (VCM)-exposed workers,” Toxicology Letters,vol.178,no.2,pp. of DNA repair genes and chromosomal damage in workers 88–94, 2008. exposed to 1,3-butadiene,” Carcinogenesis,vol.31,no.5,pp.858– [26] M. Z. Zdzienicka, G. P. van der Schans, A. T. Natarajan, L. H. 863, 2010. Thompson, I. Neuteboom, and J. W. I. M. Simons, “A Chinese [11] M. Sorsa, S. Osterman-Golkar, K. Peltonen, S. T. Saarikoski, and hamster ovary cell mutant (EM-C11) with sensitivity to simple R. Sram,ˇ “Assessment of exposure to butadiene in the process alkylating agents and a very high level of sister chromatid industry,” Toxicology,vol.113,no.1–3,pp.77–83,1996. exchanges,” Mutagenesis,vol.7,no.4,pp.265–269,1992. [12] P. Lovreglio, N. Bukvic, S. Fustinoni et al., “Lack of genotoxic [27]S.-G.Leng,J.Cheng,L.-Y.Zhangetal.,“AssociationofXRCC1 effect in workers exposed to very low doses of 1,3-butadiene,” polymorphisms and chromosomal damage levels in coke-oven Archives of Toxicology,vol.80,no.6,pp.378–381,2006. workers,” Chinese Journal of Preventive Medicine,vol.39,no.3, pp.159–163,2005. [13] S. Z. Abdel-Rahman, M. M. Ammenheuser, and J. B. Ward Jr., “Human sensitivity to 1,3-butadiene: role of microsomal [28] H. K. Lindberg, X. Wang, H. Jarventaus,¨ G. C.-M. Falck, epoxide hydrolase polymorphisms,” Carcinogenesis,vol.22,no. H. Norppa, and M. Fenech, “Origin of nuclear buds and 3, pp. 415–423, 2001. micronuclei in normal and folate-deprived human lymphocytes,” Mutation Research,vol.617,no.1-2,pp.33–45,2007. [14] S. Z. Abdel-Rahman, R. A. El-Zein, M. M. Ammenheuser et al., “Variability in human sensitivity to 1,3-butadiene: influence of [29] A. Dutra, E. Pak, S. Wincovitch, K. John, M. C. Poirier, and the allelic variants of the microsomal epoxide hydrolase gene,” O. A. Olivero, “Nuclear bud formation: a novel manifestation Environmental and Molecular Mutagenesis,vol.41,no.2,pp. of zidovudine genotoxicity,” Cytogenetic and Genome Research, 140–146, 2003. vol. 128, no. 1–3, pp. 105–110, 2010. [15] J. K. Wickliffe, M. M. Ammenheuser, J. J. Salazar et al., “Amodel [30]M.Fenech,M.Kirsch-Volders,A.T.Natarajanetal.,“Molec- of sensitivity: 1,3-Butadiene increases mutant frequencies and ular mechanisms of micronucleus, nucleoplasmic bridge and genomic damage in mice lacking a functional microsomal epox- nuclear bud formation in mammalian and human cells,” Muta- ide hydrolase gene,” Environmental and Molecular Mutagenesis, genesis, vol. 26, no. 1, pp. 125–132, 2011. vol.42,no.2,pp.106–110,2003. [16] H. Tan, Q. Wang, A. Wang et al., “Influence of GSTs, CYP2E1 and mEH polymorphisms on 1, 3-butadiene-induced micronucleus frequency in Chinese workers,” Toxicology and Applied Pharmacology,vol.247,no.3,pp.198–203,2010. [17] M. Xiang, L. Ao, H. Yang et al., “Chromosomal damage and polymorphisms of metabolic genes among 1,3-butadiene-exposed workers in a matched study in China,” Mutagenesis,vol. 27,no.4,pp.415–421,2012. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 515798, 11 pages http://dx.doi.org/10.1155/2015/515798

Review Article Applying a Weight-of-Evidence Approach to Evaluate Relevance of Molecular Landscapes in the Exposure-Disease Paradigm

Sherilyn A. Gross1 and Kristen M. Fedak2

1 Cardno ChemRisk, 4840 Pearl East Circle 300 W., Boulder, CO 80304, USA 2Department of Environmental & Radiological Health Sciences, Colorado State University, Fort Collins, CO 80523, USA

Correspondence should be addressed to Sherilyn A. Gross; [email protected]

Received 20 December 2014; Accepted 16 March 2015

Academic Editor: Jia Cao

Copyright © 2015 S. A. Gross and K. M. Fedak. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Information on polymorphisms, mutations, and epigenetic events has become increasingly important in our understanding of molecular mechanisms associated with exposures-disease outcomes. Molecular landscapes can be developed to illustrate the molecular characteristics for environmental carcinogens as well as associated disease outcomes, although comparison of these molecular landscapes can often be difficult to navigate. We developed a method to organize these molecular data that usesa weight-of-evidence approach to rank overlapping molecular events by relative importance for susceptibility to an exposure-disease paradigm. To illustrate the usefulness of this approach, we discuss the example of benzene as an environmental carcinogen and myelodysplastic syndrome (MDS) as a causative disease endpoint. Using this weight-of-evidence method, we found overlapping polymorphisms in the genes for the metabolic enzymes GST and NQO1, both of which may infer risk of benzene-induced MDS. Polymorphisms in the tumor suppressor gene, TP53, and the inflammatory cytokine gene, TNF-𝛼, were also noted, albeit inferring opposing outcomes. The alleles identified in the DNA repair gene RAD51 indicated an increased risk for MDS in MDS patients and low blood cell counts in benzene-exposed workers. We propose the weight-of-evidence approach as a tool to assist in organizing the sea of emerging molecular data in exposure-disease paradigms.

1. Introduction the manifestation of liver cancer [3]. Unfortunately, the aflatoxin biomarker of exposure-evidence of disease paradigm is In an idealistic view of carcinogenesis, the molecular initi- unique and does not represent the typical template for most ation of a disease process can be directly tied to a genetic exposure-disease paradigms. mutation or chromosome event caused by environmental Herein,weaskedthefollowing:ifexposuretoaparticular exposure to a carcinogen. However, a single event is usually environmentalcarcinogenisimplicatedinthecauseofa insufficient to induce cancer and other secondary events such disease but the exposure is not evident at the time of disease as additional gene mutations and/or chromosome changes diagnosis, what additional molecular events (e.g., genetic are usually required [1]. These additional events occur during polymorphisms, genetic mutations, and/or epigenetic events) thelatencyperiodofthediseaseandtheprogressionofthe can be linked to the exposure-disease paradigm? Further, disease is a function of individual susceptibility and gene- what additional mutational events can be linked to disease environment interactions [2]. progression, as not every and not all exposures lead to Further, in an idealistic view of a cause and effect disease outcomes? We propose that a “weight-of-evidence” relationship between an environmental carcinogen and dis- (WoE) approach can be applied to compile evidence from ease endpoints, exposure to the environmental carcinogen multiplesourcesinthepublishedliteraturetocreatea wouldremainevidentatthetimethatthediseasemanifests. molecular landscape for the environmental carcinogen and For example, a causal link has been established between for the disease endpoint in question. Herein, we describe environmental exposure to aflatoxin, DNA modification, and the proposed WoE approach for navigating the molecular 2 BioMed Research International landscapes of the exposure-disease paradigm. We discuss hazards, and set regulatory action levels for exposure or this application in the context of benzene-induced myelodys- contamination. For example, the International Agency for plastic syndrome (MDS), to demonstrate how this approach Research on Cancer (IARC) applies a WoE approach to eval- canbeusedtoidentifyandassignrelevancetomolecular uating carcinogenic risk, which they describe in the preamble events associated with both exposure and outcome, taking to their Monographs on the Evaluation of Carcinogenic Risk to the multistep process of carcinogenesis into consideration. Humans [7, 8]. Similarly, the U.S. Environmental Protection Findingsfromthisapproachmaybeusefulinidentifyinga Agency’s (EPA’s) Integrated Risk Information System (IRIS) biomarker for a specific environmental chemical, identifying program utilizes WoE approaches within their toxicological a molecular endpoint to be used in future molecular epidemi- profiles and health risk assessments, which categorize car- ology studies, providing support for traditional epidemiology cinogenicity potential using a 5-point WoE scale that was in establishing causal inference, and/or identifying molecular establishedin1986[9, 10]. events that are important for those individuals susceptible to Yet, while WoE is not a new concept, scientists have a specific exposure-disease paradigm. faced challenges in the fact that the meaning of “weight-of- evidence” in its practical application is not necessarily clear 2. Background or well defined [11]. Recently, researchers have pushed the methodological discussions of WoE one step further into a While there are multiple types of mutations identified in quality evaluation, thereby pulling theoretical frameworks cancers, it has been suggested that frank carcinogenesis into practical application [12]. For example, the Office of occurs only when cells acquire defects in the following six key Health Assessment and Translation (OHAT) within NIEHS areas of cellular control [2]: recently integrated traditional WoE concepts with systematic review guidelines to develop a complex framework for con- (1) sustainable cellular growth independent of a growth ducting literature-based health assessments [13]. The OHAT signal, approach provides transparent, nonsubjective guidelines and (2) abnormality in proteins that regulate the cell cycle, methodology for comparing and contrasting data that comes (3) loss of the ability to respond to programmed cell from a wide range of study design types (e.g., in vitro and in death, vivo toxicology, various epidemiological designs, and molecular studies), considering broader aspects of quantitative (4) cellular immortalization marked by the retention of study quality [13]. WoE approaches have also been used telomeres, to provide direction for research protocols when causation (5) continuous blood supply through sustained angio- has already been established. For example, Zelenka et al. genesis, [14]proposedaWoEframeworkforselectingthemost (6) loss of adhesion ability resulting in tissue invasion and appropriate biomarkers of exposure to use for biomonitoring metastasis. analysis. The authors presented an example wherein they used the framework to evaluate six biomarkers for benzene expo- The interval between the exposure to a carcinogen and the sures over 1 ppm over an 8-hour time frame in occupational manifestation of disease allows time for these molecular settings. changes to occur. Latency is technically defined as the period Herein, we suggest another application of WoE, which of apparent inactivity between the time of first exposure to a combines traditional methods for establishing causal infer- causativeagentandthetimeforresponse,orthefirstclinical ence with more recent uses in providing direction for manifestation of the disease [4, 5]. The outwardly observable molecular research. While no interpretive WoE method can effects of many carcinogens in humans are typically not seen eliminate the need for some level of expert judgment, WoE until after 15- or 20-year latency periods5 [ , 6]. Latency frameworks such as the one we suggest herein can help periods have also been described as the “time interval reduce subjectivity and increase transparency in the vast body between disease occurrence and detection of disease either of literature, which can lead to greater understanding of a by medical testing or by emergence of symptoms,”a definition particular exposure-disease paradigm than might be possible that suggests that latency can be shortened with improvement without the aid of such a tool. on detection methods [6]. Thus, cancer is a disease of latency Benzeneisaknownenvironmentalcarcinogengiven and in many cases a natural artifact of evolution without any adequate dose and duration of exposure [15]. Crude oil known cause. Molecular landscapes have been used to help normally contains a varying composition of petroleum establish the six key defects that are the hallmark of frank hydrocarbons, including benzene [16]. As such, benzene carcinogenesis; however, understanding the timing and order remains an unavoidable component of gasoline products and of occurrence of these molecular events during the latency is a major product in the petroleum refining industry [17, period is ongoing. 18]. The recent increase in unconventional exploration and WoE refers to the interpretive methods commonly production of oil and gas near residential communities across applied to bodies of literature when conducting hazard and the United States may lead to increased opportunities for risk assessments. These types of approaches have been used benzene contamination of valued air and water resources [19]. for decades by academics, practitioners, and regulatory agen- Benzene is also a known leukemogen but the molecular cies in both formal and informal risk assessment processes events required for the development of benzene-induced to establish understandings of causality, discuss chemical leukemiaoccuroveran8-to15-yearlatencyperiod[20, 21]. BioMed Research International 3

As part of this latency period, however, investigators have theDNAitself.Epigeneticeventscanresultinachangein described a benzene-induced “preleukemic state” occurring the expression of the underlying genetic trait by altering the in benzene-exposed individuals prior to diagnosis of a timing and quantity of expression at key points in time (i.e., leukemic state [22]. We now know that the preleukemia variation in timing of the expression of a functional protein). described in benzene-exposed individuals decades earlier is now considered as MDS [23]. In turn, the leukemias have 3. Methods been classified as myeloid and lymphoid with numerous heterogeneous groups of leukemia subtypes within each We developed a WoE approach as a tool to navigate the classification, all of which have distinct molecular landscapes, molecular landscape of any given exposure-disease para- clinical features, etiologies, prognoses, and therapy [23]. digm. The approach is demonstrated in Figure 1. Recent investigation has revealed that MDS is the most The first tier of the WoE approach involves collecting sensitive dose-dependent carcinogenic endpoint following baseline information on the outcome and exposures of benzene exposure in the occupational setting, with MDS interest. The goal of this tier is to determine if enough occurring in some petroleum workers at peak exposure levels baseline evidence exists to illustrate that a relationship may aslowas3partspermillion(ppm)[22]. MDS represents a exist between an exposure and outcome to warrant more smallsetofheterogeneousclonaldiseasesofthebonemarrow detailed, intensive molecular investigation. Without docu- that have been associated with a variety of exposures beyond mented indications that the outcome of interest may be benzene, including solvents, agricultural chemicals, cigarette associated with the suspected environmental carcinogen, and smoke, chemotherapy, and ionizing radiation, but MDS can that exposure to that agent likely occurred or is occurring also occur spontaneously [23]. in the study population, molecular investigations related to Regarding markers for benzene-induced disease, several those exposures and outcomes are not a useful tool. The biomarkers of exposure to benzene exist, including urinary first step within this tier is to perform a literature search to metabolites of benzene s-phenylmercapturic acid (sPMA), determine whether the disease endpoint in question has been benzene-induced depression of peripheral blood parameters associated with the environmental carcinogen of interest in (e.g., neutrophils and MPV), and adducts to hemoglobin published, peer-reviewed literature. This can include both and albumin [24–28]. While these biomarkers are useful for epidemiological and toxicological studies, though one should monitoring exposure in occupational cohorts or other types generally be cautious of study quality and consider the ofsuspected-exposurescenarios,thebiomarkersaretransient overall strength of the body of literature on this association. and usually disappear sometime shortly after cessation of Simultaneously, one should determine whether the individual exposure [14, 24]. Therefore, benzene represents an envi- or group of individuals in question had a historic exposure ronmentalcarcinogeninwhichthemolecularmechanisms or potential/likely exposure to a suspected toxic agent prior associated with disease progression are not clearly defined to the development of the disease endpoint. At the most or easily elucidated using these transient biomarkers of simplistic level, this can be achieved through establishing exposure. work history summaries or definitions of cohort groups In an effort to examine molecular events associated with based on job categories (using oral recall, social security MDS disease progression, large scale genome-wide associated records, or jobsite employment records). More detailed, studies have been conducted with hundreds of MDS patients specific methods to establish an exposure history should also [29–31]. In turn, genome-wide associated studies have also be employed if data is available, such as exposure monitoring been conducted in benzene-exposed workers [32, 33]. Find- ings from these studies and other published literature can or industrial hygiene records. Additionally, if exposure to provide an opportunity to compare the respective molecular the environmental carcinogen was documented at the time landscapes (e.g., genetic polymorphisms, somatic genetic of alleged exposure, records indicating biological evidence mutations, and epigenetic changes). of exposure (e.g., evidence of metabolism in bodily fluids) In considering the molecular landscapes, it is important add an additional level of certainty. An understanding of to differentiate a clear definition for each molecular piece of available biomarkers of exposure to the exposure agent and the landscape. Polymorphisms are inherent to the genome the residence time of those markers following exposure and were originally used to describe variations in shape cessationisalsousefulatthistime. and form that distinguish normal proteins within a species Once there is an understanding of a reasonably assumed from each other which makes each human genome unique. exposure that occurred in the study population as well as Because polymorphism can cause extreme variation in pro- evidence of an association between the exposure agent and tein function, some polymorphisms can infer genetic suscep- the outcome, one can confidently move into the second tier of tibility to a certain agent or disease outcomes whereas other the WoE approach. This tier involves conducting a literature polymorphisms can infer resistance. A mutation, conversely, search to establish the molecular landscape associated with is a permanent change of the DNA sequence following an the exposure-outcome paradigm of interest. A comprehen- exposure or event. Mutations result from unrepaired damage sive,systematicliteraturesearchshouldbeconductedto to DNA that occurs due to errors in the process of replication identify peer-reviewed published literature related to molec- or from the insertion or deletion of segments of DNA by ular events (e.g., genetic polymorphisms, genetic mutations, exogenous agents. Finally, epigenetic events are those that and epigenetic changes) for both the environmental car- cause modification of the genome without modification of cinogen and the disease endpoint. Identified information on 4 BioMed Research International

Tier: Goal of step Step: Type of information to identify Conclusion: How to move forward

(1a) General disease-outcome association Tier 1: Collect baseline information (1b) Individual/cohort-specific If evidence of association + on the outcome and historic exposure scenario history of exposure exists, exposures of interest move to Tier 2 (1c) Biomarkers of exposure

(2a) Associated with exposure Tier 2: Conduct literature If mutational landscape search to establish is established/supported, molecular landscape (2b) Associated with effect/ move to Tier 3 outcome

(3a) The same mutational piece exists on both exposure and outcome sides

Tier 3: Compare 3 If mutational landscape ( b) Mutational pieces with 4 molecular landscapes counterbalanced functions on pieces overlap, move to Tier . for overlapping exposure versus outcome sides Priority for Tier 4 is exposure/outcome pieces 3a, then 3b, and then 3c. (3c) Mutational piece on exposure side has function related to target organ of outcome, even if a similar/ counterbalanced piece does not exist on outcome side

(1) Activation of deactivation enzymes Tier 4: Rank overlapping (2) Proteins that function specifically in pieces of molecular the target organ of toxicity 3 landscape by value of ( ) Proliferation and inﬂammatory cytokines Molecular pieces 4 functional protein in ( ) DNA repair enzymes of most interest (5) Tumor suppressor proteins disease process (6) Proteins responsible for cell adhesion

Figure 1: Weight-of-evidence approach for navigating the molecular landscape. This figure illustrates the stepwise approach used to sort and weigh evidence related to the molecular landscape of an exposure-disease paradigm, starting with the top, left-hand block and moving sequentially to the bottom, right-hand block. the molecular event and the function or effect of that event is there a single nucleotide polymorphism (SNP) on the can be organized into lists or tables, such as those shown in exposure side that represents the same change in protein Tables 1(a) and 1(b). If no information is identified related function as the SNP from the outcome side (e.g., DNA to either the exposure or the effect/outcome side of the codes for the same amino acid sequence used to build the molecular landscape, one cannot move any further through protein)? Or, on the other hand, are there SNPs that are the WoE framework, as this indicates that more research must counterbalanced across the exposure and outcome sides? be conducted related to identifying and characterizing the Direct overlaps of the same molecular events on both the molecular landscapes. exposure and outcome side represent the highest priority The goal of the third tier is to compare the molecular for further investigation in the fourth tier. Counterbalanced landscapes for both the environmental carcinogen and the or complementary overlaps that result in counterbalancing disease endpoint identified during the second tier. To do this, functions represent a second priority level for the fourth thepiecesofthemolecularlandscapesshouldbecompared tier. Finally, there may be mutational pieces seen on the side by side to identifying overlapping molecular events. exposuresidethatresultinafunctionalchangerelatedto Theseoverlapscanoccurinanumberofways.Forexample, the target organ in which the outcome is seen. Even if BioMed Research International 5

Table 1: (a) MDS molecular landscape. (b) Benzene molecular landscape. (a) Gene Function Reference Somatic mutations CTCF Zinc finger protein [30] FANCL DNA cross-link repair in Fanconi anemia [30] BRCC3 Cohesin [30] MPL Cohesin [30] RAD21 Cohesin complex-sister chromatid separation [30, 31] SMC1A Cohesin complex-sister chromatid separation [30, 31] SMC3 Cohesin complex-sister chromatid separation [30, 31] STAG2 Cohesin complex-sister chromatid separation [30, 31] TET2 DNA hydroxymethylation [30, 31] IDH1/2 DNA methylation [30] DNMT3A DNA methylation [30, 31] SETBP1 Gain of function [31] ASXL1 Histone modification [30, 31] EZH2 Histone modification [30, 31] LAMB4 Lossofexpressionincancerwithmicrosatelliteinstability [30] NF1 Ras pathway [30] RIT1 Ras pathway activation [31] JAK2 Signal transduction [30, 31] N-/K-RAS Signal transduction [30, 31] LUC7L2 Spicing [30] SF3B1 Spliceosome [30, 31] ZRSR2 Spliceosome [30, 31] SRSF2 Spliceosome [30, 31] U2AF1 Spliceosome [30, 31] ETV6 Transcription factor [30, 31] IRF1 Transcription factor [30] RUNX1 Transcription factor [30, 31] CEBPA Transcription factor myeloid differentiation [31] WT1 Transcription factor myeloid differentiation [31] TP53 Transcription factor, tumor suppressor [30, 31] BCOR/L1 Transcription repressor [30, 31] PHF6 Transcription factor [30] ATM Ataxia telangiectasia mutated gene [30] Polymorphisms ATM Recognizing and repairing DNA lesions [42] JAK3 Variants unrelated to MDS [42] KDR Mediates VEGF’s responses to angiogenesis [42] STK11 Variants unrelated to MDS [42] VEGF/VEGFR Controversial findings with cancer risk42 [ ] RAD51 DNA repair [48] XRCC5 DNA repair [48] XRCC6 DNA repair [48] TGF MDS disease progression [49] TNF-𝛼 Increase anemia and thrombocytopenia in MDS [38] GSTP1 Increased risk in MDS [35] GSTT1 Increased risk MDS [50] RAD51 Increased risk MDS [41] 6 BioMed Research International

(a) Continued. Gene Function Reference MDR-1 Multidrug resistant, protective against MDS [36] TNF-𝛼 No effect in MDS [49] NQO1 No effect in MDS [35] TP53 Polymorphism not involved in MDS [51] BCL2L10 Reduced risk MDS [42] (b) Gene Function Reference Somatic mutations caused by benzene DNMT1 Decreased mRNA expression [45] DNMT3A Decreased mRNA expression [45] DNMT3B Decreased mRNA expression [45] MBD2 Decreased mRNA expression [45] PARP1 Decreased mRNA expression [45] p15 Hypermethylation [52] MAGE-1 Hypomethylation [52] Glycophorin A Induction of gene duplication [47] RUNX1 Transcription factor [43] Polymorphism of benzene susceptibility BLM Modulation of DNA repair [29, 39] RAD51 Modulation of DNA repair [29, 39] TP53 Modulation of DNA repair [29, 39] WDR79 Modulation of DNA repair [29, 39] WNR Modulation of DNA repair [22] XRCC1 Modulation of DNA repair [37] VCAM1 Altered adhesion [53] IL-12 Altered function polymorphism [43] MPO Altered function polymorphism [43] NQO1 Altered function polymorphism [52] IL-10 Cytokine activity [32] IL-12A Cytokine activity [32] IL-1a Cytokine activity [32] IL-4 Cytokine activity [32] GSTM1 Detoxification of exogenous compounds [54] VEGF Endothelial cytokine [53] TNF-𝛼 Inflammatory cytokine39 [ ] APEX1 Male restricted DNA repair mechanism [46] p14 p53 dependent modulation [55] p21 p53 dependent modulation [55] MSH2 Repair of mismatched DNA [56] Biomarkers of benzene exposure in blood Urinary sPMA Increases in urine of exposed individuals [27] Hemoglobin adducts 4-month duration in blood [25] Albumin adducts Duration in blood unclear [26, 28] a similar or counterbalanced/complementary piece is not changes associated with the agent of exposure and/or the seen on the outcome side, these molecular pieces could be of disease endpoint are warranted, or that the exposure-disease interest and should be carried into the fourth tier of the WoE association of interest is not supported by the molecular approach.Ifcomparisonofthemolecularlandscapeatthetier landscapes. 3 level reveals no overlaps, one must consider whether this Finally, once the molecular landscapes have been nar- implies that the molecular landscape has not yet been fully rowed down to only the overlapping pieces, the fourth tier developed, in which case more researches into the molecular oftheWoEapproachcanbeimplemented.Thistierinvolves BioMed Research International 7 ranking overlapping pieces of the molecular landscape by Table 2: Molecular landscapes (polymorphisms, mutations, epige- value of the functional protein affected by that molecular netic events) for MDS and benzene ranked by WoE. event in the disease process. When investigating exposure Molecular landscapes and effects, the priority for the functionalities is determined based on the importance of that change to susceptibility to Gene WoE rank carcinogenesis/disease progresses and the specificity that the GSMT1 1 change implies for the specific exposure-disease paradigm NQO1 2 versus a generic change that is seen in all cancers. TNF-𝛼 2 Base on our professional judgment, we suggest an initial RAD51 3 priority level for polymorphisms and mutations as follows (from greatest to least relevant): TP53 4 RUNX1 5 (1) activation and deactivation enzymes of environmen- DNMTs 6 tal carcinogens, MPO 7 (2) proteins that function specifically in the target organ Interleukins 8 of toxicity, GPA 9 (3) proliferation and inflammatory cytokines, (4) DNA repair enzymes, and MDS. In this specific scenario, we chose to rank the (5) tumor suppressor proteins, overlapping molecular landscape on the basis of susceptibility (6) proteins responsible for cell adhesion. to benzene-induced MDS. We started our WoE approach at the second tier, because the first tier step of determining Similarly, we suggest an initial priority level epigenetic event whether evidence exists for an assumed association between as follows (from greatest to least relevant): benzene and MDS has been done by other researchers, and (1) hypo- or hypermethylation (modify timing of DNA this was a hypothetical exercise so there was no individual expression into protein), or cohort for which exposure assumptions needed to be established [33]. A literature search was conducted using (2) hypo- or hyperacetylation (modify DNA expression PubMed, to identify published findings that discussed the into protein), molecular landscape associated with environmental expo- (3) histone modification (open DNA reading frame). sure to benzene and molecular mechanisms associated with benzene-induced toxicity. An independent literature search The result of the fourth tier, as well as the overall result of was performed to identify the molecular landscape associated the WoE approach, is a list of related and relevant molecular with the disease progression to MDS. The PubMed searches events associated with the exposure and the outcome side of a were conducted for illustration purposes and were not meant given exposure-disease paradigm that can be used to indicate to be a comprehensive search but rather to provide enough a more likely association has occurred between that exposure collective information to be useful for demonstrating how and the outcome. These molecular events warrant the highest the WoE approach we developed can be applied to help bring level of further consideration within the sea of information order and relevance to molecular information. relatedtothegivenrelationship. We suggest that common polymorphisms, genetic muta- The known molecular landscapes for MDS and benzene, tions, and epigenetic events be given the highest WoE rank- respectively, as determined by our literature search are dis- ing if the environmental carcinogen in question is directly played in Tables 1(a) and 1(b). or indirectly toxic through DNA adduct formation (e.g., Wherepossible,weusedthepublishedliteratureto through reactive oxygen), and then DNA repair mechanisms determine if the molecular events identified for benzene and should be elevated, as genetic changes in DNA repair genes MDSrepresentedthesametypeofchangeinproteinfunction would be expected in the process of carcinogenesis in any (e.g., enhanced function, inhibition of function, or no change tissue [2]. We anticipate that epigenetic events will become in function). Molecular landscapes (e.g., polymorphisms, more important within the next few years as researchers mutations, epigenetic events) for MDS and benzene ranked develop new methods of analysis to correlate changes in bytheWoEapproachareshowninTable2. methylation state and alterations in the timing of expression of a functional protein (e.g., phenotype). We trust that gaps 4.1. Common Polymorphisms. By overlapping the molecular in our current understanding of this process such as how landscapes for MDS and benzene, we identified common epigenetic events relate to downstream proteins involved in polymorphisms and ranked them by molecular event of the exposure-disease paradigm will be revealed [31]. interest (MEoI) based on the criteria outlined in Meth- ods. Our finding showed overlap in the genes for the 4. Results metabolic enzymes glutathione-S-transferase (GST) (GSTT1, GSTM1, and GSTP1) and NAD(P)H:quinone oxidoreductase ToillustratehowthisWoEapproachworks,weapplied 1 (NQO1). We also found additional polymorphisms in the methods to the exposure-disease paradigm of benzene benzene-exposed workers in cytochrome P4502E1 (CYP2E1) 8 BioMed Research International andmyeloperoxidase(MPO)butnotinMDSpatients. study showed that only the −238 TNF-𝛼 and not the −308A These metabolic enzymes have been examined in assays for TNF-𝛼 polymorphism was significantly associated with the variations in activity associated with benzene toxicity [34]. development of benzene-induced dysplasia and not with an For example, the metabolic enzyme CYP2E1 in the liver increased risk of MDS [39]. Although −308A TNF-𝛼 and is responsible for transformation of benzene into its major −238 TNF-𝛼 showedtheoppositeeffectinthesestudies,TNF- metabolites hydroquinone (HQ) and catechol (CAT). MPO 𝛼 does play a significant role in the target organ of toxicity in bone marrow progenitor cells has been demonstrated to [39]. Therefore, TNF-𝛼 would also rank as 𝑅=2on the WoE further metabolize HQ to the bone marrow toxin, para- scale. benzoquinone (pBQ). Detoxification of benzene metabo- The RAD51 protein plays an important role in DNA lites in the liver is controlled by GST genotypes whereas repair, meiosis, chromosome segregation, and chromosome detoxification of pBQ in the bone marrow is thought to stability, and its dysregulation has been associated with occurthroughtheNQO1enzyme[34]. We found that the multiple diseases [40]. A meta-analysis that was performed MEoI in metabolic enzymes associated with MDS was the on a total of ten studies with MDS patients and controls NQO1 germline polymorphism C609T, which results in a indicated that −135G/C RAD51 protein was associated with lowering of NQO1 enzyme activity, which in turn may result an increased susceptibility to MDS [41]. Similarly, a study in an increase in susceptibility to MDS in benzene-exposed conducted in 250 benzene-exposed workers and controls individuals.However,aNQO1C609Tpolymorphismhas indicated that the −135G/C RAD51 allele was associated with beenshowntohavenoeffectinMDS[35]. In addition, white blood cell (WBC) counts lower than 4000 𝜇L[32]. Since GSTT1 and GSTM1 are genetic polymorphisms of GST in the −135G/C RAD51 polymorphism has been linked to an humans, and a homozygous deletion in these enzymes leads increased susceptibility to MDS and the identical polymor- to a complete absence of enzyme activity [36]. It was noted phism was also associated with changes in WBC counts in in one report that the GSTM1 genotype may contribute benzene-exposed workers, 135G/C RAD51 could rank high towards progression of MDS [36]. In an evaluation of various on the WoE scale. However, RAD51 polymorphisms have polymorphisms in metabolic enzymes in Chinese workers been associated with multiple diseases so this lowers the rank occupationally exposed to benzene, NQO1C609T, GSTT1, to 𝑅=3. and GSTM1 inferred an increased risk of benzene poisoning [37]. Taken together, these findings suggest that GST 4.2. MDS Gene Mutations. The RUNX gene, also known as variant GSTM1 increases the risk for MDS and increase the AML1, codes for an important transcription factor, “core toxicity of benzene by decreasing the ability to detoxify and binding factor” alpha subunit, which is a transcription factor eliminate the active metabolite. Although the polymorphism that regulates commitment to erythroid and granulocytic lin- NQO1C609TshowednoincreasedriskofMDS,NQO1 eages and initiates the terminal differentiation of the myeloid functions in the target organ of toxicity for benzene, the bone lineage [31]. AML1 is commonly fused to RUNTX1T1 (ETO) marrow [34]. Therefore, GSTM1 and NQO1 polymorphisms 𝑅=1 𝑅=2 in a chromosome translocation t(8;21)(q22;q22), which is one rank as and ,respectively,ontheWoEscale. of the most frequent karyotypes in AML [42]. RUNX gene Mutations in the tumor suppressor gene TP53 have been mutationshavebeenidentifiedinsomecasesofMDSand shown to be frequent mutations in human cancers [29]. generally infer unfavorable prognosis [30, 31]. Several polymorphisms in TP53 have been studied in MDS. For example, it has been shown that the TP53Arg72Pro 4.3. Benzene-Induced Epigenetic Events. Interestingly, when polymorphism did not differ between MDS and healthy RUNX methylation status was examined in a cell line treated controls, and this particular polymorphism was not associ- with the benzene metabolite HQ, researchers found that HQ ated with clinical and laboratory parameters, disease progression, or overall survival of MDS patients [29]. This induced hypermethylation in RUNX as well as hypomethy- suggests that TP53 polymorphism is not involved in increased lation in RUNTX1T1 [43]. Taken together, the mutation in risk for MDS. However, in one report on benzene-exposed RUNX1 in some MDS patients and the RUNX1 epigenetic workers, the rs1042522 TP53 polymorphism was associated event shown in in vitro cell culture treated with benzene with decreased granulocytes, decreased CD4 T cells, and metabolites may be important if the RUNX1 finding can decreased B-cells [32]. Because the TP53 polymorphism in be reproduced in workers exposed to benzene, especially benzene-exposed workers did not correlate directly with since this mutation occurs in the target organ of toxicity. We toxicityandshowednoinvolvementinincreasedriskfor propose that the RUNX1 mutation in MDS and epigenetic MDS, TP53 common genetic polymorphism was ranked as changes shown in an in vitro cell culture treated with benzene 𝑅=4on the WoE scale. metabolites rank as 𝑅=5on the WoE scale. Tumor Necrosis Factor alpha (TNF-𝛼)proteinisamajor regulatory cytokine that plays a role in many immune-medi- 4.4. Common Epigenetic Events. DNA methyltransferases ated diseases and hematologic malignancies [38]. In MDS, (DNMTs) play a key role in establishing and maintaining the −308A TNF-𝛼 genetic polymorphism, which increases methylation. DNA methylation is considered to be the initial the transcription level of this inflammatory cytokine, was step in establishing the inactive chromatin state and is associated with MDS patients [38]. Overexpression of −308A critical for maintaining silence (i.e., no gene expression) in TNF-𝛼 protein may also be responsible for promoting a protooncogenes [44]. Reduced DNA methylation has been proinflammatory state in benzene-exposed workers; one correlated with shorter survival times and transformation BioMed Research International 9 from MDS to AML [31]. Among Chinese workers exposed worker susceptibility in an occupational setting. For exam- to a mixture of benzene, toluene, and xylene, those who ple, an individual with the potential for high occupational exhibited a loss of function mutation in DNMT3A also exposure to a specific chemical with a known association showed a downregulation in all DNMTs and there was a with a specific disease outcome could undergo molecular dose-dependent decrease in DNMT3A gene expression [45]. testing to determine whether they are (a) susceptible to initial The epigenetic events demonstrated in MDS patients and in toxicity from exposure to a specific chemical, (b) susceptible Chinese workers exposed to benzene are interesting, but since to initiation of molecular events associated with disease pro- the DNMTs function is nonspecific, the common DNMT gression, or (c) susceptible to a specific disease regardless of gene mutation ranks as 𝑅=6on the WoE scale. exposure. In this scenario, (a) the worker could be monitored for markers of exposure and toxicity to a specific chemical, (b) the worker could be monitored for the development of 4.5. Benzene Polymorphisms. Polymorphisms in the meta- an identified marker indicative of early disease, or (c) the bolic enzyme myeloperoxidase (MPO) within the bone mar- worker could be evaluated for the presence of molecular row, specifically, the balance between MPO and NQO1 and events specific to the manifestation to the disease in question. the benzene metabolites, have been studied. Because both Armed with this type of knowledge, one could take action of these polymorphic enzymes occur in the target organ to prevent or avoid subsequent exposure to the specific for benzene toxicity, they may have a significant effect chemical or gain awareness and practice avoidance to all on susceptibility to benzene toxicity [43]. Interleukin (IL) known environmental causes of the particular disease in cytokines, including IL-10, IL-4, IL-12, and IL-1a, play a question. pivotal role in growth, maturation, and differentiation of blood cells. Therefore, polymorphisms in these regulators Further, in situations of outward observable disease, thatcauseanenhancedproduction,interferewithreceptor manyaretemptedtoaskthefollowingquestion:“Isthis binding, or inhibit cell function could have a profound effect disease due to my past exposures?” With advanced under- on the regulation of hematopoietic system as a whole [46]. standing of the relevance of molecular landscapes and of Given that similar polymorphisms, mutations, or epigenetic overlapping molecular events between an environmental events have not been reported in MDS patients, their role in carcinogen and a specific disease outcome, it may be possible this specific disease process is questionable. Therefore these to answer this question in a more definitive way. Caution polymorphisms rank as 𝑅=7and 𝑅=8,respectively. must be taken, however, when interpreting these molecular events as we do not suggest that the WoE approach to assigning significance to an overlapping molecular landscape 4.6. Benzene Gene Mutations. The glycophorin A (GPA) should be used as a surrogate for standard epidemiologic locus codes for an erythroid lineage specific protein with two methods for determining causation. Rather, we view this allelic forms (M and N). When GPA alleles were examined method as evidentiary support for causal inference. Further in a small group of benzene-exposed workers and control research is needed to understand how molecular landscapes subjects, it was shown that lifetime cumulative exposure to correlate with individual exposures to carcinogens, molecular benzene was associated with the NN variant of GPA but not mechanisms for disease progression, and disease etiology. with the N0 variant. It was suggested that the NN mutation occurredinlonger-livedbonemarrowcellsandthattheNN The true usefulness of this approach is to further our variantresultedfromalossoftheGPAMallele,possibly understanding of molecular epidemiology on a population through benzene-induced duplication of the N allele. The level. As the field of molecular epidemiology progresses N0 variant was presumed to occur through point mutations quickly, biomarkers for exposure, disease progression, and or deletions [47]. Although these results are interesting for disease outcome are becoming more and more prevalent. benzene-exposed individuals, the role this mutation plays in Given the increasing amount of molecular information that MDS remains elusive. Therefore, the benzene-induced gene researchers have access to through the published literature, mutation ranks as 𝑅=9. a system such as the WoE approach illustrated herein is a useful tool for sorting, categorizing, and prioritizing the 5. Discussion most meaningful information. Using such a framework, researchers can determine ways to take advantage of best We developed this WoE process in an effort to understand practices in identifying exposure scenarios and defining the commonalities between the molecular landscape of an biomarker(s) relevant to the exposure-disease paradigm. environmental carcinogen and the molecular landscape of a knownorsuspecteddiseaseoutcome.Weshowedhowthe Disclaimer WoEapproachcouldbeusedtoidentifyandassignrelevance (e.g., rank) to overlapping genetic information associated The opinions and perspectives presented here are those of the with susceptibility in an exposure-disease paradigm. We authors. envision that this approach can be modified to identify and rank the most relevant overlapping epigenetic events in an Conflict of Interests exposure-disease paradigm. On an individual level, the approach might have practical The authors are employed or contracted by Cardno Chem- application for purposes of prevention such as identifying Risk, a consulting firm that provides scientific advice to 10 BioMed Research International

the government, regulatory agencies, corporations, law firms, [15] WHO, IARCMonographsontheEvaluationofCarcinogenic and various scientific and professional organizations. Cardno Risks to Humans: 100-F (Benzene), World Health Organization, ChemRisk has been engaged by entities involved in benzene Geneva, Switzerland, 2012, http://monographs.iarc.fr/ENG/ litigation. Some authors may serve as expert witnesses in Monographs/vol100F/mono100F-24.pdf. litigation regarding the potential health hazards posed by [16]S.G.Osborn,A.Vengosh,N.R.Warner,andR.B.Jackson, benzene in the future. “Methane contamination of drinking water accompanying gas-well drilling and hydraulic fracturing,” Proceedings of the National Academy of Sciences of the United States of America, Acknowledgments vol. 108, no. 20, pp. 8172–8176, 2011. The authors would like to thank Jennifer Sahmel and Whitney [17] B. Holmberg and P. Lundberg, “Benzene: standards, occurrence, and exposure,” American Journal of Industrial Medicine, Christian for their review of this paper. The research support- vol. 7, no. 5-6, pp. 375–383, 1985. ingthispaperandthetimeneedtowritethismanuscriptwere funded solely by Cardno ChemRisk. [18] OSHA, “29 CFR part 1910: occupational exposure to benzene, final rule,” Federal Register,vol.52,no.176,1987. [19] S. A. Gross, H. J. Avens, A. M. Banducci, J. Sahmel, J. M. References Panko,andB.E.Tvermoes,“AnalysisofBTEXgroundwater concentrations from surface spills associated with hydraulic [1] W. Guo, J. L. Lasky, C.-J. Chang et al., “Multi-genetic events fracturing operations,” Journal of the Air and Waste Manage- collaboratively contribute to PTEN-null leukaemia stem-cell ment Association,vol.63,no.4,pp.424–432,2013. formation,” Nature,vol.453,no.7194,pp.529–533,2008. [20] M. N. Lobato, M. Metzler, L. Drynan, A. Forster, R. Pannell, [2] P. Vineis, A. Schatzkin, and J. D. Potter, “Models of carcinogen- and T. H. Rabbitts, “Modeling chromosomal translocations esis: an overview,” Carcinogenesis,vol.31,no.10,pp.1703–1709, using conditional alleles to recapitulate initiating events in 2010. human leukemias,” Journal of the National Cancer Institute— [3]G.N.Wogan,T.W.Kensler,andJ.D.Groopman,“Present Monographs, no. 39, pp. 58–63, 2008. and future directions of translational research on aflatoxin and hepatocellular carcinoma. A review,” Food Additives and [21] S.-N. Yin, M. S. Linet, R. B. Hayes et al., “Cohort study among Contaminants, Part A, Chemistry, Analysis, Control, Exposure workers exposed to benzene in China: I. General methods and and Risk Assessment,vol.29,no.2,pp.249–257,2012. resources,” American Journal of Industrial Medicine,vol.26,no. 3,pp.383–400,1994. [4] The American Heritage Dictionaires, Ed., Stedman’s Medical Dictionary, Houghton Mifflin, Boston, Mass, USA, 2nd edition, [22] A. R. Schnatter, D. C. Glass, G. Tang, R. D. Irons, and L. 2002. Rushton, “Myelodysplastic syndrome and benzene exposure among petroleum workers: an international pooled analysis,” [5] J.L.Kelsey,A.S.Whittemore,A.S.Evans,andW.D.Thompson, Journal of the National Cancer Institute,vol.104,no.22,pp. Methods in Observational Epidemiology, Oxford University 1724–1737, 2012. Press, New York, NY, USA, 2nd edition, 1996. [23] WHO, Classification of Tumours of Haematopoietic and Lym- [6] K. J. Rothman and S. Greenland, Eds., Modern Epidemiology, phoid Tissues, WHO, 4th edition, 2008. Lippincott-Raven Publishers, Philadelphia, Pa, USA, 2nd edition, 1998. [24] A. Robert Schnatter, P. J. Kerzic, Y. Zhou et al., “Peripheral [7] WHO, IARC Mongraphs on the Evaluation of Carcinogenic Risks blood effects in benzene-exposed workers,” Chemico-Biological to Humans, World Health Organization, Geneva, Switzerland, Interactions,vol.184,no.1-2,pp.174–181,2010. 2014, http://monographs.iarc.fr/. [25]W.E.Funk,S.Waidyanatha,S.H.Chaing,andS.M.Rappa- [8] WHO, IARC Monographs on the Evaluation of Carcinogenic port, “Hemoglobin adducts of benzene oxide in neonatal and Risks to Humans: Preamble,2006,http://monographs.iarc.fr/ adult dried blood spots,” Cancer Epidemiology Biomarkers and ENG/Preamble/CurrentPreamble.pdf. Prevention,vol.17,no.8,pp.1896–1901,2008. [9] US EPA, Integrated Risk Information System (IRIS): Frequent [26] S. M. Rappaport, S. Waidyanatha, K. Yeowell-O’Connell et al., Questions,2014,http://www.epa.gov/iris/help ques.htm#tox. “Protein adducts as biomarkers of human benzene metabolism,” [10] US EPA, Guidelines for Carcinogenic Risk Assessment, 1986, Chemico-Biological Interactions,vol.153-154,pp.103–109,2005. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=54933# [27]P.B.Farmer,B.Kaur,J.Roachetal.,“TheuseofS- Download. phenylmercapturic acid as a biomarker in molecular epidemi- [11] D. L. Weed, “Weight of evidence: a review of concept and ology studies of benzene,” Chemico-Biological Interactions,vol. methods,” Risk Analysis,vol.25,no.6,pp.1545–1557,2005. 153-154, pp. 97–102, 2005. [12] L. R. Rhomberg, J. E. Goodman, L. A. Bailey et al., “A survey of [28] Y. S. Lin, R. Vermeulen, C. H. Tsai et al., “Albumin adducts frameworks for best practices in weight-of-evidence analyses,” of electrophilic benzene metabolites in benzene-exposed and Critical Reviews in Toxicology,vol.43,no.9,pp.753–784,2013. control workers,” EnvironmentalHealthPerspectives,vol.115,no. [13]A.A.Rooney,A.L.Boyles,M.S.Wolfe,J.R.Bucher,and 1,pp.28–34,2007. K. A. Thayer, “Systematic review and evidence integration [29] C. A. Larsson, G. Cote, and A. Quintas-Cardama,´ “The for literature-based environmental health science assessments,” changing mutational landscape of acute myeloid leukemia and Environmental Health Perspectives,vol.122,no.7,pp.711–718, myelodysplastic syndrome,” Molecular Cancer Research, vol. 11, 2014. no. 8, pp. 815–827, 2013. [14] M. P. Zelenka, D. B. Barr, M. J. Nicolich et al., “A weight [30] T. Haferlach, Y. Nagata, V. Grossmann et al., “Landscape of evidence approach for selecting exposure biomarkers for of genetic lesions in 944 patients with myelodysplastic syn- biomonitoring,” Biomarkers,vol.16,no.1,pp.65–73,2011. dromes,” Leukemia,vol.28,no.2,pp.241–247,2014. BioMed Research International 11

[31] R. Itzykson, O. Kosmider, and P. Fenaux, “Somatic mutations glycophorin A locus in exposed humans,” Proceedings of the and epigenetic abnormalities in myelodysplastic syndromes,” National Academy of Sciences of the United States of America, Best Practice and Research: Clinical Haematology,vol.26,no.4, vol.92,no.9,pp.4069–4073,1995. pp. 355–364, 2013. [48] H. L. Ribeiro Jr., R. T. G. de Oliveira, A. R. S. Maia et al., “Poly- [32] Q. Lan, L. Zhang, M. Shen et al., “Large-scale evaluation of morphisms of DNA repair genes are related to the pathogenesis candidate genes identifies associations between DNA repair and of myelodysplastic syndrome,” Hematological Oncology,2014. genomic maintenance and development of benzene hematotox- [49]Z.Gyulai,A.Balog,Z.Borbenyi,´ and Y. Mandi,´ “Genetic icity,” Carcinogenesis,vol.30,no.1,pp.50–58,2009. polymorphisms in patients with myelodysplastic syndrome,” [33] M. T. Smith, “Advances in understanding benzene health effects Acta Microbiologica et Immunologica Hungarica,vol.52,no.3-4, and susceptibility,” Annual Review of Public Health,vol.31,pp. pp. 463–475, 2005. 133–148, 2010. [50] M. Fang, W. Zeng, L. Huang et al., “Association between glu- [34] R. Snyder and C. C. Hedli, “An overview of benzene tathione S-transferase T1 null genotype and risk of myelodys- metabolism,” Environmental Health Perspectives,vol.104,sup- plastic syndromes: a comprehensive meta-analysis,” Tumor plement 6, pp. 1165–1171, 1996. Biology,vol.34,no.5,pp.2747–2752,2013. [35] S. Zachaki, C. Stavropoulou, T. Koromila et al., “High frequency [51] J. A. Machado-Neto, F. Traina, P. de Melo Campos, M. F. 609 of NAD(P)H:quinone oxidoreductase 1 (NQO1)C T germline Andreoli-Risso,F.F.Costa,andS.T.OlallaSaad,“Lack polymorphism in MDS/AML with trisomy 8,” Leukemia of association between MDM2 SNP309 and TP53 Arg72Pro Research,vol.37,no.7,pp.742–746,2013. polymorphisms with clinical outcomes in myelodysplastic syn- [36] B. Palodetto, P. de Melo Campos, B. D. Benites et al., “MDR- drome,” Neoplasma,vol.59,no.5,pp.530–535,2012. 1 and GST polymorphisms are involved in myelodysplasia [52] V. Bollati, A. Baccarelli, L. Hou et al., “Changes in DNA progression,” Leukemia Research,vol.37,no.8,pp.970–973, methylation patterns in subjects exposed to low-dose benzene,” 2013. Cancer Research, vol. 67, no. 3, pp. 876–880, 2007. [37] S. Kim, Q. Lan, S. Waidyanatha et al., “Genetic polymorphisms [53] Q. Lan, L. Zhang, M. Shen et al., “Polymorphisms in cytokine and benzene metabolism in humans exposed to a wide Range and cellular adhesion molecule genes and susceptibility to of air concentrations,” Pharmacogenetics and Genomics,vol.17, hematotoxicity among workers exposed to benzene,” Cancer no. 10, pp. 789–801, 2007. Research, vol. 65, pp. 9574–9581, 2005. [38] C. B. Belli, Y. Bestach, Y. Sieza et al., “The presence of - [54]M.D.C.Silva,J.Gaspar,I.D.Silva,A.Faber,andJ.Rueff, 308A TNF𝛼 is associated with anemia and thrombocytope- “GSTM1, GSTT1, and GSTP1 genotypes and the genotoxicity nia in patients with myelodysplastic syndromes,” Blood Cells, of hydroquinone in human lymphocytes,” Environmental and Molecules, and Diseases,vol.47,no.4,pp.255–258,2011. Molecular Mutagenesis,vol.43,no.4,pp.258–264,2004. [39] L. Lv, P. Kerzic, G. Lin et al., “The TNF-𝛼 238A polymorphism [55] P. Sun, Y. Qiu, Z. Zhang et al., “Association of genetic poly- is associated with susceptibility to persistent bone marrow morphisms, mRNA expression of p53 and p21 with chronic dysplasia following chronic exposure to benzene,” Leukemia benzene poisoning in a Chinese occupational population,” Research, vol. 31, no. 11, pp. 1479–1485, 2007. Cancer Epidemiology Biomarkers and Prevention,vol.18,no.6, [40] I. Tomlinson, P.Sasieni, and W.Bodmer, “How many mutations pp.1821–1828,2009. in a cancer?” TheAmericanJournalofPathology,vol.160,no.3, [56] S. Angelini, F. Maffei, J. L. Bermejo et al., “Environmental expo- pp. 755–758, 2002. sure to benzene, micronucleus formation and polymorphisms [41] Y.-Z. He, X. Hu, X.-S. Chi et al., “Association between RAD51 in DNA-repair genes: a pilot study,” Mutation Research: Genetic gene polymorphism (-135G/C) and susceptibility of myelodys- Toxicology and Environmental Mutagenesis,vol.743,no.1-2,pp. plastic syndrome and acute leukemia: evidence based on a 99–104, 2012. meta-analysis,” Tumor Biology,vol.35,no.1,pp.615–621,2014. [42] E. Fabiani, L. Fianchi, G. Falconi et al., “The BCL2L10 Leu21Arg variant and risk of therapy-related myeloid neoplasms and de novo myelodysplastic syndromes,” Leukemia & Lymphoma,vol. 55,no.7,pp.1538–1543,2014. [43] L. Zhang, C. M. McHale, N. Rothman et al., “Systems biology of human benzene exposure,” Chemico-Biological Interactions,vol. 184,no.1-2,pp.86–93,2010. [44] L. Liu, X. Ling, H. Liang et al., “Hypomethylation mediated by decreased DNMTs involves in the activation of proto-oncogene MPL in TK6 cells treated with hydroquinone,” Toxicology Letters,vol.209,no.3,pp.239–245,2012. [45]Y.Sha,W.Zhou,Z.Yangetal.,“Changesinpoly(ADP- ribosyl)ation patterns in workers exposed to BTX,” PLoS ONE, vol. 9, no. 9, Article ID e106146, 2014. [46] C. Paz-Y-Mino,A.L˜ opez-Cort´ es,M.Ar´ evalo,´ and M. E. Sanchez,´ “Monitoring of DNA damage in individuals exposed topetroleumhydrocarbonsinEcuador,”Annals of the New York Academy of Sciences,vol.1140,pp.121–128,2008. [47] N. Rothman, R. Haas, R. B. Hayes et al., “Benzene induces gene-duplicating but not gene-inactivating mutations at the Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 293408, 6 pages http://dx.doi.org/10.1155/2015/293408

Research Article Heterogeneity of Genetic Damage in Cervical Nuclei and Lymphocytes in Women with Different Levels of Dysplasia and Cancer-Associated Risk Factors

Carlos Alvarez-Moya,1 Mónica Reynoso-Silva,1 Alejandro A. Canales-Aguirre,2 José O. Chavez-Chavez,3 Hugo Castañeda-Vázquez,4 and Alfredo I. Feria-Velasco1

1 Cellular and Molecular Biology Department, Guadalajara University, Juarez 976, Colonia Centro, 44100 Guadalajara, JAL, Mexico 2Unit of Medical and Pharmaceutical Biotechnology, CIATEJ, Avenida Normalistas 800, Colinas de la Normal, 44270 Guadalajara, JAL, Mexico 3Dysplasias Unit, Institute of Cancerology Jaliscience, General Coronado 515, Colonia Centro, 44100 Guadalajara, JAL, Mexico 4Mastitis and Molecular Diagnosis, Guadalajara University, Juarez 976, Colonia Centro, 44100 Guadalajara, JAL, Mexico

Correspondence should be addressed to Monica´ Reynoso-Silva; [email protected]

Received 7 November 2014; Revised 12 February 2015; Accepted 9 March 2015

Academic Editor: Alessandra Pulliero

Copyright © 2015 Carlos Alvarez-Moya et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The comet assay can be used to assess genetic damage, but heterogeneity in the length of the tails is frequently observed. The aims of this study were to evaluate genetic damage and heterogeneity in the cervical nuclei and lymphocytes from patients with different levels of dysplasia and to determine the risk factors associated with the development of cervical cancer. The study included 97 females who presented with different levels of dysplasia. A comet assay was performed in peripheral blood lymphocytes and cervical epithelial cells. Significant genetic damage𝑃 ( ≤ 0.05) was observed only in patients diagnosed with nuclei cervical from dysplasia III (NCDIII) and lymphocytes from dysplasia I (LDI). However, the standard deviations of the tail lengths in the cervical nuclei and lymphocytes from patients with dysplasia I were significantly different (𝑃 ≤ 0.0001) from the standard deviations of the tail lengths in the nuclei cervical and lymphocytes from patients with DII and DIII (NCDII, NCDIII and LDII, LDIII), indicating a high heterogeneity in tail length. Results suggest that genetic damage could be widely present but only manifested as increased tail length in certain cell populations. This heterogeneity could obscure the statistical significance of the genetic damage.

1. Introduction presence of micronuclei, sister chromatid exchange [7–10], and a significant increase in DNA damage in cervical epithe- Cervical uterine cancer (CCU) is the second most common lial nuclei [11]. Genetic instability is a necessary condition cancer among women and is responsible for over 250,000 for cancer development and the accumulation of DNA deaths annually [1]. CCU is related to human papillomavirus damage is the molecular basis of cancer progression [12, (HPV), which activates early genes E6 and E78 of the host 13] so genetic damage in dysplastic cells is an important giving rise to benign or premalignant cervical lesions. There criterion for the diagnosis of CCU [7]. Because both genetic are other etiological factors besides human papillomavirus damage and progression of dysplasia may be influenced by (HPV) and its risk of occurrence varies from one population orden, quantity, and combination of exposure to various risk to another [2–6]. CCU begins as a mild hyperplasia (I) and factors and, due to that several studies have reported higher then progresses through moderate hyperplasia (II), severe frequencies of chromosomal damage in the lymphocytes of hyperplasia (III), and, ultimately, cervical carcinoma [6]. In individuals exposed occupationally or environmentally [14, women with cervical dysplasia, genomic instability has been 15], it is totally possible to use as monitors of genetic damage observed and is manifested as chromosomal abnormalities, to lymphocytes of patients with varying degrees of dysplasia 2 BioMed Research International

∘ and simultaneously visualize genetic damage in the nuclei of at 4 C until use. This small amount of whole blood was cells from dysplastic tissue. sufficient to study DNA damage using the comet assay and The comet assay is frequently used to detect DNA damage viability using the trypan blue exclusion test. The mean because the method is sufficiently simple and sensitive to percentage of viability for each group was >89%: (B) cervical detectdamageinindividualcells.Inthistechnique,cells epithelial cells. The cervix was gently wiped with a sterile embedded in agarose are placed on a microscope slide, swab, and the shaped end of a wooden spatula was used ∘ detergents and high salts to lyse the cells and make DNA to scrape the cervix in a 360 fashion. Two scrapings were electrophoresis under alkaline or neutral conditions. Cells collected from each patient. A sample from the first scraping with an increased frequency of DNA double-strand breaks wasspreadonaglassslideandfixedimmediatelyinethyl show increased migration of DNA toward the anode. The alcohol; this sample was used in a Pap smear test to determine migrating DNA is quantitated by staining with ethidium the pathological status of the cervix. A sample from the bromide and by measuring the intensity of fluorescence at second scraping was suspended in PBS and was later used two fixed positions within the migration pattern using a inthecometassaytostudyDNAdamageandthetrypan microscope photometer [16, 17]. The comet tail is the result blue exclusion test to determine cell viability. The samples of the migration of DNA fragments, broken strands, and were brought to the lab under cold conditions and processed sites sensitive to alkali. Variations in the tail length of comet within 1 h of sampling to avoid artefact damage to the DNA. nuclei among cells in the same study indicate differences in the degree of damage [18]; however, although comet assay 2.4. Classification of Samples. The Pap smear test was used detects genetic damage which is an important criterion for because it is an ideal method to detect cervical neoplasia. diagnosing cancer risk, the frequent heterogeneity of tail Based on the results of the Pap smears, the women were lengths observed using the comet assay [18]couldprevent divided into four categories: negative control (cervical cells the significant detection of genetic damage. In this work, and lymphocytes of healthy women without dysplasia), dys- genetic damage and heterogeneity in cervical nuclei and plasia low, dysplasia moderate, dysplasia severe, and cervical lymphocytes from patients with different levels of dysplasia carcinoma in accordance with the Koss and Durfee cytomor- were determined. Additionally, the risk factors associated phological classification [20]. The samples were graded as with the development of cervical cancer were evaluated. follows: nuclei cervical from women with dysplasia I (NCDI), dysplasia II (NCDII), and dysplasia III (NCDIII) and negative 2. Materials and Methods control (NCNC) and lymphocytes from dysplasia I (LDI), dysplasia II (LDII), and dysplasia III (LDIII) and negative 2.1. Reagents. All reagents were purchased from Sigma control (LNC). Chemical Co. (Guadalajara, Jalisco, Mexico),´ except dimethyl sulphoxide (DMSO) and disodium EDTA, which were 2.5. Comet Assay. To perform the comet assay, the cell obtained from J.T. Baker (D.F., Mexico).´ suspension of cervical epithelial cells and lymphocytes was placed on different slides and mixed with low melting-point ∘ 2.2. Study Population. The study included a total of 97 women agarose at 37 C, to a final concentration of 0.7%. The mixture (sample size sufficient) diagnosed with different stages of (100 𝜇L) was pipetted onto slides with 0.5% normal-melting- cervical neoplasia who received a gynaecological check-up at point agarose, to retain the agarose cell suspension. The drop containing the cells was covered with a glass cover the Hospital General de Occidente Zoquipan de la Secretar´ıa ∘ de Salubridad del Estado de Jalisco in the period between slip (24 mm × 24 mm) and left at 4 Cfor5min.Thecover 1 July 2012 and 30 June 2013. Subjects were coded and at slips were gently removed and the slides were then ready for the time of sampling to perform the study blinded. Details processing.Thealkalinecometassaywasperformedusingthe including the patients’ age, whether they smoked or lived with basic rationale of Singh et al. [21], with modifications. a smoker, family history of cancer, medical treatment and To induce nuclear lysis and facilitate DNA unfolding, the exposure to environmental, and occupational or household slides were immersed in a lysis buffer (2.5 M NaCl, 100 mM Na2EDTA, 10 mM Tris-HCl, 1% sodium lauryl sarcosine, 1% chemicals were noted in a questionnaire at the time of ∘ sampling. Triton X-100, and 10% DMSO, pH 10) for 2 h at 4 C. The slides were then placed in a horizontal electrophoresis system 2.3. Sampling. Two types of samples were collected from each with a high pH buffer (30 mM NaOH, 1 mM2 Na EDTA, pH individual, as performed in Udumudi et al. [19]. Informed 13) for 45 min to allow unwinding of the DNA prior to consent was obtained from the patients and the ethical electrophoresis, which was performed for 15 min at 1.0 V/cm guidelines of the hospital were followed. The two types of with an accompanying amperage of approximately 300 mA. cell samples and their collection methods were as follows: The same electrophoresis unit and power supply were used (A) peripheral blood lymphocytes (PBL). For each individual, throughout the study [22]. To avoid additional DNA damage, bloodfromafingerprickwascollectedin2conicaltubes all of the steps described above were performed under containing 10 mL of phosphate buffer solution (PBS) (160 mM yellow light. Following electrophoresis, the slides were gently NaCl, 8 mM Na2HPO4,4mMNaH2PO4,and50mMEDTA; washed to remove the alkali and then were immersed in a pH 7) and immediately transported to the laboratory and cen- neutralisation buffer (0.4 M Tris-Base, pH 7.5) for 5 min. Gels trifuged at 3000 rpm for 5 min. The supernatant was removed, were stained with ethidium bromide (100 mL at 20 mg/mL) and the pellet was suspended in PBS and immediately stored for 3 min and then rinsed three times with distilled water. BioMed Research International 3

100 Table 1: Correlation coefficients between the average tail length of 90 cervical nuclei and lymphocytes from women with different levels of dysplasia. The bold numbers indicate correlation coefficients 80 between lymphocytes and cervical cells in the same grade of 70

) dysplasia. 𝜇 60 50 NCDI NCDII NCDIII NCNC 0.9606 40 LD I 0.9039 0.7473 0.8996 0.8735 Tail length ( length Tail 30 LD II 0.9080 0.9090 0.9793 LD III 0.8284 0.9579 0.9360 0.8751 20 LNC 0.9115 0.8632 0.7311 0.8856 10 0 ∗47.63 1 3 5 7 9 ∗ 39 41 43 45 47 49 51 53 25 27 29 31 33 35 37 55 57 13 15 17 19 21 23 11 45.94 ± 44.82 60 43.26 ± 44.55 42.745 Number of women studied ± 3.18 ± ± ± 6.84 ) 50 14.10 31.00 18.34 3.35 6.95 33.43 𝜇 ± Cervical nuclei dysplasia I 40 ± 8.20 8.30 Cervical nuclei dysplasia II 30 Cervical nuclei dysplasia III 20 Cervical nuclei negative control Tail length ( length Tail Lymphocytes dysplasia I 10 Lymphocytes dysplasia II 0 Lymphocytes dysplasia III Lymphocytes negative control control control Negative Negative Dysplasia I Dysplasia I Dysplasia II Figure 1: Distribution of the average tail length of cervical nuclei Dysplasia II Dysplasia III Dysplasia III and lymphocytes from women with different degrees of dysplasia. Cervical nuclei Lymphocytes On the 𝑥-axis the number women screened is displayed. The color Figure 2: Average tail length (genetic damage) of cervical nuclei and refers to the cell type studied and the level of dysplasia. The 𝑦-axis lymphocytes from women with different grades of dysplasia. The indicates the average tail length of lymphocytes or cervical nucleus number in the bars corresponds to the average tail length of cervical in each of women. cells or nuclei from women with a certain level of dysplasia. Cervical cells and nuclei were placed separately. The negative control is also ∗ ∗ visualized. 𝑃 ≤ 0.05. Bartlett’s test suggests that the SD in both The preparation was then cover-slipped. The fluorescence cervical tissue nuclei and lymphocytes from cervical dysplasia I are microscopy used to examine the slides employed a light significantly different (𝑃 ≤ 0.0001) from those of lymphocytes and microscope equipped with a 515–560 nm excitation filter. cervical tissue nuclei from dysplasia II and III. Comets were observed at 40x magnification, and migration was determined by visually scoring the tail length according to published protocols [22]. Approximately 50 comets per are shown in Table 1.NCDIandLDIshowedthehighest slide and two slides for each individual, including controls, degree of correlation (0.9606). NCDII-LDII, NCDIII-LDIII, were evaluated. and NCNC-LNC had correlations of 0.8735, 0.9360, and 0.8856, respectively. 2.6. Statistical Analysis. The slides were coded at the time of Averages of the DNA migration for both the lymphocytes preparation, scoring, and analysis. The results were expressed and the cervical nuclei were obtained and are presented as the mean ± SDandwereanalyzedbyone-wayanalysis in Figure 2 and the comparisons with the corresponding of variance (ANOVA) using the CoStat program [23]. All negative controls were statistically significant (𝑃 ≤ 0.05) of the study groups were compared with the corresponding only for NCDIII and LDI (Figure 2). However, the standard negative control using the Dunnett test. Pearson’s correlation deviations of the cervical nuclei and lymphocytes from the was performed with SPSS 10.0 (Software, SPSS Inc.). Fifty cells dysplasia I group were significantly different (𝑃 ≤ 0.0001) were evaluated from each individual slide. A value of 𝑃 ≤ 0.05 from the cervical nuclei and lymphocytes from the DII indicated significance. and DIII groups. The comparison between negative controls showed no significant difference. 3. Results Table 2 shows the percentages of the participants who were exposed to various risk factors for cervical cancer The distribution and average tail lengths of lymphocytes and (except for the presence of HPV). Women reported having cervical nuclei from women with different levels of dysplasia exposure to multiple risk factors; Table 2 also shows the sum are presented in Figure 1. All of the study groups showed of the percentages of women exposed to each risk factor. heterogeneity in the average migration tail length, although cervicalnucleiandlymphocytesfromtheDIgroupshowed 4. Discussion a higher degree of heterogeneity. The correlation coefficients between the average migration of the cervical nuclei and The HPV is considered to be the most important risk factor lymphocytes from women with varying degrees of dysplasia in the aetiology of cervical cancer [24, 25]; however DNA and 4 BioMed Research International

Table 2: Cumulative percentage of women exposed to one (first-line) or more (second-line) onward risk factors for cervical cancer. Percentages of women exposed to more risk factors were obtained by adding individual percentages of exposure.

Recent Family Occupational Home Environmental Living with Takes Smoker exposure history of exposure to exposure to exposure to asmoker medications (S) to X-ray cancer chemicals chemicals chemicals (LS) (TM) (RERX) (FHC) (OECH) (HECH) (EECH) 15.8∗ 40.9∗ 29.5∗ 2.2∗ 20.4∗ 6.8∗ 70.4∗ 22.7∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ Smoker 56.7 45.3 18 36.2 22.6 86 38.5 ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ Living with a smoker 70.4 43.1 61.3 47.7 111.3 63.6 ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ Takes medications 31.7 49.9 30.3 99.9 52.2 ∗∗ ∗∗ ∗∗ ∗∗ Recent exposure to X-ray 22.6 9 72.6 24.9 ∗∗ ∗∗ ∗∗ Family history cancer 27.2 90.8 43.1 ∗∗ ∗∗ Occupational exposure to chemicals 77.2 29.5 ∗∗ Home exposure to chemicals 93.1 Note: percentages correspond to the responses of the women studied regarding the types of exposure to cancer risk factors. Clearly, there is simultaneous exposure to more than one risk factor. ∗ Refer to exposure to an agent. ∗∗ Refer to the simultaneous exposure to two or more agents. cells in the human body are constantly exposed to oxidative of genetic damage in the cervical nuclei and lymphocytes attack by both exogenous and endogenous agents that are because of the differential exposure of these cell types, which capable of inducing genetic damage associated with cancer contributes to the substantial heterogeneity of tail length, for development [26] and may play an important role in cervical example, a large difference in tail length since 15–80 microns carcinogenesis; these agents should thus be considered as in NCDI were observed. Similar situation occurred in LDI risk factors: smoking and other factors also increase the with variations since 22–90. As shown, the group of patients risk of this disease through oncogenic activation [27–29]. with dysplasia I was the most frequent and large during Additionally, it has been reported that there is a synergistic the time period of this studio. The cases of patients with effect between smoking and HPV infection on cervical dysplasia II and dysplasia III were less frequent, but also great carcinogenesis [30, 31]. Considering that in our study the heterogeneityoflengthtailwasobservedinbothlymphocytes percentage of women who smoked was low (15.8%) and that and cervical nuclei. Importantly, the similarity of migration all the patients showed some form of dysplasia, it follows that and heterogeneity between lymphocytes and cervical nuclei there are other risk factors that collaborate with HPV, as also (see correlation coefficients in Table 1) highlighting the case reported by Schiffman and Castle [27]. The data in Table 2 of LDI and NCDI (0.9606), LDII-CNDII (0.8735), LDIII- show a high percentage of nonsmoking women who were CNDIII (0.9360), and negative controls for lymphocytes and exposed to several other risk factors. The order of the rates of cervical nuclei (0.8856). Therefore it is understood as the use- exposure (home exposure to chemicals > living with a smoker fulness of lymphocytes as biomonitors of genetic damage in > takes medications > environmental exposure to chemicals cervical dysplasia (Figure 1). Another factor that contributes > family history of cancer > smoker > occupational exposure to the heterogeneity of tail lengths is the difference in DNA to chemicals > recent exposure to X-ray) suggested that some damage repair between the cell types studied, as previously risk factors and their combinations could induce cervical reported [32]. There are also established differences in diet carcinogenesis, although we were unable to determine the associated with the development of different types of cancer extent of genetic damage caused by each factor. [33]. Ostling and Johanson demonstrated in vitro that these Strikingly, in addition to the 15.8% of participants who differences can contribute to the aetiology of heterogeneity were smokers, 40.9% of the women studied lived with in the same population of treated cells [34]. Those authors smokers; therefore, the percentage of the participants who hypothesized that the heterogeneity of the tail length was were exposed to cigarette smoke was actually much higher the result of differential induction of genetic damage among (Table 2). A similar situation could be observed with the cells. Differential genetic damage and genetic heterogeneity other risk factors; then it is possible to observe an increase in were subsequently demonstrated in animal cells from tumor both heterogeneity and genetic damage in women exposed to biopsies and clinical models, explaining the resistance of different risk factors; further the frequency and duration of some tumors to cancer treatment [35–37]. the exposure to various chemical agents were varied. Addi- Because of the heterogeneity of the tail length, many tionally, several other indirect factors, including low socioe- authors have suggested that different types of comets should conomic status, early marriage and first child, multiparity, be recognized and considered separately [38]. In our study, poor personal hygiene, and genital infections, which can several types of comets were observed, which contributed also increase the exposure to risk factors and consequently to the heterogeneity; it was clear that there were different lead to damage genetic, were not considered in this study cells or nuclei with different magnitudes of genetic damage [31]. All of the above risk factors induce different amounts in vivo [16]. In this context, the age of the lymphocytes and BioMed Research International 5 nuclei studied is an important factor to be considered [39– The comet assay is an excellent tool for determining 42]; increased age is correlated with the heterogeneity in the presence of genetic damage. However, to contribute to the average migration (see negative control (Figure 1)). The the reduction of heterogeneity in tail length, only cells with ability of cells to repair DNA is lower in older lymphocytes an increased tail length would ideally be included in the compared with younger cells [39, 43]. measurements. Udumudi et al. [19] studied the relationship between genetic damage and the development of cervical neoplasia Conflict of Interests in nuclei and lymphocytes from women with cervical cancer using the comet assay and showed that genetic damage The authors declare that there is no conflict of interests was proportional to the degree of dysplasia [13]. In our regarding the publication of this paper. study, only samples from the NCDIII and LDI groups had sufficient cumulative damage to be detected by the comet Acknowledgments assay (Figure 2). The other groups did not show a significant difference with the corresponding negative controls; however, The authors thank the Instituto Jaliscience de Cancerolog´ıa cervical nuclei appeared to have damage proportional with (SSJ) and the Cl´ınica de Displasias-Hospital Zoquipan for the degree of dysplasia (Figure 2). The LDI group showed their support and assistance in contacting the patients with significant genetic damage (𝑃 ≤ 0.05). In the LDII and LDIII dysplasia. groups, despite an increase in migration that approached significance, there was no statistically significant difference in genetic damage (LDI 44.82 𝜇m, LDII 44.55 𝜇m, and LDII References 𝜇 42.74 m). Genetic damage could be present but only man- [1] WHO, Control Integral del Cancer´ Cervicouterino,WorldHealth ifested as increased tail length in certain cell populations, as Organization, Geneva, Switzerland, 2007. has been found in previous studies [34]. This possibility could [2] M. Shah, B. Parikh, C. A. Shah, and D. Jani, “Epidemiological haveledtothedilutionofthestatisticalsignificanceofthe study of the cancer of the uterine cervix in adivasi people migration averages (Figure 2) in our study, so our data are of five different states of India,” The Journal of Obstetrics and different from those reported by Udumudi et al. [19]. Gynecology of India,vol.335,pp.355–360,1985. The comet assay is very efficient and allows forthe [3]P.K.Dutta,A.Upadhyay,M.Dutta,A.C.Urmil,M.P. direct visualization of genetic damage in individual cells Thergaonkar, and S. S. Ganguly, “A case control study of cancer [21], typically by measuring the average migration of the cervix patients attending Command Hosptial, Pune,” Indian tail (average genetic damage). In our study, the average Journal of Cancer,vol.27,no.2,pp.101–108,1990. geneticdamageisnotsignificantinallofthestudygroups [4] J. Paavonen, “Chlamydia trachomatis and cancer,” Sexually (Figure 2). Some populations of cells showed more damage, Transmitted Infections,vol.77,no.3,pp.154–156,2001. which caused heterogeneity in the migration of the tail [5] M. D. Schiffman and S. Wacholder, “From India to the world—a length among study subjects (Figure 1). Bartlett’s test suggests better way to prevent cervical cancer,” The New England Journal that the SD obtained from the CNDI group is significantly of Medicine,vol.360,no.14,pp.1453–1455,2009. different (𝑃 ≤ 0.0001) from the SD of the nuclei from [6] F. de la Torre-Rendon,´ “Lesion´ premaligna escamosa del cuello the cervical dysplasia II and III groups, and this difference uterino, un enfoque actualizado,” Patolog´ıa,vol.46,no.4,pp. manifests itself in the high heterogeneity in tail length. 332–342, 2008. Heterogeneity could obscure the statistical significance of [7]T.C.Hsu,L.M.Cherry,andN.A.Samaan,“Differential the damage. The NCDI group appears to have more genetic mutagen susceptibility in cultured lymphocytes of normal individuals and cancer patients,” Cancer Genetics and Cytogenetics, damage than the NCDII and NCDIII groups (Figure 1), but vol. 17, no. 4, pp. 307–313, 1985. the average genetic damage is greater in the NCDIII group [8] E. I. Cortes-Guti´ errez,´ R. M. Cerda-Flores, and C. H. Leal- (Figure 2).Thesimilarlevelsofgeneticdamageandhigh Garza, “Sister chromatid exchanges in peripheral lymphocytes correlation coefficients between lymphocytes and cervical from women with carcinoma of the uterine cervix,”´ Cancer nuclei suggest that lymphocytes may serve as sentinel cells Genetics and Cytogenetics,vol.122,no.2,pp.121–123,2000. (Table 1).However,thepresenceofsubstantialheterogeneity [9]C.H.Leal-Garza,R.M.Cerda-Flores,E.Leal-Elizondo,and in the tail length should be considered. Ideally, only the cells E. I. Cortes-Guti´ errez,´ “Micronuclei in cervical smears and in each sample that have increased genetic damage would peripheral blood lymphocytes from women with and without be considered in the measurement, including those in the cervical uterine cancer,” Mutation Research—Genetic Toxicology negative controls. and Environmental Mutagenesis,vol.515,no.1-2,pp.57–62, 2002. [10] G. Gandhi and B. Kaur, “Elevated frequency of Micronuclei in 5. Conclusion uterine smears of cervix cancer patients,” Caryologia,vol.56,no. 2,pp.217–222,2003. There is differential genetic damage manifested as increased [11] E. I. Cortes-Guti´ errez,´ M. I. Davila-Rodr´ ´ıguez, E. A. Zamudio- heterogeneity in the lengths of the tail length of cervical nuclei Gonzalez,´ M. E. Aguado-Barrera, J. Vargas-Villarreal, and R. M. and peripheral blood lymphocytes in women with cervical Cerda-Flores, “DNA damage in Mexican women with cervical dysplasia. This observed heterogeneity may confound the dysplasia evaluated by comet assay,” Analytical and Quantitative detection of significant genetic damage. Cytology and Histology,vol.32,no.4,pp.207–213,2010. 6 BioMed Research International

[12] E. Prieto and N. Llopiz,´ “Normalizacion´ de la electroforesis [29] S. Duensing and K. Munger,¨ “Centrosomes, genomic instability, de celulas´ individuales (ensayo cometa),” Revista Cubana de and cervical carcinogenesis,” Critical Reviews in Eukaryotic Investigacion´ Biomedica´ ,vol.18,no.1,pp.34–36,1999. Gene Expression,vol.13,no.1,pp.9–23,2003. [13] C. G. LeonandM.E.Faxas,“C´ ancer´ de cuello uterino: aspec- [30] X. CastellsagueandN.Mu´ noz,˜ “Chapter 3: cofactors in human tos inmunologicos´ y geneticos´ de mayor relevancia,” Revista papillomavirus carcinogenesis—role of parity, oral contracep- Cubana de Medicina,vol.43,no.1,2004. tives, and tobacco smoking,” Journal of the National Cancer [14]R.Scarpato,A.Lori,G.Panasiuk,andR.Barale,“FISHanalysis Institute Monographs, no. 31, pp. 20–28, 2003. of translocations in lymphocytes of children exposed to the [31] H. zur Hausen, “Human genital cancer: synergism between two Chernobyl fallout: preferential involvement of chromosome 10,” virus infections or synergism between a virus infection and Cytogenetics and Cell Genetics, vol. 79, no. 1-2, pp. 153–156, 1997. initiating events?” The Lancet,vol.2,no.8312,pp.1370–1372, [15] S. Wang-Wuu, J. J. Tai, J. Wuu, S.-Y. Lin, W.-L. Chen, and K.- 1982. D. Wuu, “Chromosome aberrations in lymphocytes of residents [32]E.Cemeli,A.Baumgartner,andD.Anderson,“Antioxidants living in buildings constructed with radioactively contaminated and the comet assay,” Mutation Research,vol.681,no.1,pp.51– rebars,” Journal of Biomedical Science,vol.8,no.5,pp.411–415, 67, 2009. 2001. [33] J. V. Higdon, B. Delage, D. E. Williams, and R. H. Dashwood, [16] D. W. Fairbairn, P. L. Olive, and K. L. O’Neill, “The comet “Cruciferous vegetables and human cancer risk: epidemiologic assay: a comprehensive review,” Mutation Research—Reviews in evidence and mechanistic basis,” Pharmacological Research,vol. Genetic Toxicology,vol.339,no.1,pp.37–59,1995. 55,no.3,pp.224–236,2007. [17] R. Tice, “The single cell gel / comet assay: a microgel elec- [34] O. Ostling and K. J. Johanson, “Bleomycin, in contrast to gamma trophoretic technique for the detection of DNA damage and irradiation, induces extreme variation of DNA strand breakage repair in individual cells,” in Environmental Mutagenesis,D.H. from cell to cell,” International Journal of Radiation Biology,vol. Phillis and S. Venitt, Eds., pp. 315–339, Bios Scientific Publishers, 52,no.5,pp.683–691,1987. Oxford, UK, 1995. [35] G. M. Almeida, T. L. Duarte, W. P. Steward, and G. D. D. Jones, [ 1 8 ] P. L . O l i v e an d J. P. B an ath,´ “The comet assay: a method to “Detection of oxaliplatin-induced DNA crosslinks in vitro and measure DNA damage in individual cells,” Nature Protocols,vol. in cancer patients using the alkaline comet assay,” DNA Repair, 1, no. 1, pp. 23–29, 2006. vol. 5, no. 2, pp. 219–225, 2006. [19] A. Udumudi, M. Jaiswal, N. Rajeswari et al., “Risk assessment [36]P.L.Olive,R.E.Durand,S.M.Jacksonetal.,“Thecometassay in cervical dysplasia patients by single cell gel electrophoresis in clinical practice,” Acta Oncologica,vol.38,no.7,pp.839–844, assay: a study of DNA damage and repair,” Mutation Research— 1999. Genetic Toxicology and Environmental Mutagenesis,vol.412,no. 2, pp. 195–205, 1998. [37] M. J. Dorie, M. S. Kovacs, E. C. Gabalski et al., “DNA damage measured by the comet assay in head and neck cancer patients [20] L. D. Koss and G. B. Durfee, “Diagnostic cytology and its treated with tirapazamine,” Neoplasia,vol.1,no.5,pp.461–467, histopathologic basis,” The American Journal of Medical Science, 1999. vol. 242, no. 6, pp. 166–790, 1961. [38] P. L. Olive and J. P. Banath, “Detection of DNA double- [21] N. P. Singh, M. T. McCoy, R. R. Tice, and E. L. Schneider, “A strand breaks through the cell cycle after exposure to X-rays, simple technique for quantitation of low levels of DNA damage bleomycin, etoposide and 125IdUrd,” International Journal of in individual cells,” Experimental Cell Research,vol.175,no.1, Radiation Biology, vol. 64, no. 4, pp. 349–358, 1993. pp. 184–191, 1988. [22] A. Hartmann, E. Agurell, C. Beevers et al., “Recommendations [39] F.Hazane,S.Sauvaigo,T.Douki,A.Favier,andJ.-C.Beani,“Age- for conducting the in vivo alkaline comet assay,” Mutagenesis, dependent DNA repair and cell cycle distribution of human vol.18,no.1,pp.45–51,2003. skin fibroblasts in response to UVA irradiation,” Journal of Photochemistry and Photobiology B: Biology,vol.82,no.3,pp. [23] T.-H. Ma, G. L. Cabrera, A. Cebulska-Wasilewska et al., “Trades- 214–223, 2006. cantia stamen hair mutation bioassay,” Mutation Research,vol. 310,no.2,pp.211–220,1994. [40]T.Yu,S.H.MacPhail,J.P.Banath,´ D. Klokov, and P. L. Olive, “Endogenous expression of phosphorylated histone H2AX in [24] J.M.M.Walboomers,M.V.Jacobs,M.M.Manosetal.,“Human tumors in relation to DNA double-strand breaks and genomic papillomavirus is a necessary cause of invasive cervical cancer instability,” DNA Repair,vol.5,no.8,pp.935–946,2006. worldwide,” The Journal of Pathology,vol.189,no.1,pp.12–19, 1999. [41] P. J. Mayer, C. S. Lange, M. O. Bradley, and W. W. Nichols, [25] H. zur Hausen, “Molecular pathogenesis of cancer of the cervix “Age-dependent decline in rejoining of X-ray-induced DNA and its causation by specific human papillomavirus types,” double-strand breaks in normal human lymphocytes,” Mutation Current Topics in Microbiology and Immunology,vol.186,pp. Research,vol.219,no.2,pp.95–100,1989. 131–156, 1994. [42] N. P. Singh, D. B. Danner, R. R. Tice, L. Brant, and E. L. [26] T. B. Kryston, A. B. Georgiev, P. Pissis, and A. G. Georgakilas, Schneider, “DNA damage and rpair with age in individual “Role of oxidative stress and DNA damage in human carcino- human lymphocytes,” Mutation Research DNAging,vol.237,no. genesis,” Mutation Research, vol. 711, no. 1-2, pp. 193–201, 2011. 3-4, pp. 123–130, 1990. [27] M. Schiffman and P. E. Castle, “Human papillomavirus: epi- [43]N.P.Singh,C.E.Ogburn,N.S.Wolf,G.VanBelle,andG.M. demiology and public health,” Archives of Pathology and Lab- Martin, “DNA double-strand breaks in mouse kidney cells with oratory Medicine,vol.127,no.8,pp.930–934,2003. age,” Biogerontology,vol.2,no.4,pp.261–270,2001. [28] S. Alam, M. J. Conway, H.-S. Chen, and C. Meyers, “The cigarette smoke carcinogen benzo[a]pyrene enhances human papillomavirus synthesis,” Journal of Virology,vol.82,no.2,pp. 1053–1058, 2008. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 257090, 11 pages http://dx.doi.org/10.1155/2015/257090

Research Article Oxidative Stress Alters miRNA and Gene Expression Profiles in Villous First Trimester Trophoblasts

Courtney E. Cross,1 Mai F. Tolba,2 Catherine M. Rondelli,1 Meixiang Xu,1 and Sherif Z. Abdel-Rahman1

1 Department of Obstetrics and Gynecology, The University of Texas Medical Branch, Galveston, TX 77555-1066, USA 2Department of Pharmacology and Toxicology, Faculty of Pharmacy, Ain Shams University, Cairo 11566, Egypt

Correspondence should be addressed to Sherif Z. Abdel-Rahman; [email protected]

Received 10 December 2014; Revised 13 March 2015; Accepted 20 March 2015

Academic Editor: Luciana dos Reis Vasques

Copyright © 2015 Courtney E. Cross et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The relationship between oxidative stress and miRNA changes in placenta as a potential mechanism involved in preeclampsia (PE) is not fully elucidated. We investigated the impact of oxidative stress on miRNAs and mRNA expression profiles of genes associated with PE in villous 3A first trimester trophoblast cells exposed to H2O2 at 12 different concentrations (0-1 mM) for 0.5, 4, 24, and 48 h. Cytotoxicity, determined using the SRB assay, was used to calculate the IC50 of H2O2. RNA was extracted after 4 h exposure to H2O2 for miRNA and gene expression profiling.2 H O2 exerted a concentration- and time-dependent cytotoxicity on 3A trophoblast cells. Short-term exposure of 3A cells to low concentration of H2O2 (5% of IC50) significantly altered miRNA profile as evidenced by significant changes in 195 out of 595 evaluable miRNAs. Tool for annotations of microRNAs (TAM) analysis indicated that these altered miRNAs fall into 43 clusters and 34 families, with 41 functions identified. Exposure to H2O2 altered mRNA expression of 22 out of 84 key genes involved in dysregulation of placental development. In conclusion, short-term exposure of villous first trimester trophoblasts to low concentrations of H2O2 significantly alters miRNA profile and expression of genes implicated in placental development.

1. Introduction 4]. H2O2 levels are also significantly higher in preeclamptic placentas compared to normotensive placentas at term [5]. Preeclampsia (PE), which affects 3% to 8% of pregnant women, Moreover, evidence supports an early increase in oxidative remains a major cause of short- and long-term maternal and stress in the placenta by the end of the first trimester before neonatal morbidity and mortality. It is a medical condition the clinical development of PE [6–8]. Oxidative stress can characterized by de novo hypertension in pregnancy (dias- induce endothelial dysfunction and vasoconstriction [9]. > tolic 90 mm Hg) after 20-week gestation with high protein- Therefore, the relationship between oxidative stress and PE is > uria ( 300 mg) [1].PEisthoughttoresultfromacombination a vicious cycle where increased oxidative stress can induce PE of many factors including shallow trophoblast invasion, and the occurrence of PE also exacerbates oxidative stress. failed maternal spiral artery remodeling, and an increase in The age-adjusted incidence of PE in the United States endothelial activation leading to placental hypoxia, reactive increased almost 25% from 1987 to 2004 [10]. With disease oxygen species (ROS) generation, apoptosis and necrosis of riskontherisebutnoeffectivewaytopredictitsdevelopment, trophoblasts, and systemic activation of inflammatory proc- itiscrucialtounderstandtheearlyetiologicmechanismsof esses in the mother [1]. Superoxide anions generated endoge- PE to develop early detection biomarkers and possible pre- nously or exogenously can rapidly be converted to hydrogen ventive measures for the disease. peroxide (H2O2)[2] and studies showed a significant eleva- MicroRNAs (miRNAs) are short 20–22 single strand reg- tion of H2O2 levelsinthebloodstreamofwomenwithPE[3, ulatory RNAs that function by inhibiting translation of their 2 BioMed Research International targets or promoting target RNA degradation [11, 12]. There is were fixed with 10% tricholoroacetic acid for one hour and a temporal and placental-specific pattern of miRNA expres- then stained with 0.4% SRB dissolved in 1% acetic acid. Excess sion [13]. This pattern includes two large imprinted miRNA dye was washed and wells were air-dried. Bound dye was solu- clusters, one located at chromosome 19q13.41 (C19MC) and bilizedin10mMTris(pH7.4)andabsorbancewasmeasured another at 14q32 (C14MC). Expression of C14MC decreases at 560nm using a microplate reader (TECAN GENios Pro, while expression of C19MC increases as pregnancy progresses Mannedorf,¨ Switzerland). Results were expressed as the rela- [14]. Women with PE, eclampsia, and HELLP (Hemolysis, tive percentage of absorbance compared to control. All exper- Elevated Liver enzymes, Low Platelets) syndrome have signif- iments were performed in triplicate. Half-maximal inhibitory icant alterations in the placental miRNA profile. For example, concentration (IC50) was calculated using SigmaPlot, version dysregulation of C19MC expression is seen in preeclamptic 12.3 (Systat Software Inc., San Jose, CA, USA). placentas [15]. In addition, miR-210, which regulates hypoxia- inducible factor 1, is commonly found to be upregulated in PE 2.4. Assessment of Oxidative Stress Markers. Reduced glu- women [16]. Dysregulated expression of a number of other tathione (GSH) levels and superoxide dismutase (SOD) miRNAs, such as the oncogenic miR-17 family, occurs in activity were assessed in the cell lysates as markers for preeclamptic placentas but the temporal change is unknown oxidative stress [22, 23]. To determine the level of GSH, an [17–20]. Given the reported changes in miRNAs observed in aliquot (0.2 mL) of cell lysate was added to a tube containing PE, early detection, protection, and regulation of miRNAs 1.7mLphosphatebufferand0.1mLEllman’sreagent;thenthe mayhelpdecreasetheimpactofthisdisease. absorbance was read at 412 nm within 5 min [24]. The results The relationship between oxidative stress and miRNA were expressed as mg/𝜇g protein. SOD activity was assessed changesinplacentaasapotentialmechanisminvolvedinPE in the cell lysates using an assay that relies on the ability isnotfullyelucidated.Inthepresentstudy,weinvestigatethe of SOD enzyme to inhibit the phenazine methosulphate- impact of oxidative stress on miRNA and mRNA expression (PMS-) mediated reduction of nitroblue tetrazolium dye. The profiles, with specific emphasis on mRNAs of genes known change in absorbance over 5 min was measured at 560 nm to be associated with PE, in villous first trimester 3A cytotro- [25]. SOD activity was expressed as U/mg protein. Protein phoblast cell line. levels were determined using the BCA protein assay (Pierce Biotechnology, Rockford, IL, USA). Student’s 𝑡-test was used 2. Materials and Methods for statistical comparisons and significant differences were established at 𝑃 < 0.05. 2.1. Chemicals and Reagents. The villous 3A cytotrophoblast first trimester placental cell line (CRL-1584) was purchased 2.5.DeterminationofEffectsofH2O2 Exposure on miRNA and from American Type Culture Collection (ATCC) (Manassas, mRNA Expression Profiles. Cells were grown to 90% conflu- VA). The miRNA and mRNA kits with appropriate real- 2 𝜇 time reagents and miRNEasy were purchased from Qiagen ency in 75 cm flasks and exposed to 25 MH2O2 (equivalent (Valencia, CA). CytoScan SRB cell cytotoxicity assay kit was to 5% of the IC50)incompletemediumfor4h.TotalRNAwas purchased from G-biosciences (St. Louis, MO). Eagle’s Min- isolated using the Qiagen miRNEasy Mini Kit and quality/ imum Essential Medium (EMEM) and 0.25% trypsin were quantity was measured in the Molecular Genomics Core at Gibco brand (Life technologies, Grand Island, NY). Fetal UTMB. RNA was quantitated spectrophotometrically using a bovine serum (FBS) was acquired from Atlanta Biologicals NanoDrop ND-1000 (NanoDrop Technologies, DE). Quality (Lawrenceville, GA). All other chemicals were purchased of the purified RNA was assessed by visualization of 18S and from Sigma-Aldrich Co. (St. Louis, MO, USA) and were of 28S RNA bands using an Agilent Bioanalyzer 2100 (Agilent the highest purity available. Technologies, CA). Resulting electropherograms were used in the calculation of the 28S/18S ratio and the RNA Integrity Number. Reverse transcription was carried out using either 2.2. Cell Culture and H2O2 Treatment. Villous 3A cytotro- 2 2 ∘ phoblast cells were maintained in 75 cm flasks at 37 C/5% the miScript II RT kit or RT First Strand and subsequent SYBR green based real-time PCR on a Bio-Rad Chromo4 CO2 in complete medium consisting of EMEM supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. Real-Time PCR Detector per the manufacturer’s recom- mendation. The miRNA profile screening was performed Cells were passaged at ∼90% confluency.2 H O2 exposures were performed in complete medium for the times and using miScript Human miRNome PCR Array (MIHS-3216Z, ∘ 2 concentrations specified in a humidified incubator at37 C Qiagen, Valencia, CA). RT Profiler Human Preeclampsia supplemented with 5% CO2. PCR Array (PAHS-163Z, Qiagen, Valencia, CA) was used for gene expression profiling. Data was analyzed using the ΔΔCT

2.3.DeterminationofH2O2 IC50. Cells were seeded in 12-well method with either the miScript miRNA PCR Array Data 5 2 plates at a density of 0.25 × 10 cells/well and were allowed to Analysis version 3.5 or with the RT Profiler PCR Array Data attach for 24 hours. Growth media were replaced either with Analysis (SABiosciences, Valencia, CA). fresh media for mock exposure or with one of 12 different concentrations of H2O2 (0 to 1 mM) in complete medium for 0.5, 3. Results 4, 24, or 48 h without further media replacement. Cytotoxic- ity was assessed at the end of H2O2 exposure using the SRB 3.1. H2O2-Induced Cytotoxicity in Villous 3A Trophoblasts. In assay as previously described [21]. Briefly, after exposure, cells order to determine the appropriate H2O2 concentration for BioMed Research International 3

=592±49.49 = 487.3 ± 107 120 H2O2 IC50 140 H2O2 IC50

100 120

100 80 80 60 60 Viability (%) Viability 40 (%) Viability 40

20 20

0 0 1 10 100 1000 1 10 100 1000 Concentration (𝜇M) Concentration (𝜇M) (a) 30 min exposure (b) 4 h exposure = 14.28 ± 0.67 120 H2O2 IC50 = 90 ± 17.5 120 H2O2 IC50

100 100

80 80

60 60 Viability (%) Viability 40 (%) Viability 40

20 20

0 0 1 10 100 1000 1 10 100 1000 Concentration (𝜇M) Concentration (𝜇M) (c) 24 h exposure (d) 48 h exposure

Figure 1: IC50 calculation for H2O2 after 30 min, 4 h, 24 hr, or 48 hr in villous 3A trophoblasts: 3A cells were exposed to varying concentrations (0,2,3,6,10,20,30,60,100,200,300,and1000𝜇M) of H2O2 and IC50 determined using the SRB assay. Data represent 𝑛=3for each time point and concentration. IC50 was calculated using SigmaPlot v12.3.

further experiments, cells were exposed to varying micro- 3.2. H2O2 Alters Normal miRNA Expression Profile in Villous molar concentrations of H2O2 in complete medium for up 3A Trophoblasts. To investigate the potential effect of oxida- to 48 h and cytotoxicity was determined using the SRB assay. tive stress on miRNA expression profile in villous 3A tropho- Exposure of 3A placental cells to H2O2 resulted in a time- and blasts, cells were exposed to 25 𝜇MH2O2 for 4 h in complete concentration-dependent cytotoxic effect. The50 IC values medium and total RNA (including miRNA) was isolated for H2O2 were 592, 487, 90, and 15 𝜇Mafter30min,4,24, using the Qiagen miRNeasy Mini Kit. The Qiagen v16 miRNA and 48 h exposures, respectively (Figure 1). A concentration Array was used to evaluate the expression of 1008 miRNAs equivalent to 5% of the IC50 concentration was used for the after H2O2 exposure compared to unexposed control. In our miRNAs and mRNA expression profiling experiments. The 4- study, 417 miRNAs were not expressed (unevaluable) in the hour exposure was selected for short-term exposure studies tested cell line. Out of the 591 evaluable miRNAs, 195 were of the effect of H2O2 on miRNA and mRNA expression up- or downregulated by at least twofold after H2O2 exposure profile. The levels of GSH as well as SOD activity were (Supplemental Table 1 available online at http://dx.doi.org/ assessed after exposure of the cells to 25 𝜇MH2O2 for 4 h as markers for oxidative stress status [22, 23]. GSH level in 10.1155/2015/257090). The majority of altered miRNAs (95.5%) were upregulated by at least 2-fold and only 4% were H2O2 challenged cells was significantly reduced by 30% (from 361.44 ± 10.01mg/mgproteininuntreatedcellsto251.35± downregulated by at least 2-fold (Figure 2). Mir-21, -770, 39.23 in the treated cells; 𝑃 < 0.01). Similarly, SOD activity and -596 were downregulated by more than 5-fold, while in cells exposed to H2O2 was significantly reduced by over mir-3907 was downregulated by more than 50-fold. Mir-637, 35% (208.9 ± 5.12 U/mg protein in treated cells compared to mir-1911, mir-26b, mir-615, let-7a, and let-7f were upregulated 321.71 ± 6.78 in the untreated cells; 𝑃 < 0.01). more than 50-fold (Supplemental Table 1). 4 BioMed Research International

Table 2: Functional pathways regulated by miRNAs altered by H2O2 in villous 3A trophoblasts. 41% Functiona Count Percentb 𝑃 valuec Angiogenesis 14 61 7.9𝑒 − 3 20% 19% Apoptosis 22 56 3.10𝑒 − 3 8% Bone regeneration 17 61 3.41𝑒 − 3 Cell cycle related 29 50 7.50𝑒 − 3 1% 31% Cell proliferation 16 62 3.76𝑒 − 3 Epithelial-mesenchymal 24 63 1.72𝑒 − 4 miRNA profiling transition Changed < 2-fold Downregulated Folliculogenesis 71005.67𝑒 − 4 Upregulated Unexpressed 4.48𝑒 − 4 Unchanged Granulopoiesis 990 Hormones regulation 29 54 1.77𝑒 − 3 Figure 2: Pie chart representation for miRNome array results after Human embryonic stem 36 51 1.80𝑒 − 3 4hexposuretoH2O2 in villous 3A trophoblasts. cell (hESC) regulation Immune response 29 73 3.36𝑒 − 7 2.02𝑒 − 4 Table 1: Clusters and families of miRNAs significantly altered by Inflammation 22 65 2.08𝑒 − 3 H2O2 in villous 3A trophoblasts. Adipocyte differentiation 16 64 Anticell proliferation 11 100 1.07𝑒 − 6 Count Percentb 𝑃 valuec Brain development 12 63 9.45𝑒 − 3 Clustera 31 62 2.54𝑒 − 5 hsa-let-7e cluster 3 100 0.0418 Cell death 2.29𝑒 − 3 hsa-let-106b cluster 3 100 0.0418 Cell division 11 73 0.0415 hsa-let-23b cluster 3 100 0.0418 Cell fate determination 12 55 1.77𝑒 − 3 Familya Hematopoiesis 18 62 let-7 family 9 100 6.40𝑒 − 5 Lipid metabolism 11 58 0.0304 mir-15 family 4 100 0.0144 miRNA tumor suppressors 24 65 9.08𝑒 − 5 mir-17 family 6 75 0.0237 Onco-miRNAs 19 61 1.61𝑒 − 3 mir-181 family 4 100 0.0144 aAnalysis of altered miRNAs was performed by TAM (tool for annotations mir-29 family 3 100 0.0418 of miRNAs, version 2). Size of miRNA category was set as 1–100. b mir-329 family 3 100 0.0418 Percent of miRNA changed between treated and controls cells within a pathway/function. mir-368 family 3 100 0.0418 c𝑃 < 0.05 indicates a significant number of miRNAs altered within a mir-99 family 3 100 0.0418 pathway/function. aAnalysis of overexpressed miRNAs was performed by TAM (tool for annotations of miRNAs, version 2). Size of miRNA category was set as 1–100. b Percent of miRNA changed between treated and controls cells within the well as miRNA tumor suppressors, and onco-miRNAs which cluster or family. c𝑃 < 0.05 indicates a significant number of miRNAs altered within a can affect cell proliferation and invasionTable ( 2). Notably, cluster/family. 21outof30(70%)ofevaluablemiRNAsofthematernally imprinted miRNA cluster on chromosome 14 (C14MC) that is predominantly expressed in placenta and developing Using TAM (tool for annotations of microRNAs, version embryonic tissues were altered. Expression of the paternally 2; http://202.38.126.151/hmdd/tools/tam.html/) analysis, our imprinted chromosome 19 cluster (C19MC) was too low to data indicate that the altered miRNAs fall into 43 clusters be evaluated in this study. and 34 families, with 41 functions identified. Three clusters of miRNAs were significantly altered (𝑃 < 0.05)after4h 3.3. H2O2 Alters mRNA Expression in Villous 3A Trophoblasts. exposure to 25 𝜇MH2O2 including hsa-let-7e, -let-106b, and - In order to investigate the effect of oxidative stress on mRNA let-23b clusters (Table 1). Eight miRNA families (let-7,mir-15, expression profile, we determined mRNA expression levels mir-17, mir-181, mir-29, mir-329, mir-368, and mir-99) were of 84 genes potentially involved in preeclamptic pregnancies. significantly altered𝑃 ( < 0.05)afterH2O2 challenge (Table 1). Villous 3A cells were mock-exposed in complete medium or Significant alterations (𝑃 < 0.05) occurred after exposure exposed to 25 𝜇MH2O2 in complete medium for 4 hours and to H2O2 in miRNAs involved in critical cellular functions total RNA was isolated using the Qiagen miRNeasy Mini Kit. 2 such as angiogenesis, apoptosis, cell proliferation, epithelial- The Qiagen RT Preeclampsia Array was used to evaluate the mesenchymal transition, folliculogenesis, granulopoiesis, expression of these 84 genes after2 H O2 exposure. As shown hormone regulation, human embryonic stem cell regulation, in Figure 3,ofthese84genes,22wereup-ordownregulated immune response, inflammation, anticell proliferation as by at least twofold after H2O2 exposure. Of these, only BioMed Research International 5

Layout 1 2 3 4 5 6 7 8 9 10 11 12 A ABCC1 ABCG2 ADM AGTR1 ANGPT2 APLN ATP1B1 ATP2A2 BCL6 BHLHE40 C3 CAV1 1.46−1.58 −1.07 −1.01 −2.49 −1.87 1.96 1.89 1.34 −1.04 −1.25 −1.4 B CCL2 CD40LG CDH13 CFD CLU COL14A1 CP CRH CRHBP CXCL10 CXCL9 CYP26A1 −2.38 −2.21 −1.35 −1.711.05 −1.94 −2.27 −2.1 1.24 −1.87 −1.62 −1.87 C DCN DUSP1 EDN1 ENG F5 FABP4 FLT1 FLT4 FSTL3 HBEGF HGF HIF1A −2 1.64 1.01 1.19 −1.87 −1.7 1.62 −1.69 −1.61 5.76 −1.94 1.57 D HLA-G HP HSD17B1 HSP90AA1 HTR3A HTRA1 IFNG IGF1 IGFBP3 IL10 IL11 IL15 −1.37 −2.41 −1.87 1.38 −1.8 −1.06 −1.87 −2.6 1.07 −1.18 −1.63 −1.38 E IL18 IL1A IL2 IL6 IL8 INHA INHBA ITGB3 KIT KRT19 LEP LPL −2.64 −2.07 −1.87 −2.02 −1.64 −1.87 −1.04 −1.98 −1.22 −1.87 −8.42 −2.22 F MAS1 MMP12 MMP9 NCAM1 NDRG1 NOS3 NTRK2 PAPPA2 PDGFD PGF PGR QPCT −1.87 −1.8 −2.07 −1.42 −2.59 −1.76 −2.67 −2.31 −1.87 1.21 −1.87 −2.54 G SERPINA3 SOD1 SPP1 STAT1 TAC1 TAC3 TEK TGFB1 TNF TREM1 VCAN VEGFA −2.17 −1 −1.64 −1.08 −2.54 −1.87 −1.87 1.26 −5.13 −1.71 −1.87 1.22

01 02 03 04 05 06 07 08 09 10 11 12 A

−3.073 0 3.073 Magnitude of log 2 (fold change)

Figure 3: Heat map representation for the differential expression of genes associated with PE after 4 h exposure toH2O2 in villous 3A trophoblasts. Gene expression is represented in the heat map in the color scale of −3.073–3.073 in green-red color scheme (𝑛=2). Genes evaluated and their locations on the heat map are depicted in the associated table above the heat map. the HBEGF gene was found to be upregulated while the other the variation in sensitivity of the cells studied to determine 21 genes were downregulated (Table 3). H2O2 toxic effects. The villous trophoblast cell line 3A used 2 We used the Qiagen online RT Profiler PCR Array Data in our study, which is a first trimester placental cell line, Analysis v3.5 to determine the correlation between the altered seems to be more sensitive to H2O2 damage compared to miRNAs and mRNAs in our study. Our analysis indicates that HTR-8/SVneo cells or term primary cells used in these 53 of the overexpressed miRNAs in our study putatively target investigations [5, 26]. Our studies were carried out using 5% 11 of the downregulated genes (Table 4). of the IC50 of H2O2 (25 𝜇M at 4 h) which is a subcytotoxic concentration. This level of exposure was chosen to mimic 4. Discussion H2O2 concentration in PE placentas previously reported by Zhou et al. [5]. Given the reported duplicitous nature of ROS In the current study, H2O2 exerted a cytotoxic effect in depending on their levels [28], we evaluated the oxidative 3A trophoblasts in a concentration- and time-dependent stress status of the cells at this exposure level by assessing GSH manner. These data are consistent with previous reports [5, level and SOD activity in the challenged cells as markers of 26]. However, the median inhibitory concentration (IC50)of oxidative stress [22, 23]. Both GSH and SOD are of major H2O2 in our study after 48 h exposure was 15 𝜇Mwhichis importance in intracellular redox regulation [29, 30]. Our 10 times lower than the IC50 reported by Zhou et al. using results revealed a significant reduction in GSH level and SOD HTR-8/SVneo cells [5].Inaseparatestudy,Molletal.[26] activity indicating oxidative stress at this exposure level. used H2O2 concentrations up to 1000 𝜇M, which is 66 times Reports indicate that miRNAs regulate migration, inva- higher than our IC50 concentration, to evaluate apoptosis sion,apoptosis,andproliferationoftrophoblastsaswellas and proliferation in human term placentas. Murata et al. [27] angiogenesis within the placenta, although the functions of used a concentration of 100 𝜇MH2O2 to determine apoptotic only a few miRNAs have been characterized [18]. Expression and invasion rates in term extravillous trophoblasts at 24 changesofmiRNAsthatregulatethesefunctionshavebeen hours. This concentration is comparable to our IC50 of 90 𝜇M reported in PE. Wang et al. [20] demonstrated an increase at 24 hours. This variability in results can be attributed to in miR-17, -20a, and -20b in PE. These miRNAs target 6 BioMed Research International

Table3:mRNAsalteredbyatleasttwofoldafter4hH2O2 exposure in villous 3A trophoblasts. Gene symbol Gene description Fold changeb HBEGFa Heparin-binding EGF-like growth factor 5.7637 ANGPT2 Angiopoietin 2 −2.4932 CCL2 Chemokine (C-C motif) ligand 2 −2.3834 CD40LG CD40 ligand −2.2084 CP Ceruloplasmin (ferroxidase) −2.2705 CRH Corticotropin releasing hormone −2.1038 DCN Decorin −2.0042 HP Haptoglobin −2.4083 IGF1 Insulin-like growth factor 1 (somatomedin C) −2.5991 IL18 Interleukin 18 (interferon-gamma-inducing factor) −2.6445 IL1A Interleukin 1, alpha −2.0748 IL6 Interleukin 6 (interferon, beta 2) −2.0181 LEP Leptin −8.4152 LPL Lipoprotein lipase −2.2238 MMP9 Matrix metallopeptidase 9 (gelatinase B, 92 kDa gelatinase, 92 kDa type IV collagenase) −2.0748 NDRG1 N-myc downstream regulated 1 −2.5901 NTRK2 Neurotrophic tyrosine kinase, receptor, type 2 −2.6721 PAPPA2 Pappalysin 2 −2.3102 QPCT Glutaminyl-peptide cyclotransferase −2.5368 SERPINA3 Serpin peptidase inhibitor, clade A (alpha-1 antiproteinase, antitrypsin), member 3 −2.1705 TAC1 Tachykinin, precursor 1 −2.5368 TNF Tumor necrosis factor −5.1266 aUpregulated gene is in bold font. b Fold change in H2O2 treated cells compared to nontreated control. the angiogenesis factors EPHB4 and ephrin-B2 and could knowntobecell-andtrimester-specific[14, 41]. In addition, be responsible for the decreases in angiogenesis seen in PE PE often occurs concurrently with IUGR (intrauterine grown [20]. Upregulation of miR-29b results in increased apoptosis restriction) and preterm labor, both of which have their own with a corresponding decrease in invasion and angiogenesis miRNA pattern which can thus produce conflicting data. in trophoblast cells [31]. MiR-155 targets cyclin D1 and the The placental mammal specific miRNA cluster on chromo- angiogenic regulating factor CYR61; both are downregulated some 14 (C14MC) predominantly expressed in placenta and in PE [32, 33]. Overexpression of miR-210 in extravillous embryonic tissues and maternally imprinted had 21 out of 30 trophoblasts in culture results in decreased cell invasion [34]. (70%) of evaluable miRNAs altered in our study. Meanwhile, MiR-376c, which is lower in both placenta and serum of expression of the paternally imprinted C19MC cluster was not preeclamptic women, reduces protein levels of ALK5 and detected. This expression pattern is in agreement with that ALK7, thus increasing cell proliferation [35]. found by Morales-Prieto et al. [14], showing higher expres- Inourstudy,after4hofH2O2 exposure, expression of 195 sion of the C14 cluster in the first trimester with very little of evaluable miRNAs (CT < 35) was altered by at least twofold. expression of C19 miRNAs. Although there are some data Of these, only seven were downregulated. We chose to focus indicating an increase in C19MC miRNAs in PE, this increase on the short-term exposure to H2O2 for the determination is most likely due to a loss of methylation as this miRNA of miRNA profile as other studies show a rapid (<4h) cluster is paternally imprinted [15]. Evidence suggests that miRNA expression response and subsequent downregulation hypomethylation of multiple genes contributes to early onset of protein targets [36]. Our data are in partial agreement with PE [42]. Data is mostly lacking for the role of the C14MC published studies that examined miRNA changes in PE, with cluster in PE. MiR-483-5p was not altered while miR-377 was some conflicts. There were 6 miRNAs reported to be either significantly upregulated in one study of preeclamptic women up- or downregulated in PE that were not altered in our [16]. Yet, neither of these miRNAs could be evaluated in our system. These include miR-210 [37], -34a [38], -149 [19], -19b study due to the high cycle number in our assay (CT > 35). [39], -92b, and -197 [17]. Two miRNAs, miR-194 and 195, were The expression profile of 84 genes known or suspected to upregulated in our system in contrast to reported downregu- bealteredinPEwasexaminedinthisstudy(Figure 3). These lation in the literature [19, 40]. The discrepancy in results is genes are involved in different pathways including pregnancy not surprising since most studies were conducted on term maintenance, oxidative stress, hormones, growth factors, placental tissue while we used a first trimester cytotro- endothelial cell signaling, and signal transduction. As shown phoblast cell line. Expressions of placental miRNAs are in Table 3,ofthese84genes,22wereup-ordownregulated BioMed Research International 7

Table 4: Altered miRNAs with putative altered mRNAs targets after 4 h exposure to2 H O2 in villous 3A trophoblasts. miRNA name Number of genes Number of target sites Range of strength scoresa Target genes targeted by this miRNA identified in target genes hsa-miR-181a-5p 33−0.4593 to −0.234 DCN, IL1A, TNF hsa-miR-181c-5p 33−0.4573 to −0.219 DCN, IL1A, TNF hsa-miR-181b-5p 33−0.4593 to −0.234 DCN, IL1A, TNF hsa-miR-181d-5p 33−0.4573 to −0.234 DCN, IL1A, TNF hsa-miR-27b-3p 33−0.203 to −0.0872 IGF1, LEP, LPL hsa-miR-29a-3p 34−0.4102 to −0.094 IGF1,LEP,LPL hsa-miR-29c-3p 34−0.4102 to −0.09 IGF1, LEP, LPL hsa-miR-875-3p 34−0.504 to −0.2443 CD40LG, NDRG1, TNF hsa-miR-9-5p 33−0.453 to −0.0467 IGF1, LEP, NDRG1 hsa-miR-454-3p 23−0.3152 to −0.219 IGF1, TNF hsa-miR-543 22−0.39 to −0.259 ANGPT2, IL1A hsa-miR-30e-5p 22−0.452 to −0.1342 IGF1, IL1A hsa-miR-30b-5p 22−0.452 to −0.1236 IGF1, IL1A hsa-miR-30a-5p 22−0.452 to −0.1342 IGF1, IL1A hsa-miR-30c-5p 22−0.452 to −0.1236 IGF1, IL1A hsa-miR-30d-5p 22−0.452 to −0.1342 IGF1, IL1A hsa-miR-424-5p 11−0.0766 to −0.0766 IGF1 hsa-miR-152-3p 11−0.3972 to −0.3972 IGF1 hsa-miR-148b-3p 11−0.3972 to −0.3972 IGF1 hsa-miR-425-5p 11−0.4961 to −0.4961 IGF1 hsa-miR-16-5p 11−0.0977 to −0.0977 IGF1 hsa-miR-497-5p 11−0.0766 to −0.0766 IGF1 hsa-miR-490-5p 11−0.197 to −0.197 NTRK2 hsa-miR-193a-5p 11−0.115 to −0.115 NTRK2 hsa-miR-151a-5p 11−0.401 to −0.401 NTRK2 hsa-miR-330-3p 12−0.05 to −0.01 NTRK2 hsa-miR-17-5p 12−0.149 to −0.148 NTRK2 hsa-miR-1271-5p 11−0.108 to −0.108 NDRG1 hsa-miR-20a-5p 12−0.17 to −0.148 NTRK2 hsa-miR-154-5p 11−0.263 to −0.263 NTRK2 hsa-miR-26b-5p 11−0.1205 to −0.1205 IGF1 hsa-miR-766-3p 11−0.161 to −0.161 IGF1 hsa-miR-190a-5p 11−0.1862 to −0.1862 IGF1 hsa-miR-452-5p 11−0.3111 to −0.3111 IGF1 hsa-miR-320a 11−0.1466 to −0.1466 IGF1 hsa-miR-320b 11−0.1466 to −0.1466 IGF1 hsa-miR-18b-5p 11−0.2185 to −0.2185 IGF1 hsa-miR-374a-5p 11−0.511 to −0.511 CCL2 hsa-miR-26a-5p 11−0.105 to −0.105 IGF1 hsa-miR-222-3p 11−0.1612 to −0.1612 IGF1 hsa-miR-221-3p 11−0.1612 to −0.1612 IGF1 hsa-miR-192-5p 11−0.2211 to −0.2211 IGF1 hsa-miR-196a-5p 11−0.0267 to −0.0267 IGF1 hsa-let-7f-5p 11−0.1633 to −0.1633 IGF1 hsa-let-7a-5p 11−0.1592 to −0.1592 IGF1 8 BioMed Research International

Table 4: Continued. miRNA name Number of genes Number of target sites Range of strength scoresa Target genes targeted by this miRNA identified in target genes hsa-miR-196b-5p 11−0.0267 to −0.0267 IGF1 hsa-let-7b-5p 11−0.1592 to −0.1592 IGF1 hsa-let-7i-5p 11−0.1592 to −0.1592 IGF1 hsa-let-7e-5p 11−0.1592 to −0.1592 IGF1 hsa-let-7d-5p 11−0.1612 to −0.1612 IGF1 hsa-let-7c-5p 11−0.1592 to −0.1592 IGF1 hsa-let-7g-5p 11−0.1592 to −0.1592 IGF1 hsa-miR-576-5p 11−0.1892 to −0.1892 IGF1 aStrength scores are the 𝑍 scores derived from the TargetScan algorithm. A more negative number indicates a stronger score and an increased likelihood that the gene is a bona fide target for the miRNA evaluated.

by at least twofold after H2O2 exposure. While it is difficult The levels of matrix metalloproteinase-9 (MMP9) were to provide mechanistic explanations to all changes observed decreased in trophoblast cells exposed to H2O2. MMP9 is in the expression profiles of the 22 genes, it is possible to an important factor for extracellular matrix remodeling and provide some explanations based on the putative function of is responsible for the invasiveness of trophoblasts [53]. Its these genes in PE and through the correlation of our miRNA deficiency induced phenotypic changes that mimic PEin and mRNA data. For example, expression of Heparin-binding mice [54] and its levels are reduced in pregnancies com- EGF-like growth factor (HBEGF) was upregulated by 5.8- plicated with PE [55]. Moreover, the level of corticotrophin fold in cells challenged with H2O2. HBEGF is expressed in releasing hormone (CRH) was also downregulated by H2O2 both villous and extravillous trophoblasts through the first stress. CRH is released by trophoblasts to promote embryo trimester [43]andisknowntoactasasurvivalfactorthat implantation [56]. hampers apoptosis [44] triggered by oxidative stress or other Leptin (LEP) gene expression was downregulated in our factors [26, 45]. Therefore, its induction under the current study. In silico analysis predicted the LEP to be a target experimental conditions could be a cellular defense against of four miRNAs (mir-27b-3p, mir-29a-3p, mir-29c-3p, and oxidative stress known to be associated with PE. mir-9-5p; Table 4). All of those miRNAs were overexpressed Our data showed that the expression of monocyte che- under the current experimental conditions. Our findings moattractant protein-1 (CCL2) was downregulated, while the contrast other reports indicating that LEP is overexpressed expression of its putative regulator mir-374a-5p was upreg- in placental tissues at term in PE [57]. It should be noted ulated. CCL2 is known to be expressed in first trimester tro- howeverthatmostpublisheddataisderivedfromstudiesof phoblasts [46] and is responsible for the activation and term trophoblasts or placental tissues while our study focuses recruitment of macrophages to the developing placenta to aid on first trimester trophoblasts, which mimics the early stage in tissue remodeling after implantation47 [ ]. The expression of pregnancy. This can explain some of the discrepancies in of CCL2 in first trimester trophoblasts was reported to be reg- gene expression data observed between our study and others. ulated by tumor necrosis factor (TNF-𝛼)[48]. In our study, Our study has some limitations which we acknowledge. expression level of TNF-𝛼 was also downregulated as was We examined only short-term effects of high H2O2 exposure angiopoietin-2 (ANGPT2) when exposed to increased oxida- on cell viability and short-term effects of low H2O2 expo- tive stress. ANGPT2 is crucial in regulating vascular remodel- sure on miRNA and mRNA gene expression profiles. Gene ing through its interaction with endothelial cell Tie-2 receptor expression changes can differ due to hypoxic versus normal [49]. The levels of ANGPT2 were found to be lower in PE growth conditions or short- versus long-term exposure. This compared to normal pregnancy [50]. Similarly, we found study focuses only on short-term exposure and cytotoxicity the insulin-like growth factor-1 (IGF-1) expression to be due to the reported rapid changes in miRNA expression after decreased.IGF-1wasreportedtobedecreasedinplacentaltis- initial exposure [36] but will be expanded in the future to sues from women suffering PE51 [ ]. Of note is that 38 miRNAs include long-term genetic changes. In addition, a change in that putatively regulate IGF-1 (Table 4) were upregulated miRNA or mRNA expression does not necessarily lead to a in our study. The pregnancy-associated plasma protein A2 change in encoded protein levels, protein modification, or (PAPPA2) is an insulin-like growth factor-binding protein protein function. Such information would be generated from (IGFBP) protease expressed at high levels in the placenta. specifically designed studies based on our findings. Lastly, we IncreasedlevelsofPAPPA2inPEsuggestacompensatory are using an SV40-transformed first trimester cell line rather response to abnormal placentation, which might increase than primary cells or choriocarcinoma cancer cells, which, insulin-like growth factor (IGF) availability and promote to our knowledge, is one of the best available models for fetoplacental growth [52]. However, PAPPA2 was downreg- studies such as ours. Although the in vitro approach provides ulated in our study, which might also explain the observed conformity of cell type over multiple experiments, in vitro reduction in IGF-1 expression in the current work. response to ROS may not totally mimic the in vivo response. BioMed Research International 9

In summary, our data indicate that short-term exposure [9]G.S.Prins,L.Birch,W.Y.Tang,andS.-M.Ho,“Developmental of 3A villous first trimester trophoblasts to H2O2 signifi- estrogen exposures predispose to prostate carcinogenesis with cantly alters miRNA profile and mRNA expression of genes aging,” Reproductive Toxicology,vol.23,no.3,pp.374–382,2007. implicated in defective placental development. Our data, [10] A. B. Wallis, A. F. Saftlas, J. Hsia, and H. K. Atrash, “Secular which indicate that oxidative stress alters miRNAs and RNAs trends in the rates of preeclampsia, eclampsia, and gestational expression, could partially explain some of the early changes hypertension, United States, 1987–2004,” The American Journal in gene expression profiles and miRNA observed in PE. of Hypertension,vol.21,no.5,pp.521–526,2008. [11] R. I. Gregory, T. P. Chendrimada, and R. Shiekhattar, “Mic- roRNA biogenesis: isolation and characterization of the micro- Conflict of Interests processor complex,” Methods in Molecular Biology,vol.342,pp. 33–47, 2006. The authors declare that there is no conflict of interests [12] L. He and G. J. Hannon, “MicroRNAs: small RNAs with a big regarding the publication of this paper. role in gene regulation,” Nature Reviews Genetics,vol.5,no.7, pp. 522–531, 2004. [13] J.-F. Mouillet, T. Chu, and Y. Sadovsky, “Expression patterns of Authors’ Contribution placental microRNAs,” Birth Defects Research, Part A—Clinical Courtney E. Cross and Mai F. Tolba contributed equally to and Molecular Teratology,vol.91,no.8,pp.737–743,2011. this paper. [14] D. M. Morales-Prieto, W. Chaiwangyen, S. Ospina-Prieto et al., “MicroRNA expression profiles of trophoblastic cells,” Placenta, vol.33,no.9,pp.725–734,2012. Acknowledgments [15]S.Buckberry,T.Bianco-Miotto,andC.T.Roberts,“Imprinted and X-linked non-coding RNAs as potential regulators of This work was supported by the National Institute of Environ- human placental function,” Epigenetics,vol.9,no.1,pp.81–89, mental Health Sciences (T32ES007254 to Courtney E. Cross 2014. and Catherine M. Rondelli) and the John Sealy Memorial [16] K. Mayor-Lynn, T. Toloubeydokhti, A. C. Cruz, and N. Chegini, Endowment Fund for Biomedical Research. Additional sup- “Expression profile of microRNAs and mRNAs in human port was provided by the NIEHS Center in Environmental placentas from pregnancies complicated by preeclampsia and Toxicology at the University of Texas Medical Branch funded preterm labor,” Reproductive Sciences,vol.18,no.1,pp.46–56, through P30 ES006676. 2011. [17] S.-Y. Choi, J. Yun, O.-J. Lee et al., “MicroRNA expression profiles in placenta with severe preeclampsia using a PNA-based References microarray,” Placenta, vol. 34, no. 9, pp. 799–804, 2013. [18] G. Fu, J. Brkic,´ H. Hayder, and C. Peng, “MicroRNAs in [1] E. A. P. Steegers, P. von Dadelszen, J. J. Duvekot, and R. Pijnen- human placental development and pregnancy complications,” borg, “Pre-eclampsia,” The Lancet,vol.376,no.9741,pp.631– International Journal of Molecular Sciences,vol.14,no.3,pp. 644, 2010. 5519–5544, 2013. [2]E.A.Veal,A.M.Day,andB.A.Morgan,“Hydrogenperoxide [19] L. Guo, S. Q. Tsai, N. E. Hardison et al., “Differentially expressed sensing and signaling,” Molecular Cell,vol.26,no.1,pp.1–14, microRNAs and affected biological pathways revealed by 2007. modulated modularity clustering (MMC) analysis of human [3] M. Emberton, G. L. Andriole, J. de La Rosette et al., “Benign preeclamptic and IUGR placentas,” Placenta,vol.34,no.7,pp. prostatic hyperplasia: a progressive disease of aging men,” 599–605, 2013. Urology,vol.61,no.2,pp.267–273,2003. [20]W.Wang,L.Feng,H.Zhangetal.,“Preeclampsiaup-regulates [4]M.S.SteinerandS.Raghow,“Antiestrogensandselective angiogenesis-associated microRNA (i.e., miR-17, -20a, and - estrogen receptor modulators reduce prostate cancer risk,” 20b) that target ephrin-B2 and EPHB4 in human placenta,” The World Journal of Urology,vol.21,no.1,pp.31–36,2003. Journal of Clinical Endocrinology & Metabolism,vol.97,no.6, pp. E1051–E1059, 2012. [5]X.Zhou,G.-Y.Zhang,J.Wang,S.-L.Lu,J.Cao,andL.-Z.Sun, [21] P. Skehan, R. Storeng, D. Scudiero et al., “New colorimetric “A novel bridge between oxidative stress and immunity: the cytotoxicity assay for anticancer-drug screening,” Journal of the interaction between hydrogen peroxide and human leukocyte National Cancer Institute,vol.82,no.13,pp.1107–1112,1990. antigen G in placental trophoblasts during preeclampsia,” The [22] C. Monteiller, L. Tran, W.MacNee et al., “The pro-inflammatory American Journal of Obstetrics and Gynecology,vol.206,no.5, effects of low-toxicity low-solubility particles, nanoparticles and pp. 447.e7–447.e16, 2012. fineparticles,onepithelialcellsinvitro:theroleofsurfacearea,” [6] G.S.PrinsandK.S.Korach,“Theroleofestrogensandestrogen Occupational and Environmental Medicine,vol.64,no.9,pp. receptors in normal prostate growth and disease,” Steroids,vol. 609–615, 2007. 73,no.3,pp.233–244,2008. [23] M. K. Kale, “Thyroid gland in free radical induced oxidative [7]G.P.Risbridger,J.J.Bianco,S.J.Ellem,andS.J.McPherson, stress,” in Free Radicals in Human Health and Disease,Springer, “Oestrogens and prostate cancer,” Endocrine-Related Cancer, New Delhi, India, 2015. vol.10,no.2,pp.187–191,2003. [24] G. L. Ellman, “Tissue sulfhydryl groups,” Archives of Biochem- [8] Q. Wu, J. Shi, L. Chen et al., “Regulation of proliferation istry and Biophysics,vol.82,no.1,pp.70–77,1959. and differentiation of prostatic stromal cells by oestradiol [25]M.Nishikimi,N.A.Roa,andK.Yogi,“Measurementof through prostatic epithelial cells in a paracrine manner,” BJU superoxide dismutase,” Biochemical and Biophysical Research International,vol.101,no.4,pp.497–502,2008. Communications,vol.46,p.849,1972. 10 BioMed Research International

[26]S.J.Moll,C.J.P.Jones,I.P.Crocker,P.N.Baker,andA.E.P. [42] J. D. Blair, R. K. C. Yuen, B. K. Lim, D. E. McFadden, P. von Heazell, “Epidermal growth factor rescues trophoblast apopto- Dadelszen, and W.P.Robinson, “Widespread DNA hypomethy- sis induced by reactive oxygen species,” Apoptosis,vol.12,no.9, lation at gene enhancer regions in placentas associated with pp. 1611–1622, 2007. early-onset pre-eclampsia,” Molecular Human Reproduction, [27]M.Murata,K.Fukushima,T.Takao,H.Seki,S.Takeda,andN. vol. 19, no. 10, pp. 697–708, 2013. Wake, “Oxidative stress produced by xanthine oxidase induces [43] R. E. Leach, R. Khalifa, N. D. Ramirez et al., “Multiple roles apoptosis in human extravillous trophoblast cells,” The Journal for heparin-binding epidermal growth factor-like growth factor of Reproduction and Development,vol.59,no.1,pp.7–13,2013. are suggested by its cell-specific expression during the human [28] J. D. Lambeth and A. S. Neish, “Nox enzymes and new thinking endometrial cycle and early placentation,” Journal of Clinical on reactive oxygen: a double-edged sword revisited,” Annual Endocrinology and Metabolism,vol.84,no.9,pp.3355–3363, Review of Pathology,vol.9,pp.119–145,2014. 1999. [29] A. V.Peskin, “Cu,Zn-superoxide dismutase gene dosage and cell [44] M. P.Michalsky, A. Kuhn, V.Mehta, and G. E. Besner, “Heparin- resistance to oxidative stress: a review,” Bioscience Reports,vol. binding EGF-like growth factor decreases apoptosis in intesti- 17,no.1,pp.85–89,1997. nal epithelial cells in vitro,” Journal of Pediatric Surgery,vol.36, [30] H. J. Forman, H. Zhang, and A. Rinna, “Glutathione: overview no. 8, pp. 1130–1135, 2001. of its protective roles, measurement, and biosynthesis,” Molecu- [45] M. Mackova, R. T. Kilani, S. T. Davidge, and L. J. Guilbert, “The lar Aspects of Medicine,vol.30,no.1-2,pp.1–12,2009. effect of oxygen tension on intracellular survival signalling in [31]P.Li,W.Guo,L.Duetal.,“MicroRNA-29bcontributesto primary villous trophoblasts,” Placenta,vol.24,no.6,pp.627– pre-eclampsia through its effects on apoptosis, invasion and 637, 2003. angiogenesis of trophoblast cells,” Clinical Science,vol.124,no. [46] H. O. D. Critchley, R. L. Jones, R. G. Lea et al., “Role of 1,pp.27–40,2013. inflammatory mediators in human endometrium during pro- [32] Y. Dai, Z. Qiu, Z. Diao et al., “MicroRNA-155 inhibits prolifera- gesterone withdrawal and early pregnancy,” Journal of Clinical tion and migration of human extravillous trophoblast derived Endocrinology and Metabolism, vol. 84, no. 1, pp. 240–248, 1999. HTR-8/SVneo cells via down-regulating cyclin D1,” Placenta, [47]V.M.Abrahams,Y.M.Kim,S.L.Straszewski,R.Romero, vol.33,no.10,pp.824–829,2012. and G. Mor, “Macrophages and apoptotic cell clearance during [33]Y.Zhang,Z.Diao,L.Suetal.,“MicroRNA-155contributesto pregnancy,” American Journal of Reproductive Immunology,vol. preeclampsia by down-regulating CYR61,” American Journal of 51, no. 4, pp. 275–282, 2004. Obstetrics & Gynecology, vol. 202, no. 5, pp. 466.e1–466.e7, 2010. [48] C.J.Lockwood,P.Matta,G.Krikunetal.,“Regulationofmono- [34] L. Anton, A. O. Olarerin-George, N. Schwartz et al., “miR- cyte chemoattractant protein-1 expression by tumor necrosis 210 inhibits trophoblast invasion and is a serum biomarker for factor-alpha and interleukin-1beta in first trimester human preeclampsia,” The American Journal of Pathology,vol.183,no. decidual cells: implications for preeclampsia,” The American 5, pp. 1437–1445, 2013. Journal of Pathology,vol.168,no.2,pp.445–452,2006. [35] G. Fu, G. Ye, L. Nadeem et al., “MicroRNA-376c impairs trans- [49] P. C. Maisonpierre, C. Suri, P. F. Jones et al., “Angiopoietin-2, a forming growth factor-𝛽 and nodal signaling to promote natural antagonist for Tie2 that disrupts in vivo angiogenesis,” trophoblast cell proliferation and invasion,” Hypertension,vol. Science,vol.277,no.5322,pp.55–60,1997. 61,no.4,pp.864–872,2013. [50] E. G. Zhang, S. K. Smith, P.N. Baker, and D. S. Charnock-Jones, [36]L.S.Chan,P.Y.K.Yue,Y.Y.Wong,andR.N.S.Wong,“Mic- “The regulation and localization of angiopoietin-1, -2, and their roRNA-15b contributes to ginsenoside-Rg1-induced angiogen- receptor Tie2 in normal and pathologic human placentae,” esis through increased expression of VEGFR-2,” Biochemical Molecular Medicine,vol.7,no.9,pp.624–635,2001. Pharmacology,vol.86,no.3,pp.392–400,2013. [51] E. A. Dubova, K. A. Pavlov, V. M. Lyapin, G. V. Kulikova, [37] B. L. Pineles, R. Romero, D. Montenegro et al., “Distinct A. I. Shchyogolev, and G. T. Sukhikh, “Expression of insulin- subsets of microRNAs are expressed differentially in the human like growth factors in the placenta in preeclampsia,” Bulletin of placentas of patients with preeclampsia,” American Journal of Experimental Biology and Medicine,vol.157,no.1,pp.103–107, Obstetrics and Gynecology,vol.196,no.3,pp.261.e1–261.e6, 2014. 2007. [52] J. K. Christians and A. Gruslin, “Altered levels of insulin-like [38] L. Doridot, D. Houry, H. Gaillard, S. T. Chelbi, S. Barbaux, and growth factor binding protein proteases in preeclampsia and D. Vaiman, “miR-34A expression, epigenetic regulation, and intrauterine growth restriction,” Prenatal Diagnosis,vol.30,no. function in human placental diseases,” Epigenetics,vol.9,no.1, 9,pp.815–820,2010. pp. 142–151, 2014. [53] Y.Yang,J.Zhang,Y.Gongetal.,“Increasedexpressionofprosta- [39] P. Kumar, Y. Luo, C. Tudela, J. M. Alexander, and C. R. sin contributes to early-onset severe preeclampsia through inhib- Mendelson, “The c-myc-regulated microRNA-17∼92 (mir-17∼ iting trophoblast invasion,” Journal of Perinatology,vol.35,pp. 92) and miR-106a∼363 clusters target hCYP19A1 and hGCM1 16–22, 2015. to inhibit human trophoblast differentiation,” Molecular and [54] V. Plaks, J. Rinkenberger, J. Dai et al., “Matrix metalloprote- Cellular Biology,vol.33,no.9,pp.1782–1796,2013. inase-9 deficiency phenocopies features of preeclampsia and [40] Y. Bai, W. Yang, H.-X. Yang et al., “Downregulated miR-195 intrauterine growth restriction,” Proceedings of the National detected in preeclamptic placenta affects trophoblast cell inva- Academy of Sciences of the United States of America,vol.110,no. sion via modulating ActRIIA expression,” PLoS ONE,vol.7,no. 27, pp. 11109–11114, 2013. 6, Article ID e38875, 2012. [55] C. H. Graham and K. R. McCrae, “Altered expression of gelat- [41] D. Wang, W. Song, and Q. Na, “The emerging roles of placenta- inase and surface-associated plasminogen activator activity specific microRNAs in regulating trophoblast proliferation dur- by trophoblast cells isolated from placentas of preeclamptic ing the first trimester,” The Australian & New Zealand Journal of patients,” The American Journal of Obstetrics and Gynecology, Obstetrics & Gynaecology, vol. 52, no. 6, pp. 565–570, 2012. vol. 175, no. 3, pp. 555–562, 1996. BioMed Research International 11

[56] A. M. Bamberger, V. Minas, S. N. Kalantaridou et al., “Corti- cotropin-releasing hormone modulates human trophoblast invasion through carcinoembryonic antigen-related cell adhesion molecule-1 regulation,” The American Journal of Pathology, vol. 168, no. 1, pp. 141–150, 2006. [57] V. Sitras, R. H. Paulssen, H. Grønaas et al., “Differential placental gene expression in severe preeclampsia,” Placenta,vol.30, no. 5, pp. 424–433, 2009. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 698795, 11 pages http://dx.doi.org/10.1155/2015/698795

Review Article Bisphenol A Effects on Mammalian Oogenesis and Epigenetic Integrity of Oocytes: A Case Study Exploring Risks of Endocrine Disrupting Chemicals

Ursula Eichenlaub-Ritter1 and Francesca Pacchierotti2 1 Faculty of Biology, Gene Technology/Microbiology, University of Bielefeld, 33601 Bielefeld, Germany 2Laboratory of Toxicology, Unit of Radiation Biology and Human Health, ENEA CR Casaccia, 00123SantaMariadiGaleria,Rome,Italy

Correspondence should be addressed to Ursula Eichenlaub-Ritter; [email protected]

Received 16 January 2015; Revised 5 May 2015; Accepted 5 May 2015

Academic Editor: Dong-Wook Han

Copyright © 2015 U. Eichenlaub-Ritter and F. Pacchierotti. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Bisphenol A (BPA), originally developed as a synthetic oestrogen, is nowadays extensively used in the production of polymeric plastics. Under harsh conditions, these plastics may release BPA, which then can leach into the environment. Detectable concentrations of BPA have been measured in most analysed samples of human serum, plasma, or urine, as well as in follicular fluid, foetal serum, and amniotic fluid. Here we summarize the evidence about adverse BPA effects on the genetic and epigenetic integrity of mammalian oocytes. We conclude that increasing evidence supports the notion that low BPA concentrations adversely affect the epigenome of mammalian female germ cells, with functional consequences on gene expression, chromosome dynamics in meiosis, and oocyte development. Specific time windows, during which profound chromatin remodelling occurs and maternal imprints are established or protected, appear particularly vulnerable to epigenetic deregulation by BPA. Transgenerational effects have been also observed in the offspring of BPA-treated rodents, although the epigenetic mechanisms of inheritance still need to be clarified. The relevance of these findings for human health protection still needs to be fully assessed, but they warrant further investigation in both experimental models and humans.

1. Introduction epigenetic regulation. Recently, it has been shown in rodents that in utero undernourishment alters the germline DNA According to the “Developmental Origin of Health and Dis- methylome of 𝐹1adultmalesinalocus-specificmanner. ease” hypothesis (DOHaD) that was proposed over a decade Although altered DNA methylation did not persist in 𝐹2tis- ago [1, 2], lifestyle, nutrition, and exposures during pregnancy sues, dysregulated expression of genes neighbouring affected can influence the health of the offspring from birth to much loci was observed, suggesting the possibility of intergenera- laterinlife.TheDOHaDhypothesisproposesthatfamine, tional transmission of environmentally induced disease not nutritional deficits, or diabetes in the mother can predispose mediated by Mendelian inheritance [6]. the offspring to diseases or reduce its fertility, especially if According to the US Environmental Protection Agency, exposures occur during critical periods of embryogenesis or endocrine disrupting chemicals (EDCs) are defined as foetal development [3–5]. Prenatal exposure of the develop- “exogenous agent(s) that interfere(s) in synthesis, secretion, ing germline may entail the additional risk to transmit the transport, metabolism, binding action, or elimination of nat- induced damage to the following generation. ural blood-borne hormones that are present in the body and Environmental exposures that affect metabolism or hor- are responsible for homeostasis, reproduction, and develop- monal homeostasis do not necessarily induce DNA muta- mental processes” [7]. As such they may transiently alter gene tions but may influence gene expression by disturbances in expression patterns in exposed cells, organs, and individuals 2 BioMed Research International

CH CH3 3 HO OH O O

CH3 CH3 (n) (n) Monomer bisphenol A, BPA; BPA polymer 4,4󳰀-(propane-2,2-diyl)diphenol

Figure 1: BPA monomer and polymer.

by interfering in hormonal homeostasis, for example, by measured in some areas of Great Britain [17]. BPA can be acting as agonist or antagonist in hormone receptor-mediated efficiently biodegraded in water and soil by microorganisms signalling. Moreover, it has recently been shown that expo- and by photolysis in water at wavelengths above 290 nm [17, sure to EDCs may induce transgenerational phenotype alter- 19, 20]. However, in spite of environmental biodegradation, ations, possibly caused by differential DNA methylation in 0.1–790 𝜇g/kg BPA were detected in fresh weight (f.w.) food 3 gene promoter regions, termed epimutations [8–12]. and up to 0.86 mg/m in drinking water or commercial drinks [17]; biomonitoring studies detected BPA in human serum, plasma, or urine of over 90% US and Canadian citizens 2. Bisphenol A [21]. Daily dietary BPA intakes of about 0.02–0.08 𝜇g/kg/day 󸀠 𝜇 Monomer bisphenol A (4,4 -(propane-2,2-diyl)diphenol) and 0.22–0.33 g/kg/day have been estimated for adults (BPA) (Figure 1) was developed in 1891 as a synthetic oestro- and infants, respectively (for references see [17]). BPA is gen (xenoestrogen). BPA indeed binds to oestrogen receptors rapidly metabolised to bisphenol A-glucuronide in liver, but in vivo and in vitro [13], but, due to its low oestrogenic activ- unconjugated BPA has been detected in serum and blood 3 5 ity (≈10 –10 less than the natural steroid, oestradiol), it of the general population at concentrations of 4.4 mg/L and was replaced by diethylstylbestrol (DES), which had much 2.5 mg/L, respectively, and over 50 mg/L BPA were measured stronger oestrogenic properties. DES is sadly known for the in workers [18, 22]. BPA has been also detected in follicular teratogenic and carcinogenic effects observed in the genital fluid, foetal serum, and amniotic fluid (average 1-2 ng/mL) organs of daughters of women using DES in pregnancy to [23] and in umbilical cord serum of human mid gestation prevent spontaneous abortion [14]. Recently, an epigenetic embryos [24]. influence of DES on the regulation of histone [15]andDNA [16] methyltransferases has been shown that could play a role 4. BPA Effects on Mammalian Oogenesis in the induction of its reproductive effects. The case of DES could be considered an alarming sentinel of the importance Although potential adverse effects of low BPA concentrations of epigenetic mechanisms in EDCs adverse effects. on reproduction are still a matter of debate, most studies Even though BPA was not marketed as a hormonal active suggest that BPA is an ovarian toxicant and reduces oocyte substance, in the last decades, it found application as plasti- quality in animal models and in humans [23, 25]. Poten- cizer in the production of polymeric plastics, mainly polycar- tial mechanisms of BPA action on hormonal homeostasis bonate (71%) and epoxy resins (29%) [17]. For a long time, include binding to nonclassic membrane oestrogen recep- polymericBPAwasconsideredharmless,asitdoesnotinter- tors (mERs), binding to glucuronide receptor, activation of 𝛾 act with steroid receptors (Figure 1). Over the last 50 years, nuclear oestrogen-related receptor gamma (ERR ), suppres- the use of BPA-containing polymers in common items such sion of thyroid hormone receptor transcription, decrease of as plastic bottles, toys, lining of aluminium cans and pipes, cholesterol transport through the mitochondrial membrane, dental sealants, and thermal receipt paper led to increasing increase of fatty acid oxidation, stimulation of prolactin BPA production which reached about 5 million tons in 2010 release, and impairment of aromatase expression (reviewed [17]. Unfortunately, polycarbonate plastics damaged by heat, in [18, 25–27]). UV, harsh alkaline treatment, or after vigorous washing were Mammalian oocytes are amongst the most long-lived shown to release monomeric BPA. By now, it is estimated cells in the body. Primordial germ cells start differentiating that the worldwide release of BPA into the environment is alreadyintheearlypostimplantationembryo,aftermigration exceeding one million pounds/year [18]. to the genital ridges (Figure 2)[28]. Nests of primary oocytes, entering meiosis I, are formed in the human ovary by the 3rd month of pregnancy. Pairing and recombination between 3. Environmental and homologous chromosomes take place in the foetal ovary Human BPA Contamination before birth. Around the 7th month of pregnancy, oocytes finally develop to the late dictyotene stage, when they become In the USA, average BPA groundwater concentrations range meiotically arrested. By that time, the synaptonemal com- 3 3 between 0.0041 and 1.9 mg/m ,andupto20mg/m BPA were plexes have disappeared and the homologous chromosomes BioMed Research International 3

OogenesisFertilization Embryogenesis

Imprinted genes Zfp57 Kap1 High Stella 5 󳰀 Passive mC Active AID 5 󳰀 →5 mC

󳰀 →5 Dnmt1 hmC Dnmt3a Dnmt1 Tet3 󳰀 C Global methylation Dnmt3b Dnmt3a →5 ELP3 CpG methylation CpG (maternal) Imprinted genes Dnmt3l AID 󳰀 C Nonimprinted genes Dnmt1− Low

PGC Immature GV Mature GV MII Zygote 2-cell Blastocyst

Growth Proliferation

Erasure Establishment Maintenance

Figure 2: Critical stages for epigenetic reprogramming of chromatin in mammalian female germ cells, including periods of imprint erasure during formation of primordial germ cells, maternal imprint establishment during oocyte growth, and imprint maintenance after fertilization of the egg and development to the blastocyst. Statuses of maternal imprints are indicated by red lines, of global methylation in maternal chromatin in dotted blue lines, of global paternal methylation in green dotted lines, and of development and tissue-specific methylation in orange dotted lines. Some enzymes that participate in demethylation (activation induced cytidine deaminase), DNA methylation (Dnmts), or maintenance of methylation (zinc finger protein 57, Zfp57; tripartite motive containing, Kap1/Trim28; developmental pluripotency associated 3, Stella/Dppa3) are indicated next to the respective lines showing changes in DNA methylation, adapted from [84].

remain attached by one or more meiotic exchanges and chi- layer of flattened granulosa cells around the dictyate-arrested asmata until much later in oogenesis when oocytes resume oocytes. Primordial follicles will develop to the secondary, meiosis in the sexually adult female; in the meantime the tertiary, and finally large antral stage only from puberty sister chromatids of each chromosome remain tightly linked onwards, when folliculogenesis proceeds under the influence by cohesion complexes (reviewed in [29]). of gonadotropic hormones (follicle stimulating hormone, Exposure of mice, from midgestation until birth, to daily FSH, and luteinizing hormone, LH). Mature follicles contain doses of 400 ng BPA resulted in synaptic abnormalities and multilayered outer mural granulosa cells and layers of cumu- increased rates of recombination between homologous chro- lus granulosa cells surrounding a fully grown, meiotically and mosomes in the oocytes. Interestingly, these effects resembled developmentally competent oocyte, within a large fluid-filled thoseobservedinmicehomozygousforatargeteddisruption space, the antrum. The follicle is surrounded by a basal mem- of the gene encoding for oestrogen receptor 𝛽 [30]. Increased brane and by layers of luteal cells that are involved in steroido- recombination was also observed in oocytes of rhesus mon- genesis. Growth of the oocyte and full development of the fol- keys prenatally exposed to BPA [31] and in human oocytes licle is a lengthy process that takes about 120 days in humans. treated in vitro with BPA [32]. Moreover, the expression of In rhesus monkeys, chronic exposure to BPA during preg- genes involved in recombination and DNA repair was altered nancy, leading to serum concentrations of 2.2–3.3 ng/mL, in the BPA-exposed human foetal oocytes [33]. It is known caused in the offspring a significant increase in the fre- that alterations in the number and localization of chiasmata quency of abnormal follicles containing multiple oocytes can adversely affect chromosome segregation and predispose [31]. Disturbances in nest breakdown and primordial follicle the oocytes to aneuploidy (reviewed in [29]). The studies on formationwerealsonotedinmiceuponin utero exposure altered recombination in BPA-exposed foetal oocytes there- [34]. In vitro experiments supported the notion that BPA fore suggest that the female offspring of BPA-exposed moth- impairs follicular development. Germ cell nest breakdown ers might be at risk for meiotic chromosome nondisjunction. and primordial follicle assembly were significantly reduced In addition to effects on chromosome synapsis in the when newborn mouse ovaries were exposed to 10 or 100 𝜇M oocytes, BPA was shown to induce also alterations of follicle BPA in culture medium [35]. Similarly, 100 𝜇g/mL BPA maturation. Normally, at birth, the oocyte nests break down (440 𝜇M) inhibited follicle growth and induced atresia in a and primordial follicles are formed by recruitment of a single mouse follicle in vitro model [36], by mechanisms that were 4 BioMed Research International independent of the genomic oestrogenic pathway [37]. Con- b.w. BPA, for 6–8 days prior to isolation and in vitro matura- tinuous 12-day exposure of mouse follicles in culture, from tion of oocytes, induced meiotic arrest, spindle abnormalities the early preantral up to the large antral stage, to 30 𝜇M and misalignment of metaphase II chromosomes. These BPA led to reduced granulosa cell proliferation and arrest experimental results supported the hypothesis that a sudden, of some oocytes at meiosis I [38], whereas lower BPA con- unexpected increase of aneuploid oocytes that had previously centrations did not significantly affect follicular development occurred in the mouse colony had been caused by accidental and hormone release. Finally, in vitro treatment of human release of BPA from damaged plastic bottles and cages. oocytes with low BPA concentrations was shown to increase An independent study, conducted under similar exposure the frequency of oocyte degeneration [32]. conditions to verify these findings, showed BPA induction Since the entire pool of primary oocytes that can ever be of subtle spindle abnormalities in metaphase II oocytes, but ovulated is already formed in the foetal ovary, disturbances not aneuploidy [48]. Similarly, no evidence of aneuploidy in oocyte meiosis and follicle formation, recruitment and induction was obtained in metaphase II oocytes collected survival, induced by any mechanism, can contribute to from mice treated with a single BPA dose, with 7 daily premature ovarian depletion, a clinically recognized condi- administrations or exposed for 7 weeks to BPA in drinking tion in women (termed “premature ovarian insufficiency,” water [49]. Differences in the animal diet were suggested POI). In addition, the fidelity of chromosome segregation at to explain these inconsistencies when it was shown that the first and second meiosis might be compromised when the phytoestrogen content in animal feed could influence the rate oocyte pool becomes prematurely depleted. Whether BPA of spindle aberrations induced in metaphase II oocytes by exposure may influence follicle pool size in humans is still 7 daily low dose administrations of BPA [50]. Other studies controversial. High urinary BPA levels were associated with showed an influence of the diet on BPA-induced changes in reducedantralfolliclecountsinacohortof209women DNA methylation [51], supporting an interaction between undergoing infertility treatments [39], whereas no correlation BPA biological activity and dietary factors. wasfoundbetweenserumBPAlevelsandantralfollicle In vitro experiments in mouse oocytes showed that high counts in another study on a smaller cohort of 44 patients concentrations of BPA induced spindle aberrations, chro- [40]. Nevertheless, several data suggest a negative impact of mosome congression abnormalities, and meiotic arrest, but BPA on woman fertility. Urinary BPA levels were negatively not aneuploidy [48, 52], suggesting that an efficient spindle correlated with numbers and quality of oocytes retrieved in assembly checkpoint was able to prevent chromosome segre- stimulated cycles for assisted reproduction [41, 42]. Increased gation errors in healthy young oocytes. An inverse relation- urinary [41, 42]orserum[43]BPAconcentrationswerealso ship between BPA concentration and percentage of oocytes associated with decreased peak oestradiol levels. Finally, a that progressed to metaphase II and a dose-dependent study on 137 patients undergoing assisted reproduction sug- increase in aberrant spindles and unaligned chromosomes at gested that high urinary BPA levels might be associated with metaphase II were also reported for human oocytes exposed up to 50% higher chance of implantation failures, in compari- in vitro to 20, 200 ng/mL, or 20 𝜇g/mL BPA (88, 880 nM, son to patients with low or no evidence of BPA exposure [44]. 𝜇 Once follicles have developed to the large antral, Graafian 88 M) [53]. stage,releaseofone(inmonoovulatoryspecies,likehumans) Fertilization triggers release of metaphase II oocytes from or of multiple (in multiovulatory species like rodents) oocytes second meiotic arrest. This entails second polar body extru- from meiotic arrest occurs under the influence of gona- sion and completion of oocyte second meiosis during which dotropins and the LH surge, downstream from signalling by sister chromatids separate from each other (reviewed in [29]). complex and redundant pathways. Resumption of meiosis I, There is a paucity of data about BPA effects on the second normally, results in gene expression changes and molecular meiotic division in oocytes. Chronic exposures of mice to signalling in cumulus cells and oocyte via epidermal growth 0.5 mg/L BPA in drinking water resulted in the premature factor-like hormones and critical changes in the concen- separation of sister chromatids in their metaphase II oocytes, tration of cyclic nucleotides. Fully grown, developmentally which, however, had no consequence upon the fidelity of competent oocytes then become transcriptionally quiescent, chromosome segregation during the second meiotic division, and their chromatin is remodelled to surround the nucleolus as demonstrated by the normal chromosome constitution of in a characteristic fashion [45]. Following the LH surge, zygotes under the same exposure condition [49]. the maturation promoting factor/cyclin-dependent kinase 1 pathway of the oocyte becomes activated, chromatin becomes condensed, and histones are characteristically deacetylated 5. BPA Epigenetic Effects on Female Germ and posttranslationally modified in specific ways [46]. A Cells and Their Consequences spindleisthenformedintheooplasm,andoocytescomplete the first meiosis with the reductional division of homologous Oogenesis, from primordial germ cell differentiation to ferti- chromosomes, reach the metaphase II stage, when they arrest lization, and preimplantation embryonic development entail again, and are finally ovulated, surrounded by the expanded profound epigenetic changes (Figure 2). After global DNA cumulus complex. demethylation in primordial germ cells, female specific In 2003, it was reported that low, chronic BPA doses might genomic imprinting is set during oocyte development in a induce aneuploidy in oocytes exposed prior to resumption of site-specific sequential fashion in imprinting control regions meiosis [47]. Oral treatment of mice with 20, 40, or 100 ng/g (ICRs). The whole process is completed prior to resumption BioMed Research International 5 of meiosis [54–56]. The zygote and the early preimplanta- develop in the ovary. Chao and coworkers exposed CD-1 tion embryo are further subjected to extensive chromatin mice to low doses of BPA (20 or 40 𝜇g/kg b.w.) either by remodelling and DNA methylation changes: active hydrox- daily hypodermal injections from postnatal day 7 to postnatal ymethylation and global DNA demethylation in the male day 14 or by intraperitoneal injections administered each chromatin and passive global demethylation in the female fifth day between postnatal days 5 and66 20[ ]. BPA not chromatin (Figure 2). Most of the enzymes for these events only dose-dependently inhibited methylation of Igf2r and are maternally provided by the oocyte before full zygotic Peg3 differentially methylated regions, but also lowered the gene activation. The remodelling and epigenetic changes transcription of DNA methyltransferase Dmnt1, Dmnt3a, in male and female chromatin proceed according to a Dmnt3b and Dnmt3l genes in the oocytes. Since the oestrogen highly regulated sex-specific program [57], which spares the receptor (ER) may recruit coactivator complexes with histone removal of DNA methylation on genomic imprints from acetyltransferase or methyltransferase activities to activate sperm and oocyte that are important for normal preimplan- downstream target genes [67], Chao and coworkers examined tation development and are retained in tissues to regulate the expression of ERs and found a significant increase in monoallelic gene expression from paternal or maternal alleles ER𝛼 mRNAandproteinlevelsatthehighestBPAdose.ER [58, 59]. Disturbances in genomic imprinting and in DNA inhibitor ICI182780 abolished the reduction in Dnmt gene methylation pattern or histone pattern and chromatin con- expression in the ovary of BPA-exposed mice. These obser- formation can contribute to epigenetic diseases such as the vations suggested that ER signalling mediated the epigenetic Angelman, Beckwith-Wiedemann, Prader-Willi, and Russell- effects induced by BPA in the oocytes66 [ ]. Silver syndromes [60–62] and predispose the offspring, or Further evidence for BPA effects on methylation in dif- even the following generations, to cancer and other diseases ferentially methylated regions of imprinted genes during related to epigenetic instability. oogenesis and early embryogenesis came from [68]. Exposure The first report showing BPA epigenetic effects came of mice to BPA (10 𝜇g/kg or 10 mg/kg b.w.) during late stages vy from studies in the agouti viable yellow (𝐴 )mutantmouse of meiosis and oocyte growth, from 2 weeks prior to mating vy model. The 𝐴 allele carries an Intracisternal A Particle until day 9.5 of gestation, resulted in significant alterations in (IAP) retrotransposon insertion upstream of the locus. The the expression of imprinted genes Peg3, Snrpn, H19/Igf2,and Agouti gene is usually expressed during a narrow window Kcnq1 in embryonal and placental tissues and affected foetal, in embryogenesis and codes for a signalling molecule that placental, and postnatal development. The higher BPA dose canleadtoeitherproductionofblackeumelaninoryellow disrupted the parental specific, monoallelic expression of the pheomelanin from a hair-cycle-specific promoter in exon Snrpn, Igf2 and Kcnq1ot1 genes in a tissue-specific manner, vy 2[63, 64]. In the 𝐴 allele, transcriptional control of the and resulted in the biallelic expression of the paternally Agouti coding sequence is driven by promoter elements in expressed Snrpn gene in the placenta, suggesting that mater- the retrotransposon, containing 9 CpGs, whose methylation nal imprinting of Snrpn was disturbed before fertilization level can vary among individual isogenic mice [51, 64, 65]. or that Snrpn was susceptible to loss of imprinting during Hypomethylation of the 9 CpGs leads to the binding of the early embryogenesis. Expression of the normally repressed agouti protein to the melanocortin 4 receptor in all tissues, maternal allele of Kcnq1ot1 in the placentas ranged from 12.3 ectopic gene expression, and shifting of the coat colour from to 72.3% of total expression, whereas no differences were wild type pseudoagouti brown, to mottled, to yellow. In addi- found between control and BPA-treated mice on the Kcnq1ot1 tion, the overproduction of the agouti protein associated to maternal allele expression in the embryo [68], suggesting hypomethylation causes obesity, diabetes, and tumorigenesis loss of methylation control in a tissue-specific fashion. The in adult mice through its multiple actions on gene expression. low BPA dose did not significantly affect imprinted gene When 𝑎/𝑎 female mice received a phytoestrogen-free expression except for Snrpn and Kcnq1ot1 loci in the placenta. diet doped with 50 mg/kg b.w. BPA in the weeks prior to To determine whether the altered expression pattern of vy mating with 𝐴 /𝑎 males and throughout the gestation and imprinted genes was linked to abnormal DNA methylation, lactation time periods, coat colour shift towards yellow and DNA methylation was analysed from the placentas and vy obesity were observed in the 𝐴 /𝑎 heterozygous offspring. embryos of mice exposed to the high BPA dose during oocyte These effects were related to decreased methylation in the growth and early embryogenesis. A small but significant (𝑃< vy 9CpGsiteswithinthe 𝐴 allele. Dietary supplementation 0.05 through ANOVA) reduction in the mean methylation with either methyl donors (folic acid, betaine, vitamin B12, level of the Snrpn imprinting control region was detected and choline) or phytoestrogen prevented hypomethylation, by pyrosequencing; furthermore, the decrease of methylation resulting in more offspring with brown or mottled brown was attributed to the normally hypermethylated maternal coat colour and normal weight [51]. Although the altered allele by bisulfite mutagenesis sequencing, which allows phenotypeofBPA-exposedmicecouldbetransgenerationally assaying allele-specific methylation levels. Pyrosequencing transmitted, the CpG methylation pattern was not inherited analysis of the H19/Igf2 imprinting control region in BPA- in the blastocyst, suggesting that other epigenetic mecha- exposed embryos revealed also a slightly, but significantly, nisms, like histone-mediated chromatin alterations, might be reduced average methylation of the 6 analysed CpG sites (𝑃< responsible for the transgenerational effects in this model. 0.001). Analysis of global DNA methylation by Luminometric The insulin-like growth factor II receptor (Igf2r)andthe Methylation Assay (LUMA) in control and BPA-exposed paternally expressed gene 3 (Peg3) are first imprinted in samples found a significant difference in the placentas, but female mice after birth when follicles and oocytes start to not in 9.5-day embryos, after exposure to the high dose only 6 BioMed Research International

(𝑃 < 0.05). Exposure of females to the high BPA dose 6. From Epigenetic Alterations to only from day 5.5 to day 12.5 of gestation, that is, outside Chromosome Segregation Errors in Oocytes of the critical windows of DNA methylation acquisition in Exposed to BPA the oocytes and epigenetic reprogramming in embryos, did not significantly affect expression of imprinted genes. As An impact of BPA on the oocyte epigenome was confirmed mightbeexpectedfromtheimportantroleofimprinted by in vitro experiments using the same preantral follicle cul- genes in placental development, aberrant imprinting induced ture model in which nonlinear negative effects had been by BPA exposure was associated with abnormal placental shown on spindle integrity, chromosome congression, and phenotypes [68]. In conclusion, this study revealed that expo- meiotic progression [38]. Follicles were chronically exposed sure to environmentally relevant doses of BPA during crit- in vitro to 3 or 300 nM BPA for 12 days, during which they ical windows of oocyte development and growth and early maturedundertheinfluenceoffolliclestimulatinghormone embryogenesis can perturb expression and methylation of up to the large antral stage, when stimulation of ovulation imprinted genes with the most significant effects observed by recombinant hCG and recombinant EGF caused resump- in the placenta. It remains to be established whether loss of tion of oocyte maturation and development of oocytes to imprinting per se and/or disturbance of imprint maintenance metaphase II, at day 13 of culture [38, 76]. Follicle survival and wereduetodirecteffectsontheearlyembryo/placenta,or development, oocyte growth and maturation rates, and chro- were preprogrammed in the oocyte, prior to conception. mosome alignment on the metaphase II plate were compared Other studies, although not specifically focused on oo- between controls and BPA-exposed groups. Concomitantly, cytes, support the hypothesis that BPA exposure may affect possible BPA-induced epigenetic alterations at the level of methylation of cytosines in DNA of imprinted and nonim- DNA methylation and posttranslational histone modifica- printed genes, outside and within coding regions. A genome- tionswereanalysedinsingleoocytes.Thespecificculture wide analysis showed that perinatal exposure to 50 𝜇g/kg conditions were shown not to affect the physiological DNA or 50 mg/kg BPA in diet induced nonmonotonic dose- methylation pattern of maternally imprinted genes in the dependent alterations of DNA methylation patterns in liver. oocytes [77]. Thus, the methylation patterns of differentially Altered methylation was predominantly found within CpG methylated regions in the maternally imprinted, Snrpn, Igf2r, island shores, and, overall, several hundred novel BPA-sen- Mest genes and in the paternally imprinted H19 gene were sitive methylation sites were identified involving pathways in analysed using limiting dilution bisulfite pyrosequencing metabolism and stimulus response [69]. [78]. A cut-off of at least 50% abnormally methylated CpG In another study it was shown that exposure of pregnant sites was established to define epimutations. mice throughout gestation to low doses of BPA (20 𝜇g/kg Changes of posttranslational histone modifications had b.w.) altered the epigenome in the forebrain of the offspring, been previously imputed to BPA [79], and alterations of H3K9 inducing hypomethylation at NotI locus, and deregulation of trimethylation in pericentromeric heterochromatin had been gene expression [70]. Transgenerational changes in behaviour associated in cultured oocytes to meiotic arrest, unaligned were also noted in mice upon gestational BPA exposures [71]. chromosomes, and spindle defects [80], a phenotype similar An epigenetic impact of BPA was demonstrated also on to that observed after treatment with a low BPA concentration male germ cells. Male offspring of rats perinatally exposed to [38]. Furthermore, biallelically different histone posttransla- BPA had reduced sperm counts and other changes in phe- tional epigenetic marks are functionally relevant for a correct notypes not only in the first but also in the 𝐹3 generation expression of imprinted genes, as supported by the evidence [9, 10]. Induction of sperm epimutations and male-mediated that Beckwith-Wiedemann syndrome patients exhibit bial- transgenerational inheritance of obesity and reproductive lelic instead of monoallelic gene expression and similar marks disturbances were also shown after BPA exposure of rats for trimethylated histone H3 lysine 9 (H3K9me3) [81]. Based [12, 72]. When female mice were exposed during gestation onthesenotions,relativehistoneH3K9trimethylationand and lactation to low BPA doses (40 𝜇g/kg b.w.) deregulated H4K12 acetylation were assessed by quantitative confocal glucose homeostasis in the 𝐹2 generation was observed; microscopy of control or BPA-exposed mouse metaphase II decreased global methylation and differential methylation of oocytes from preantral follicle cultures. aspecificCpGsiteintheglucokinasepromoterinthe𝐹1 Only the low BPA concentration (3 nM) caused a slight sperm suggested that the 𝐹2phenotypecouldbecausedby but significant acceleration of follicular growth. Overall, 7.5% epigenetic alterations induced in the male paternal germline of all analysed maternally imprinted alleles were abnormally by BPA prenatal exposure [73]. demethylated in the group exposed to 3 nM BPA, a percent- Finally, exposure to BPA appears to affect DNA methy- age significantly higher𝑃 ( < 0.05) compared to the control lation also in humans; a study in human foetuses found an and 300 nM BPA group [76]. The specific rates of abnormally organ-specific association between changes of global DNA demethylated alleles were 16.7% in Mest,7.4%inIgf2r,and methylation and BPA exposure [74]; a cross-sectional study 4.8% in Snrpn alleles. No BPA effect was detected on the of epigenomic alterations in prepubescent girls from Egypt paternally imprinted H19 allele. Single changes in cytosine revealed that increasing urinary BPA levels were associated methylation, presumably not relevant for gene expression, withchangesinmethylation,inparticularreducedmethyla- were not significantly affected by BPA exposures. The obser- tion in genes involved in immune function, metabolism, and vations suggest that low, chronic BPA exposure during oocyte on the X chromosome [75]. growth can either adversely influence maternal imprinting BioMed Research International 7

In vitro control 3 nM BPA In vitro control 3 nM BPA DAPI DAPI DAPI DAPI

(a) (b) (c) (d) H3K9me3 H3K9me3 H4K12ac H4K12ac

5𝜇m 5𝜇m

(a󳰀)(b󳰀)(c󳰀)(d󳰀)

In vitro control 3 nM BPA DAPI CREST DAPI CREST m m : 1.28 ± 0.3 𝜇 : 1.22 ± 0.3 𝜇 iKD iKD 2𝜇m

(e)

Ac HDAC G9a

M M Dnmt M DNA ATRX M M CG H3K9 P M CG control Epigenetic Epigenetic maturation CG Regulation at at Regulation H3S10 Aurora kinase

Histone

M Methylation P Phosphorylation Ac Acetylation (f)

󸀠 Figure 3: Changes in histone posttranslational modifications in low dose BPA-exposed metaphase II mouse oocytes. (a-b ) Images of histone H3K9 trimethylation in control and BPA-exposed oocytes: some unaligned chromosomes (b) and reduced trimethylated pericentromeric 󸀠 󸀠 heterochromatin (b )intheBPAgroup.(c-d) Unchanged pattern of histone H4K12 acetylation. (e) Decreased distance between centromeres of sister chromatids in metaphase II chromosomes of BPA-exposed oocytes (blue) shown in fixed oocytes that were stained by CREST autoantibodies for centromeres (red). (f) Model indicating relevance of H3K9 trimethylation for recruitment of Dnmts (right side) and other factors like ATRX (ATP-dependent helicase that belongs to SWI/SNF family of chromatin remodelling factors) and Aurora kinase that might play a role in centromere regulation, microtubule attachment, and chromosome alignment through phosphorylation of different target proteins and histone H3S10 (left side). For further explanation, see text and [76].

per se or affect imprint stability. The particular sensitivity of methylation of Mest was also detected in mouse oocytes of the maternal Mest allele to epimutations by low BPA con- connexin 37 deficient transgenic mice82 [ ]. centrations may relate to influences of BPA on bidirectional After exposure to 3 nM BPA, congression failures and signalling between the oocyte and its surrounding cumulus loosely aligned chromosomes at the metaphase II plate were granulosa cells by gap junctional communication, as altered observed (Figures 3(a) and 3(b)). Concomitantly, the relative 8 BioMed Research International

H3K9me3 fluorescence was significantly lower compared to Acknowledgments 󸀠 󸀠 the control group (𝑃 < 0.001) (Figures 3(a )and3(b )), whereas there was no difference in the intensity of fluo- The authors are grateful to Tom Trapphoff for help with prep- 󸀠 rescence associated with H4K12 acetylation (Figures 3(c ) aration of the figures. The studies were supported by the 󸀠 German Research Foundation (FOR 1041). and 3(d )). Interestingly, under the same treatment conditions, the average interkinetochore distance was slightly, but significantly, reduced from 1.28 ± 0.3 𝜇mto1.22± References 0.3 𝜇m(Figure 3(e)). H3K9 is acetylated in immature Ger- [1]D.J.P.Barker,J.G.Eriksson,T.Forsen,´ and C. Osmond, “Fetal minal Vesicle pig oocytes but becomes deacetylated and origins of adult disease: strength of effects and biological basis,” trimethylated during meiosis I and at metaphase II [83]. International Journal of Epidemiology,vol.31,no.6,pp.1235– Trimethylated H3K9 is also a hallmark of heterochromatin 1239, 2002. ofmaturemetaphaseIIoocytesinthemouseandonly [2]P.Bateson,D.Barker,T.Clutton-Brocketal.,“Developmental gradually disappears from chromatin after fertilization, to plasticity and human health,” Nature,vol.430,no.6998,pp.419– increase again later in preimplantation development [57]. 421, 2004. One can speculate that the changes in histone methylation [3]P.D.GluckmanandM.A.Hanson,“Developmentalandepi- of metaphase II chromosomes, induced by low chronic BPA genetic pathways to obesity: an evolutionary-developmental treatment, may influence recruitment of DNA methyltrans- perspective,” International Journal of Obesity,vol.32,supple- ferases and chromatin remodelling proteins, like ATRX or ment 7, pp. S62–S71, 2008. Aurora kinase, that are critical for heterochromatin forma- [4] T. J. Roseboom and E. D. Watson, “The next generation of disease risk: are the effects of prenatal nutrition transmitted tion at centromeres (schematically depicted in Figure 3(f)), across generations? Evidence from animal and human studies,” thereby affecting microtubule attachment and chromosome Placenta,vol.33,no.2,pp.e40–e44,2012. segregation [76]. In other words, the changes in histone post- [5] P. Iozzo, M. Holmes, M. V. Schmidt et al., “Developmental translational modification and DNA methylation detected ORIgins of healthy and unhealthy AgeiNg: the role of maternal in oocytes after BPA exposure could represent the “missing obesity—introduction to DORIAN,” Obesity Facts,vol.7,no.2, link” explaining the effects of low BPA concentrations on pp. 130–151, 2014. chromosome meiotic segregation. [6]E.J.Radford,M.Ito,H.Shietal.,“In utero undernourishment perturbs the adult sperm methylome and intergenerational metabolism,” Science, vol. 345, no. 6198, Article ID 1255903, 7. Conclusions 2014. [7] E. Diamanti-Kandarakis, J.-P. Bourguignon, L. C. Giudice et By now there is increasing evidence that low BPA concentra- al., “Endocrine-disrupting chemicals: an Endocrine Society sci- tions adversely affect the epigenome of mammalian female entific statement,” Endocrine Reviews,vol.30,no.4,pp.293–342, germ cells, with functional consequences on gene expression, 2009. and oocyte development and quality. There are specific time [8] A. Baccarelli and V. Bollati, “Epigenetics and environmental chemicals,” Current Opinion in Pediatrics,vol.21,no.2,pp.243– windows, during which profound chromatin remodelling 251, 2009. occurs and maternal imprints are established or protected [9] S. Salian, T. Doshi, and G. Vanage, “Perinatal exposure of rats to that appear particularly vulnerable to epigenetic deregulation Bisphenol A affects the fertility of male offspring,” Life Sciences, by BPA: these correspond to primordial germ cell formation vol.85,no.21-22,pp.742–752,2009. and oocyte meiotic prophase in the foetus, oocyte matura- [10] S. Salian, T. Doshi, and G. Vanage, “Impairment in protein tion after puberty, and early preimplantation development expression profile of testicular steroid receptor coregulators in after fertilization. Transgenerational effects have also been male rat offspring perinatally exposed to Bisphenol A,” Life observed in the offspring of BPA-treated rodents, although Sciences,vol.85,no.1-2,pp.11–18,2009. the epigenetic mechanisms of inheritance still need to be [11] V.Bollati and A. Baccarelli, “Environmental epigenetics,” Hered- clarified. Thus BPA exposure might have long-lasting con- ity,vol.105,no.1,pp.105–112,2010. sequences on the female reproductive health, ranging from [12] M. Manikkam, R. Tracey, C. Guerrero-Bosagna, and M. K. reduced fertility to offspring defects. Finally, the studies on Skinner, “Plastics derived endocrine disruptors (BPA, DEHP and DBP) induce epigenetic transgenerational inheritance of histone posttranslational modifications suggest that BPA can obesity, reproductive disease and sperm epimutations,” PLoS alsopredisposetheoocytestoalteredchromosomebehaviour ONE,vol.8,no.1,ArticleIDe55387,2013. in meiosis, particularly at low concentrations. The relevance [13] J. B. Matthews, K. Twomey, and T. R. Zacharewski, “In vitro and of these findings for human health protection still needs to be in vivo interactions of bisphenol A and its metabolite, bisphenol fully assessed, but they warrant further investigation in both A glucuronide, with estrogen receptors 𝛼 and 𝛽,” Chemical experimental models and humans. Research in Toxicology,vol.14,no.2,pp.149–157,2001. [14] S. H. Swan, “Intrauterine exposure to diethylstilbestrol: long- term effects in humans,” Acta Pathologica, Microbiologica, et Conflict of Interests Immunologica Scandinavica,vol.108,no.12,pp.793–804,2000. [15] T. G. Bredfeldt, K. L. Greathouse, S. H. Safe, M.-C. Hung, M. T. The authors declare no conflict of interests. Bedford, and C. L. Walker, “Xenoestrogen-induced regulation BioMed Research International 9

of EZH2 and histone methylation via estrogen receptor signal- monkey,” Proceedings of the National Academy of Sciences of the ing to PI3K/AKT,” Molecular Endocrinology,vol.24,no.5,pp. United States of America,vol.109,no.43,pp.17525–17530,2012. 993–1006, 2010. [32] M. A. Brieno-Enr˜ ´ıquez, P. Robles, N. Camats-Tarruella et al., [16]K.Sato,H.Fukata,Y.Kogo,J.Ohgane,K.Shiota,andC.Mori, “Human meiotic progression and recombination are affected by “Neonatal exposure to diethylstilbestrol alters expression of Bisphenol A exposure during in vitro human oocyte develop- DNA methyltransferases and methylation of genomic DNA in ment,” Human Reproduction,vol.26,no.10,pp.2807–2818,2011. the mouse uterus,” Endocrine Journal,vol.56,no.1,pp.131–139, [33] M. A. Brieno-Enr˜ ´ıquez, R. Reig-Viader, L. Cabero et al., “Gene 2009. expression is altered after bisphenol A exposure in human fetal [17] A. Careghini, A. F. Mastorgio, S. Saponaro, and E. Sezenna, oocytes in vitro,” Molecular Human Reproduction,vol.18,no.4, “Bisphenol A, nonylphenols, benzophenones, and benzotria- Article ID gar074, pp. 171–183, 2012. zoles in soils, groundwater, surface water, sediments, and food: [34]W.Wang,K.S.Hafner,andJ.A.Flaws,“Inuterobisphenol a review,” Environmental Science and Pollution Research,vol.22, A exposure disrupts germ cell nest breakdown and reduces no. 8, pp. 5711–5741, 2015. fertility with age in the mouse,” Toxicology and Applied Phar- [18] G.Mileva,S.L.Baker,A.T.Konkle,andC.Bielajew,“Bisphenol- macology,vol.276,no.2,pp.157–164,2014. A: epigenetic reprogramming and effects on reproduction and [35] T. Zhang, L. Li, X.-S. Qin et al., “Di-(2-ethylhexyl) phthalate behaviour,” International Journal of Environmental Research and and bisphenol A exposure impairs mouse primordial follicle Public Health,vol.11,no.7,pp.7537–7561,2014. assembly in vitro,” Environmental and Molecular Mutagenesis, [19]Y.Yang,Z.Wang,T.He,Y.Dai,andS.Xie,“Sedimentbacte- vol.55,no.4,pp.343–353,2014. rial communities associated with anaerobic biodegradation of [36]J.Peretz,R.K.Gupta,J.Singh,I.Hernandez-Ochoa,´ and J. bisphenol A,” Microbial Ecology,2014. A. Flaws, “Bisphenol A impairs follicle growth, inhibits ster- [20] J. Michałowicz, “Bisphenol a—sources, toxicity and biotransfor- oidogenesis, and downregulates rate-limiting enzymes in the mation,” Environmental Toxicology and Pharmacology,vol.37, estradiol biosynthesis pathway,” Toxicological Sciences,vol.119, no. 2, pp. 738–758, 2014. no. 1, pp. 209–217, 2011. [21] A. M. Calafat, Z. Kuklenyik, J. A. Reidy, S. P. Caudill, J. Ekong, [37]J.Peretz,Z.R.Craig,andJ.A.Flaws,“Bisphenolainhibits and L. L. Needham, “Urinary concentrations of bisphenol A and follicle growth and induces atresia in cultured mouse antral fol- 4-Nonylphenol in a human reference population,” Environmen- licles independently of the genomic estrogenic pathway,” Biol- tal Health Perspectives,vol.113,no.4,pp.391–395,2005. ogy of Reproduction,vol.87,no.3,article63,ArticleIDArticle [22] J. R. Rochester, “Bisphenol A and human health: a review of the 63, 2012. literature,” Reproductive Toxicology,vol.42,pp.132–155,2013. [38] S. Lenie, R. Cortvrindt, U. Eichenlaub-Ritter, and J. Smitz, [23] J.Peretz,L.Vrooman,W.A.Rickeetal.,“BisphenolAandrepro- “Continuous exposure to bisphenol A during in vitrofollic- ductive health: update of experimental and human evidence ular development induces meiotic abnormalities,” Mutation 2007–2013,” Environmental Health Perspectives,vol.122,no.8, Research—Genetic Toxicology and Environmental Mutagenesis, pp. 775–786, 2014. vol. 651, no. 1-2, pp. 71–81, 2008. [24] R.R.Gerona,T.J.Woodruff,C.A.Dickensonetal.,“Bisphenol- [39] I. Souter, K. W. Smith, I. Dimitriadis et al., “The association of A (BPA), BPA glucuronide, and BPA sulfate in midgestation bisphenol-A urinary concentrations with antral follicle counts umbilical cord serum in a northern and central California pop- and other measures of ovarian reserve in women undergoing ulation,” Environmental Science and Technology,vol.47,no.21, infertility treatments,” Reproductive Toxicology,vol.42,pp.224– pp. 12477–12485, 2013. 231, 2013. [25] R. Machtinger and R. Orvieto, “Bisphenol A, oocyte matura- [40] M. S. Bloom, D. Kim, F.S. vom Saal et al., “Bisphenol A exposure tion, implantation, and IVF outcome: review of animal and reduces the estradiol response to gonadotropin stimulation human data,” Reproductive BioMedicine Online,vol.29,no.4, during in vitro fertilization,” Fertility and Sterility,vol.96,no. pp. 404–410, 2014. 3, pp. 672.e2–677.e2, 2011. [26]J.Mathieu-Denoncourt,S.J.Wallace,S.R.deSolla,and [41] E. Mok-Lin, S. Ehrlich, P. L. Williams et al., “Urinary bisphenol V. S. Langlois, “Plasticizer endocrine disruption: highlight- A concentrations and ovarian response among women under- ing developmental and reproductive effects in mammals and going IVF,” International Journal of Andrology,vol.33,no.2,pp. non-mammalian aquatic species,” General and Comparative 385–393, 2010. Endocrinology,2014. [42] S. Ehrlich, P. L. Williams, S. A. Missmer et al., “Urinary bisphe- [27]D.E.Cantonwine,R.Hauser,andJ.D.Meeker,“BisphenolA nol A concentrations and early reproductive health outcomes and human reproductive health,” Expert Review of Obstetrics among women undergoing IVF,” Human Reproduction,vol.27, and Gynecology,vol.8,no.4,pp.329–335,2013. no. 12, pp. 3583–3592, 2012. [28] P. Hajkova, S. Erhardt, N. Lane et al., “Epigenetic reprogram- [43] V. Y. Fujimoto, D. Kim, F. S. vom Saal, J. D. Lamb, J. A. Taylor, ming in mouse primordial germ cells,” Mechanisms of Develop- and M. S. Bloom, “Serum unconjugated bisphenol A concen- ment,vol.117,no.1-2,pp.15–23,2002. trations in women may adversely influence oocyte quality [29] U. Eichenlaub-Ritter, “Meiosis: how to get a good start in life,” during in vitro fertilization,” Fertility and Sterility,vol.95,no. in Textbook of Human Reproductive Genetics, K. Sermon and S. 5, pp. 1816–1819, 2011. Viville, Eds., pp. 33–51, Cambridge University Press, 2014. [44] S. Ehrlich, P. L. Williams, S. A. Missmer et al., “Urinary [30]M.Susiarjo,T.J.Hassold,E.Freeman,andP.A.Hunt,“Bisphe- bisphenol A concentrations and implantation failure among nol A exposure in utero disrupts early oogenesis in the mouse,” women undergoing in vitro fertilization,” Environmental Health PLoS genetics,vol.3,no.1,articlee5,2007. Perspectives,vol.120,no.7,pp.978–983,2012. [31] P.A. Hunt, C. Lawson, M. Gieske et al., “Bisphenol A alters early [45] M. Zuccotti, V. Merico, S. Cecconi, C. A. Redi, and S. Garagna, oogenesis and follicle formation in the fetal ovary of the rhesus “What does it take to make a developmentally competent 10 BioMed Research International

mammalian egg?” Human Reproduction Update,vol.17,no.4, induced pluripotent stem cell,” Journal of Human Genetics,vol. pp.525–540,2011. 57, no. 2, pp. 84–91, 2012. [46] S.-I. Kageyama, H. Liu, N. Kaneko, M. Ooga, M. Nagata, and [63]D.M.J.Duhl,H.Vrieling,K.A.Miller,G.L.Wolff,andG.S. F. Aoki, “Alterations in epigenetic modifications during oocyte Barsh, “Neomorphic agouti mutations in obese yellow mice,” growth in mice,” Reproduction,vol.133,no.1,pp.85–94,2007. Nature Genetics,vol.8,no.1,pp.59–65,1994. [47]P.A.Hunt,K.E.Koehler,M.Susiarjoetal.,“BisphenolA [64] R. A. Waterland and R. L. Jirtle, “Transposable elements: targets exposure causes meiotic aneuploidy in the female mouse,” for early nutritional effects on epigenetic gene regulation,” Current Biology,vol.13,no.7,pp.546–553,2003. Molecular and Cellular Biology, vol. 23, no. 15, pp. 5293–5300, [48] U. Eichenlaub-Ritter, E. Vogt, S. Cukurcam, F. Sun, F. Pac- 2003. chierotti, and J. Parry, “Exposure of mouse oocytes to bisphenol [65] R. L. Jirtle, “The Agouti mouse: a biosensor for environmental A causes meiotic arrest but not aneuploidy,” Mutation Research, epigenomics studies investigating the developmental origins of vol. 651, no. 1-2, pp. 82–92, 2008. health and disease,” Epigenomics,vol.6,no.5,pp.447–450,2014. [49] F. Pacchierotti, R. Ranaldi, U. Eichenlaub-Ritter, S. Attia, and [66]H.-H.Chao,X.-F.Zhang,B.Chenetal.,“BisphenolAexposure I.-D. Adler, “Evaluation of aneugenic effects of bisphenol A in modifies methylation of imprinted genes in mouse oocytes via somatic and germ cells of the mouse,” Mutation Research,vol. the estrogen receptor signaling pathway,” Histochemistry and 651, no. 1-2, pp. 64–70, 2008. Cell Biology,vol.137,no.2,pp.249–259,2012. [50] A. Muhlhauser, M. Susiarjo, C. Rubio et al., “Bisphenol A effects [67] J. M. Hall and D. P. McDonnell, “Coregulators in nuclear estro- on the growing mouse oocyte are influenced by diet,” Biology of gen receptor action: from concept to therapeutic targeting,” Reproduction,vol.80,no.5,pp.1066–1071,2009. Molecular Interventions,vol.5,no.6,pp.343–357,2005. [51] D. C. Dolinoy, D. Huang, and R. L. Jirtle, “Maternal nutri- [68]M.Susiarjo,I.Sasson,C.Mesaros,andM.S.Bartolomei, ent supplementation counteracts bisphenol A-induced DNA “Bisphenol A exposure disrupts genomic imprinting in the hypomethylation in early development,” Proceedings of the mouse,” PLoS Genetics, vol. 9, no. 4, Article ID e1003401, 2013. National Academy of Sciences of the United States of America, [69] J. H. Kim, M. A. Sartor, L. S. Rozek et al., “Perinatal bisphenol vol. 104, no. 32, pp. 13056–13061, 2007. A exposure promotes dose-dependent alterations of the mouse [52] A. Can, O. Semiz, and O. Cinar, “Bisphenol-A induces cell cycle methylome,” BMC Genomics,vol.15,no.1,article30,2014. delay and alters centrosome and spindle microtubular organ- [70]T.Yaoi,K.Itoh,K.Nakamura,H.Ogi,Y.Fujiwara,andS. ization in oocytes during meiosis,” Molecular Human Reproduc- Fushiki, “Genome-wide analysis of epigenomic alterations in tion, vol. 11, no. 6, pp. 389–396, 2005. fetal mouse forebrain after exposure to low doses of bisphenol [53] R. Machtinger, C. M. H. Combelles, S. A. Missmer et al., A,” Biochemical and Biophysical Research Communications,vol. “Bisphenol-A and human oocyte maturation in vitro,” Human 376, no. 3, pp. 563–567, 2008. Reproduction,vol.28,no.10,pp.2735–2745,2013. [71]J.T.Wolstenholme,M.Edwards,S.R.J.Shettyetal.,“Ges- [54]D.Lucifero,M.R.W.Mann,M.S.Bartolomei,andJ.M. tational exposure to bisphenol a produces transgenerational Trasler, “Gene-specific timing and epigenetic memory in oocyte changes in behaviors and gene expression,” Endocrinology,vol. imprinting,” Human Molecular Genetics,vol.13,no.8,pp.839– 153, no. 8, pp. 3828–3838, 2012. 849, 2004. [72] T. Doshi, C. D’Souza, and G. Vanage, “Aberrant DNA methyla- [55] H. Kobayashi, T. Sakurai, M. Imai et al., “Contribution of intra- tion at Igf2-H19 imprinting control region in spermatozoa upon genic DNA methylation in mouse gametic DNA methylomes to neonatal exposure to bisphenol A and its association with post establish Oocyte-specific heritable marks,” PLoS Genetics,vol.8, implantation loss,” Molecular Biology Reports,vol.40,no.8,pp. no.1,ArticleIDe1002440,2012. 4747–4757, 2013. [73]G.Li,H.Chang,W.Xia,Z.Mao,Y.Li,andS.Xu,“F0maternal [56] K. Shirane, H. Toh, H. Kobayashi et al., “Mouse oocyte methy- BPA exposure induced glucose intolerance of F2 generation lomes at base resolution reveal genome-wide accumulation of throughDNAmethylationchangeinGck,”Toxicology Letters, non-CpG methylation and role of DNA methyltransferases,” vol. 228, no. 3, pp. 192–199, 2014. PLoS Genetics,vol.9,no.4,ArticleIDe1003439,2013. [74]M.S.Nahar,C.Liao,K.Kannan,C.Harris,andD.C.Dolinoy, [57] A. Burton and M. Torres-Padilla, “Chromatin dynamics in the “In utero bisphenol A concentration, metabolism, and global regulation of cell fate allocation during early embryogenesis,” DNA methylation across matched placenta, kidney, and liver in Nature Reviews Molecular Cell Biology,vol.15,no.11,pp.723– the human fetus,” Chemosphere,vol.124,pp.54–60,2015. 735, 2014. [75] J. H. Kim, L. S. Rozek, A. S. Soliman et al., “Bisphenol A- [58] D. Bourc’his and C. Proudhon, “Sexual dimorphism in parental associated epigenomic changes in prepubescent girls: a cross- imprint ontogeny and contribution to embryonic development,” sectional study in Gharbiah, Egypt,” Environmental Health,vol. Molecular and Cellular Endocrinology,vol.282,no.1-2,pp.87– 12, no. 1, article 33, 2013. 94, 2008. [76]T.Trapphoff,M.Heiligentag,N.ElHajj,T.Haaf,andU. [59] C. W. Hanna and G. Kelsey, “The specification of imprints in Eichenlaub-Ritter, “Chronic exposure to a low concentration of mammals,” Heredity,vol.113,no.2,pp.176–183,2014. bisphenol A during follicle culture affects the epigenetic status [60]H.Judson,B.E.Hayward,E.Sheridan,andD.T.Bonthron,“A of germinal vesicles and metaphase II oocytes,” Fertility and global disorder of imprinting in the human female germ line,” Sterility, vol. 100, no. 6, pp. 1758.e1–1767.e1, 2013. Nature,vol.416,no.6880,pp.539–542,2002. [77] E. Anckaert, T. Adriaenssens, S. Romero, S. Dremier, and J. [61] B. Horsthemke, “Mechanisms of imprint dysregulation,” The Smitz, “Unaltered imprinting establishment of key imprinted American Journal of Medical Genetics. Part C: Seminars in genes in mouse oocytes after in vitro follicle culture under Medical Genetics,vol.154,no.3,pp.321–328,2010. variable follicle-stimulating hormone exposure,” International [62] S.-I. Tomizawa and H. Sasaki, “Genomic imprinting and its Journal of Developmental Biology,vol.53,no.4,pp.541–548, relevance to congenital disease, infertility, molar pregnancy and 2009. BioMed Research International 11

[78] N. El Hajj, T. Trapphoff, M. Linke et al., “Limiting dilution bisulfite (pyro)sequencing reveals parent-specific methylation patterns in single early mouse embryos and bovine oocytes,” Epigenetics, vol. 6, no. 10, pp. 1176–1188, 2011. [79] L.F.Doherty,J.G.Bromer,Y.Zhou,T.S.Aldad,andH.S.Taylor, “In utero exposure to diethylstilbestrol (DES) or bisphenol-A (BPA) increases EZH2 expression in the mammary gland: an epigenetic mechanism linking endocrine disruptors to breast cancer,” Hormones and Cancer,vol.1,no.3,pp.146–155,2010. [80]E.Vogt,A.Kipp,andU.Eichenlaub-Ritter,“AurorakinaseB, epigenetic state of centromeric heterochromatin and chiasma resolution in oocytes,” Reproductive Biomedicine Online,vol.19, no. 3, pp. 352–368, 2009. [81] R. Nativio, A. Sparago, Y. Ito, R. Weksberg, A. Riccio, and A. Murrell, “Disruption of genomic neighbourhood at the imprinted IGF2-H19 locus in Beckwith-Wiedemann syndrome and Silver-Russell syndrome,” Human Molecular Genetics,vol. 20, no. 7, Article ID ddr018, pp. 1363–1374, 2011. [82] M. M. Denomme, C. R. White, C. Gillio-Meina et al., “Compro- mised fertility disrupts Peg1 but not Snrpn and Peg3 imprinted methylation acquisition in mouse oocytes,” Frontiers in Genet- ics, vol. 3, article 129, Article ID Article 129, 2012. [83]T.Endo,K.Naito,F.Aoki,S.Kume,andH.Tojo,“Changes in histone modifications during in vitro maturation of porcine oocytes,” Molecular Reproduction and Development,vol.71,no. 1,pp.123–128,2005. [84] S. A. Smallwood and G. Kelsey, “De novo DNA methylation: a germ cell perspective,” Trends in Genetics,vol.28,no.1,pp.33– 42, 2012. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 123484, 23 pages http://dx.doi.org/10.1155/2015/123484

Review Article Environmental Impact on DNA Methylation in the Germline: State of the Art and Gaps of Knowledge

Francesca Pacchierotti and Marcello Spanò

Laboratory of Toxicology, Technical Unit of Radiation Biology and Human Health, CR Casaccia, ENEA, Via Anguillarese 301, 00123 Rome, Italy

Correspondence should be addressed to Francesca Pacchierotti; [email protected]

Received 14 January 2015; Accepted 3 May 2015

Academic Editor: Heide Schatten

Copyright © 2015 F. Pacchierotti and M. Spano.` This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The epigenome consists of chemical changes in DNA and chromatin that without modifying the DNA sequence modulate gene expression and cellular phenotype. The epigenome is highly plastic and reacts to changing external conditions with modifications that can be inherited to daughter cells and across generations. Whereas this innate plasticity allows for adaptation to a changing environment, it also implies the potential of epigenetic derailment leading to so-called epimutations. DNA methylation is the most studied epigenetic mark. DNA methylation changes have been associated with cancer, infertility, cardiovascular, respiratory, metabolic, immunologic, and neurodegenerative pathologies. Experiments in rodents demonstrate that exposure to a variety of chemical stressors, occurring during the prenatal or the adult life, may induce DNA methylation changes in germ cells, which may be transmitted across generations with phenotypic consequences. An increasing number of human biomonitoring studies show environmentally related DNA methylation changes mainly in blood leukocytes, whereas very few data have been so far collected on possible epigenetic changes induced in the germline, even by the analysis of easily accessible sperm. In this paper, we review the state of the art on factors impinging on DNA methylation in the germline, highlight gaps of knowledge, and propose priorities for future studies.

1. Introduction regions. DNA methylation, operated by a family of DNA methyltransferases [5], is implicated in many life-essential Epigenetics formally refers to heritable changes in gene cellular and developmental processes, such as embryonic expression and in phenotype occurring without changes reprogramming, cellular differentiation, silencing of genes in the underlying DNA sequence. Alterations in epigenetic and transposons, parental imprinting, X chromosome inac- markshavebeeninvolvedintheetiologyofcomplexsyn- tivation, and genomic stability [6–8]. dromes and diseases, including cancer, infertility, cardiovas- It is largely accepted that exposure to a variety of environ- cular, respiratory, metabolic, immunologic, and neurodegen- mental toxicants has a negative impact on human health and erative pathologies [1–3]. The main epigenetic mechanisms contributes to the development of a large array of diseases. responsible for these alterations are represented by DNA The epigenome is more plastic and flexible than the genome. methylation, posttranslational histone modifications, and Changes of epigenetic marks, such as DNA methylation, regulation by noncoding microRNAs [4]. In particular, DNA can affect the chromatin structure and modify binding of methylation, the most studied epigenetic mark so far, involves transcription factors and gene expression. The theoretical the enzymatically mediated covalent addition of a methyl framework of a changing environment and a modifiable group to the C5 position of cytosine, forming 5-methyl epigenome might offer unexplored and unsuspected ways to cytosine (5-mC). Cytosine methylation primarily happens understand gene-environment interactions and potentially in CpG dinucleotides in CpG-rich sequences, dubbed as mitigate the impact of environmental toxicants on human CpGislands,oftenoccurringnearorinthegenepromoter health [9, 10]. 2 BioMed Research International

A rapidly growing number of epidemiological studies especially those belonging to the variegated and heteroge- have been carried out throughout the world, in which neous class of compounds collectively defined as endocrine environmental exposure and lifestyle (an umbrella term disruptors (EDs), has been linked to the increased incidence including diet, behavior, stress, physical activity, working of male reproductive pathologies [28–32]. The interference habits, voluntary alcohol and tobacco consumption, etc.), with developmental epigenetic processes has been evoked as meshed with the genetic background, have been associated oneofthepotentialmechanismsofEDsactionaffectingthe with epigenetic changes, mostly DNA methylation (reviewed integrity of the male reproductive system [33]. Recent evi- in [11–18]).Thetrendinthefieldappearstoshiftfrom dence that unbalanced one carbon metabolism may impact the introduction of a novel “proof-of-principle” approach in male reproductive health [34] and that a variety of epigenetic toxicology to a more systematic scientific specialty [19, 20]. markers, including global or gene-specific DNA methylation, Environmental epidemiology research is addressing epi- can be altered in infertility patients [35, 36]isinagreement genetic mechanisms as mediators of environmental exposure with this hypothesis. on disease risk or just as biosensors of exposure even if In spite of the extensive DNA demethylation occurring not mechanistically relevant. Because stable methylation in preimplantation embryonic cells, there are sequences, marks at differentially methylated regions (DMRs) regulating corresponding primarily, but not exclusively, to parentally imprinted genes are acquired before gastrulation, they may imprinted genes that escape global demethylation [37, 38]. serveasarchivesofearlyexposurewiththepotentialto This means that changes of DNA methylation induced by improve our understanding of developmental origin of adult environmental stressors in germ cells could not only have diseases. consequences for the reproductive health of the exposed DNA methylation has been by far the most extensively individual but also might be potentially heritable from one measured epigenetic mark because of its obviously funda- generation to another and might cause transgenerational mental biological interest, its mitotic stability, the availability adverse effects by a nongenetic mechanism of inheritance. of methods for its quantification, globally or in targeted Notwithstanding the knowledge about epigenetic reg- regions, its stability during the DNA extraction and purifi- ulation of gonadal development and the evidence about cation procedures, and its durability in archival biological epigenetic changes induced in rodent germline by several materials. By and large, the strategy consists of comparatively chemicals or dietary conditions, the number of studies aimed assessing the methylation differences at CpG islands in gene at testing possible effects of lifestyle or chemical exposure promoters or DMRs, between control and exposed groups. on human sperm DNA methylation is extremely limited in The information on DNA methylation status and changes comparison to the number of studies carried out on blood in association with environmental exposure and lifestyle cells. Sperm can be obtained by a similarly noninvasive has been mostly collected from peripheral blood leukocytes procedure; they represent the target cell for male reproductive (PBL), which can be sampled by a minimally invasive effects and not merely a surrogate of it; their ultimate DNA approach. However, tissue specificity, together with purity methylation pattern is acquired by a multistep process started of cells for DNA methylation determination, represents a in PGCs and completed during the spermatogonia and relevant issue in epigenetic studies as each tissue and, within a spermatocyte differentiation phases [39, 40], which might tissue, probably each cell type have its own epigenetic profile. therefore be repeatedly exposed to environmental insults. During our lifetime, the genome undergoes two main For all these reasons it would be very valuable to extend epigenomic reprogramming periods, each of which involves the analysis of sperm epigenetics beyond infertility clinical waves of DNA demethylation and de novo methylation. These investigation to environmental biomonitoring studies. precise and coordinated genome-wide reprogramming steps Focus of this literature review will be on data linking are associated with pivotal developmental stages like the various exogenous factors, from specific chemical exposure establishment of cell totipotency and the differentiation of the to psychological stress, to DNA methylation changes in germ cell lineages [21, 22]. The first wave occurs, with notable the germline and their consequences in the offspring. We differences between sexes, in all cells of the preimplantation have taken into consideration both experimental rodent and embryo.Thesecondwaveoccursinprimordialgermcells human studies. In addition, due to the very limited amount (PGCs) only; this time the demethylation events are more of data from human biomonitoring investigations, we have radical and involve imprinted genes whose allelic-specific decided to include a survey of papers reporting human popu- methylation is first erased and then reset according to the lation studies which have shown an environmental impact on sex of the germline [23]. Conceivably, these phases of mam- DNA methylation of somatic cells, to highlight those sources malian development are especially sensitive to environmental of exposure that would be worth further germline-oriented stressors, which can impact epigenetic plasticity with poten- investigations, and the present gaps of knowledge. tial enduring effects on metabolic pathways and disease susceptibility. Indeed, such scenario would be in agreement with the theory of the fetal basis of adult onset disease [24–26]. 2. Human Studies on Environmentally The early fetal period of life is particularly critical Linked DNA Methylation Changes in for gonadal development, and many common reproductive Somatic and Germ Cells disorders of the adult male, such as infertility and testis cancer,havebeenproposedtohaveafetalorigin[27]. In Several recent and excellent reviews [12, 13, 15, 17, 41, 42] addition, prenatal exposure to environmental contaminants, have been published on this subject. Here, we try to offer an BioMed Research International 3 original aggregation pattern of the data, which are mostly [60]. These results strongly point to the need for long-term very recent but often incomplete and methodologically follow-ups to determine whether the observed DNA methyla- heterogeneous, to highlight research trends and gaps of tion changes may be associated with specific health outcomes. knowledge, also in relation to experimental rodent studies. Chromium VI [61], mercury [62, 63], lead [64, 65], The epidemiological studies specifically reporting the effects cadmium [62, 66], and selenium [67]areothermetalsfor of environmental chemicals on DNA methylation are which association studies between human exposure and summarized in Table 1. DNA methylation changes, mainly in peripheral blood cells, have been conducted. Cadmium can cross the placental 2.1. Metals: Arsenic, Chromium, Cadmium, Lead, Mercury, barrier and its potential as a developmental toxicant has been and Selenium. Environmental toxic metals have been asso- studied in an American survey by comparing maternal blood ciated with important human pathologies like cancer, cardio- cadmium levels during pregnancy and genome-wide DNA vascular and autoimmune diseases, and neurological disor- methylation in leukocyte DNA collected from cord blood ders and, recently, their impact on the epigenome has started cells [68]. A variety of genes showed methylation changes to be explored [43]. associated with maternal cadmium concentrations. The set Inorganic arsenic is a carcinogenic metal. Several mil- was enriched in genes involved in transcriptional regulation lions of people around the world are exposed to arsenic control and apoptosis. Conserved DNA motifs with sequence concentrations in their drinking water that exceed the World similarity to specific transcription factor binding sites were Health Organization’s recommended limit of 10 ppb. The identified within the CpG islands of the gene set. Altogether mechanism(s) of arsenic toxicity and carcinogenicity are the results pointed to a possible functional impact of cad- not fully clarified; recently, epigenetic alterations have been mium on fetal DNA methylation. proposed to play a role and have been explored in cohort and Overall, the number of studies on the epigenetic impact of case-control studies, especially in Asian populations living in environmental metal exposure is limited. The study designs, highly contaminated areas (reviewed in [44]). Exposure, gen- the number of people enrolled, the genomic sequences inves- erally assessed by the metal concentration in drinking water tigated, and the methods used to assess methylation changes and/or in biological fluids or tissues, has been associated (locus-specific, global locus-independent, epigenome-wide) with dose-dependent global DNA hypermethylation [45–47] are quite heterogeneous. Widely different exposure levels and with hypermethylation of specific oncosuppressor genes have been evaluated in occupational studies and in studies [48–52]. Genome-wide comparisons of DNA methylation on the general population. Therefore, any attempt to draw patterns from people who developed skin lesion and a control general conclusions is still premature. However, the expected group in Bangladesh have evidenced 6 CpG sites with greatest decrease of costs of epigenome-wide analytical methods will changes of DNA methylation among cases, one of which likely allow acquiring a wealth of data in the near future. belongs to the RHBDF1 gene, previously reported to be In addition to such unsupervised studies, more focused hypermethylated in arsenic-exposed cases [53]. Similarly, by investigations on global hypomethylation and downregu- using high throughput approaches, specific DNA methylation lation of the methylation machinery, hypomethylation in changes in particular genes were detected between arsenic- regions controlling transposons or oncogene expression, or induced and non-arsenic-induced urothelial carcinomas in hypermethylation at oncosuppressor genes could offer the Taiwan [54]. best contribution to unravel epigenetic mechanisms under- The epigenetic effects of in utero arsenic exposure were lying environmental cancer and to develop novel predictive investigated in umbilical cord blood to find out a mechanistic biomarkers. basis for possible arsenic-induced alterations of fetal developmental programming. Hypermethylation of the transposonic 2.2. Air Pollution (Particulate Matter, Polycyclic Aromatic repeat LINE-1, p16 promoter, and other specific sequences was Hydrocarbons, Benzene, and Volatile Organic Compounds). associated with arsenic concentration in maternal drinking Exposure to air pollution is a side-product of urbaniza- water [55, 56]. Other studies also showed some effects of tion and industrialization representing a dramatic health arsenic maternal exposure on cord blood DNA methylation problem, associated with childhood asthma, wheeze, and [57, 58], although the involved sequences were not always increased cardiovascular morbidity and mortality. It is gen- consistent. erally assessed by measuring the levels of particulate matter The newborn blood DNA methylation pattern seems to with aerodynamic diameter ≤2.5 (PM2.5) or ≤10 𝜇m(PM10) be affected also by in utero exposure to low concentrations, together with the levels of other air pollutants like black as shown by the results of a prospective American birth carbon, ozone, polycyclic aromatic hydrocarbons (PAHs), cohort study using high throughput arrays [59]. In another sulfur, and nitrogen dioxide. There have been several studies large Mexican cohort, a total of 2,705 genes in cord blood carried out across the globe, which have considered possible leukocytes showed differences in DNA methylation that were impacts of air pollution on DNA methylation with sometime associated with maternal exposure to arsenic in drinking contrasting results [69]. water. The gene set was highly enriched in binding sites of InarecentEuropeanstudyonacohortofyoungnon- the early growth response and CTCF transcription factors. smoking subjects, the exposure to ambient concentrations Furthermore, DNA methylation levels of seven of these genes of NO2, PM10, PM2.5, and O3 and traffic parameters were were associated with differences in birth outcomes including associated with a decreased global DNA methylation level in gestational age, placental weight, and head circumference blood cells [70]. 4 BioMed Research International ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] 51 61 53 52 55 57 58 59 63 65 62 45 47 56 50 67 54 49 68 48 60 66 46 64 [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ Reference in and and , p16 p53 p16 SEPP1 LINE-1 LINE-1 NPY2R , RHBDF1 KCNK17 KLK7 , ,and ; Pb: hypomethylation COL1A2 Hypermethylation in CTNNA2 ZNF132 Differential methylation patterns 6 differentially methylated sites including Hypermethylation at MLH1 Hypermethylation Hypomethylation Hypermethylation at Differential methylation in thousands of sites Hypermethylation at severalloci Differential methylation in thousands of sites Hypomethylation at Differential methylation at subsets of genes Hg: hypermethylation at GSTM1 at Hypermethylation (sex-dependent) Hypermethylation at LINE-1 males :COBRA : methylation-specific Arrays Bisulfite PCRsequencing Hypermethylation Arrays Arrays Bisulfite methyl specific Hypermethylation PCR Bisulfite methyl specificBisulfite PCRpyrosequencing Hypermethylation PCR Methylation-specific PCR and thermal denaturation Method DNA methylation changes ELISA Arrays Arrays Arrays Global: total 5-mC by HPLC-MS; LINE-1 p53 restriction endonuclease digestion Bisulfite PCRpyrosequencing Bisulfite PCRpyrosequencing Hypomethylation at Arrays Arrays Methyl incorporation assay; bisulfite PCRpyrosequencing; LUMA (Luminometric Methylation Assay) Bisulfite PCRpyrosequencing Hypomethylation at Bisulfite PCRpyrosequencing , Alu MLH1 SEPW1 LINE-1 , ,and Alu Alu , , p53 ,and , SEPP1 ,and p16 p16 p53 DAPK Global Methyl acceptance assay Hypermethylation Global Methyl acceptance assay Hypermethylation Global Methyl acceptance assay Hypermethylation Global Global Methyl acceptance assay Hypomethylation p16 LINE-1 Global Global p16 LINE-1 DNMT1 , , Metals and , LINE-1 LINE-1 MLH1 p53 , Epigenome-wide Epigenome-wide Epigenome-wide Epigenome-wide Epigenome-wide Epigenome-wide Epigenome-wide , p16 LINE-1 p16 LINE-1 PBL (maternal and umbilical cord samples) Tumor biopsies Epigenome-wide PBL PBL Tumor biopsies PBL PBL PBL PBL PBL PBL PBL PBL PBL PBL (maternal and newborn) PBL PBL Target tissuePBL Target DNA region PBL PBL PBL PBL (maternal and umbilical cord samples) Table 1: Selected epidemiological studies on the effects ofenvironmental chemical exposureshuman on methylation. DNA 113 mother-child pairs, Bangladesh 28 patients with urothelial carcinoma, Taiwan 16 individuals of which 8with arsenicosis, were Mexico 10 individuals with skin lesions and 10 controls, Bangladesh 38 patients with urothelial carcinoma, Taiwan 163 patients with arseniasis and 110 controls, China 202 women, Argentina 96 adults, India 320 adults, Bangladesh 44 newborns, Bangladesh 134 infants, USA newborns, Mexico 9 exposed individuals 64 adults, India 17 mother-child pairs, USA Study population 294 adults, India 115 workers and 60China controls, 517 elderly men, USA 202 women, Argentina 113 mother-child pairs, Bangladesh 71 newborns, Thailand PBL(cord blood) Exposure (chemical) Arsenic (As) Chromium (Cr) Mercury (Hg) 131Lead dental (Pb) professionals, USA Buccal mucosa Cadmium (Cd) Selenium (Se)Hg, Pb, 286 and adults, Cd Bangladesh 43 women undergoing IVF, USA BioMed Research International 5 ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] 83 – 71 91 73 75 74 72 85 93 78 76 87 92 62 79 70 88 89 98 80 84 94 90 [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ 81 [ Reference , , , IL6 p16 , and , 𝛾 IL-6 and , 𝛼 𝛾 ,and LINE-1 IFN p15 APC LINE-1 Sat LINE-1 p53 iNOS LINE-1 Alu Alu Alu p53 F3 ; IFN- , TLR-2 ; ; TLR2 NOS3 ,and , MAGE1 IL-6 RASSF1A and GCR , hypermethylation p15 LINE-1 ICAM-1 ,and ,and ,and , Hypomethylation at and Hypermethylation Differential methylation at NOS2A Differential methylation at TLR2 hypermethylation at F3 hypomethylation; IL-6 Hypomethylation at NBL2 Hypomethylation at Alu Hypomethylation Hypermethylation at Alu hypomethylation at Hypomethylation at and Hypomethylation at the TSP50 Hypomethylation at p53 hypermethylation at Hypomethylation at severalloci Hypomethylation Bisulfite PCRpyrosequencing ? Bisulfite PCRpyrosequencing Bisulfite PCRpyrosequencing Bisulfite PCRpyrosequencing Bisulfite PCRpyrosequencing Bisulfite PCRpyrosequencing MethodHPLC Bisulfite PCRpyrosequencing Hypomethylation at DNA methylation changes Bisulfite PCRpyrosequencing Bisulfite PCRpyrosequencing Differential methylation Bisulfite PCRpyrosequencing Bisulfite PCRpyrosequencing Hypomethylation Bisulfite PCRpyrosequencing Hypomethylation at Arrays Bisulfite PCRpyrosequencing Methylation-specific PCR Hypermethylation Bisulfite PCRpyrosequencing Hypomethylation at Bisulfite PCRpyrosequencing Hypomethylation at Arrays ELISA Bisulfite PCRsequencing Hypermethylation at , , , , , , ); , , , ); 𝛼 ,and APC HIC1 ) IL-6 iNOS L1HS , NOS3 , L1Ta , MLT1D Sat ,and 𝛾 AluSx , Alu HIC1 - p53 iNOS , IL-4 D4Z4 ,and TLR-2 p15 , Alu Alu Alu Alu ERV9 , AluYd6 Alu Alu ,and GCR , , , , , , ,and , MAGE1 , IFN , IL-6 , OGG1 p16 ,and IL-6 and Alu , , Alu , ,and IL-6 Foxp3 𝛾 ,and Global High-resolution GC-MS Hypomethylation Global ,and ACSL3 , MAGE1 ICAM-1 , hTERT and RASSF1A ,and and L1PA2 𝛾 , subfamilies ( Alu p53 CRAT LINE-1 LINE-1 LINE-1 LINE-1 NOS2A LINE-1 , LINE-1 , , subfamilies ( , IFN , subfamilies ( LINE-1 p15 NBL2 Epigenome-wide ERV1 ICAM-1 p53 LINE-1 , IFN- , p16 LINE-1 Tandem repeats AluYb8 , Table 1: Continued. Alu F3 L1PA5 NOS1 LINE-1 TLR-2 p16 F3 LINE-1 HERV Treg cells from PBL PBL PBL PBL PBL PBL PBL PBL PBL PBL PBL Target tissuePBL Target DNA region PBL PBL PBL PBL (cord blood and atyears) 9 PBL PBL 141 asthmatic children and 70 controls, USA 940 children, USA Buccal cells 158 petrochemical workers and 50 control office workers, Bulgaria 63 steel workers, Italy 78 gasoline filling attendants, 58 control office worker, Italy 776 elderly individuals, USA 60truckdriversand60controls, China 718 elderly individuals, USA 120maleworkers,Italyand China 78 gasoline filling attendants, 77 urban traffic officers, control 57 office workers, Italy 777 elderly individuals, USA 706 elderly individuals, USA Study population 48 adult nonsmokers, Belgium 70 Inuits, Greenland 43 women undergoing IVF, USA 49 nonsmoking coke-oven workers and 43 controls, Poland 56 mother-child pairs, USA PBL (cord blood) 358 mother-child pairs, USA 86 adults, South Korea 46 preadolescent girls, Egypt Saliva Epigenome-wide 67 industrial workers, 65 residents, and 45 rural controls, Thailand 6mte-hlpisUAPBL(cordblood)164mother-childpairs,USA 53 mother-child pairs, USA Global PBL (cord blood) Exposure (chemical) Air pollution (including PM and PAH) POPs Bisphenol A 6 BioMed Research International ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] ] 95 97 99 96 111 113 113 115 117 112 118 121 119 114 110 103 122 105 120 109 106 [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ [ Reference ; , , MYO1G RUNX3 AluYb8 F2RL3 F2RL3 AHRR AHRR AluYb8 AHRR GFI1 GPR15 PTPRO hypomethylation in ,and and , and other loci DMR differential and No consistent changes Hypomethylation Differential methylation at several loci Hypomethylation at Differential methylation 10 in genes including CYP1A1 Hypermethylation at loci IGF2 methylation Hypermethylation at AluYb8 boys Hypomethylation at AHRR Hypomethylation Hypomethylation at hypomethylation at CNTNAP2 Hypomethylation at Differential methylation at 910 loci; hypomethylation at AHRR Differential methylation at 19 Differential methylation in 8 genes; Hypermethylation at AXL Flow cytometry immunodetection of 5-mC; bisulfite PCRpyrosequencing ELISA Method DNA methylation changes Arrays Bisulfite PCRpyrosequencing Hypermethylation Arrays Arrays Arrays Arrays Bisulfite PCRpyrosequencing Arrays MALDI-TOF MS Arrays Bisulfite PCRpyrosequencing Differential methylation Bisulfite PCRpyrosequencing Hypermethylation at Arrays Arrays Arrays Bisulfite PCRpyrosequencing Arrays Hypomethylation at Bisulfite PCRpyrosequencing Hypermethylation Bisulfite PCRpyrosequencing , , ); , ); H19 ) ,and L1HS L1Ta MLT1D AluSx Alu and , Alu ERV9 AluYd6 , ,and AluYb8 , IGF2 AXL 𝑆𝑎𝑡𝛼 F2RL3 ,and ,and LINE-1 6p21 (1 locus) LINE-1 L1PA2 subfamilies ( LINE-1 2q37 (2 loci) , subfamilies ( subfamilies ( LINE-1 Epigenome-wide Epigenome-wide Epigenome-wide Epigenome-wide Epigenome-wide ERV1 AHRR AluYb8 Table 1: Continued. Alu Global; L1PA5 LINE-1 HERV Sperm Target tissuePBL Target DNA region PBL Placenta Epigenome-wide PBL PBL PBL PBL PBL PBL Child PBL Epigenome-wide 30 mother-child pairs, USA PBL (cord blood) Global 262 fertile men, Greenland/Poland/Ukraine Study population 685 adults, USA 192 mother-child pairs, Spain Placenta 177 adults, Germany 36 mother-child pairs (18 nonsmokers and 18 smokers), USA 1,062 mother-child pairs, Norway PBL (cord blood)132 mother-child Epigenome-wide pairs, Canada PBL (adolescents) Epigenome-wide 418 mother-child pairs, USA206 mother-child pairs, USA PBL (cord blood) Placenta DMRs of Epigenome-wide 374 adults, Italy 3,588 adults, Germany 399 African American youths, USA 261 adults, Italy 8mte-hlpisUAPBL(cordblood) 380mother-childpairs,USA 107 African American young men, USA 348 mother-child pairs, USA Child buccal cells 111 adult African American women, USA 272 mother-child pairs, USA Child buccal cells Epigenome-wide 173 mother-child pairs, USA527 mother-asthmatic child pairs, USA Child buccal cells Exposure (chemical) PFASs EDs Tobacco smoke BioMed Research International 7

Numerous studies focused on the vulnerable subpopula- promoter in blood cells was associated with BPA exposure tion of elderly people [71–74]. Hypomethylation of repeated in a cohort of women undergoing in vitro fertilization sequences (LINE-1 and, in some cases, also Alu)wasreported (IVF) [62]. In a survey of prepubescent Egyptian girls [94] to be associated with increased pollutant concentrations. higher urinary BPA concentrations were associated with In addition, methylation changes of specific genes involved lower genomic methylation and, interestingly, many affected in inflammatory and immune response pathways were genes were among those whose expression changes had been observed, which were regarded as modifiers of the association previously associated with BPA exposure. between air pollutants and reduced lung function [74]. Another class of emerging POPs is represented by per- Other studies investigated effects of air pollution on DNA fluoroalkyl substances (PFASs), which include a variety of methylation in children, showing changes in genes involved compoundswidelyusedinmanyindustrialprocessesand in asthma morbidity [75] or nitric oxide metabolism in products. Cross-sectional associations between serum PFASs airways [76, 77]. and LINE-1 DNAmethylationwereevaluatedinanAmerican Lower global DNA methylation [78] and hypermethyla- population highly exposed via contaminated drinking water tion of specific genes [79, 80]inumbilicalcordwhiteblood [95]. A significant association was found for some but not all specific PFASs. To explore the possible effects on male cells were shown to be associated with maternal exposure to 𝛼 airborne PAHs, pointing to a possible prenatal environmental reproduction, global methylation and LINE-1, Alu,andSat epigenetic origin of childhood diseases. methylation levels were directly assessed in sperm DNA from fertile men from Greenland, Poland, and Ukraine Workers of different job sectors, exposed to particulate characterized by a wide contrast to PFASs plasma levels. No matter, PAHs, and benzene, have been also monitored for strong consistent associations between PFASs exposure and possible DNA methylation changes in peripheral blood cells any of the sperm methylation biomarkers could be detected [81–91]. In general, DNA repetitive elements, such as LINE- [96]. 1, Alu,andHERV, have been analyzed, in addition to Three studies explored the influence of maternal POPs specific genes including p53, p15, p16, APC, RASSF1A, HIC1, serum concentrations on DNA methylation of umbilical cord iNOS, hTERT,andIL-6. Each specific gene and subfamily of blood cells. Global DNA hypomethylation appeared to be repetitive sequences seem to respond independently to the associated with the serum level of specific PFASs97 [ ]. Inter- exposure and no set of sequences has yet emerged as an ideal estingly, two studies examining the effects on various families reporting system of epigenetic effects. In addition, exposure of repeated DNA sequences [98, 99] showed that the asso- conditions and methods of assessment were quite heteroge- ciation between serum xenoestrogen contamination and Alu neous making any overall conclusion impossible. These stud- hypomethylation in cord blood was influenced by the baby ies support the notion that epigenetic biomonitoring is still a gender, in agreement with the hypothesis of a differential, new area of environmental health studies that necessitates of gender-dependent, susceptibility to prenatal EDs exposure. international coordination, methodological harmonization, and mechanistically sound interpretation of results. 2.4. Antibiotics. Low birth weight (LBW) has been associated 2.3. Persistent Organic Pollutants (POPs) and Endocrine with common adult-onset chronic diseases. Its etiology is Disruptors (EDs). This heterogeneous class of chemicals is multifactorial and exposure to antibiotics has been reported strongly suspected to interfere with the human hormonal to increase LBW risk. Among possible mechanisms underly- homeostasis and to hamper reproductive integrity, especially ing this association, epigenetic changes have been proposed. when exposure occurs during the pre- and perinatal In the US Newborn Epigenetics Study (NEST), the life stages. In a cohort of Greenland Inuits, in DNA methylation status of the DMRs of a variety of growth reg- extracted from blood samples, Alu sequences showed ulatory imprinted genes (IGF2,H19,MEST,PEG3,PLAGL1, significant hypomethylation as a function of increasing SGCE/PEG10, NNAT, and MEG3) was analyzed, in umbilical 󸀠 blood concentration of p,p -DDT [1,1,1-trichloro-2,2-bis(p- cord blood cells, in relation to the infant birth weight and 󸀠 chlorophenyl)ethane], its main metabolites p,p -DDE maternal (self-reported) antibiotic use. Methylation at IGF2, [1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene], 𝛽-HCH H19, PLAGL1, MEG3,andPEG3 was associated with maternal (hexachlorocyclohexane), oxy- and 𝛼-chlordane, mirex, sum antibiotic use, although only methylation at the PLAGL1 of PCBs (polychlorinated biphenyls), and sum of POPs; DMR was also associated with birth weight [100]. the methylation level of LINE-1 sequences showed a similar inverse trend with the exposure, albeit not statistically 2.5. Tobacco Smoke. The potential epigenetic links between significant92 [ ]. In agreement with these data, the blood current and prenatal smoking and smoking-related diseases concentrations of various organochlorine pesticides in a are extensively discussed in recent review papers [101, 102]. cohortofhealthyKoreanswereinverselyassociatedwiththe Smokers and nonsmokers have been compared by high methylation level of Alu but not of LINE-1 sequences [93]. throughput methods in several cohorts of adult and young Another widely debated chemical is the ubiquitous people, with some consistent alterations detected involving bisphenol A (BPA), a monomer used in epoxy resins DNA methylation differences at specific positions in the and polycarbonate plastics. Exposure to BPA and possi- F2RL3 [103–107], in the AHRR [108–112], and in GPR15 ble methylation alterations were studied with genome-wide genes [112], which emerged as strong candidates to pre- approaches. Significant hypomethylation of the TSP50 gene dict smoking-related negative health outcomes. Epigenetic 8 BioMed Research International changes in the offspring of mothers smoking during preg- potential consequences on paternal imprinting [130]. nancy have been characterized by genome wide approaches, Notably, a trend of increased demethylation with alcohol to contribute unraveling the mechanistic pathways of some, consumption was shown, in sperm of male volunteers, at the well-known, prenatal smoking-related adverse effects. Accu- H19 and IG DMRs, with a significant difference observed at mulating data indicate that prenatal exposure to tobacco the IG-DMR between the nondrinkers and heavy alcohol smoke is associated with reproducible epigenetic changes at a consumers [131]. global and gene-specific level that persist well in childhood and adolescence [97, 113–122]. Changes have been found, 2.6.2. Maternal Effects. Newborn methylomes contain mole- among others, in genes involved in transcription, in oxidative cular memory of the individual in utero experience [132]. stress and detoxification pathways, and in repetitive elements, In the above chapters we have discussed various examples even though the biological significance of a variety of altered linking maternal environmental exposure to newborn DNA loci remains to be understood. methylation, as assessed through cord blood cell analysis. However, the maternal impact appears to extend beyond that 2.6. Parental Influence of specific chemical contaminants, as also metabolism, nutrition, and stress seem to influence the offspring methylome. 2.6.1. Paternal Effects. In humans, there is sparse evidence In an American black mother-child cohort study, linking lifestyle paternal factors with the offspring epigenome genome-wide analysis of cord blood cells showed that about [123]. Paternal obesity has been associated with hypomethy- 20 CpG sites in cancer and cardiovascular disease relevant lation at the IGF2 [124]andMEST, PEG3 and NNAT [125] genes were highly significantly associated with maternal BMI DMRs in the offspring cord blood cells, independently of [133]. maternal obesity and other potential confounders. The epigenetic consequences of prenatal famine and It is noteworthy that global sperm DNA methylation caloric restriction have been evaluated in a cohort of people has been shown to increase, on average, by 1.76% per year conceived in the winter 1944-45 during a severe famine at the [126],andanin-depthanalysisofthemethylomeintwo endofWorldWarII(DutchHungerWinterFamiliesStudy). sperm samples collected 9–19 years apart from 17 fertile These people appear to bear the consequences of prenatal American men has shown several age-related changes [127]. stress later in life, including an adverse metabolic profile One hundred and thirty-nine regions were significantly and (suboptimal glucose handling, higher BMI, and elevated consistently hypomethylated and 8 regions were significantly total and low-density lipoprotein cholesterol) and increased hypermethylated with age; 117 genes were associated with risk of schizophrenia. While the overall global methylation these regions of methylation alterations (promoter or gene levels in their blood cells appear to be unaffected134 [ ], body), with a portion of them surprisingly located at genes significant DNA methylation changes have been shown, at previously associated with schizophrenia and bipolar disor- several specific loci corresponding to imprinted genes or to der. In the same samples, LINE-1 showed global hyperme- genes implicated in growth and metabolic diseases, including thylation with age, while another study, aimed at relating IGF2, IL-10, LEP, ABCA1, GNASAS,andMEG3 [135–139]. numerous variables with sperm DNA methylation [128], A genome-scale analysis has demonstrated that differential did not show age-dependent changes in LINE-1, Alu,and DNA methylation preferentially occurs at regulatory regions Sat𝛼 methylation level, probably because of the narrow age and maps to genes enriched for differential expression during contrast of studied populations. In the latter study, personal early development [140]. Changes have been also shown characteristics and habits, body mass index (BMI), semen to depend on the sex of the exposed individual and the quality parameters, sperm chromatin integrity, biomarkers of gestational timing of the exposure [137]. accessory gland function, and the plasma concentration of Folate plays an essential role in one-carbon metabolism reproductive hormones were related to sperm DNA methy- involving remethylation of homocysteine to methionine, lation in a cohort of 224 men of proven fertility, living in which is a precursor of S-adenosylmethionine, the primary threeEuropeanregions,Greenland,Warsaw,andKharkiv. methyl group donor for most biological methylations, includ- The geographical location emerged as the main determinant ing DNA methylation. A few pilot studies considered possible of the methylation level in repetitive sequences and no other effects of maternal intake of methyl-donor compounds on consistent associations between methylation markers and the their infant DNA methylation. Compared to infants born assessed variables were identified across countries [128]. to women reporting no folic acid intake before or during Until now, only three human biomonitoring studies pregnancy, methylation levels at the H19 DMR in umbilical addressed the impact of environmental factors on sperm cord blood leukocytes decreased with increasing folic acid DNA methylation. One is the already cited investigation on intake, the decrease most pronounced in the male offspring the possible effects of PFASs exposure on LINE-1, Alu,and [141]. In another study [142], increased methylation at the Sat𝛼 methylation level [96], which did not show consistent maternally IGF2 imprinted gene and decreased methylation PFASs-associated alterations. The other two studies at the maternally imprinted gene PEG3 and at the repetitive focused on occupational radiation exposure and alcohol transposonic sequence LINE-1 were associated with folic acid consumption. An increase of hypermethylated spermatozoa supplementation after the 12th week of gestation but not was shown in radiation-exposed workers [129]. Some years during the first trimester or before conception. Finally, a third ago,alcoholhadbeenshowntoreducethemethyltransferase study [143] did not detect any major association between mRNA levels in sperm of chronically treated rats with intake of methyl donor nutrients (vitamin B12, betaine, BioMed Research International 9 choline, folate, Cd, Zn, and Fe) during pregnancy and LINE-1 [149–151], but, due to their still very small number, they will DNA methylation. not be further addressed. The results of a human epidemiological study, conducted These studies were prompted by the observations of heri- in rural populations in Gambia experiencing pronounced table traits unexplained by Mendelian inheritance. However, seasonal fluctuations in nutritional status and diet, support only a subset of studies showing epigenetic transgenera- a role for periconceptional maternal plasma concentration tional effects also provided evidence of potentially heritable of key micronutrients involved in one-carbon metabolism epigenetic changes in the exposed gametes. In this review, on infant DNA methylation [144, 145]. The study focused only those studies that analysed possible DNA methylation on the analysis of human candidate metastable epiallelic loci changes in the male or female germline of exposed animals [25, 146]. Metastable epialleles are genomic regions where have been considered. epigenetic patterning occurs before gastrulation in a stochas- Overall, 24 papers were reviewed, 19 reporting studies tic fashion leading to systematic interindividual variation in mice and 5 in rats. Studies on the possible induction of within one species. Their existence is well documented in epigenetic alterations in germ cells were grouped according themousesincethepioneeringstudiesontheagouti viable to the exposure time window: either prenatally, during the V𝑦 yellow (𝐴 ) locus, and, in this model, maternal diet has been critical period of germline differentiation (10 studies), or shown to modulate the establishment of the epigenetic marks postnatally, in prepuberal or adult male (12 studies) or female (reviewedin[38]). The human survey has shown that DNA (5 studies) animals. One of the 5 studies on exposure of the methylation at metastable epialleles in lymphocytes and hair female germline was carried out in vitro (Table 2). follicle cells of infants conceived during the rainy (“hungry”) From the emerging overview, the research objectives season is significantly different from that of infants conceived still appear rather sparse: a group of studies evaluated the in the dry (“harvest”) season, providing first evidences of a impact of metabolic changes, due to undernourishment [152], lasting and systemic effect of periconceptional environment low-protein [153], folate-deficient154 [ ], zinc-deficient155 [ ], on human epigenotype. obesogenic, and/or diabetogenic diets [149, 156–158]. Other Maternal depression has been associated with a higher studies investigated the effects of specific compounds, many risk of LBW and hypermethylation at the MEG3 DMR of of which belong to the class of so-called endocrine disrupters, infants [147]. Furthermore, LBW infants had lower methy- including vinclozolin [159–161], methoxychlor [162], dioxin lation at the IGF2 DMR, while high birth weight infants [163], and bisphenol A [164–166]. The remaining studies had higher methylation at the PLAGL1 DMR compared deal with a heterogeneous group of potentially epigenetics withnormalbirthweightinfants.Thus,imprintedgene disrupting agents: particulate air pollution [167], ethanol plasticity may play a role in the observed association between [168, 169], tamoxifen [170, 171], fenvalerate [172], and sodium depressive mood in pregnancy and LBW. fluoride [173]. Preliminary human studies are providing first evidence Regarding the genomic targets, several studies focused supporting the conclusion that traumatic experiences can on a few loci, either maternally (Mest, Snrpn, Igf2r,and result in lasting, broad, and functionally organized DNA Peg3)orpaternally(H19, Meg3, and Rasgrf1)imprinted methylation signature in several tissues in offspring. A (methylated). Other studies evaluated changes of methyla- Canadian study (Project Ice Storm) was set up some months tion in metabolism-related genes, such as the LPLase,the after the 1998 Quebec ice storm by recruiting women who Ppar𝛼,ortheLep gene. One paper included the analysis of had been pregnant during the disaster, scoring their degrees methylation of Line-1 repeated sequences [173]. A few studies of objective hardship and subjective distress. Thirteen years assessed possible changes of total DNA methylation, whereas later, genome-wide DNA methylation profiling in T cells the most recent papers report analyses at a genome-wide obtained from 36 of the children was assessed. Prenatal level. With a few exceptions [153, 160, 163, 170, 173], the maternal objective hardship (but not maternal subjective studies showed some kind of exposure-related effect. Both distress) was correlated with DNA methylation levels in increase and decrease of methylation levels were reported. 1675 CpGs affiliated with 957 genes predominantly related to As pointed out before, the studies are too scattered and too immune function [148]. heterogeneous in the analytical methods to allow drawing general conclusions; however, some hints are emerging, like increased methylation of maternally imprinted and decreased 3. Rodent Experimental Studies on methylation of paternally imprinted genes in the exposed the Induction of Epigenetic Changes male germline. Interestingly, the few studies on oocyte in the Germline exposure show an opposite effect of treatment on the same maternally imprinted genes, which seem to respond with a In the last few years, experiments in rodents started to test decreased methylation level. the hypothesis that exogenous exposure to some measur- In some studies the impact of exposure on germ cells has able factor, during a controlled time window, could induce been compared with the impact on somatic cells. Depending epigenetic changes in the germline. The large majority of on the type and time of exposure, some studies showed that these experiments investigated changes of DNA methylation the methylation control mechanisms were more robust in at a global, gene-specific, or genome-wide level and will be the somatic than in the germ cells, as in the case of prenatal discussed in this section. A few considered also other types treatment with ethanol [168]ormethoxychlor[162], but of epigenetic changes such as the sperm microRNA content a reversed sensitivity was also observed, as in the case of 10 BioMed Research International ] ] ] ] ] ] ] ] ] ] ] 161 152 163 159 162 162 154 168 160 160 160 [ [ [ [ [ [ [ [ [ [ [ Reference , and Snrpn and , Peg3 H19 Meg3 Mest Meg3 ,and and H19 Snrpn Igf2r , ; no effect inthe other in sperm; no effect in Mest es were analyzed, they are specified). H19 Peg3 ,and in sperm; no effect in somatic tissues and significantly increased methylation of Snrpn , Peg3 No detectable DNA methylationsperm; changes in in liver and muscleincreases cells of significant methylation in Altered methylation detected at multiple sequences involving both hypo- and hypermethylation events 57 genomic regions had alteredprofiles methylation in sperm fromfolate-deficient males diet; exposed both hypo- to and hypermethylation were observed;differences observed methylation promoter in regionsimplicated in genes of development and withthe functions central in nervous system, kidney,digestive spleen, tract, and muscle tissue,associated and with of diabetes, genes autoimmune diseases, neurological diseases, autism, schizophrenia, and cancer 166 differentially methylated regions,were of which hypomethylated 111 and 55 hypermethylatedundernourished in relative to control mice; the bisulfite pyrosequencing validation confirmed 17/24 hypomethylated and 0/8 hypermethylated regions In sperm, highly significant hypomethylation and Significantly decreased methylation of Meg3 Mest Significantly decreased methylation of significantly increased methylation of and hypermethylation; less evident effects tissues somatic in somatic tissues genes and in somatic tissues Highly significant 3% decreasemethylated in CpGs number the of of Arrays Method DNA methylation changes digestion Bisulfite PCR Bisulfite PCR Bisulfite PCR Bisulfite PCR Bisulfite PCR pyrosequencing pyrosequencing pyrosequencing pyrosequencing pyrosequencing Methylation-specific restriction Endonuclease Bisulfite PCRsequencing detectable No DNA methylation changes Bisulfite PCRsequencing detectable No DNA methylation changes Bisulfite PCRsequencing detectable No DNA methylation changes Mass spectrometry; arrays , , , , Peg3 Peg3 Peg3 Peg3 ,and Mest Mest Mest Mest , , , , Peg3 ,and ,and ,and ,and , Igf2r Meg3 Meg3 Meg3 Meg3 LPLase LPLase LPLase Global; , , , , Multiple, unspecified Snrpn Snrpn Snrpn Snrpn Snrpn epigenome-wide H19 H19 H19 H19 Epigenome-wide Sperm Sperm animals animals and skeletal Target tissue Target DNA region cells in adult liver, skeletal Testis cells of Testis cells of Testis cells of Testis cells of adult animals adult animals adult animals adult animals muscle cells in muscle cells in muscle cells in muscle cells in sperm of adult Sperm, plus tail, Sperm, plus tail, Sperm, plus tail, Sperm, plus tail, Sperm, plus liver liver, and skeletal liver, and skeletal liver, and skeletal muscle, and brain 6-day-old animals 6-day-old animals 6-day-old animals 6-day-old animals; Dose and time of exposure In utero exposure: i.p. injection of 10 mg/kg/day, between day 8 and day 15 of gestation In utero exposure: i.p. injection of 100 mg/kg/day, between day 8 and day 15In of utero gestation exposure: i.p. injection of 100 mg/kg/day, between day 8 and day 15In of utero gestation exposure: i.p. injection of 100 mg/kg/day, between day 8 and day 15 of gestation In utero exposure: i.p. injection of 50 mg/kg/day, between day 10 and day 18 of gestation In utero exposure: i.p. injection of 10 mg/kg/day, between day 10 and day 18 of gestation In utero exposure: p.o. administration of 0.5between g/kg/day day 10 and day 18 of gestation In utero exposure: treatment of dams with folate-deficient diet from two weeks beforethroughout mating gestation and lactation In utero exposure: nutritional restriction between day 12.5 andgestation day 18.5 of Adult exposure: i.p. injection of 10 mg/kg/day forconsecutive 8 days In utero exposure: i.p. injection of 2 orbetweenday9andday19ofgestation 10 ng/kg/day Rats Rats Rats Rats Mice Mice Mice Mice Mice Mice animals Exposed Exposure Flutamide Procymidone Vinclozolin Vinclozolin Vinclozolin DioxinMethoxychlor Mice Ethanol Folate-deficient diet Undernourishment Methoxychlor Table 2: Synopsis of papers reporting effects of experimental treatments DNA on methylation of rodent male or female germ cells additional(when tissu BioMed Research International 11 ] ] ] ] ] ] ] ] ] ] ] ] ] 171 153 173 175 172 157 158 170 167 156 149 169 166 [ [ [ [ [ [ [ [ [ [ [ [ [ Reference , Ace and ICR H19 Igf2-H19; methylation H19 hypermethylation Lep DMR on day 35 after methylation in sperm; Ptgfrn Peg3 𝑃𝑝𝑎𝑟𝛼 ,and methylation level was hypomethylation; significant 𝛼 Pml , Ppar Hap1 Nr3c2 , DMRs were not significantly altered by Significant Significantly increased global methylation in 10-week exposed samples, which persistedexposure after interruption; methylation changes abolished by use of air filters Paternal prediabetes altered overall methylome patterns in sperm, with adifferentially large methylated portion genes of overlapping with that of pancreatic islets in offspring Foxo3a Evident demethylation was observed inmethylation the pattern of No detectable DNA methylation changes Significantly reduced methylation at global methylation level unaffected No effects of diet on treatment; the methylation patterns of Snrpn maternal diabetes No detectable DNA methylation changes About 25% reduction in globalboth methylation whole in testicular cells and spermatids overall, sperm cytosine methylation patternslargely were conserved under various dietary regimes DNA methylation of imprinted genes wasaltered; the not significantly decreased and the level was significantly increasedfed in high-fat mice diet compared with control mice assay arrays ELISA Arrays Arrays Method DNA methylation changes COBRA COBRA antibody Bisulfite PCR pyrosequencing 5-methyl cytosine testis sections with anti-5-mC antibody flow cytometry after immunostaining with and methyl acceptance ELISA; semiquantitative Cytosine extension assay Bisulfite PCRsequencing Significant hypomethylation at the Bisulfite PCRsequencing Hypomethylation Bisulfite PCRsequencing detectable No DNA methylation changes Bisulfite PCR sequencing; Bisulfite PCR sequencing; Bisulfite PCR sequencing; immunohistochemistry of , , , , 𝛼 , Peg3 ,and Peg5 Ppar ; , Peg3 global Grb10 , ,and , global , , Lep Rasgrf1 Ascl2 , Peg3 , Mest Snrpn Plagl1 Igf2r , Global Global Olfr151 Cdkn1c 𝑃𝑝𝑎𝑟𝛼 and Snrpn , Rasgrf1 Igf2-H19 , Line-1; , , , H19 H19 Igf2-H19; H19 epigenome-wide Epigenome-wide Epigenome-wide Kcnq1 Dlk1 and Peg3 H19 Igf2r Table 2: Continued. Sperm Sperm Sperm Sperm Sperm Sperm Sperm Sperm Sperm animals Oocytes Oocytes elongated spermatids Testis cells; Target tissue Target DNA region cells in adult Sperm, plus liver Dose and time of exposure Adult exposure: p.o. administration of 5.9days g/kg/day over for a 29 period ofAdult 5 weeks exposure: p.o. administration of 10 mg/kg/day30 days for Adult exposure: p.o. administration of 100drinking mg/L water in for 35 days Adult exposure: p.o. administration of 0.4days mg/kg/day a 5 week for 60 days adult exposure: p.o. administration of 0.4days mg/kg/day a 5 week for 60 days Neonatal exposure: 5 daily subcutaneous injections0.4 of mg/kg/day BPA starting one daybirth after Adult exposure: 3 or 10 weeks ofin exposure polluted to sites ambient air Sperm sampling either immediately orweeks 6 after the end of 10-week exposure Adult exposure: 10-week exposure to high-fat dietweeks from of 5 age Adult exposure: 13-week exposure to high-fat diet,3weeksofage,plusstreptozotocini.p. from injection at 12 weeks ofAdult age exposure: 9-week exposure to low-protein diet from 3 weeks of age Adult exposure to acetophenone Adult exposure: single i.p. injection of 230streptozotocin,15,25,or35daysbefore mg/kg oocyte collection after superovulation Adult exposure: 12 weeks of feeding withprior high-fat to diet oocyte collection by superovulation Rats Rats Rats Mice Mice Mice Mice Mice Mice Mice Mice Mice Mice animals Exposed Exposure Ethanol Fenvalerate Sodium fluoride Tamoxifen Tamoxifen Bisphenol A Particulate air pollution High-fat diet High-fat diet plus streptozotocin subdiabetogenic treatment Low-protein diet Olfactory fear conditioning Streptozotocin diabetogenic treatment High-fat diet 12 BioMed Research International ] ] ] 155 165 164 [ [ [ Reference H19 ,and Igf2r , no effect on ; Mest genes H19 Peg3 , Igf2r genes; no effect on DNA methylation was reduced to aboutcontrol 60% levels of in zinc-deficient oocytes Significant dose-dependent reduction of methylation in Significantly decreased methylation at thenot but low, at the high, BPASnrpn dose, in Method DNA methylation changes Bisulfite PCR pyrosequencing Immunocytochemistry with anti-5-mC antibody Bisulfite PCRsequencing , Peg3 Igf2r , ,and Snrpn Mest Global , Igf2r and , H19 H19 Table 2: Continued. Oocytes Oocytes Oocytes Target tissue Target DNA region g/kg b.w. either by daily 𝜇 Dose and time of exposure Adult exposure: 5 days of feeding withprior zinc-deficient to diet oocyte collection by superovulation Neonatal exposure: 20 or 40 hypodermal injections from postnatal day 7 to postnatal dayintraperitoneal 14 injections or administered by each fifth day betweenpostnatal and20,priortocollectionofovarian days 5 oocytes In vitro exposure to 312 or days 300 nM of during follicleto culture antral from stage preantral Mice Mice Mice animals Exposed Exposure Zinc-deficient diet Bisphenol A Bisphenol A BioMed Research International 13 prenatal exposure to dioxin [163]. It is conceivable that each sired offspring. However, differentially expressed genes in tissue might respond accordingly to its specific developmen- the placenta did not match differentially methylated genes in tal program; therefore, studies of exposure during the period sperm, suggesting that mechanisms other than DNA methy- of prenatal germline differentiation are especially important lation might be involved in the transgenerational transmis- for assessing any environmental epigenetic impact on the sion of epigenetic messages, like sperm histone modifications. germline. Similarly, in utero undernourishment induced hypome- The evaluation of an epigenetic impact of exogenous thylationofseveralgenesinsperm,aswellaschangesof factors on the germline is still in its infancy. A critical issue expression of metabolic genes in the brain and liver of late that has not yet been thoroughly addressed is if and how gestation fetuses sired by the exposed animals; interestingly, much the DNA methylation changes have a functional impact the genes whose expression was altered in the fetuses mapped on the gene expression level and have any causal role on the close to differentially methylated regions in sperm, although male germ cell toxicity that is sometimes induced, as after differential methylation was not transgenerationally retained prenatal vinclozolin [161]orethanol[168] exposure. [152]. The authors concluded that “... it is unlikely that A group of studies aimed mainly to evaluate possible these expression changes are directly mediated by altered transgenerational consequences of epigenetic alterations in methylation; rather, the cumulative effects of dysregulated the germline. The simplest hypothesis was that methylation epigenetic patterns earlier in development may yield sus- changes in gametes could resist zygotic reprogramming tained alterations in chromatin architecture, transcriptional andhavefunctionalconsequencesintheoffspringsiredby regulatory networks, cell type, or tissue structure.” exposed animals. Postweaning growth delay and decreased methylation Indeed, in mice, ethanol exposure in utero was shown to at the H19 ICR CTCF binding sites were observed in the induce H19 demethylation in sperm of adults, as well as in offspring of adult mice treated with ethanol, although no the brain cells of their offspring, with a good concordance decrease of H19 DNA methylation was detectable in the between the CpG demethylation patterns across cell types sperm DNA [169]. Methylation was significantly increased and generations [168]. In another study testing possible in Peg3 and significantly decreased in H19 8-cell embryos transgenerational effects of tamoxifen in rats171 [ , 174], the siredbymalemicetreatedwithsodiumfluoride,whileno offspring sired by tamoxifen-treated animals showed an change of methylation was detected in sperm [173]. Increased increased incidence of embryonic resorptions, and resorbed methylation of a number of imprinted genes, associated with embryos (but not normal ones) carried methylation errors downregulation of transcription, was detected in the resorb- similar to those detected in the sperm of exposed fathers. ing embryos sired by tamoxifen-treated male rats, in spite In male mice a prediabetic condition closely resembling of the fact that their sperm did not show DNA methylation the metabolic abnormalities of human prediabetes can be changes in any of the 9 analyzed imprinted genes [170]. induced by high-fat diet and chemical treatment; these mice The complexity of the interplay between environmen- transmit to their offspring glucose intolerance and insulin tally sensitive epigenetic markers in sperm and epigenetic resistance [156]. Epigenomic profiling in offspring pancreatic modulation of development in the following generation is islets identified changes in cytosine methylation at several further illustrated by a recently published report on the insulin signaling genes, and these changes correlated with the transgenerational consequence of paternal exposure to a expression of these genes. The analysis of cytosine methyla- conditioning olfactory experience [175]. The F1 progeny of tion profiles in sperm of prediabetic fathers showed several conditioned mice reacted just like the fathers with enhanced alterations and a large proportion of differentially methylated response, in spite of never being conditioned themselves. genes overlapped with that of the offspring pancreatic islets. The F1 neuroanatomy was also affected. Behavioral sensitivity Bisulfite sequencing of some of these genes in blastocysts and neuroanatomical alterations in the nervous system were showed that they resisted global postfertilization demethyla- present also in the IVF-derived F1 generation and persisted tion and largely inherited cytosine methylation from sperm, until at least the F2 generation. Hypomethylation in specific suggesting that there might be intergenerational transmission CpG islands of the Olfr151 gene encoding for the specific odor of methylation profiles. receptor was detected in sperm of exposed mice. These find- However, other studies demonstrated that epigenetic ings led the authors “to hypothesize that relative hypomethy- inheritance via the gametes can be more complex than the lation of Olfr151 in F0 sperm may lead to inheritance of direct transmission of DNA methylation alterations, and a the hypomethylated Olfr151 in F1 Main Olfactory Epithelium crosstalk might exist between different levels of epigenetic (MOE) and F1 sperm, creating an inheritance cascade.” regulation across generations. However, the epigenetic mark was found in the sperm but not A genome-wide analysis of methylation changes in sperm in the MOE of F1 mice. Noting that DNA methylation and of mice exposed to a folate-deficient diet showed altered histone modifications are known to be dependent on each methylation profiles in genes implicated in development, other, the authors suggested that changes in the methylation chronic diseases such as cancer, diabetes, autism, and pattern in sperm DNA might have resulted in histone schizophrenia [154].Inthesamestudy,atwofoldgreater modifications around the olfactory gene in MOE DNA. resorption rate and an increased frequency of developmental The literature on effects of experimental exposure upon abnormalities were observed in pregnancies sired by exposed DNA methylation in rodent oocytes is less abundant com- males. Moreover, significant changes in the expression of over paredwiththatoneffectsinducedinsperm.Twopapers 300 genes were detected in the placenta of exposed-animals report undermethylation of maternally imprinted genes in 14 BioMed Research International oocytes exposed to bisphenol-A either in vivo [164]orin background. The best explanation for 75% of VMRs was the culture [165]. In addition to the challenge posed by working interaction of genotype with different in utero environments, with a little number of cells, experimental studies on female- including maternal smoking, maternal depression, maternal mediated epigenetic inheritance also face the difficulty of BMI, infant birth weight, gestational age, and birth order. strictly distinguishing a mechanism of epigenetic inheritance A prevalence of genetic over environmental determinants via the gametes from other mechanisms of epigenetic inheri- of interindividual variation of CpGs methylation has been tance, such as those based on adverse uterine environment or recently reported in large Scottish and Australian cohorts lactation-mediated effects. This issue emerged, for example, [184]. in a recent paper [158] showing functional alterations of Finally, age is expected to be a major variable affecting the methylation patterns in the Lep and Ppar𝛼 metabolic genes DNA methylation profiles in different tissues. In fact, recent in oocytes of mice treated for 12 weeks with a high-fat diet, as studies aimed at exploring the importance of epigenetic well as in the liver cells of their offspring. changes to the ageing process highlighted age-signatures of DNA methylation [185–187]. 4. Discussion One of the problems in drawing an overall pattern from published literature on environmental epigenetic effects DNA methylation is a life-essential process that modulates is due to the heterogeneity of detection methods and gene expression and drives cell differentiation in multicellular approaches. Several different methods have been developed organisms. Synergistically with other epigenetic mechanisms, for DNA methylation analysis and their advantages and draw- it allows cells and organisms to adapt to external changes, in a backs are discussed in excellent, recent reviews [188–191]. We timely way that mutational mechanisms could never meet. As havewitnessedinashorttimethepassagefromtheanalysis such, DNA methylation is unsurprisingly sensitive to external restricted to single specific regions to a global and genome- stimuli. At the same time and in contrast to mutations, wide scale. Even if, on purely theoretical considerations, the DNA methylation changes are reversible. This duality poses ideal choice would point at a technique able to measure a challenge to researchers who aim to establish possible links the entire methylome at a single-base-pair resolution in a between environmental exposure and epigenetic changes that particular cell system, researchers have to face other issues may have a long-lasting impact on cell function and ulti- related to time- and cost-effectiveness and make reasonable mately on health. Cancer, in all its forms, is the most typical compromises with their own scientific questions and the example of a disease associated with aberrant epigenetics, technology available. By and large, cost-affordable technolo- which may be triggered by environmental exposure [2, 176], gies are limited in their sensitivity to DNA methylation but ample evidence exists where erroneous epigenetic marks detection,likethoserelyingontheglobalimmunostaining also play prominent roles in neurological disorders such as of the 5-mCs or like pyrosequencing that analyzes only a Alzheimer’s disease, autoimmune diseases such as rheuma- limited amount of informative cytosines [192, 193]. On the toid arthritis, and cardiovascular diseases, among others [2]. other hand, technologies based on high-resolution methy- The path of environmental epigenetics will necessarily have to lation arrays [194] are able to measure countless sequences move from initially sparse association studies towards causal acrossthegenomebutarecostlyanddemandsophisticated relationships supported by biological plausibility. bioinformatics. The methylation analysis at targeted genes, The plasticity of the human epigenome and the difficulty likethoseimprinted,involvedinsomemetabolicpathway, to sort out major environmental effects from “background or supposedly metastable, and/or in repetitive elements noise” can be appreciated from the studies showing a sea- (transposonic or not) is a frequently used approach in sonality and weather influence on some DNA methylation environmental epigenetics. Interestingly, it is emerging that biomarkers analyzed in recent human biomonitoring studies repetitive elements, such as Alu and LINE-1,whichwere [177, 178] or the findings of genome-wide analyses that initially chosen simply as a proxy of the global methylation showed an influence of long-term shiftwork on DNA methy- level due to their abundance throughout the genome, respond lation at several loci [179–182]. These latter studies, prompted to environmental stress in a sequence-specific manner and by the evidence of an association between exposure to light have to be considered as separate entities [96, 98, 120, at night, circadian rhythms, and cancer risk, demonstrated 195]. An international methodological standardization and indeed methylation changes in many cancer-relevant genes harmonization effort would contribute to reaching more and pathways, but they need to be confirmed by independent solid evidence on the epigenetic impact of environmental replication in larger samples and supported by fundamental stressors. It certainly represents a Herculean task as the mechanistic research, before any firm conclusion can be human haploid DNA methylome contains approximately 30 drawn. million CpGs that exist in a methylated, hydroxymethylated, In addition, the fact that interindividual variation in or unmethylated state. methylation may also be a consequence of DNA sequence Notwithstanding such difficulties, environmental epige- polymorphisms that result in methylation quantitative trait netics may become a potent concept to fully assess the impact loci should not be overlooked. Teh and coworkers [183] of the exposome on human health [196]. In particular, the have investigated the genotypes and DNA methylomes of notionthat,inmammals,tissuedifferentiationismainly 237neonatesandfoundsome1500punctuateregionsof established during prenatal life, and fundamental DNA the methylome highly variable across individuals, termed methylation changes occur in preimplantation embryo and variably methylated regions (VMRs), against a homogeneous during gonadal differentiation, may support the hypothesis BioMed Research International 15 of prenatal origin of adult-onset diseases. To push science epigenetic changes in the germline of experimental rodents forward, epidemiological mother-child cohort studies and want to be related with human biomonitoring data. maternal exposure assessment will be instrumental. Infact,aspreviouslyshown,whereasthedatabaseon Also the bases of reproductive health are founded during environmental factors impinging on DNA methylation in prenatal life, with primordial germ cell differentiation and human leukocytes is already abundant, very few data exist gonad development, although the process of gametogenesis on the variables affecting DNA methylation in human germ will only be completed after puberty. This means that multiple cells, even in the most easily accessible sperm. Acknowl- exposure windows must be considered to assess possible edging the limitation of a comparison between two large, environmental effects on the gamete genetic and epigenetic but independent, studies, carried out in different cohorts, integrity. the increase of LINE-1 methylation reported in blood cells The results of the studies in rodents that we have in association with perfluorooctane sulfonate (PFOS) serum described in the previous section show that DNA methylation level [95], and the lack of an association between PFOS serum in germ cells can be altered by many different kinds of level and LINE-1 methylation in sperm [96]exemplifiesthe exposure during the fetal as well as the adult life. Still, these difficulty of any extrapolation between somatic and germline studies suffer of some limitations: more data are available on environmental epigenetics. themalethanonthefemalegermline,andonlyfewofthem Much more fruitful has been until now the field of male carriedouttheanalysesatthemostinformativegenomescale, reproductive clinical epigenetics [35, 36]. The review of data addressed the functional impact of epigenetic changes on the showing DNA methylation and other epigenetic changes gene expression level and related cell pathways, and took into in the sperm of subfertile patients was out of the scope consideration dose-effect relationships. Nevertheless, their of this paper. However, these data are important also for results are very important because they establish proof of reproductive environmental epigenetics because they seem principle demonstration that a variety of exogenous stressors to indicate a functional significance of DNA methylation may alter DNA methylation at developmentally important changes in the male germline. At the same time, they evidence imprinted or metabolic genes. the need to conduct specific epigenetic analyses on the As a target of environmental exposure, the germline sperm of men exposed to reproductive toxicants, with the meets a double risk, of compromising the reproduction awareness that their PBLs could not surrogate the relevant capacity of the exposed individual and transmitting possible target cells. Recently, the entire methylome of human sperm damage to the following generation. Some of the studies in has been analyzed at high resolution thanks to the most rats and mice indeed showed that treatment induced not advanced technologies [127, 197, 198]. While this dataset only DNA methylation changes in sperm but also phenotype will be consolidated by repeated analyses and the degree alterations in the sired offspring. These observations are of interindividual variation will be assessed, it will provide consistent with the notion that DNA methylation profiles of an essential reference for future studies on the impact of the gametes are not completely reset after fertilization but environmental stressors. can be partly transmitted across generations. Actually, few From an overall assessment of the current database on experiments tested this notion in the specific conditions, human somatic environmental epigenetics, rodent germline with some showing apparent inheritance of gamete methy- epigenetic toxicological studies, and the most environmen- lation [156, 168], while others not showing the same result tally relevant human reprotoxic agents, a priority list of envi- [152]. Nevertheless, several authors agree in pointing out ronmental stressors on which directing future human sperm that direct transmission of methylation changes is not the epigenetic biomonitoring studies might be proposed: dys- only mechanism through which altered sperm methylation metabolism as a consequence of environmental and genetic might affect the offspring phenotype and that sustained factors, including their possible interactions, endocrine dis- alterations of transcriptional regulatory networks early in rupting compounds, and major lifestyle toxicants like tobacco development may likely result from a complex interplay smoke and alcohol. In addition, emphasis should be on between DNA methylation changes, chromatin modifica- prenatal exposure, and mother child cohorts should be tions, and other epigenetic mechanisms. One implication of studied more actively. Finally, prospective, long-term, multi- epigenetic inheritance systems is that they provide a potential generation follow-up surveys should be possibly set up to take mechanism by which parents could transfer information to into account grandparental effects. their offspring about the environment they experienced. In other words, mechanisms exist that could allow organisms Abbreviations to “inform” their progeny about prevailing environmental conditions. ABCA1: ATP-binding cassette, subfamily A In some of the experimental studies [162, 163, 168, 169], (ABC1), member 1 changes of DNA methylation in the germline and somatic ACE: Angiotensin I-converting enzyme cells of exposed animals were compared. On the basis of the ACSL3: Acyl-CoA synthetase long-chain family few available data, it is not possible to draw any general con- member 3 clusion, but, much more than for induced genetic changes, it AHRR: Aryl hydrocarbon receptor repressor is conceivable that each cell type, with its own transcriptional APC: Adenomatous polyposis coli program, would be specifically affected at an epigenetic level. ASCL2: Achaete-scute family bHLH This consideration poses a problem when data on induced transcription factor 2 16 BioMed Research International

AXL: AXL receptor tyrosine kinase MAGE1: Melanoma antigen family A, 1 BMI: Body mass index MALDI-TOF MS: Matrix-assisted laser desorption CDKN1C: Cyclin-dependent kinase inhibitor 1C ionization time-of-flight mass CNTNAP2: Contactin associated protein-like 2 spectrometry COBRA: Combined bisulfite restriction analysis MEG3: Maternally expressed 3 (nonprotein COL1A2: Collagen, type I, alpha 2 coding) CRAT: Carnitine O-acetyltransferase MEST: Mesoderm specific transcript CTCF: CCCTC-binding factor (zinc finger protein) MLH1: MutL homolog 1 CTNNA2: Catenin (cadherin-associated protein), alpha MYO1G: Myosin IG 2 NNAT: Neuronatin CYP1A1: Cytochrome P450, family 1, subfamily A, NOS1: Nitric oxide synthase 1 polypeptide 1 NOS2A: Nitric oxide synthase 2A DAPK: Death-associated protein kinase NOS3: Nitric oxide synthase 3 DLK1: Delta-like 1 homolog NPY2R: Neuropeptide Y receptor Y2 DMR: Differentially methylated region NR3C2: Nuclear receptor subfamily 3, group C, DNMT1: DNA (cytosine-5-)-methyl transferase 1 member 2 EDs: Endocrine disruptors OGG1: 8-Oxoguanine DNA glycosylase 1 ELISA: Enzyme linked immunosorbent assay OLFR151: Olfactory receptor 151 F2RL3: Coagulation factor II (thrombin) p15: Cyclin-dependent kinase inhibitor 2B receptor-like 3 p16: Cyclin-dependent kinase inhibitor 2A F3: Coagulation factor III p53: Tumor protein p53 FOXO3A: Forkhead box O3 PAHs: Polycyclic aromatic hydrocarbons FOXP3: Forkhead box transcription factor 3 PBL: Peripheral blood lymphocytes GC-MS: Gas chromatography-mass spectroscopy PCBs: Polychlorinated biphenyls GCR: Glucocorticoid receptor PCR: Polymerase chain reaction GFI1: Growth factor independent 1 transcription PEG3: Paternally expressed 3 repressor PEG5: Paternally expressed 3 (alias NNAT: GNASAS: GNAS antisense RNA neuronatin) GPR15: G protein-coupled receptor 15 PEG10: Paternally expressed 10 GRB10: Growth factor receptor-bound protein 10 PFASs: Perfluoroalkyl substances GSTM1: Glutathione S-transferase mu 1 PGC: Primordial germ cell H19: Imprinted maternally expressed transcript PLAGL1: Pleomorphic adenoma gene-like 1 (nonprotein coding) PM: Particulate matter HAP1: Huntingtin-associated protein 1 POPs: Persistent organic pollutants HERV: Human endogenous retrovirus PPAR𝛼: Peroxisome proliferator-activated HIC1: Hypermethylated in cancer 1 receptor alpha HPLC-MS: High performance liquid PTGFRN: Prostaglandin F2 receptor negative chromatography-mass spectrometry regulator hTERT: Human telomerase reverse transcriptase PTPRO: Protein tyrosine phosphatase, receptor ICAM-1: Intercellular adhesion molecule 1 type, O ICR: Imprinting control center RASGRF1: Ras protein-specific guanine IFN𝛾:Interferongamma nucleotide-releasing factor 1 IGF2: Insulin-like growth factor 2 RASSF1A: Ras association (RalGDS/AF-6) domain IGF2R: Insulin-like growth factor 2 receptor family member 1 IL-4: Interleukin-4 RHBDF1: Rhomboid family member 1 IL-6: Interleukin-6 RUNX3: Runt-relatedtranscriptionfactor3 IL-10: Interleukin-10 SEPP1: Selenoprotein P, plasma, 1 iNOS: Inducible nitric oxide synthase SEPW1: Selenoprotein W, 1 IVF: In vitro fertilization SGCE: Sarcoglycan, epsilon KCNK17: Potassium channel, subfamily K, member 17 SNRPN: Small nuclear ribonucleoprotein KCNQ1: Potassium voltage-gated channel, KQT like polypeptide N subfamily, member 1 TLR-2: Toll-like receptor 2 KLK7: Kallikrein-related peptidase 7 TSP50: Testes-specific protease 50 LBW: Low birth weight ZNF132: Zinc finger protein 132. LEP: Leptin LINE-1: Long interspersed nuclear element 1, retrotransposable element 1 Conflict of Interests LPLASE: Lysophospholipase 5-mC: 5-methyl cytosine, 5-methyl deoxycytidine The authors declare that there is no conflict of interests. BioMed Research International 17

Acknowledgment [18] U. Lim and M.-A. Song, “Dietary and lifestyle factors of DNA methylation,” Methods in Molecular Biology,vol.863,pp.359– The authors wish to apologize to those authors whose 376, 2012. contribution was missed by our collections or was not quoted [19] K. M. Bakulski and M. D. Fallin, “Epigenetic epidemiology: because of space limitations. promises for public health research,” Environmental and Molec- ular Mutagenesis,vol.55,no.3,pp.171–183,2014. References [20] H. H. Burris and A. A. Baccarelli, “Environmental epigenetics: from novelty to scientific discipline,” Journal of Applied Toxicol- [1] C. B. Santos-Rebouc¸as and M. M. G. Pimentel, “Implication ogy,vol.34,no.2,pp.113–116,2014. of abnormal epigenetic patterns for human diseases,” European [21] I. Cantone and A. G. Fisher, “Epigenetic programming and Journal of Human Genetics,vol.15,no.1,pp.10–17,2007. reprogramming during development,” Nature Structural and [2] A. Portela and M. Esteller, “Epigenetic modifications and Molecular Biology,vol.20,no.3,pp.282–289,2013. human disease,” Nature Biotechnology,vol.28,no.10,pp.1057– [22] S. Seisenberger, J. R. Peat, and W. Reik, “Conceptual links 1068, 2010. between DNA methylation reprogramming in the early embryo [3] H. Heyn and M. Esteller, “DNA methylation profiling in the andprimordialgermcells,”Current Opinion in Cell Biology,vol. clinic: applications and challenges,” Nature Reviews. Genetics, 25,no.3,pp.281–288,2013. vol.13,no.10,pp.679–692,2012. [23] G. Kelsey and R. Feil, “New insights into establishment and [4] S. Feng, S. E. Jacobsen, and W. Reik, “Epigenetic reprogram- maintenance of DNA methylation imprints in mammals,” ming in plant and animal development,” Science,vol.330,no. Philosophical Transactions of the Royal Society of London, Series 6004, pp. 622–627, 2010. B: Biological Sciences,vol.368,no.1609,ArticleID20110336, [5]H.Denis,M.N.Ndlovu,andF.Fuks,“Regulationofmam- 2013. malian DNA methyltransferases: a route to new mechanisms,” [24] D. J. P.Barker, P.D. Gluckman, K. M. Godfrey, J. E. Harding, J. A. EMBO Reports,vol.12,no.7,pp.647–656,2011. Owens, and J. S. Robinson, “Fetal nutrition and cardiovascular [6]J.A.HackettandM.A.Surani,“DNAmethylationdynamics disease in adult life,” The Lancet, vol. 341, no. 8850, pp. 938–941, during the mammalian life cycle,” Philosophical Transactions of 1993. the Royal Society of London. Series B, Biological sciences,vol.368, [25] P. Dominguez-Salas, S. E. Cox, A. M. Prentice, B. J. Hennig, no. 1609, Article ID 20110328, 2013. andS.E.Moore,“Maternalnutritionalstatus,C1 metabolism [7] S.Seisenberger,J.R.Peat,T.A.Hore,F.Santos,W.Dean,andW. and offspring DNA methylation: a review of current evidence Reik, “Reprogramming DNA methylation in the mammalian in human subjects,” Proceedings of the Nutrition Society,vol.71, life cycle: building and breaking epigenetic barriers,” Philo- no. 1, pp. 154–165, 2012. sophical transactions of the Royal Society of London, Series B: [26] M. Pembrey, R. Saffery, L. O. Bygren, and Network in Epigenetic Biological sciences,vol.368,no.1609,ArticleID20110330,2013. Epidemiology, “Human transgenerational responses to early- [8] Z. D. Smith and A. Meissner, “DNA methylation: roles in life experience: potential impact on development, health and mammalian development,” Nature Reviews Genetics,vol.14,no. biomedical research,” Journal of Medical Genetics,vol.51,no.9, 3, pp. 204–220, 2013. pp. 563–572, 2014. [9] M. Szyf, “The implications of DNA methylation for toxicology: [27]A.Juul,K.Almstrup,A.M.Anderssonetal.,“Possiblefetal toward toxicomethylomics, the toxicology of DNA methyla- determinants of male infertility,” Nature Reviews Endocrinology, tion,” Toxicological Sciences,vol.120,no.2,pp.235–255,2011. vol.10,no.9,pp.553–562,2014. [10] J. R. McCarrey, “The epigenome as a target for heritable [28] R. M. Sharpe, “Environmental/lifestyle effects on spermatogen- environmental disruptions of cellular function,” Molecular and esis,” Philosophical Transactions of the Royal Society B: Biological Cellular Endocrinology,vol.354,no.1-2,pp.9–15,2012. Sciences,vol.365,no.1546,pp.1697–1712,2010. [11] V.Bollati and A. Baccarelli, “Environmental epigenetics,” Hered- [29] J. D. Meeker, “Exposure to environmental endocrine disrupting ity,vol.105,no.1,pp.105–112,2010. compounds and men’s health,” Maturitas,vol.66,no.3,pp.236– [12] J. A. Alegr´ıa-Torres, A. Baccarelli, and V. Bollati, “Epigenetics 241, 2010. and lifestyle,” Epigenomics,vol.3,no.3,pp.267–277,2011. [30] A. Giwercman and Y. L. Giwercman, “Environmental factors [13] B. C. Christensen and C. J. Marsit, “Epigenomics in environ- and testicular function,” Best Practice and Research: Clinical mental health,” Frontiers in Genetics,vol.2,article84,2011. Endocrinology and Metabolism,vol.25,no.2,pp.391–402,2011. [14] M. B. Terry, L. Delgado-Cruzata, N. Vin-Raviv, H. C. Wu, and R. [31] J. P.Bonde and A. Giwercman, “Environmental xenobiotics and M. Santella, “DNA methylation in white blood cells: association male reproductive health,” Asian Journal of Andrology,vol.16, with risk factors in epidemiologic studies,” Epigenetics,vol.6, no. 1, pp. 3–4, 2014. no.7,pp.828–837,2011. [32] A. Vested, A. Giwercman, J. P. Bonde, and G. Toft, “Persistent [15] V.K.Cortessis,D.C.Thomas,A.J.Levineetal.,“Environmental organic pollutants and male reproductive health,” Asian Journal epigenetics: prospects for studying epigenetic mediation of of Andrology,vol.16,no.1,pp.71–80,2014. exposure-response relationships,” Human Genetics,vol.131,no. [33]J.DelMazo,M.A.Brieno-Enr˜ ´ıquez, J. Garc´ıa-Lopez,´ L. A. 10, pp. 1565–1589, 2012. Lopez-Fern´ andez,´ and M. De Felici, “Endocrine disruptors, [16] R. Feil and M. F. Fraga, “Epigenetics and the environment: gene deregulation and male germ cell tumors,” International emerging patterns and implications,” Nature Reviews Genetics, Journal of Developmental Biology,vol.57,no.2-4,pp.225–239, vol. 13, no. 2, pp. 97–109, 2012. 2013. [17] L. Hou, X. Zhang, D. Wang, and A. Baccarelli, “Environmental [34] K. Singh and D. Jaiswal, “One-carbon metabolism, spermatoge- chemical exposures and human epigenetics,” International Jour- nesis, and male infertility,” Reproductive Sciences,vol.20,no.6, nal of Epidemiology,vol.41,no.1,pp.79–105,2012. pp. 622–630, 2013. 18 BioMed Research International

[35] C. C. Boissonnas, P. Jouannet, and H. Jammes, “Epigenetic [52] W.-T. Chen, W.-C. Hung, W.-Y. Kang, Y.-C. Huang, and C.-Y. disorders and male subfertility,” Fertility and Sterility,vol.99, Chai, “Urothelial carcinomas arising in arsenic-contaminated no. 3, pp. 624–631, 2013. areas are associated with hypermethylation of the gene pro- [36] R. Klaver¨ and J. Gromoll, “Bringing epigenetics into the diag- moter of the death-associated protein kinase,” Histopathology, nostics of the andrology laboratory: challenges and perspec- vol. 51, no. 6, pp. 785–792, 2007. tives,” Asian Journal of Andrology, vol. 16, no. 5, pp. 669–674, [53] W. J. Seow, M. L. Kile, A. A. Baccarelli et al., “Epigenome- 2014. wide DNA methylation changes with development of arsenic- [37]J.Borgel,S.Guibert,Y.Lietal.,“Targetsanddynamicsof induced skin lesions in Bangladesh: a case-control follow-up promoter DNA methylation during early mouse development,” study,” Environmental and Molecular Mutagenesis,vol.55,no. Nature Genetics, vol. 42, no. 12, pp. 1093–1100, 2010. 6, pp. 449–456, 2014. [38] L. Daxinger and E. Whitelaw, “Understanding transgenera- [54] T.-Y.Yang, L.-I. Hsu, A. W.Chiu et al., “Comparison of genome- tional epigenetic inheritance via the gametes in mammals,” wide DNA methylation in urothelial carcinomas of patients Nature Reviews Genetics, vol. 13, no. 2, pp. 153–162, 2012. with and without arsenic exposure,” Environmental Research, vol. 128, pp. 57–63, 2014. [39] J. M. Trasler, “Epigenetics in spermatogenesis,” Molecular and Cellular Endocrinology,vol.306,no.1-2,pp.33–36,2009. [55] M. L. Kile, A. Baccarelli, E. Hoffman et al., “Prenatal arsenic exposure and DNA methylation in maternal and umbilical cord [40] D. T. Carrell, “Epigenetics of the male gamete,” Fertility and blood leukocytes,” EnvironmentalHealthPerspectives,vol.120, Sterility,vol.97,no.2,pp.267–274,2012. no.7,pp.1061–1066,2012. [41] R. Guerrero-Preston, J. Herbstman, and L. R. Goldman, “Epige- [56] M. L. Kile, E. A. Houseman, A. A. Baccarelli et al., “Effect of nomic biomonitors: global DNA hypomethylation as a bio- prenatal arsenic exposure on DNA methylation and leukocyte dosimeter of life-long environmental exposures,” Epigenomics, subpopulations in cord blood,” Epigenetics,vol.9,no.5,pp.774– vol.3,no.1,pp.1–5,2011. 782, 2014. [42] R. R. Kanherkar, N. Bhatia-Dey, and A. B. Csoka, “Epigenetics [57] J. R. Pilsner, M. N. Hall, X. Liu et al., “Influence of prenatal across the human lifespan,” Frontiers in Cell and Developmental arsenic exposure and newborn sex on global methylation of Biology,vol.2,article49,2014. cord blood DNA,” PLoS ONE,vol.7,no.5,ArticleIDe37147, [43] P. D. Ray, A. Yosim, and R. C. Fry, “Incorporating epigenetic 2012. data into the risk assessment process for the toxic metals [58] P.Intarasunanont,P.Navasumrit,S.Waraprasitetal.,“Effectsof arsenic, cadmium, chromium, lead, and mercury: strategies and arsenic exposure on DNA methylation in cord blood samples challenges,” Frontiers in Genetics,vol.5,article201,2014. from newborn babies and in a human lymphoblast cell line,” [44]K.A.BaileyandR.C.Fry,“Arsenic-associatedchangestothe EnvironmentalHealth:AGlobalAccessScienceSource,vol.11, epigenome: what are the functional consequences?” Current no.1,article31,2012. Environmental Health Reports,vol.1,no.1,pp.22–34,2014. [59] D. C. Koestler, M. Avissar-Whiting, E. A. Houseman, M. R. [45] J. R. Pilsner, X. Liu, H. Ahsan et al., “Genomic methylation Karagas, and C. J. Marsit, “Differential DNA methylation in of peripheral blood leukocyte DNA: influences of arsenic and umbilical cord blood of infants exposed to low levels of arsenic folate in Bangladeshi adults,” The American Journal of Clinical in utero,” Environmental Health Perspectives,vol.121,no.8,pp. Nutrition, vol. 86, no. 4, pp. 1179–1186, 2007. 971–977, 2013. [46] S. Majumdar, S. Chanda, B. Ganguli, D. N. G. Mazumder, S. [60] D. Rojas, J. E. Rager, L. Smeester et al., “Prenatal arsenic expo- Lahiri, and U. B. Dasgupta, “Arsenic exposure induces genomic sure and the epigenome: identifying sites of 5-methylcytosine hypermethylation,” Environmental Toxicology,vol.25,no.3,pp. alterations that predict functional changes in gene expression 315–318, 2010. in newborn cord blood and subsequent birth outcomes,” Toxi- [47] M. M. Niedzwiecki, M. N. Hall, X. Liu et al., “A dose-response cological Sciences,vol.143,no.1,pp.97–106,2015. study of arsenic exposure and global methylation of peripheral [61] T.-C. Wang, Y.-S. Song, H. Wang et al., “Oxidative DNA blood mononuclear cell DNA in Bangladeshi adults,” Environ- damageandglobalDNAhypomethylationarerelatedtofolate mental Health Perspectives,vol.121,no.11-12,pp.1306–1312,2013. deficiency in chromate manufacturing workers,” Journal of [48] S. Chanda, U. B. Dasgupta, D. Guhamazumder et al., “DNA Hazardous Materials, vol. 213-214, pp. 440–446, 2012. hypermethylation of promoter of gene p53 and p16 in arsenic- [62] C. W. Hanna, M. S. Bloom, W. P. Robinson et al., “DNA exposed people with and without malignancy,” Toxicological methylation changes in whole blood is associated with exposure Sciences,vol.89,no.2,pp.431–437,2006. to the environmental contaminants, mercury, lead, cadmium [49] A.-H. Zhang, H.-H. Bin, X.-L. Pan, and X.-G. Xi, “Analysis of and bisphenol A, in women undergoing ovarian stimulation for p16 gene mutation, deletion and methylation in patients with IVF,” Human Reproduction,vol.27,no.5,pp.1401–1410,2012. arseniasis produced by indoor unventilated-stove coal usage [63] J.M.Goodrich,N.Basu,A.Franzblau,andD.C.Dolinoy,“Mer- in Guizhou, China,” Journal of Toxicology and Environmental cury biomarkers and DNA methylation among michigan dental Health, Part A,vol.70,no.11,pp.970–975,2007. professionals,” Environmental and Molecular Mutagenesis,vol. [50] L.Smeester,J.E.Rager,K.A.Baileyetal.,“Epigeneticchangesin 54,no.3,pp.195–203,2013. individuals with arsenicosis,” Chemical Research in Toxicology, [64] R. O. Wright, J. Schwartz, R. J. Wright et al., “Biomarkers of vol.24,no.2,pp.165–167,2011. lead exposure and DNA methylation within retrotransposons,” [51] M. B. Hossain, M. Vahter, G. Concha, and K. Broberg, “Envi- Environmental Health Perspectives,vol.118,no.6,pp.790–795, ronmental arsenic exposure and DNA methylation of the tumor 2010. suppressor gene p16 and the DNA repair gene MLH1: effect of [65] L. Kovatsi, E. Georgiou, A. Ioannou et al., “P16 promoter 2+ arsenic metabolism and genotype,” Metallomics,vol.4,no.11, methylation in Pb -exposed individuals,” Clinical Toxicology, pp. 1167–1175, 2012. vol.48,no.2,pp.124–128,2010. BioMed Research International 19

[66] M. B. Hossain, M. Vahter, G. Concha, and K. Broberg, “Low- interferon-𝛾 in cord white blood cells,” Environmental Health level environmental cadmium exposure is associated with Perspectives,vol.120,no.8,pp.1195–1200,2012. DNA hypomethylation in Argentinean women,” Environmental [81] L. Tarantini, M. Bonzini, P. Apostoli et al., “Effects of partic- Health Perspectives,vol.120,no.6,pp.879–884,2012. ulate matter on genomic DNA methylation content and iNOS [67] J. R. Pilsner, M. N. Hall, X. Liu et al., “Associations of plasma promoter methylation,” Environmental Health Perspectives,vol. selenium with arsenic and genomic methylation of leukocyte 117, no. 2, pp. 217–222, 2009. DNA in bangladesh,” EnvironmentalHealthPerspectives,vol. [82] L. Hou, X. Zhang, L. Tarantini et al., “Ambient PM exposure and 119, no. 1, pp. 113–118, 2011. DNA methylation in tumor suppressor genes: a cross-sectional [68] A. P. Sanders, L. Smeester, D. Rojas et al., “Cadmium exposure study,” Particle and Fibre Toxicology,vol.8,article25,2011. and the epigenome: exposure-associated patterns of DNA methylation in leukocytes from mother-baby pairs,” Epigenetics, [83] L. Dioni, M. Hoxha, F. Nordio et al., “Effects of short-term vol. 9, no. 2, pp. 212–221, 2014. exposure to inhalable particulate matter on telomere length, telomerase expression, and telomerase methylation in steel [69] H. Ji and G. K. Khurana Hershey, “Genetic and epigenetic workers,” Environmental Health Perspectives,vol.119,no.5,pp. influence on the response to environmental particulate matter,” 622–627, 2011. JournalofAllergyandClinicalImmunology,vol.129,no.1,pp. 33–41, 2012. [84] S. Pavanello, V. Bollati, A. C. Pesatori et al., “Global and gene-specific promoter methylation changes are related to [70] S. De Prins, G. Koppen, G. Jacobs et al., “Influence of ambient anti-B[a]PDE-DNA adduct levels and influence micronuclei airpollutiononglobalDNAmethylationinhealthyadults:a levels in polycyclic aromatic hydrocarbon-exposed individuals,” seasonal follow-up,” Environment International,vol.59,pp.418– International Journal of Cancer,vol.125,no.7,pp.1692–1697, 424, 2013. 2009. [71] A. Baccarelli, R. O. Wright, V. Bollati et al., “Rapid DNA methylation changes after exposure to traffic particles,” The [85] L. Guo, H.-M. Byun, J. Zhong et al., “Effects of short-term American Journal of Respiratory and Critical Care Medicine,vol. exposure to inhalable particulate matter on DNA methylation 179, no. 7, pp. 572–578, 2009. of tandem repeats,” Environmental and Molecular Mutagenesis, vol.55,no.4,pp.322–335,2014. [72] J. Madrigano, A. Baccarelli, M. A. Mittleman et al., “Prolonged exposure to particulate pollution, genes associated with glu- [86] L. Hou, X. Zhang, Y. Zheng et al., “Altered methylation in tathione pathways, and DNA methylation in a cohort of older tandem repeat element and elemental component levels in men,” EnvironmentalHealthPerspectives,vol.119,no.7,pp.977– inhalable air particles,” Environmental and Molecular Mutage- 982, 2011. nesis, vol. 55, no. 3, pp. 256–265, 2014. [73] M. A. Bind, J. Lepeule, A. Zanobetti et al., “Air pollution [87] H.-M. Byun, V.Motta, T.Panni et al., “Evolutionary age of repet- and gene-specific methylation in the Normative Aging Study: itive element subfamilies and sensitivity of DNA methylation to association, effect modification, and mediation analysis,” Epige- airborne pollutants,” Particle and Fibre Toxicology,vol.10,no.1, netics,vol.9,no.3,pp.448–458,2014. article 28, 2013. [74] J. Lepeule, M.-A. C. Bind, A. A. Baccarelli et al., “Epigenetic [88] M. Peluso, V. Bollati, A. Munnia et al., “DNA methylation influences on associations between air pollutants and lung func- differences in exposed workers and nearby residents of the tion in elderly men: the normative aging study,” Environmental Ma Ta phut industrial estate, rayong, Thailand,” International Health Perspectives,vol.122,no.6,pp.566–572,2014. Journal of Epidemiology,vol.41,no.6,pp.1753–1760,2012. [75] K. Nadeau, C. McDonald-Hyman, E. M. Noth et al., “Ambient [89] V. Bollati, A. Baccarelli, L. Hou et al., “Changes in DNA air pollution impairs regulatory T-cell function in asthma,” methylation patterns in subjects exposed to low-dose benzene,” Journal of Allergy and Clinical Immunology,vol.126,no.4,pp. Cancer Research, vol. 67, no. 3, pp. 876–880, 2007. 845.e10–852.e10, 2010. [90] S. Fustinoni, F. Rossella, E. Polledri et al., “Global DNA [76]C.V.Breton,M.T.Salam,X.Wang,H.-M.Byun,K.D.Sieg- methylation and low-level exposure to benzene,” La Medicina mund, and F. D. Gilliland, “Particulate matter, DNA methyla- del Lavoro,vol.103,no.2,pp.84–95,2012. tion in nitric oxide synthase, and childhood respiratory disease,” Environmental Health Perspectives,vol.120,no.9,pp.1320–1326, [91] W. J. Seow, A. C. Pesatori, E. Dimont et al., “Urinary benzene 2012. biomarkers and DNA methylation in Bulgarian petrochemical workers: study findings and comparison of linear and beta [77] M. T. Salam, H. M. Byun, F. Lurmann et al., “Genetic and regression models,” PLoS ONE,vol.7,no.12,ArticleIDe50471, epigenetic variations in inducible nitric oxide synthase pro- 2012. moter, particulate pollution, and exhaled nitric oxide levels in children,” Journal of Allergy and Clinical Immunology,vol.129, [92]J.A.Rusiecki,A.Baccarelli,V.Bollati,L.Tarantini,L.E.Moore, no. 1, pp. 232.e7–239.e7, 2012. and E. C. Bonefeld-Jorgensen, “Global DNA hypomethylation [78] J. B. Herbstman, D. Tang, D. Zhu et al., “Prenatal exposure is associated with high serum-persistent organic pollutants in to polycyclic aromatic hydrocarbons, benzo[a]pyrene-DNA Greenlandic inuit,” Environmental Health Perspectives,vol.116, adducts, and genomic DNA methylation in cord blood,” Envi- no. 11, pp. 1547–1552, 2008. ronmental Health Perspectives,vol.120,no.5,pp.733–738,2012. [93] K.-Y. Kim, D.-S. Kim, S.-K. Lee et al., “Association of low- [79]F.Perera,W.-Y.Tang,J.Herbstmanetal.,“RelationofDNA dose exposure to persistent organic pollutants with global DNA 󸀠 methylation of 5 -CpG island of ACSL3 to transplacental hypomethylation in healthy Koreans,” Environmental Health exposure to airborne polycyclic aromatic hydrocarbons and Perspectives,vol.118,no.3,pp.370–374,2010. childhood asthma,” PLoS ONE,vol.4,no.2,ArticleIDe4488, [94] J. H. Kim, L. S. Rozek, A. S. Soliman et al., “Bisphenol A- 2009. associated epigenomic changes in prepubescent girls: a cross- [80] W.-Y. Tang, L. Levin, G. Talaska et al., “Maternal exposure to sectional study in Gharbiah, Egypt,” Environmental Health: A polycyclic aromatic hydrocarbons and 5󸀠-CpG methylation of Global Access Science Source, vol. 12, article 33, 2013. 20 BioMed Research International

[95] D. J. Watkins, G. A. Wellenius, R. A. Butler, S. M. Bartell, T. [111] N. S. Shenker, P.M. Ueland, S. Polidoro et al., “DNA methylation Fletcher, and K. T. Kelsey, “Associations between serum per- as a long-term biomarker of exposure to tobacco smoke,” fluoroalkyl acids and LINE-1 DNA methylation,” Environment Epidemiology,vol.24,no.5,pp.712–716,2013. International,vol.63,pp.71–76,2014. [112] M.V.Dogan,B.Shields,C.Cutronaetal.,“Theeffectofsmoking [96] G. Leter, C. Consales, P. Eleuteri et al., “Exposure to perflu- on DNA methylation of peripheral blood mononuclear cells oroalkyl substances and sperm DNA global methylation in from African American women,” BMC Genomics,vol.15,no.1, arctic and european populations,” Environmental and Molecular article 151, 2014. Mutagenesis,vol.55,no.7,pp.591–600,2014. [113] C. V. Breton, H.-M. Byun, M. Wenten, F. Pan, A. Yang, and F. [97] R. Guerrero-Preston, L. R. Goldman, P. Brebi-Mieville et al., D. Gilliland, “Prenatal tobacco smoke exposure affects global “GlobalDNAhypomethylationisassociatedwithinutero and gene-specific DNA methylation,” The American Journal of exposure to cotinine and perfluorinated alkyl compounds,” Respiratory and Critical Care Medicine,vol.180,no.5,pp.462– Epigenetics,vol.5,no.6,pp.539–546,2010. 467, 2009. [98] K. Huen, P. Yousefi, A. Bradman et al., “Effects of age, sex, and [114] C. V. Breton, M. T. Salam, and F. D. Gilliland, “Heritability and persistent organic pollutants on DNA methylation in children,” role for the environment in DNA methylation in AXL receptor Environmental and Molecular Mutagenesis,vol.55,no.3,pp. tyrosine kinase,” Epigenetics,vol.6,no.7,pp.895–898,2011. 209–222, 2014. [115] C. V. Breton, K. D. Siegmund, B. R. Joubert et al., “Prenatal [99] N. Vilahur, M. Bustamante, H. Byun et al., “Prenatal exposure to tobacco smoke exposure is associated with childhood DNA mixtures of xenoestrogens and repetitive element DNA methy- CpG methylation,” PLoS ONE,vol.9,no.6,ArticleIDe99716, lation changes in human placenta,” Environment International, 2014. vol. 71, pp. 81–87, 2014. [116] M. Suter, A. Abramovici, L. Showalter et al., “In utero tobacco [100]A.C.Vidal,S.K.Murphy,A.P.Murthaetal.,“Associations exposure epigenetically modifies placental CYP1A1 expression,” between antibiotic exposure during pregnancy, birth weight Metabolism,vol.59,no.10,pp.1481–1490,2010. and aberrant methylation at imprinted genes among offspring,” [117] M. Suter, J. Ma, A. Harris et al., “Maternal tobacco use modestly International Journal of Obesity,vol.37,no.7,pp.907–913,2013. alters correlated epigenome-wide placental DNA methylation [101] V. S. Knopik, M. A. MaCcani, S. Francazio, and J. E. McGeary, and gene expression,” Epigenetics,vol.6,no.11,pp.1284–1294, “The epigenetics of maternal cigarette smoking during preg- 2011. nancy and effects on child development,” Development and [118] B. R. Joubert, S. E. Haberg,˚ R. M. Nilsen et al., “450K Psychopathology,vol.24,no.4,pp.1377–1390,2012. epigenome-wide scan identifies differential DNA methylation [102]K.W.K.LeeandZ.Pausova,“CigarettesmokingandDNA in newborns related to maternal smoking during pregnancy,” methylation,” Frontiers in Genetics, vol. 4, article 132, 2013. Environmental Health Perspectives,vol.120,no.10,pp.1425– [103]L.P.Breitling,R.Yang,B.Korn,B.Burwinkel,andH.Brenner, 1431, 2012. “Tobacco-smoking-related differential DNA methylation: 27K [119] S. K. Murphy, A. Adigun, Z. Huang et al., “Gender-specific discovery and replication,” American Journal of Human Genet- methylation differences in relation to prenatal exposure to ics, vol. 88, no. 4, pp. 450–457, 2011. cigarette smoke,” Gene,vol.494,no.1,pp.36–43,2012. [104]L.P.Breitling,K.Salzmann,D.Rothenbacher,B.Burwinkel, [120] C. S. Wilhelm-Benartzi, E. A. Houseman, M. A. Maccani et and H. Brenner, “Smoking, F2RL3 methylation, and prognosis al., “In utero exposures, infant growth, and DNA methylation in stable coronary heart disease,” European Heart Journal,vol. of repetitive elements and developmentally related genes in 33, no. 22, pp. 2841–2848, 2012. human placenta,” Environmental Health Perspectives,vol.120, [105] N.S.Shenker,S.Polidoro,K.vanVeldhovenetal.,“Epigenome- no. 2, pp. 296–302, 2012. wide association study in the European Prospective Inves- [121] J.Z.J.Maccani,D.C.Koestler,E.A.Houseman,C.J.Marsit,and tigation into Cancer and Nutrition (EPIC-Turin) identifies K. T. Kelsey, “Placental DNA methylation alterations associated novel genetic loci associated with smoking,” Human Molecular with maternal tobacco smoking at the RUNX3 gene are also Genetics,vol.22,no.5,pp.843–851,2013. associated with gestational age,” Epigenomics,vol.5,no.6,pp. [106] Y. Zhang, R. Yang, B. Burwinkel, L. P.Breitling, and H. Brenner, 619–630, 2013. “F2RL3 methylation as a biomarker of current and lifetime [122] K. W. Lee, R. Richmond, P. Hu et al., “Prenatal exposure to smoking exposures,” EnvironmentalHealthPerspectives,vol. maternal cigarette smoking and DNA methylation: epigenome- 122, no. 2, pp. 131–137, 2014. wide association in a discovery sample of adolescents and [107] Y. Zhang, R. Yang, B. Burwinkel et al., “F2RL3 methylation in replication in an independent cohort at birth through 17 years blood DNA is a strong predictor of mortality,” International of age,” EnvironmentalHealthPerspectives,vol.123,no.2,pp. Journal of Epidemiology,vol.43,no.4,pp.1215–1225,2014. 193–199, 2015. [108] M. M. Monick, S. R. H. Beach, J. Plume et al., “Coordinated [123]A.Soubry,C.Hoyo,R.L.Jirtle,andS.K.Murphy,“Apater- changes in AHRR methylation in lymphoblasts and pulmonary nal environmental legacy: evidence for epigenetic inheritance macrophages from smokers,” American Journal of Medical through the male germ line,” BioEssays,vol.36,no.4,pp.359– Genetics, Part B: Neuropsychiatric Genetics,vol.159,no.2,pp. 371, 2014. 141–151, 2012. [124] A. Soubry, J. M. Schildkraut, A. Murtha et al., “Paternal obesity [109] R.A.Philibert,S.R.H.Beach,andG.H.Brody,“Demethylation is associated with IGF2 hypomethylation in newborns: results of the aryl hydrocarbon receptor repressor as a biomarker for from a Newborn Epigenetics Study (NEST) cohort,” BMC nascent smokers,” Epigenetics,vol.7,no.11,pp.1331–1338,2012. Medicine,vol.11,no.1,article29,2013. [110] R. A. Philibert, S. R. H. Beach, M.-K. Lei, and G. H. Brody, [125]A.Soubry,S.K.Murphy,F.Wangetal.,“Newbornsofobese “Changes in DNA methylation at the aryl hydrocarbon receptor parents have altered DNA methylation patterns at imprinted repressor may be a new biomarker for smoking,” Clinical genes,” International Journal of Obesity,vol.39,pp.650–657, Epigenetics,vol.5,no.1,article19,2013. 2015. BioMed Research International 21

[126]T.G.Jenkins,K.I.Aston,B.R.Cairns,andD.T.Carrell, folic acid use before and during pregnancy,” Epigenetics,vol.6, “Paternalagingandassociatedintraindividualalterationsof no.7,pp.928–936,2011. global sperm 5-methylcytosine and 5-hydroxymethylcytosine [142]P.Haggarty,G.Hoad,D.M.Campbell,G.W.Horgan,C.Piy- levels,” Fertility and Sterility,vol.100,no.4,pp.945–951,2013. athilake, and G. McNeill, “Folate in pregnancy and imprinted [127]T.G.Jenkins,K.I.Aston,C.Pflueger,B.R.Cairns,D.T.Carrell, gene and repeat element methylation in the offspring,” Ameri- and J. M. Greally, “Age-associated sperm DNA methylation can Journal of Clinical Nutrition,vol.97,no.1,pp.94–99,2013. alterations: possible implications in offspring disease suscepti- [143]C.E.Boeke,A.Baccarelli,K.P.Kleinmanetal.,“Gestational bility,” PLoS Genetics,vol.10,no.7,ArticleIDe1004458,2014. intake of methyl donors and global LINE-1 DNA methylation [128] C. Consales, G. Leter, J. P. Bonde et al., “Indices of methyla- in maternal and cord blood: prospective results from a folate- tion in sperm DNA from fertile men differ between distinct replete population,” Epigenetics,vol.7,no.3,pp.253–260,2012. geographical regions,” Human Reproduction,vol.29,no.9,pp. [144]R.A.Waterland,R.Kellermayer,E.Laritskyetal.,“Seasonof 2065–2072, 2014. conception in rural gambia affects DNA methylation at putative [129]D.Kumar,S.R.Salian,G.Kalthuretal.,“Semenabnormalities, human metastable epialleles,” PLoS Genetics,vol.6,no.12, sperm DNA damage and global hypermethylation in health Article ID e1001252, 2010. workers occupationally exposed to ionizing radiation,” PLoS [145]P.Dominguez-Salas,S.E.Moore,M.S.Bakeretal.,“Maternal ONE,vol.8,no.7,ArticleIDe69927,2013. nutrition at conception modulates DNA methylation of human [130] D. M. Bielawski, F. M. Zaher, D. M. Svinarich, and E. L. metastable epialleles,” Nature Communications,vol.5,article Abel, “Paternal alcohol exposure affects sperm cytosine methyl- 3746, 2014. transferase messenger RNA levels,” Alcoholism: Clinical and [146]R.A.Harris,D.Nagy-Szakal,andR.Kellermayer,“Human Experimental Research,vol.26,no.3,pp.347–351,2002. metastable epiallele candidates link to common disorders,” [131]L.A.Ouko,K.Shantikumar,J.Knezovich,P.Haycock,D.J. Epigenetics,vol.8,no.2,pp.157–163,2013. Schnugh, and M. Ramsay, “Effect of alcohol consumption on CpG methylation in the differentially methylated regions of H19 [147]Y.Liu,S.K.Murphy,A.P.Murthaetal.,“Depressionin and IG-DMR in male gametes—implications for fetal alcohol pregnancy, infant birth weight and DNA methylation of imprint spectrum disorders,” Alcoholism: Clinical and Experimental regulatory elements,” Epigenetics,vol.7,no.7,pp.735–746,2012. Research,vol.33,no.9,pp.1615–1627,2009. [148] L.Cao-Lei,R.Massart,M.J.Sudermanetal.,“DNAmethylation [132]M.Lane,R.L.Robker,andS.A.Robertson,“Parentingfrom signatures triggered by prenatal maternal stress exposure to a before conception,” Science,vol.345,no.6198,pp.756–760, natural disaster: project ice storm,” PLoS ONE,vol.9,no.9, 2014. Article ID e107653, 2014. [133] X. Liu, Q. Chen, H.-J. Tsai et al., “Maternal preconception [149]T.Fullston,E.M.C.O.Teague,N.O.Palmeretal.,“Paternal body mass index and offspring cord blood DNA methylation: obesity initiates metabolic disturbances in two generations of exploration of early life origins of disease,” Environmental and mice with incomplete penetrance to the F2 generation and Molecular Mutagenesis,vol.55,no.3,pp.223–230,2014. alters the transcriptional profile of testis and sperm microRNA content,” The FASEB Journal,vol.27,no.10,pp.4226–4243,2013. [134] L. H. Lumey, M. B. Terry, L. Delgado-Cruzata et al., “Adult global DNA methylation in relation to pre-natal nutrition,” [150]A.B.Rodgers,C.P.Morgan,S.L.Bronson,S.Revello,and International Journal of Epidemiology,vol.41,no.1,pp.116–123, T. L. Bale, “Paternal stress exposure alters sperm MicroRNA 2012. content and reprograms offspring HPA stress axis regulation,” [135] B. T. Heijmans, E. W. Tobi, A. D. Stein et al., “Persistent The Journal of Neuroscience,vol.33,no.21,pp.9003–9012,2013. epigenetic differences associated with prenatal exposure to [151] K. Gapp, A. Jawaid, P.Sarkies et al., “Implication of sperm RNAs famine in humans,” Proceedings of the National Academy of in transgenerational inheritance of the effects of early trauma in Sciences of the United States of America,vol.105,no.44,pp. mice,” Nature Neuroscience,vol.17,no.5,pp.667–669,2014. 17046–17049, 2008. [152] E. J. Radford, M. Ito, H. Shi et al., “In utero undernourishment [136]B.T.Heijmans,E.W.Tobi,L.H.Lumey,andP.E.Slagboom, perturbs the adult sperm methylome and intergenerational “The epigenome: archive of the prenatal environment,” Epige- metabolism,” Science, vol. 345, no. 6198, Article ID 1255903, netics,vol.4,no.8,pp.526–531,2009. 2014. [137]E.W.Tobi,L.H.Lumey,R.P.Talensetal.,“DNAmethylation [153] B. R. Carone, L. Fauquier, N. Habib et al., “Paternally induced differences after exposure to prenatal famine are common and transgenerational environmental reprogramming of metabolic timing- and sex-specific,” Human Molecular Genetics,vol.18, gene expression in mammals,” Cell,vol.143,no.7,pp.1084–1096, no. 21, pp. 4046–4053, 2009. 2010. [138] E. W. Tobi, B. T. Heijmans, D. Kremer et al., “DNA methylation [154] R. Lambrot, C. Xu, S. Saint-Phar et al., “Low paternal dietary of IGF2, GNASAS, INSIGF and LEP and being born small for folate alters the mouse sperm epigenome and is associated with gestational age,” Epigenetics,vol.6,no.2,pp.171–176,2011. negative pregnancy outcomes,” Nature Communications,vol.4, [139]E.W.Tobi,P.E.Slagboom,J.vanDongenetal.,“Prenatal article 2889, 2013. famine and genetic variation are independently and additively [155] X. Tian and F. J. Diaz, “Acute dietary zinc deficiency before associated with dna methylation at regulatory loci within conception compromises oocyte epigenetic programming and IGF2/H19,” PLoS ONE,vol.7,no.5,ArticleIDe37933,2012. disrupts embryonic development,” Developmental Biology,vol. [140] E. W. Tobi, J. J. Goeman, R. Monajemi et al., “DNA methyla- 376,no.1,pp.51–61,2013. tion signatures link prenatal famine exposure to growth and [156] Y. Wei, C.-R. Yang, Y.-P. Wei et al., “Paternally induced metabolism,” Nature Communications,vol.5,article5592,2014. transgenerational inheritance of susceptibility to diabetes in [141]C.Hoyo,A.P.Murtha,J.M.Schildkrautetal.,“Methylation mammals,” Proceedings of the National Academy of Sciences of variation at IGF2 differentially methylated regions and maternal the United States of America, vol. 111, no. 5, pp. 1873–1878, 2014. 22 BioMed Research International

[157]Z.-J.Ge,X.-W.Liang,L.Guoetal.,“Maternaldiabetescauses 2-H19 locus-specific DNA methylation in rat spermatozoa and alterations of DNA methylation statuses of some imprinted its association with embryo loss,” Fertility and Sterility,vol.91, genes in murine oocytes,” Biology of Reproduction,vol.88,no. no.5,supplement,pp.2253–2263,2009. 5, article 117, 9 pages, 2013. [172]D.Xia,N.Parvizi,Y.Zhouetal.,“Paternalfenvalerateexposure [158] Z. J. Ge, S. M. Luo, F. Lin et al., “DNA methylation in oocytes influences reproductive functions in the offspring,” Reproduc- and liver of female mice and their offspring: Effects of high-fat- tive Sciences, vol. 20, no. 11, pp. 1308–1315, 2013. diet-induced obesity,” Environmental Health Perspectives,vol. [173] J.-Q. Zhu, Y.-J. Si, L.-Y. Cheng et al., “Sodium fluoride disrupts 122, no. 2, pp. 159–164, 2014. DNA methylation of H19 and Peg3 imprinted genes during the [159] M. D. Anway, A. S. Cupp, N. Uzumcu, and M. K. Skinner, early development of mouse embryo,” Archives of Toxicology, “Toxicology: epigenetic transgenerational actions of endocrine vol. 88, no. 2, pp. 241–248, 2014. disruptors and male fertility,” Science,vol.308,no.5727,pp. [174]S.Pathak,M.Saxena,R.D’Souza,andN.H.Balasinor,“Dis- 1466–1469, 2005. rupted imprinting status at the H19 differentially methylated [160] K. Inawaka, M. Kawabe, S. Takahashi et al., “Maternal exposure region is associated with the resorbed embryo phenotype in to anti-androgenic compounds, vinclozolin, flutamide and rats,” Reproduction, Fertility and Development,vol.22,no.6,pp. procymidone, has no effects on spermatogenesis and DNA 939–948, 2010. methylation in male rats of subsequent generations,” Toxicology [175] B. G. Dias and K. J. Ressler, “Parental olfactory experience influ- and Applied Pharmacology,vol.237,no.2,pp.178–187,2009. ences behavior and neural structure in subsequent generations,” [161] C. Stouder and A. Paoloni-Giacobino, “Transgenerational Nature Neuroscience,vol.17,no.1,pp.89–96,2014. effects of the endocrine disruptor vinclozolin on the methyla- [176] C. P. Wild, “Environmental exposure measurement in cancer tion pattern of imprinted genes in the mouse sperm,” Reproduc- epidemiology,” Mutagenesis,vol.24,no.2,pp.117–125,2009. tion,vol.139,no.2,pp.373–379,2010. [177] F. Ricceri, M. Trevisan, V. Fiano et al., “Seasonality modifies [162] C. Stouder and A. Paoloni-Giacobino, “Specific transgenera- methylation profiles in healthy people,” PLoS ONE,vol.9,no. tional imprinting effects of the endocrine disruptor methoxy- 9, Article ID e106846, 2014. chlor on male gametes,” Reproduction,vol.141,no.2,pp.207– [178]M.-A.Bind,A.Zanobetti,A.Gasparrinietal.,“Effectsof 216, 2011. temperature and relative humidity on DNA methylation,” Epi- [163] E. Somm, C. Stouder, and A. Paoloni-Giacobino, “Effect of demiology,vol.25,no.4,pp.561–569,2014. developmental dioxin exposure on methylation and expression [179] V. Bollati, A. Baccarelli, S. Sartori et al., “Epigenetic effects of of specific imprinted genes in mice,” Reproductive Toxicology, shiftwork on blood DNA methylation,” Chronobiology Interna- vol.35,no.1,pp.150–155,2013. tional,vol.27,no.5,pp.1093–1104,2010. [164] H.-H. Chao, X.-F. Zhang, B. Chen et al., “Bisphenol A exposure [180]Y.Zhu,R.G.Stevens,A.E.Hoffmanetal.,“Epigenetic modifies methylation of imprinted genes in mouse oocytes via impact of long-term shiftwork: pilot evidence from circadian the estrogen receptor signaling pathway,” Histochemistry and genes and whole-genome methylation analysis,” Chronobiology Cell Biology,vol.137,no.2,pp.249–259,2012. International,vol.28,no.10,pp.852–861,2011. [165] T. Trapphoff, M. Heiligentag, N. El Hajj, T. Haaf, and U. [181] F. Shi, X. Chen, A. Fu et al., “Aberrant DNA methylation of miR- Eichenlaub-Ritter, “Chronic exposure to a low concentration of 219 promoter in long-term night shiftworkers,” Environmental bisphenol A during follicle culture affects the epigenetic status and Molecular Mutagenesis,vol.54,no.6,pp.406–413,2013. of germinal vesicles and metaphase II oocytes,” Fertility and [182] D. I. Jacobs, J. Hansen, A. Fu et al., “Methylation alterations Sterility, vol. 100, no. 6, pp. 1758.e1–1767.e1, 2013. at imprinted genes detected among long-term shiftworkers,” [166]T.Doshi,C.D’Souza,andG.Vanage,“AberrantDNAmethyla- Environmental and Molecular Mutagenesis,vol.54,no.2,pp. tion at Igf2-H19 imprinting control region in spermatozoa upon 141–146, 2013. neonatal exposure to bisphenol A and its association with post [183]A.L.Teh,H.Pan,L.Chenetal.,“Theeffectofgenotypeand implantation loss,” Molecular Biology Reports,vol.40,no.8,pp. in utero environment on interindividual variation in neonate 4747–4757, 2013. DNA methylomes,” Genome Research,vol.24,no.7,pp.1064– [167] C. Yauk, A. Polyzos, A. Rowan-Carroll et al., “Germ-line 1074, 2014. mutations, DNA damage, and global hypermethylation in mice [184]S.Shah,A.F.McRae,R.E.Marionietal.,“Geneticand exposed to particulate air pollution in an urban/industrial environmental exposures constrain epigenetic drift over the location,” Proceedings of the National Academy of Sciences of the human life course,” Genome Research, vol. 24, no. 11, pp. 1725– United States of America,vol.105,no.2,pp.605–610,2008. 1733, 2014. [168] C. Stouder, E. Somm, and A. Paoloni-Giacobino, “Prenatal [185] J. Kim, K. Kim, H. Kim, G. Yoon, and K. Lee, “Characterization exposure to ethanol: a specific effect on the H19 gene in sperm,” ofagesignaturesofDNAmethylationinnormalandcancer Reproductive Toxicology, vol. 31, no. 4, pp. 507–512, 2011. tissues from multiple studies,” BMC Genomics,vol.15,no.1,p. [169] J. G. Knezovich and M. Ramsay, “The effect of preconception 997, 2014. paternal alcohol exposure on epigenetic remodeling of the H19 [186] L.M.Reynolds,J.R.Taylor,J.Dingetal.,“Age-relatedvariations and Rasgrf1 imprinting control regions in mouse offspring,” in the methylome associated with gene expression in human Frontiers in Genetics,vol.3,article10,2012. monocytes and T cells,” Nature Communications,vol.5,p.5366, [170] N.A.Kedia-Mokashi,L.Kadam,M.Ankolkar,K.Dumasia,and 2014. N. H. Balasinor, “Aberrant methylation of multiple imprinted [187] J. L. Mcclay, K. A. Aberg, S. L. Clark et al., “A methylome- genes in embryos of tamoxifen-treated male rats,” Reproduction, wide study of aging using massively parallel sequencing of the vol. 146, no. 2, pp. 155–168, 2013. methyl-CpG-enriched genomic fraction from blood in over 700 [171] S. Pathak, N. Kedia-Mokashi, M. Saxena et al., “Effect of subjects,” Human Molecular Genetics,vol.23,no.5,pp.1175– tamoxifen treatment on global and insulin-like growth factor 1185, 2014. BioMed Research International 23

[188] S. D. Fouse, R. P. Nagarajan, and J. F. Costello, “Genome-scale DNA methylation analysis,” Epigenomics,vol.2,no.1,pp.105– 117, 2010. [189] P. W. Laird, “Principles and challenges of genome-wide DNA methylation analysis,” Nature Reviews Genetics,vol.11,no.3,pp. 191–203, 2010. [190] E. Meaburn and R. Schulz, “Next generation sequencing in epigenetics: insights and challenges,” Seminars in Cell and Developmental Biology,vol.23,no.2,pp.192–199,2012. [191] K. Mensaert, S. Denil, G. Trooskens, W. van Criekinge, O. Thas, and T. de Meyer, “Next-generation technologies and data analytical approaches for epigenomics,” Environmental and Molecular Mutagenesis,vol.55,no.3,pp.155–170,2014. [192]A.S.Yang,M.R.H.Estecio,´ K. Doshi, Y. Kondo, E. H. Tajara, andJ.-P.J.Issa,“AsimplemethodforestimatingglobalDNA methylation using bisulfite PCR of repetitive DNA elements,” Nucleic Acids Research,vol.32,no.3,articlee38,2004. [193] J. Tost and I. G. Gut, “DNA methylation analysis by pyrosequencing,” Nature Protocols,vol.2,no.9,pp.2265–2275,2007. [194] M. Bibikova, B. Barnes, C. Tsan et al., “High density DNA methylation array with single CpG site resolution,” Genomics, vol. 98, no. 4, pp. 288–295, 2011. [195]E.M.Price,A.M.Cotton,M.S.Penaherrera,˜ D. E. McFadden, M. S. Kobor, and W. P. Robinson, “Different measures of ‘genome-wide’ DNA methylation exhibit unique properties in placental and somatic tissues,” Epigenetics,vol.7,no.6,pp.652– 663, 2012. [196]T.MikeskaandJ.M.Craig,“DNAmethylationbiomarkers: cancer and beyond,” Genes,vol.5,no.3,pp.821–864,2014. [197]A.Molaro,E.Hodges,F.Fangetal.,“Spermmethylation profiles reveal features of epigenetic inheritance and evolution in primates,” Cell, vol. 146, no. 6, pp. 1029–1041, 2011. [198] C. Krausz, J. Sandoval, S. Sayols et al., “Novel insights into DNA methylation features in spermatozoa: stability and peculiari- ties,” PLoS ONE,vol.7,no.10,ArticleIDe44479,2012. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 876047, 13 pages http://dx.doi.org/10.1155/2015/876047

Research Article Putative Epimutagens in Maternal Peripheral and Cord Blood Samples Identified Using Human Induced Pluripotent Stem Cells

Yoshikazu Arai,1,2,3 Koji Hayakawa,1 Daisuke Arai,1 Rie Ito,4 Yusuke Iwasaki,4 Koichi Saito,4 Kazuhiko Akutsu,5 Satoshi Takatori,5 Rie Ishii,6 Rumiko Hayashi,7 Shun-Ichiro Izumi,8 Norihiro Sugino,9 Fumio Kondo,10 Masakazu Horie,11 Hiroyuki Nakazawa,4 Tsunehisa Makino,12 Mitsuko Hirosawa,1 Kunio Shiota,1 and Jun Ohgane1,2

1 Laboratory of Cellular Biochemistry, Animal Resource Sciences/Veterinary Medical Sciences, The University of Tokyo, Tokyo 113-8657, Japan 2 Laboratory of Genomic Function Engineering, Department of Life Science, School of Agriculture, Meiji University, Kawasaki 214-8571, Japan 3 Laboratory of Developmental Engineering, Department of Life Science, School of Agriculture, Meiji University, Kawasaki 214-8571, Japan 4 Department of Analytical Chemistry, Faculty of Pharmaceutical Sciences, Hoshi University, Tokyo 142-8501, Japan 5 Division of Food Chemistry, Osaka Prefectural Institute of Public Health, Osaka 537-0025, Japan 6 Saitama Prefectural Institute of Public Health, Saitama 355-0133, Japan 7 Department of Toxicology, Aichi Prefectural Institute of Public Health, Aichi 462-8576, Japan 8 Department of Obstetrics and Gynecology, School of Medicine, Tokai University, Kanagawa 259-1193, Japan 9 Department of Obstetrics and Gynecology, Yamaguchi University Graduate School of Medicine, Ube 755-8505, Japan 10Department of Pharmacology, School of Medicine, Aichi Medical University, Aichi 480-1195, Japan 11 Faculty of Home Economics, Otsuma Women’s University, Tokyo 102-8357, Japan 12Fuji-OyamaHospital,Shizuoka410-1326,Japan

Correspondence should be addressed to Kunio Shiota; [email protected] and Jun Ohgane; [email protected]

Received 15 November 2014; Revised 26 January 2015; Accepted 9 March 2015

Academic Editor: Francesca Pacchierotti

Copyright © 2015 Yoshikazu Arai et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The regulation of transcription and genome stability by epigenetic systems are crucial for the proper development of mammalian embryos. Chemicals that disturb epigenetic systems are termed epimutagens. We previously performed chemical screening that focused on heterochromatin formation and DNA methylation status in mouse embryonic stem cells and identified five epimutagens: diethyl phosphate (DEP), mercury (Hg), cotinine, selenium (Se), and octachlorodipropyl ether (S-421). Here, we used human induced pluripotent stem cells (hiPSCs) to confirm the effects of 20 chemicals, including the five epimutagens, detected atlow concentrations in maternal peripheral and cord blood samples. Of note, these individual chemicals did not exhibit epimutagenic activity in hiPSCs. However, because the fetal environment contains various chemicals, we evaluated the effects of combined exposure to chemicals (DEP, Hg, cotinine, Se, and S-421) on hiPSCs. The combined exposure caused a decrease in the number of heterochromatin signals and aberrant DNA methylation status at multiple gene loci in hiPSCs. The combined exposure also affected embryoid body formation and neural differentiation from hiPSCs. Therefore, DEP, Hg, cotinine, Se, and S-421 were defined asan “epimutagen combination” that is effective at low concentrations as detected in maternal peripheral and cord blood. 2 BioMed Research International

1. Introduction in Glasgow minimum essential medium containing 10% KSR, 100 mM nonessential amino acids, and 100 mM 2- Epigenetic systems are crucial for normal embryonic devel- mercaptoethanol (all from Invitrogen). PA6 was obtained opment via the transcriptional regulation of tissue and from RIKEN BioResource Center. cell-type-dependent gene expression. Epigenetic marks, The hiPSCs were cultured with the indicated chemicals such as DNA methylation and histone modification, cause at concentrations equivalent to serum levels (1x) or 10-fold dynamic changes in heterochromatic and euchromatic higher (10x) (Table 1). The serum levels (1x) were determined regions depending on the cellular conditions and cell type based on the concentrations of cord blood samples and/or [1–4]. During the differentiation process, demethylation and pregnant mothers’ serum using liquid chromatography-mass the de novo methylationofDNAoccuratgenelocitoform spectrometry (LC/MS), gas chromatography-mass spectrom- tissue-dependent and differentially methylated regions (T- etry (GC/MS), or inductively coupled plasma-mass spec- DMRs) in the mammalian genome [5–8]. Epigenetic systems trometry (ICP/MS), as described in a previous study [12]. have dual aspects of plasticity and stability depending on The chemicals were added as described previously [12], and the cellular environment and cell fate decisions, respectively. the final concentrations of solvents were 0.007% hydrochloric Therefore, the long-lasting effects of low concentrations of acid (HCl) for tin (Sn), 0.0025% nitric acid (HNO3)forSe, chemicalsonabnormalphenotypesmightbeattributableto cadmium (Cd), Hg, and lead (Pb), or 0.1% ethanol (EtOH) epigenetic alterations; chemicals that disturb the epigenetic for the other 15 chemicals and trichostatin A (TSA) (Sigma- status are termed epimutagens. Aldrich, Tokyo, Japan). The chemicals were divided into Various types of chemicals, including endocrine disrup- groups as follows: group A (pesticides), group B (tobacco), tors, dioxins, heavy metals, and tobacco, and their metabo- group C (perfluorinated compounds (PFCs)), group D lites have been detected in the fetal environment [9–11]. We (heavy metals), and group E (phthalate) (Table 1). The mix- previously performed epimutagen screening using mouse ture of chemicals in group D (heavy metals) dissolved in HCl embryonic stem cells (mESCs). Of the 25 environmental and HNO3 was added to culture medium, and the final con- chemicals detected in human blood samples, five chemicals centrations of the solvents in culture medium were 0.007% (diethyl phosphate (DEP), mercury (Hg), cotinine, selenium and 0.0025%, respectively. As to the other chemical mixtures (Se), and octachlorodipropyl ether (S-421)) disturbed epige- (groupsA,B,C,andE),thefinalconcentrationofthesolvent netic systems at relatively low concentrations (0.1–100 ppb) was0.1%EtOH.Themixtureofthefiveepimutagens,DEP, [12]. We also demonstrated that dimethyl sulfoxide (DMSO), Hg, cotinine, Se, and S-421, dissolved in HNO3 and EtOH which was previously used as a cryopreservant for fertilized was added to culture medium, and the final concentrations eggs, altered the DNA methylation status in both gene areas in culture medium were 0.0025% and 0.1%, respectively. and repetitive sequences during the differentiation of mESCs into embryoid bodies (EBs) [13]. 2.2. Immunohistochemistry. HumaniPSCsanddifferentiat- Different mammalian species display different sensitivi- ing cells were fixed in 4% paraformaldehyde for 10 min. ties to chemicals [14, 15]; therefore, the epimutagenic effects of After permeabilization with 0.2% Triton X-100 for 5 min, chemicals need to be assessed using a human model system. samples were blocked using blocking buffer (5% bovine Here, we aimed to establish a screening system for epimuta- serum albumin, 0.1% Tween-20 in PBS) for 30 min. The gens using human induced pluripotent stem cells (hiPSCs), samples were incubated with either anti-heterochromatin an in vitro model for early human embryos, to examine the protein 1𝛼 (HP1𝛼) mouse monoclonal antibodies (Cat. individual and combined effects of environmental chemicals number: MAB3584, Chemicon, Temecula, CA, USA) or on the epigenetic status of human embryos/fetuses. anti-𝛽III-tubulin mouse monoclonal antibodies (Cat. number: MMS-435P, Covance, Princeton, NJ, USA) primary 2. Materials and Methods antibodies diluted in blocking buffer (1 : 500 and 1 : 200, resp.) for 45 min, followed by washing three times in 2.1. Culture of hiPSCs and Exposure to Chemicals. Human PBS containing 0.05% Tween-20. After incubation with iPSCs (201B7) that have been established at Dr. Yamanaka’s fluorescent secondary antibodies (Alexa Fluor 594 goat laboratory [16] were obtained from RIKEN BioResource anti-mouse IgG, Invitrogen) diluted in blocking buffer Center (Tsukuba, Japan). The hiPSC line was cultured on (1 : 200) for 60 min, the samples were washed again. The SNL feeder cells with Primate ES Cell Medium (ReproCELL, sampleswerethenmountedonaglassslidewithPer- Yokohama, Japan), supplemented with 5 ng/mL bFGF (Wako, maFluor aqueous mounting medium (Thermo Scientific, 󸀠 Osaka, Japan). hiPSC colonies were detached and separated Rockford, IL, USA) containing 0.2 𝜇g/mL of 4 ,6-diamidino- into small clumps using a reagent containing 20% knock- 2-phenylindole (DAPI) (Dojindo, Kumamoto, Japan). All out serum replacement (KSR; Invitrogen, Rockville, MD, reactions were performed at room temperature. Immunoflu- USA), 0.25% trypsin (Invitrogen), 1 mg/mL collagenase IV orescent images of anti-HP1𝛼-oranti-𝛽III-tubulin staining (Wako), and 1 mM CaCl2 (Wako). To induce the formation were then acquired by confocal fluorescence microscopy of EBs, small clumps of hiPSCs were transferred to bacterial using FV10i (Olympus, Tokyo, Japan) or CellVoyager CV1000 Petri dishes in Primate ES Cell Medium without bFGF (Yokogawa Electric Corporation, Tokyo, Japan) microscopes, after the removal of feeder cells. For neural differentiation, respectively. Images obtained using anti-HP1𝛼- (5–10 visual small clumps of hiPSCs were plated on a PA6 feeder layer fields) and anti-𝛽III-tubulin antibodies (150 fields) were BioMed Research International 3

Table1:Thechemicalsusedinthepresentstudy.

aConcentration in serum cEffect of chemicals on mESCs bExposure Group Chemical Mean ± SD (ppb) (bconcentration that affected concentration (ppb) Cord blood’s (mothers’) mESCs) 1x 10x 3-PBA <0.2d (0.3e) 0.1 1.0 − TCP <0.2d (0.9e) 0.1 1.0 + (10x) DMP 4.3 ± 3.9 (8.6 ± 4.2) 0.1 1.0 + (10x) ± ± A, pesticide DEP 0.28 0.1 (0.3 0.1) 0.1 1.0 + (1x) DMTP 0.9 ± 0.8 (16.2 ± 4.5) 0.1 1.0 − DETP 2.8 ± 1.8 (7.9 ± 3.0) 0.1 1.0 + (10x) DMDTP ND (0.3e,f) 0.1 1.0 + (10x) DEDTP ND (<0.05d,f) 0.1 1.0 − S-421 ND (10.3g,h) 0.01 0.1 + (1x) Nicotine ± ± i − B, tobacco 1.4 0.57 (1.6 2.6) 100 1000 Cotinine 8.7e (43.7 ± 55.8)i 100 1000 + (1x) PFOA ± ± − C, PFCs 1.4 0.5 (1.5 0.6) 10 100 PFOS 1.4 ± 0.6 (3.9 ± 1.4) 10 100 − Sn ND (1.02 ± 0.51j)1.010 − Se ND (110 ± 18) 100 1000 + (1x) D, heavy metals Cd 0.042 ± 0.003 (0.038 ± 0.016) 0.1 1.0 − Hg ND (0.6 ± 0.34) 1.0 10 + (1x) Pb 0.3 ± 0.05 (0.3 ± 0.12) 1.0 10 − DEHP ± ± − E, phthalate 4.0 1.1 (5.3 0.8) 1.2 12 MEHP 6.3 ± 5.1 (4.3 ± 1.5) 5.2 52 + (10x) ND: not determined. aThe details are in our previous report [12]. b1x: serum level detected in cord blood samples and/or pregnant mothers’ serum. 10x: ten-fold higher level than that of the cord blood samples and/or pregnant mothers’ serum. cDescribed in our previous report [12]. dLess than detection limit level. eDetected only in one sample (n =11–22). fConcentrations determined using plasma samples in the previous report [33]. gDetected in all examined samples (𝑛=58). hConcentrations determined using human milk samples (ng/g in lipids) in our previous report [12]. iMean ± SD values were calculated using all the samples containing both smokers and nonsmokers. Nicotine and cotinine were detected at relatively high levels (appropriately 100 ppb) from smokers but were not detected from nonsmokers. Thus, the exposure concentrations were determined based on the average values of the smokers’ samples. jConcentrations determined using urine samples in our previous report [12]. analyzed from individual samples and quantified using patterns. Specifically, genomic DNA was extracted from ImageJsoftwareprovidedbytheNationalInstituteofHealth hiPSCs in lysis buffer (100 mM Tris-HCl pH 8.0, 5mM (http://rsb.info.nih.gov/ij/). Briefly, RGB images were con- EDTA, 0.2% SDS, 200 mM NaCl, and 200 𝜇g/mL pro- ∘ verted to 8-bit grayscale (0–255). Next, the thresholds of teinase K) at 55 C for 30 min. After removing proteins intensity of the HP1𝛼 images were determined using the with phenol/chloroform/isoamyl alcohol (50/49/1, v/v/v), automatic threshold setting of the ImageJ program (between genomic DNA was treated with RNase A (Roche Diagnostics, 23 and 39) and the number of HP1𝛼 signals per nucleus Mannheim, Germany) and purified using EtOH precipita- (appropriately 100 nuclei in each sample) was counted. For tion. Purified genomic DNA was digested with the restriction 𝛽III-tubulin images, the threshold was set at 25, and 𝛽III- enzyme HindIII (TaKaRa, Kyoto, Japan) and purified by tubulin-positive area was measured by ImageJ software. EtOH precipitation. After denaturing the digested genomic DNA with 0.3 M NaOH, sodium metabisulfite (pH 5.0) 2.3. Combined Bisulfite Restriction Analysis (COBRA) Assay. and hydroquinone were added to final concentrations of Genomic DNA extraction and bisulfite conversion were 2.0 M and 0.5 mM, respectively. A bisulfite reaction was performed as described previously [12]. DNA methylation then performed using a thermal cycler with the following ∘ ∘ analysis was performed using COBRA assays [17]for10 cycling conditions: 20 cycles of 95 Cfor30secand55Cfor ∘ T-DMRs that exhibited human ESC-specific methylation 15 min, followed by 55 C for 10 h. Bisulfite-treated genomic 4 BioMed Research International

Table 2: PCR primers used.

(a) 󸀠 󸀠 Bisulfite PCR primers Primers (5 to 3 ) Size (bp) From transcription start site Forward: TGGGGTTTATGAGGGTAAGGT SUSD1 214 1.5 kbp downstream Reverse: CCACACCACACACAACCAAT Forward: GGTTTAGGTGTTTGAAGATTTGGT BRD1 378 500 bp upstream Reverse: ATAAATACCCCTAATCCCCCTAAA Forward: ATTTGGAAATAGGGAGGAGTAATTTT NEBL 262 1.5 kbp upstream Reverse: TCTCAACAACTTATTTTCTTACAACACA Forward: TGTGGTTTATGTTTGGAATTG GLI3 183 2.0 kbp downstream Reverse: TCACTAACTCTTCACCCACAATTTA Forward: TTTAAATTATTTTGTTTTGGGGATG POU2F1 490 2.5 kbp downstream Reverse: TCTACCTCTCACAAACCAACTATCC Forward: TTTTATGTGGGGTTTAGTAGTTTGG HOXB6 269 1.5 kbp upstream Reverse: ACACATTCACACTCACAAACACATTA Forward: TGAAAGGGAAGGGGTTGTTT HOXA3 216 1.5 kbp downstream Reverse: TCCCTATATTATACACTATCCCAAAAA Forward: TGGGTTTGTTATTTATTTTGTTGGTA CBX8 357 1.0 kbp downstream Reverse: CTACCCCACTCTTAAAACCATCTTCT Forward: TTATGGGTTTAGGTATGGTGGAAATA NANOG 291 500 bp downstream Reverse: AAAACTACCCAATAACATCCACAAAC Forward: ATAGTTTTGGGAAAGTAGAATTTGGT BMP2 379 1.5 kbp upstream Reverse: TATTTATCTCACCCAACTCAAAAACA (b) 󸀠 󸀠 RT-PCR primers Primers (5 to 3 ) Size (bp) Forward: CAGGTGGCGGACGTGTGAAAATTGAGAGTG MAP2 212 Reverse: CACGCTGGATCTGCCTGGGGACTGTG Forward: ACCCATTATCCAGATGTGTTTGCCCGAG PAX6 317 Reverse: ATGGTGAAGCTGGGCATAGGCGGCAG Forward: CTCCAAGACTTCCCTCAGCTTT NES 163 Reverse: CTTAAGAAAGGCTGGCACAGGT Forward: CAAGATCAGCAATGCCT GAPDH 68 Reverse: CTTCCACGATACCAAAGTTGTC

DNA was then purified using a QIAquick gel extraction kit 2.4. RNA Extraction and RT-PCR. Total RNA was extracted (Qiagen GmbH, Hilden, Germany), desulfonated with 0.3 M using an RNeasy plus mini kit (Qiagen). First-strand cDNA ∘ NaOH at 37 C for 15 min, and EtOH precipitated. Purified synthesis was performed using the SuperScript III first- bisulfite-treated DNA was amplified using BioTaq HS DNA strand synthesis system for RT-PCR (Invitrogen). PCR was polymerase (Bioline, London, UK) using specific primers for performed using BioTaq HS DNA polymerase with specific T-DMRs (Table 2). Polymerase chain reaction (PCR) was primers for each gene locus (Table 2). PCR reactions were ∘ ∘ performed using the following conditions: 95 Cfor10min; performed under the following cycling conditions: 95 Cfor ∘ ∘ ∘ ∘ ∘ ∘ 40 cycles of 95 Cfor30sec,60Cfor30sec,and72Cfor 10 min; 25 cycles of 95 Cfor30sec,60Cfor30sec,and72C ∘ ∘ 1 min; and a final extension at 72 Cfor2min.Amplified for 1 min; and a final extension at 72 Cfor2min. PCRproductsweredigestedusingHpyCH4IV (New Eng- ∘ land BioLabs, Inc., Beverly, MA, USA) at 37 Cfor3hand 2.5. Statistical Analysis. Statistical comparisons of the HP1𝛼 then analyzed by microchip electrophoresis using MCE-202 signals were performed using the Wilcoxon test, and those of (MultiNA;Shimadzu,Kyoto,Japan).TheDNAmethylation DNA methylation status, expression levels of neural marker levelsanalyzedbytheCOBRAassaywerecalculatedusingthe genes, and areas detected using anti-𝛽III-tubulin antibodies formula were performed using Student’s t-test. 𝐼𝐶 ( ) = 100 × , Estimated methylation degree % (𝐼𝐶 +𝐼𝑈𝐶) 3. Results (1) 3.1. Effects of 20 Environmental Chemicals on Heterochromatin Signals in hiPSCs. The outline of the present study together 𝐶 𝑈𝐶 where 𝐼 and 𝐼 represent the sum of the intensities of with that of our previous study [12]isillustratedinFigure1. digested and undigested bands, respectively. We first examined the effects of the 20 chemicals detected in BioMed Research International 5

The 1st epimutagen screening using the mESC system (Arai et al., 2011)

Determination of cord blood samples and pregnant mothers’ serum concentration of chemicals

Identification of five chemicals at the 1× concentrations as epimutagen candidates (DEP, Hg, cotinine, Se, and S-421)

The 2nd epimutagen screening using the hiPSC system (the present study) Single exposure of 20 chemicals Combinatorial exposure (1) Serum level (1×) (1) Chemicals in groups A–E (2) Ten-fold higher level (10×) (2)The5 epimutagen candidates (DEP, Hg, cotinine, Se, and S-421)

No effects DNA methylation analysis of Long-lasting effects of chemicals except for PFOA (1×) transcriptional regulatory regions on differentiation of hiPSCs Tissue/cell-type dependently (1) EB formation methylated genes (2) Neural differentiation

Figure 1: Outline of this study.

cordbloodserumand/orthatofpregnantmothers(Table1) perfluorooctanoate (PFOA) caused a significant increase in as described in our previous report [12]bycountingthe the heterochromatin signal in hiPSCs (Figure 3(a)). The number of heterochromatin foci. In mice, heterochromatin chemical concentrations used in the present study were 1,000- can be clearly visualized by staining with both DAPI and to 10,000-fold lower than those used to show genotoxicity of HP1𝛼, a heterochromatin marker [12, 18]. In the hiPSCs, some of these chemicals (e.g., nicotine and Cd) in human cells DAPI signals were also merged with HP1𝛼 signals (Fig- [21, 22], indicating that the concentrations used in the present ure 2(a)). It has also been reported that HP1𝛼 localizes at study did not result in genotoxicity. In addition, observation the pericentromeric heterochromatin in human cells [19]. of chromosome-condensed M phase nuclei in DAPI-stained Taken together, our results suggest that DAPI and HP1𝛼 images, which were used to examine heterochromatin dots can be used to identify pericentromeric heterochromatin (Figures 2–4),canbeusedasanindicatoroflivingand in hiPSCs as well as mESCs. However, whole human cell dividing cells. Thus, we compared the number of M phase nuclei were stained more intensely and broadly than mouse nuclei in hiPSCs treated with one or combination of the 20 cell nuclei, making it difficult to identify the pericentric chemicals at serum concentrations (1x) and 10-fold higher heterochromatin dots compared to the surrounding regions concentrations (10x) with the number in solvent-exposed (Figure 2(a)), which was consistent with a previous finding control cells. Compared with solvent-exposed control cells, [20]. Therefore, we used HP1𝛼 immunostaining to detect none of the single-chemical-exposed hiPSCs or the multiple- heterochromatin signals in hiPSCs, and exposure to 20nM chemical-exposed cells exhibited significant differences in and 40 nM TSA altered the heterochromatin signals detected thenumberofMphasenuclei(seeSupplementalFig- by HP1𝛼 staining in hiPSCs in a dose-dependent manner ure1oftheSupplementaryMaterialavailableonlineat (Figure 2(b)). http://dx.doi.org/10.1155/2015/876047). This result indicates Our previous study indicated that DEP, Hg, cotinine, that the chemical concentrations used in the present study did Se, and S-421 exhibited epimutagenic activity in mESCs not cause cytotoxicity. [12]. In contrast, these chemicals did not alter heterochro- Overall, 19 out of the 20 chemicals originally tested matin signals in hiPSCs at either serum concentrations did not exhibit epimutagenic activities in hiPSCs, even at (1x) or 10-fold higher concentrations (10x) (Figure 2(c)). concentrations that were 10-fold higher than their serum An additional five chemicals (3,5,6-trichloro-2-pyridinol levels.Thesedataindicatethattheresponseandsensitivity (TCP), dimethyl phosphate (DMP), diethyl thiophosphate of human and mouse cells differ. In addition, PFOA altered (DETP), dimethyl dithiophosphate (DMDTP), and mono(2- heterochromatin formation in hiPSCs at 1x, but not at ethylhexyl)phthalate (MEHP)) also had no effect on the 10x, serum concentrations, suggesting that the epigenetic heterochromatin signals (Figure 2(d)), even though they alterations that accompany chemical exposure are not simply affected heterochromatin signals in mESCs12 [ ]. Thus, mouse dose-dependent. and human cells clearly exhibit different sensitivities to these chemicals. We also studied 10 chemicals that did 3.2. Effects of Combined Exposure to Chemicals on Heterochro- not exert epimutagenic effects in mESCs [12]. Of these, 1x matin Marks in hiPSCs. We next examined the effects of 6 BioMed Research International

DAPI a b DAPI TSA0 22040 (nM) ∗ HP1𝛼 ∗ 1𝛼 20 15 HP1𝛼 a b HP1𝛼 10

Intensity DAPI 5 signals/nuclei Number of HP of Number Merge a b 0 22040 (nM) ab TSA

(a) (b) Control DEP Hg Cotinine Se S-421 HP1𝛼

DAPI

20 20 20 20 20 1𝛼 15 15 15 15 15

10 10 10 10 10

signals/nuclei 5 5 5 5 5 Number of HP of Number

01×10× 01×10× 01×10× 01×10× 01×10× DEP Hg Cotinine Se S-421

DAPI

20 20 20 20 1𝛼 15 15 15 15 15

10 10 10 10 10

signals/nuclei 5 5 5 5 5 Number of HP of Number 0 1× 10× 0 1× 10× 0 1× 10× 0 1× 10× 0 1× 10× TCP DMP DETP DMDTP MEHP

(d)

Figure 2: Epimutagen screening of hiPSCs. (a) Visualization of heterochromatin structure in nuclei by DAPI staining and immunofluorescence using anti-HP1𝛼 antibody. The intensities of signals of DAPI and HP1𝛼 onthedottedlines(a-b)weremeasuredusingtheImageJsoftware and plotted. The strong HP1𝛼 signals (filled triangles) were confirmed to merge with the DAPI signals. (b) Altered heterochromatin formation after treating hiPSCs with the known epimutagen TSA. hiPSCs were treated with TSA (0, 2, 20, or 40 nM) for 96 h, and heterochromatin was detected using immunofluorescence with anti-HP1𝛼 antibodies (red) and DAPI counterstaining (blue). The number of HP1𝛼 signals per interphase nucleus was counted using ImageJ software. The number of signals is shown as a box plot. Statistical comparisons of signal number ∗ were performed using the Wilcoxon test. 𝑃 < 0.01.Scalebar=10𝜇m. (c) The number of HP1𝛼 signals in hiPSCs exposed to serum levels (1x) or 10-fold increased concentrations (10x) of DEP, Hg, cotinine, Se, or S-421 for 96 h were analyzed. The upper panel shows images of cells exposed to 1x chemicals, and the lower panel presents the number of signals as a box plot. Scale bar = 10 𝜇m. (d) Exposure to the 10x concentrations of TCP, DMP, DETP, DMDTP, and MEHP. All heterochromatin analyses were performed at least twice independently. BioMed Research International 7

3-PBA DMTP DEDTP Nicotine PFOA PFOS Sn Cd Pb DEHP HP1𝛼

DAPI

∗ 20 20 20 1𝛼 15 15 15 15 15 10 10 10 10 10 5 signals/nuclei 5 5 5 5 Number of HP of Number 0 1× 10× 0 1× 10× 0 1× 10× 0 1× 10× 0 1× 10× 3-PBA DMTP DEDTP Nicotine PFOA 20 20 20 20

1𝛼 15 15 15 15 15 10 10 10 10 10 5 5 5 5 5 signals/nuclei Number of HP of Number 0 1× 10 × 0 1× 10 × 0 1× 10 × 0 1× 10 × 0 1× 10 × PFOS Sn Cd Pb DEHP

(a) Group ABCDE Pesticide TobaccoPFCs Heavy metals Phthalate HP1𝛼

DAPI

∗

15 15 15 15 1𝛼 15

10 10 10 10 10

signals/nuclei 5 5 5 5 5 Number of HP of Number

−+ −+ −+ −+ −+ Group AB CDE Pesticide Tobacco PFCs Heavy metals Phthalate (9 chemicals) (2 chemicals) (2 chemicals) (5 chemicals) (2 chemicals) Combination of chemicals belonging to the same group

(b)

Figure 3: Exposure to single or multiple chemicals. (a) After 96 h exposure to either of the 10 chemicals that did not affect heterochromatin formation in mESCs at their serum levels (1x) or 10-fold higher level than serum concentrations (10x), the number of HP1𝛼 signals (red) was counted using ImageJ software. The upper panel shows images of cells exposed to 1x chemicals; the number of signals is shown as a box plot ∗ in the lower panel. Statistical comparisons of signal number were performed using the Wilcoxon test. 𝑃 < 0.01.Scalebar=10𝜇m. (b) Effects of exposing hiPSCs to groups of chemicals on heterochromatin formation. Cells were treated with serum concentrations of combinations of chemicals belonging to groups A–E for 96 h, and the heterochromatin status was evaluated by counting the number of HP1𝛼 signals. ∗ 𝑃 < 0.01. All heterochromatin analyses were performed twice independently. 8 BioMed Research International

Combination of the five chemicals ∗ (DEP, Hg, cotinine, Se, and S-421) 20 − +

HP1𝛼 1𝛼 15

signals/nuclei 5 DAPI Number of HP of Number

− + Combination of the five chemicals

(a) 100 ∗ ∗ 80 ∗

40 ∗

20 DNA methylation level (%) level methylation DNA

0 GLI3 CBX8

BRD1 NEBL

BMP2 SUSD1 HOXB6 HOXA3 POU2F1 NANOG Control (solvent) Combination of the five chemicals (DEP, Hg, cotinine, Se, and S-421)

(b) Combination of the five chemicals (DEP, Hg, cotinine, Se, and S-421) − +

(c)

Figure 4: Effects of a chemical mixture (DEP, Hg, cotinine, Se, and S-421) on the epigenetic status of hiPSCs. (a) Cells were treated with serum concentrations of a chemical mixture (DEP, Hg, cotinine, Se, and S-421) for 96 h, and the number of HP1𝛼 signals(red)wascounted ∗ using ImageJ software. Statistical comparisons of signal number were performed using the Wilcoxon test. 𝑃 < 0.01.Scalebar=10𝜇m. Heterochromatin analysis was performed twice independently. (b) The DNA methylation status of the T-DMRs of 10 gene regions obtained using COBRA assays. Human iPSCs were cultured as described in (a); the DNA methylation percentage is shown as means ± SE (𝑛=3). The white and black boxes indicate the methylation level of solvent-treated control and chemical-exposed cells, respectively. Statistical ∗ comparisons of DNA methylation were performed using Student’s t-test. 𝑃 < 0.05. (c) Impaired EB formation after exposure to the chemical mixture. Cells were treated with serum concentrations of the five chemicals for 96 h, and cells were differentiated into EBs in the presence of chemicals for 15 days. Scale bar = 250 𝜇m. Experiments were performed thrice independently.

combined exposure to chemicals belonging to the same group However, it is noteworthy that PFOA, which belonged to (A, pesticides; B, tobacco; C, PFCs; D, heavy metals; and E, group C, altered the heterochromatin signal alone (Fig- phthalate; Table 1). Combined exposure to chemicals from ure 3(a)). Therefore, these data suggest that simple mixtures group C increased heterochromatin signals, whereas those of similar types of chemicals do not affect heterochromatin from groups A, B, D, and E had no effect (Figure 3(b)). formation. BioMed Research International 9

3.3. Effects of Combined Mouse Epimutagens (DEP, Hg, Coti- for exposure to the epimutagen mixture, either throughout nine, Se, and S-421) on Heterochromatin Marks, DNA Methy- the culture period (both stem and differentiation periods, lation Status, and EB Formation in hiPSCs. Previous studies similar to condition III) or for 10 days following induction of demonstrated that DEP, Hg, cotinine, Se, and S-421 exerted EB differentiation. No significant differences in EB size were epimutagenic activities in mESCs [12]; therefore, we exam- observed in the hiPSCs treated with the chemical mixture inedtheeffectsofamixtureofthesefivechemicals(Figure4). only after differentiation induction compared with control Exposure of hiPSCs to a mixture of serum concentrations of EBs whereas EB size differences were observed as early as day DEP, Hg, cotinine, Se, and S-421 decreased heterochromatin 10 of differentiation in the hiPSCs exposed to the chemical signals (Figure 4(a)). This mixture also affected the DNA mixture throughout the culture period, as in condition III methylation status in the T-DMRs of gene loci that are (data not shown). Thus, we did not further examine EB transcriptional regulatory regions, showing differences in formation with only post-differentiation exposure to the DNA methylation levels depending on tissue/cell types, and epimutagen mixture. related to early mammalian development (Figure 4(b)). We previously performed genome-wide DNA methylation analy- 3.5. Effect of the Epimutagen Mixture on Neural Differentia- ses for human ESCs and their differentiation derivatives using tion. We next investigated the effects of the epimutagen mix- apromotertilingarrayandaCOBRAassaywithmicrochip ture on neural differentiation. On day 20 after the induction electrophoresis to confirm the reproducibility of the tiling of neural differentiation, the colonies had expanded in culture array data (unpublished data). We identified transcriptional conditions I and II. The cells grown in culture condition III regulatory regions for which the DNA methylation level had detached and died (Figure 5(b), right panel). On day 14, could be reproducibly detected depending on tissue/cell the colonies grown in culture condition III remained intact type.ThegenelociweanalyzedusingtheCOBRAassay but were smaller than those in culture condition I (data not with microchip electrophoresis in the present study were shown). also included in this gene set. Among various epigenetic The cells grown in culture conditions I and II could modifications, slight changes are most detectable with the differentiate into neurons, as confirmed by staining using highest reproducibility in DNA methylation levels. Thus, we antibodies against the neural marker 𝛽III-tubulin on day 24 decided to analyze DNA methylation level of the gene loci (Figure 5(c), left panel). There was no difference in the 𝛽III- after treatment with the five chemicals. The mixture of five tubulin-positive areas between culture conditions I and II chemicals, termed as an epimutagen mixture, also caused the (Figure 5(c), right panel). However, the expression levels of abnormal development of EBs (Figure 4(c)), whereas normal neural marker genes (NES, MAP2,andPAX6) were lower EBs with yolk-sac-like structures were observed in the vehicle in hiPSCs grown in culture condition II than those grown control. Therefore, the epimutagen mixture has the potential in culture condition I (Figure 5(d)). The expression level to affect the differentiation of cells during embryogenesis. of MAP2, a mature neural marker, in culture condition However, so far no individual serum samples showed the II was markedly decreased compared with vehicle control, presence of all five chemicals in combination, based on the suggesting that exposure to the epimutagen mixture caused maternal and cord blood data. long-lasting impairment of neural differentiation.

3.4. Disruption of Normal EB Formation after Exposure of 4. Discussion hiPSCs to the Epimutagen Mixture Only during the Stem Cell State. Human iPSCs were maintained for 4 days in In the present study, a mixture of chemicals (DEP, Hg, coti- stem culture medium followed by differentiation medium, nine, Se, and S-421) affected heterochromatin signals, DNA either with or without the epimutagen mixture (Figure 5(a)). methylation status, EB formation, and neural differentiation Three culture conditions were used. Culture condition I in hiPSCs. Various chemicals have been detected at low was a solvent-treated control. In culture conditions II and concentrations in human fetal samples, and prenatal chemical III, cells were treated with the chemicals for 4 days before exposure has been reported to cause developmental disorders differentiation. EB formation was then induced in the absence such as neural dysfunction in children after birth [23, 24]. (II) or presence (III) of the chemicals (Figure 5(a),left Exposure to multiple chemicals potentially affecting human panel). In the vehicle control, normal EBs with yolk-sac-like health is also a growing concern [25], and, in fact, fetuses are structures formed as expected. In contrast, abnormal EBs exposed to complex combinations of chemicals. For example, were observed after continuous exposure to the epimutagen polychlorinatedbiphenyls,lead,andmethylmercurywere mixture (culture condition III) (Figure 5(a), right panel). detected in identical samples of cord blood, mother’s blood, Our previous study demonstrated the irreversible effect of or lipid [26, 27], suggesting that the combinational effects DEP on mouse heterochromatin configuration even after of chemicals on epigenetic systems should be considered. its removal as an abnormal epigenetic memory [12], and Consistent with this, a mixture of the five chemicals (DEP, we examined whether chemical exposure has long-lasting Hg, cotinine, Se, and S-421) and PFOA were found to be effects after removal of the chemicals in human cell differ- epimutagenic in hiPSCs in the current study. entiation. Importantly, treating hiPSCs with the epimutagen The combined exposure to DEP, Hg, cotinine, Se, and S- mixture only during the stem cell state (culture condition 421 only before differentiation also disturbed EB formation II) was sufficient to inhibit the formation of normal EBs and neural differentiation. Because the hiPSC system is (Figure 5(a)). We first performed a preliminary experiment an in vitro model of developing early embryos, epigenetic 10 BioMed Research International

Neural differentiation Culture EB formation Culture condition Undiff. Diff. condition Undiff. Diff. I I Culture Culture 0 4 24 (days) condition04 24 (days) condition a I − − I − − II II + − II II + − a III + + III + + b III Combination of chemicals III Combination of chemicals b (DEP, Hg, cotinine, Se, and S-421) (DEP, Hg, cotinine, Se, and S-421)

(a) (b) Culture condition 30 1.0 ∗ III 25 0.8 𝛽III-tubulin 20 0.6 15 0.4 DAPI III-tubulin-

𝛽 10 0.2 positive area (%) area positive 5 Relative expression level expression Relative 0 Merge 0 I II IIIIII III NES MAP2 PAX6

Culture Chemical exposure EB Neural condition formation differentiation Undiff. Diff.

I − − Normal Normal II + − Affected Affected Affected III + + Affected (did not survive)

(e)

Figure 5: Effects of the chemical mixture (DEP, Hg, cotinine, Se, and S-421) on cellular differentiation. (a) Effects of the timing of chemical exposure on EB formation. EBs derived from hiPSCs were cultured using three culture conditions: I, solvent-treated control; II and III, cells treated with the chemical mixture for 4 days before differentiation. EB formation was then induced for up to 24 days in the absence (II) or presence (III) of the chemical mixture (left panel). The right panel shows images of EBs on day 24. Scale bar=250 𝜇m. −: solvent only; +: exposure to serum concentrations of the chemical mixture. Experiments were performed thrice independently. (b) Effects of the five chemicals on neural differentiation. The culture conditions used were the same as in (a). Differentiated cells were analyzed on day 20(right panel). Enlarged images are shown as “a” and “b” for conditions II and III, respectively. Scale bar = 200 𝜇m. Experiments were performed twice independently. (c) After 24 days of neural differentiation in culture conditions I and II, cells were stained with antibodies for theneural marker 𝛽III-tubulin, and the 𝛽III-tubulin-positive area (%) in 150 images was measured using ImageJ software. The data are presented as means ± SE. (d) Neural marker gene expression. On day 24, cells grown in culture conditions I and II were harvested, and the expression levels of the neural marker genes NES, MAP2,andPAX6 were assessed using RT-PCR. The relative expression levels were normalized to that of GAPDH. The expression levels are shown as mean ± SD (𝑛=3). Statistical comparisons of the expression level were performed using Student’s ∗ t-test. The 𝑃-value of NES, MAP2,andPAX6 was 0.081, 0.015, and 0.065, respectively. 𝑃 < 0.05. (e) Summary of cellular differentiation in chemical-exposed hiPSCs. −: solvent only; +: exposure to serum concentrations of DEP, Hg, cotinine, Se, and S-421.

errors that occur in undifferentiated cells might serve as an multiple chemicals at trace levels; it is possible that certain epigenetic memory that is sufficient to cause later develop- combinations of chemicals might have the potential to cause mental abnormalities in differentiating embryonic cells. The epigenetic dysfunction in developing early embryos. cytotoxicity of chemicals was reported to be more severe in We demonstrated previously that the serum concentra- the early stages of development than in adulthood [28, 29]. tionsofepimutagenicchemicalsdisturbedtheconfiguration In addition, prenatal exposure to pesticides was found to of heterochromatin and the DNA methylation status of T- cause long-term developmental disorders after birth [30, 31]. DMRs using mESCs [12]. However, in the present study using It is also evident that developing fetuses are exposed to hiPSCs, the single exposure to most of these chemicals did BioMed Research International 11 not alter heterochromatin signals, although it should be noted cotinine, Se, and S-421. These conclusions were formed based that the sensitivity of this heterochromatin configuration- on the epigenetic evaluation of heterochromatin marks and based screening method might not be sufficient for detection DNA methylation status, as well as the developmental poten- of slight alterations of some single epigenetic modification. tial of EB formation and neural differentiation. Combined In previous reports, cytotoxic analyses revealed that the exposure to these epimutagens at low concentrations caused sensitivities of rodents and humans to chemicals including long-lasting effects, suggesting that epigenetic alterations organophosphates and 2,3,7,8-tetrachlorodibenzo-p-dioxin exert long-term effects that result in aberrant tissue develop- differed; specifically, human cells were less sensitive to these ment and that epimutagens are harmful during human fetal chemicals than rodents [14, 15]. The fetal environment con- development. tains various chemicals, and it is important to interpret the data regarding chemical sensitivity in terms of epigenetic Conflict of Interests influence, as chemical sensitivities differ depending on the animal species. In addition, our previous data on mESCs The authors have no competing financial interests. showed both hypo- and hypermethylation by Se or Hg exposure. However, several gene loci that became hypermethylated by combinatorial exposure of the five-chemical Acknowledgments mixture could be identified in the present study using hiPSCs. The authors thank the RIKEN BioResource Center for pro- Although the mechanism underlying these changes remains viding hiPSCs (201B7) and PA6. A Health Science Research to be elucidated, the combination of the five-chemical mix- Grant from the Ministry of Health, Labor and Welfare, Japan ture is suggested to result in the abnormal upregulation of the (to Tsunehisa Makino), and a Grant-in-Aid for Scientific DNA-methylating system, including DNA methyltransferase Research (S) from the Ministry of Education, Culture, Sports, enzymes. Science and Technology of Japan (21221008 to Kunio Shiota) Our previous data indicated that the effect of 5-aza-dC, on supported this work. DNA demethylation of gene loci in particular, was not dose- dependent [32] and that DNA demethylation of tissue/cell- type-specific gene loci was caused only by relatively low- References dose treatment with 5-aza-dC and not by high-dose treat- [1] J. L. Golob, S. L. Paige, V. Muskheli, L. Pabon, and C. E. Murry, ment. This suggests that the epigenetic changes induced by “Chromatin remodeling during mouse and human embryonic chemicals are not always dose-dependent; this finding may stem cell differentiation,” Developmental Dynamics,vol.237,no. be applicable to PFOA. Although the epigenetic effects of 5, pp. 1389–1398, 2008. PFOA remain to be elucidated, our data suggest that the [2] K. Ikegami, J. Ohgane, S. Tanaka, S. Yagi, and K. Shiota, heterochromatin configuration of hiPSCs was affected by “Interplay between DNA methylation, histone modification and PFOA only when its concentration was within a certain range, chromatin remodeling in stem cells and during development,” which includes the 10 ppb concentration used in the present The International Journal of Developmental Biology,vol.53,no. study. 2-3, pp. 203–214, 2009. Several studies have reported that the five chemicals ana- [3]J.D.Lieb,S.Beck,M.L.Bulyketal.,“Applyingwhole-genome lyzed in the present study were detectable in umbilical cord studies of epigenetic regulation to study human disease,” Cyto- and/or maternal serum samples at very low concentrations genetic and Genome Research,vol.114,no.1,pp.1–15,2006. [9, 12, 33–36]. Recent clinical studies have suggested that fetal [4] K. Shiota, “DNA methylation profiles of CpG islands for cellular exposure to heavy metals, cigarette smoke, or pesticides could differentiation and development in mammals,” Cytogenetic and increase the risk of abnormal neurodevelopment, behavioral Genome Research, vol. 105, no. 2–4, pp. 325–334, 2004. problems, obesity, and metabolic disorders during childhood [5]M.O.Nakanishi,K.Hayakawa,K.Nakabayashi,K.Hata,K. [24, 37–40]. This led us to hypothesize that fetal exposure to Shiota, and S. Tanaka, “Trophoblast-specific DNA methylation environmental chemicals affects the growth and development occurs after the segregation of the trophectoderm and inner cell of children after birth. The present study suggested that mass in the mouse periimplantation embryo,” Epigenetics,vol.7, no. 2, pp. 173–182, 2012. the combined exposure to serum concentrations of the five chemicals disturbed the heterochromatin configuration of [6]S.Sato,S.Yagi,Y.Araietal.,“Genome-wideDNAmethyla- tion profile of tissue-dependent and differentially methylated pericentric regions stained with anti-HP1𝛼 antibodies and regions (T-DMRs) residing in mouse pluripotent stem cells,” the DNA methylation patterns of several genes in hiPSCs. Genes to Cells,vol.15,no.6,pp.607–618,2010. Moreover, exposure to the epimutagen mixture only prior to [7]K.Shiota,Y.Kogo,J.Ohganeetal.,“EpigeneticmarksbyDNA inducing hiPSC differentiation affected cell morphology and methylation specific to stem, germ and somatic cells in mice,” gene expression patterns in differentiated EBs or neuronal Genes to Cells,vol.7,no.9,pp.961–969,2002. cells, suggesting that the chemicals have a long-term effect [8] S. Yagi, K. Hirabayashi, S. Sato et al., “DNA methylation profile on cellular differentiation. Taken together, these data suggest of tissue-dependent and differentially methylated regions (T- that fetal exposure to environmental chemicals might cause DMRs) in mouse promoter regions demonstrating tissue- a later onset of developmental disorders after birth by specific gene expression,” Genome Research,vol.18,no.12,pp. disturbing the epigenetic memory. 1969–1978, 2008. Inconclusion,weobservedthathiPSCsweresensitive [9]I.Al-Saleh,N.Shinwari,A.Mashhour,G.E.D.Mohamed, to an epimutagenic chemical mixture consisting of DEP, Hg, and A. Rabah, “Heavy metals (lead, cadmium and mercury) 12 BioMed Research International

in maternal, cord blood and placenta of healthy women,” [24] R. Harari, J. Julvez, K. Murata et al., “Neurobehavioral deficits International Journal of Hygiene and Environmental Health,vol. and increased blood pressure in school-age children prenatally 214,no.2,pp.79–101,2011. exposed to pesticides,” Environmental Health Perspectives,vol. [10]Y.Iwasaki,M.Goto,K.Mochizukietal.,“Development 118, no. 6, pp. 890–896, 2010. and validation of a hydrophilic interaction chromatography- [25]A.Kortenkamp,M.Faust,M.Scholze,andT.Backhaus,“Low- tandem mass spectrometry for quantification of nicotine and level exposure to multiple chemicals: reason for human health its metabolites in human maternal and cord sera,” Biomedical concerns?” Environmental Health Perspectives, vol. 115, supple- Chromatography,vol.25,no.4,pp.503–510,2011. ment1,pp.106–114,2007. [11] C. Mori, M. Komiyama, T. Adachi et al., “Application of [26] O. Boucher, M. J. Burden, G. Muckle et al., “Response inhibi- toxicogenomic analysis to risk assessment of delayed long-term tion and error monitoring during a visual Go/No-Go task in effects of multiple chemicals, including endocrine disruptors in inuit children exposed to lead, polychlorinated biphenyls, and human fetuses,” EHP Toxicogenomics, vol. 111, no. 1T, pp. 7–13, methylmercury,” Environmental Health Perspectives,vol.120, 2003. no. 4, pp. 608–615, 2012. [12] Y. Arai, J. Ohgane, S. Yagi et al., “Epigenetic assessment of [27] M. R. Thompson and K. Boekelheide, “Multiple environmen- environmental chemicals detected in maternal peripheral and tal chemical exposures to lead, mercury and polychlorinated cord blood samples,” Journal of Reproduction and Development, biphenyls among childbearing-aged women (NHANES 1999– vol.57,no.4,pp.507–517,2011. 2004): body burden and risk factors,” Environmental Research, [13] M.Iwatani,K.Ikegami,Y.Kremenskaetal.,“Dimethylsulfoxide vol.121,pp.23–30,2013. hasanimpactonepigeneticprofileinmouseembryoidbody,” [28] G. M. Benke and S. D. Murphy, “The influence of age on the Stem Cells, vol. 24, no. 11, pp. 2549–2556, 2006. toxicity and metabolism of methyl parathion and parathion in male and female rats,” Toxicology and Applied Pharmacology, [14] R. A. Budinsky, E. L. LeCluyse, S. S. Ferguson, J. C. Rowlands, vol. 31, no. 2, pp. 254–269, 1975. andT.Simon,“HumanandratprimaryhepatocyteCYP1A1 and 1A2 induction with 2,3,7,8-tetrachlorodibenzo-p- [29] T. K. Chakraborti, J. D. Farrar, and C. N. Pope, “Compar- dioxin, 2,3,7,8-tetrachlorodibenzofuran, and 2,3,4,7,8- ative neurochemical and neurobehavioral effects of repeated pentachlorodibenzofuran,” Toxicological Sciences,vol.118, chlorpyrifos exposures in young and adult rats,” Pharmacology no. 1, pp. 224–235, 2010. Biochemistry and Behavior,vol.46,no.1,pp.219–224,1993. [15] B. Veronesi and M. Ehrich, “Differential cytotoxic sensitivity [30] V. Rauh, S. Arunajadai, M. Horton et al., “Seven-year neurode- in mouse and human cell lines exposed to organophosphate velopmental scores and prenatal exposure to chlorpyrifos, a insecticides,” Toxicology and Applied Pharmacology,vol.120,no. common agricultural pesticide,” EnvironmentalHealthPerspec- 2, pp. 240–246, 1993. tives,vol.119,no.8,pp.1196–1201,2011. [16] K. Takahashi, K. Tanabe, M. Ohnuki et al., “Induction of [31] C. Wohlfahrt-Veje, K. M. Main, I. M. Schmidt et al., “Lower pluripotent stem cells from adult human fibroblasts by defined birth weight and increased body fat at school age in children factors,” Cell,vol.131,no.5,pp.861–872,2007. prenatally exposed to modern pesticides: a prospective study,” Environmental Health,vol.10,no.1,article79,2011. [17] Z. Xiong and P. W. Laird, “COBRA: a sensitive and quantitative [32] H. W. Lim, M. Iwatani, N. Hattori, S. Tanaka, S. Yagi, and K. DNA methylation assay,” Nucleic Acids Research,vol.25,no.12, Shiota, “Resistance to 5-aza-2’-deoxycytidine in genic regions pp. 2532–2534, 1997. compared to non-genic repetitive sequences,” Journal of Repro- [18] B. Lehnertz, Y. Ueda, A. A. H. A. Derijck et al., “Suv39h- duction and Development,vol.56,no.1,pp.86–93,2010. mediated histone H3 lysine 9 methylation directs DNA methy- [33] T. Berman, D. Hochner-Celnikier, D. B. Barr et al., “Pesticide lation to major satellite repeats at pericentric heterochromatin,” exposure among pregnant women in Jerusalem, Israel: results Current Biology,vol.13,no.14,pp.1192–1200,2003. of a pilot study,” Environment International,vol.37,no.1,pp. [19] T. Hayakawa, T. Haraguchi, H. Masumoto, and Y. Hiraoka, 198–203, 2011. “Cell cycle behavior of human HP1 subtypes: distinct molecular [34] A. Bradman, R. Castorina, D. B. Barr et al., “Determinants of domains of HP1 are required for their centromeric localization organophosphorus pesticide urinary metabolite levels in young during interphase and metaphase,” Journal of Cell Science,vol. children living in an agricultural community,” International 116, part 16, pp. 3327–3338, 2003. Journal of Environmental Research and Public Health,vol.8,no. [20] M. Narita, S. Nunez,˜ E. Heard et al., “Rb-mediated heterochro- 4, pp. 1061–1083, 2011. matin formation and silencing of E2F target genes during [35] R. P. Parajuli, T. Fujiwara, M. Umezaki, H. Furusawa, P. H. Ser, cellular senescence,” Cell,vol.113,no.6,pp.703–716,2003. and C. Watanabe, “Cord blood levels of toxic and essential trace [21] C. Ginzkey, G. Friehs, C. Koehler, S. Hackenberg, R. Hagen, elements and their determinants in the Terai region of Nepal: a and N. H. Kleinsasser, “Assessment of nicotine-induced DNA birth cohort study,” Biological Trace Element Research,vol.147, damage in a genotoxicological test battery,” Mutation Research: no. 1–3, pp. 75–83, 2012. Genetic Toxicology and Environmental Mutagenesis,vol.751,no. [36] M. Sakamoto, H. M. Chan, J. L. Domingo, M. Kubota, and K. 1,pp.34–39,2013. Murata, “Changes in body burden of mercury, lead, arsenic, [22] S. Gateva, G. Jovtchev, and M. Stergios, “Cytotoxic and clasto- cadmium and selenium in infants during early lactation in genic activity of CdCl2 in human lymphocytes from different comparison with placental transfer,” Ecotoxicology and Environ- donors,” Environmental Toxicology and Pharmacology,vol.36, mental Safety,vol.84,pp.179–184,2012. no. 1, pp. 223–230, 2013. [37] M. Behl, D. Rao, K. Aagaard et al., “Evaluation of the association [23]B.Eskenazi,A.R.Marks,A.Bradmanetal.,“Organophosphate between maternal smoking, childhood obesity, and metabolic pesticide exposure and neurodevelopment in young Mexican– disorders: a national toxicology program workshop review,” American children,” Environmental Health Perspectives, vol. 115, Environmental Health Perspectives,vol.121,no.2,pp.170–180, no. 5, pp. 792–798, 2007. 2013. BioMed Research International 13

[38] O. Boucher, S. W. Jacobson, P. Plusquellec et al., “Prena- tal methylmercury, postnatal lead exposure, and evidence of attention deficit/hyperactivity disorder among Inuit children in Arctic Quebec,”´ EnvironmentalHealthPerspectives,vol.120,no. 10, pp. 1456–1461, 2012. [39] C.Desrosiers,O.Boucher,N.Forget-Duboisetal.,“Associations between prenatal cigarette smoke exposure and externalized behaviors at school age among Inuit children exposed to environmental contaminants,” Neurotoxicology and Teratology, vol. 39, pp. 84–90, 2013. [40] S. Llop, J. Julvez, A. Fernandez-Somoano et al., “Prenatal and postnatal insecticide use and infant neuropsychological development in a multicenter birth cohort study,” Environment International,vol.59,pp.175–182,2013. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 274852, 10 pages http://dx.doi.org/10.1155/2015/274852

Research Article Noncoding RNAs: Possible Players in the Development of Fluorosis

Atul P. Daiwile,1 Saravanadevi Sivanesan,1 Alberto Izzotti,2,3 Amit Bafana,1 Pravin K. Naoghare,1 Patrizio Arrigo,4 Hemant J. Purohit,5 Devendra Parmar,6 and Krishnamurthi Kannan1

1 Environmental Health Division, National Environmental Engineering Research Institute (NEERI), CSIR, Nagpur 440020, India 2DepartmentofHealthSciences,UniversityofGenoa,No.5,16126Genoa,Italy 3IRCCS AOU San Martino IST, No. 10, 16132 Genoa, Italy 4CNR Institute for Macromolecular Studies, No. 6, 16149 Genoa, Italy 5Environmental Genomics Division, National Environmental Engineering Research Institute (NEERI), CSIR, Nagpur 440020, India 6Developmental Toxicology, Indian Institute of Toxicology Research, Lucknow 226001, India

Correspondence should be addressed to Pravin K. Naoghare; [email protected]

Received 10 October 2014; Accepted 4 February 2015

Academic Editor: Luciana dos Reis Vasques

Copyright © 2015 Atul P. Daiwile et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Fluorosis is caused by excess of fluoride intake over a long period of time. Aberrant change in the Runt-related transcription factor 2 (RUNX2) mediated signaling cascade is one of the decisive steps during the pathogenesis of fluorosis. Up to date, role of fluoride on the epigenetic alterations is not studied. In the present study, global expression profiling of short noncoding RNAs, in particular miRNAs and snoRNAs, was carried out in sodium fluoride (NaF) treated human osteosarcoma (HOS) cells to understand their possible role in the development of fluorosis. qPCR and in silico hybridization revealed that miR-124 and miR-155 can be directly involved in the transcriptional regulation of Runt-related transcription factor 2 (RUNX2) and receptor activator of nuclear factor 𝜅-B ligand (RANKL) genes. Compared to control, C/D box analysis revealed marked elevation in the number of UG dinucleotides and D-box sequences in NaF exposed HOS cells. Herein, we report miR-124 and miR-155 as the new possible players involved in the development of fluorosis. We show that the alterations in UG dinucleotides and D-box sequences of snoRNAs could be dueto NaF exposure.

1. Introduction Available drug therapy against fluorosis includes supplementations with aluminum, magnesium, calcium, amino acid, The global prevalence of fluorosis is a matter of great and vitamins (C, D, and E) [1, 3]. These supplementations concern. Chronic exposure of high concentration of fluoride can stimulate calcium bone mineralization, activation of (>1.5 mg/L) may result in skeletal and dental deformities [1]. phagocytes/osteoclasts and antioxidant enzymes to combat At present, 25 countries around the world have endemic to oxidative stress induced by fluoride2 [ , 3]. However, the fluorosis [1]. In India, the health of approximately 62 million therapeutic effects incurred due to these supplements are people is at risk due to high concentration of fluoride in unstable and reversible [2, 3]. drinking water/diet and so forth [1]. Preventive measures No concurrent therapies are available against fluorosis against fluorosis rely on the intake of low fluoride diet/water due to limited information on the expression of aberrant to reduce the ingestion of fluoride inside the body [1, 2]. posttranscriptional cascades involved in the development of However, these preventive measures were not successful in skeletal and dental deformities [3]. Intake of high concen- most of the underdeveloped and developing counties due to tration of fluoride is reported to impede skeletal forma- high cost associated with the defluoridation techniques1 [ , 2]. tion and remodeling processes due to failure in the signal 2 BioMed Research International transduction between the osteoblast and osteoclast cells by NaF. Our results suggest the possible involvement of miR- through receptor activator of nuclear factor 𝜅-B ligand 124 and miR-155 in the development of fluorosis. In addition, (RANKL), receptor activator of nuclear factor 𝜅-B (RANK), high sensitivity of NaF to D-box of snoRNAs suggests the and osteoprotegerin (OPG) system [4]. RANKL/RANK posttranscriptional regulation of hypo/hypermethylated and signaling positively regulates osteoclast differentiation and pseudourydilated rRNAs that could be involved in bone promotes bone remodeling, whereas OPG/RANKL signal- formation and remodeling processes. ing protects the skeleton from the bone [4–7]. Aberrant variations in the RANKL/RANK/OPG system limits the 2. Material and Methods availability of free calcium required for osteocalcin (bone gamma-carboxyglutamic acid-containing protein; BGLAP) 2.1. Cell Lines. Human osteosarcoma (HOS) cell line was mediated bone mineralization, leading to the excess of purchased from National Centre for Cell Science (NCCS) fluoride deposition inside the bones [6, 8]. Pune, India. Cells were resuspended in Minimum Essential The research on posttranscriptional gene-expression reg- Medium (MEM) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, 50 𝜇g/mL streptomycin, and 25 𝜇g/mL ulation raveled the pivotal role of short noncoding RNAs ∘ (miRNA, snoRNA, etc.) in the response of cell against envi- Fungizone. Cells were maintained at 37 Cinahumidified ronmental factors. In particular the miRNAs are the deeply atmosphere with 5% CO2. investigated class of noncoding RNAs. But, limited reports entail that miRNAs can regulate bone formation and remod- 2.2. Cell Viability Assay. Cell viability was assessed using 3- eling processes [9–11]. miR-204 and miR-211 were reported as (4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide the negative regulators of the gene involved in osteoblast acti- (MTT) assay. Cells (10,000 cells/well) were seeded in a vation, RUNX2 [9]. Similarly, the RUNX2-targeting miRNAs 96-well plate for 24h. After 24h of seeding, HOS cells were reported to impede osteoblast differentiation [10, 11]. were exposed to different concentrations of NaF (10 mg/L Thus, it is likely that miRNAs can be involved in the posttran- to 250 mg/L) and further incubated for 24 h. After 24 h of exposure, MTT solution (5 mg/L in PBS) (Invitrogen, USA) scriptional regulation of other important signaling cascades ∘ (e.g., RANKL/RANK/OPG system and BGLAP) required for was added and cells were incubated at 37 Cfor3h.Media bone formation and remodeling processes. Less information were discarded and cells were resuspended in 200 𝜇Lof is available on the effects of environmental factors on short DMSO (Sigma Aldrich, USA) and analyzed at 565 nm. Lethal noncoding RNAs, the small nucleolar RNAs (snoRNAs) concentration 50 (LC50) of NaF was calculated using the dose [12]. Small nucleolar RNAs (snoRNAs) are a group of small response curve analysis. RNA that principally regulate chemical alterations in other noncoding RNAs such as transfer RNAs (tRNAs), ribosomal 2.3. NaF Treatment. HOS cells were exposed to sublethal RNAs (rRNAs), and small nuclear RNAs (snRNAs) [13]. concentrations (8 mg/L (1/5th) and 20 mg/L (1/2nd) of 24 h snoRNAs are divided into two main classes depending upon LC50) of NaF (Sigma Aldrich, USA) for 30 days; subsequent the presence of C/D box and H/ACA box sequences [13]. controls were maintained. 50,000 cells/mL were seeded in cell 󸀠 C/D box snoRNAs are primarily involved in 2 -O-ribose flask,followedbythetreatmentofsodiumfluoride.Cellswere methylation, whereas H/ACA box snoRNAs are involved in subcultured after 3 days and NaF treatment was given after the pseudouridylation of noncoding RNAs [13]. every passage and culture media change. Finally, cells were Deleterious effects of fluoride on the expression levels of harvestedinTrizolreagent(Invitrogen,USA). noncoding RNAs have not been studied. Herein, for the first time, we report a global expression profiling of noncoding 2.4. miRNA Microarray. Total RNA was isolated using Abso- RNAs (miRNAs and small nucleolar RNAs (snoRNAs)) in lutely RNA Kit (Stratagene, USA), according to the manufacturer’s recommendations. Quality of RNA was checked using sodium fluoride (NaF) exposed human osteosarcoma (HOS) Agilent Bioanalyzer 2100 (Agilent Technologies Inc., USA). cells. Compared to control, analysis using Affymetrix miRNA Affymetrix miRNA 3.0 array platform was used to elucidate 3.0 array platform (5639 human probes) revealed differen- the variations in the expression profiles of 5639 human tial expressions of signature miRNAs and snoRNAs after noncoding mature RNA probes after NaF treatment to HOS sodium fluoride treatment (8 mg/L and 20 mg/L) to HOS ∘ cells. Hybridization was carried out at 48 Cfor16hat60rpm. cells. Further validation using real time PCR illustrated the Probe intensities were measured using GeneChip Scanner upregulation of miR-124 and miR-155, whereas the expression 3000 7G (Affymetrix, USA). All the original microarray data levels of RANKL, BGLAP, and RUNX2 genes were decreased (CEL files) for the experiments was preprocessed using RMA in NaF treated cells. OPG, a soluble decoy of RANKL, was (Robust Multichip Average) algorithm that consists of three upregulated in NaF treated cells. Data obtained from in steps: a background adjustment, quantile normalization, and silico hybridization analysis revealed that miR-124 and miR- summarization. The raw data normalization has been per- 155 could be directly involved in the posttranscriptional formed by selecting “organism human” in Expression Con- regulation of RUNX2 and RANKL genes. Compared to the soletool.Allaboveproceduresweredonebyselectingdefault terminal C/D box, microarray analysis revealed more marked RMA algorithm, data adjustment, and background correc- effect of fluoride exposure on the expression levels of internal tion (GABG) in Affymetrix Expression Console 1.2.1.20. Box C/D boxes of snoRNAs. In addition, D-box, a pivotal element whisker plot (see Supplementary data; S1 available online in methylation process, was found to be strongly influenced at http://dx.doi.org/10.1155/2015/274852)andqualitycontrol BioMed Research International 3

Table 1: qPCR primers.

󸀠 󸀠 Sr. number Gene name Primer Sequence (5 -3 ) miRNAs 1 miRNA-155 Forward primer TGTTAATGCTAATATGTAGGAG 2 miRNA-124 Forward primer GGCATTCACCGCGTGCCTTA Forward primer AGTGATGATGACCCCAGGTAACTC 3RUN48 Reverse primer CTGCGGTGATGGCATCAG Genes Forward primer GGCAGGCACAGTCTTCCC 4RUNX2 Reverse primer GGCCCAGTTCTGAAGCACC Forwardprimer TCGTTGGATCACAGCACATCA 5RANKL Reverse primer TATGGGAACCAGATGGGATGTC Forward primer ATGAGAGCCCTCACACTCCTC 6BGLAP Reverse primer GCCGTAGAAGCGCCGATAGGC Forward primer CTGGAACCCCAGAGCGAAAT 7OPG Reverse primer GCCTCCTCACACAGGGTAAC Forward primer GCACCACGTCCAATGACAT 5RPII(internalcontrol) Reverse primer GTGCGGCTGGTTCCATAA Forward primer ACGAAGTGTTGGATATAAGC 6HPRT(internalcontrol) Reverse primer ATAATTTTACTGGCGATGTC analyses were carried out using Expression Console software. SYBRGreenERqPCRSuperMixcontaininguniversalreverse Normalized intensity files were exported from Expression primers (Invitrogen, USA) and forward customized primers Consoles tool in txt format. In next step, all normalized exper- of miR-124/miR-155 (Table 1). RUN48 was used as internal imental data were imported in GeneSpring GX 12.5 software controls as described earlier [14]. Fold change was calculated ∧ for the differential miRNA expression, fold change analysis, using ddCT method (2 -ddCt). and cluster analysis. Statistical analysis was performed for For validation of genes, total RNA was isolated using Tri- the identification of differentially expressed miRNA from zol reagent. 2 𝜇g of total RNA was converted to cDNA using two groups (Dose 1/5 (8 mg/L) versus control and Dose 1/2 High Capacity Reverse Transcription kit (Applied Biosys- (20 mg/L) versus control) (Supplementary Table; S2). One- tems, USA). qPCR primers for RUNX2, RANKL, BGLAP, way ANOVA method was applied for assessing the statis- and OPG were used as described earlier (Table 1)[15–17]. tically significant differentially expressed miRNAs for two RNA polymerase II (RPII) and hypoxanthine phosphori- different treated experiment groups with reference to control. bosyl transferase (HPRT) were used as internal controls as The 𝑃 valuecut-off0.05wasconsideredstatisticallysignifi- described earlier [18]. Relative quantification was performed cant. miRNA microarray data has been submitted in NCBI on Applied Biosystems 7300 real time PCR system using GEO, accession number GSE57550. EXPRESS SYBR GreenER qPCR SuperMix. All the primers were validated before relative quantification. Fold change was ∧ 2.5. In Silico Hybridization Analysis. mirTarBase and calculated using ddCT method (2 -ddCt). miRANDA software were used to identify functionally validated targets of miR-124 and miR-155. Furthermore, differ- 2.7. C/D Box Analysis of snoRNA. Differentially expressed entially expressed miRNAs were screened for their possible snoRNAs obtained by microarray analysis were further ana- putative targets using in silico hybridization analysis. Possible lyzed for the expression levels of C/D boxes using an in silico binding sites of miR-124 and miR-155 with the RUNX2 and approach. Probes mapped on the array showed high length RANKL were identified. variability encompassed in the range from 48 to 237 bp. This great variability suggested partitioning the C/D box data set on the basis of their length. Box plot (Supplementary data; 2.6. Real Time qPCR. Screened miRNAs and genes were val- S3A) shows a statistical analysis of length outliers of C/D box idated using quantitative qPCR. miRNAs were isolated using snoRNAs. The length outliers in snoRNA were determined by mirPremier microRNA Isolation Kit (Sigma Aldrich, USA). interquartile range method, described below: NaF exposed cells were lyzed using microRNA lysis buffer = 𝑞̂ − 1.5 , supplemented with RNA binding solution. miRNAs were Lower Limit 0.25 IQR isolated as per the instructions given in the manufactures (1) Upper Limit = 𝑞̂0.75 + 1.5 IQR. protocol. 500 ng of miRNA was converted to cDNA using NCode VILO miRNA cDNA Synthesis Kit (Invitrogen, USA). Lower and upper limits set for outlier detection were LL = Relative quantification of miRNA was performed on Applied 45.87 and UL = 112.87, respectively (please refer to supple- Biosystems 7300 real time PCR system using EXPRESS mentary data; S3 for the details). 4 BioMed Research International

100 level, fold change in the expression levels of the RUNX2 was 90 − ± 80 observed to be 3.25 ( 0.27), whereas at higher dose level 70 thefoldchangewasobservedtobe−2.15 (±0.13) (Figure 5). 60 Validation of other osteoblastic lineage markers (RANKL, 50 40 BGLAP, and OPG) through qPCR suggested their possible

Viability (%) Viability 30 correlation with the expression levels of miR-124 and miR- 20 155. Expression levels of the RANKL and BGLAP were 10 𝑃 < 0.05 0 found to be significantly decreased ( ), whereas the expression level of the OPG was found to be significantly 10 20 80 40 60 100 125 200 225 150 175 250 increased (𝑃 < 0.05) in NaF treated HOS cells. At lower

Control dose level, fold change in the expression levels of the RANKL, Concentration (mg/L) BGLAP, and OPG was observed to be −2.85 (±0.18), −2.1 (±0.20), and 2 (±0.24), respectively. At higher dose level, fold Figure 1: LC50 (24 h) of NaF against HOS cells. Graph represents the cytotoxicity caused by sodium fluoride after 24 h of exposure. change in the expression levels of the RANKL, BGLAP, and OPG was observed to be −2.04 (±0.16), −1.5 (±0.14), and 1.5 (±0.02), respectively (Figure 5). Effects of sublethal concentrations of NaF on the expres- 3. Results sion profiles of 347 C/D box snoRNAs were analyzed using a quantile approach. C-box (UGAUGA) is typically located 󸀠 Cell viability was found to be decreased with increasing near the 5 terminalofsnoRNA,whereastheD-box(CUGA) 󸀠 concentrations of NaF. LC50 (24 h) concentration of NaF is located in proximity of 3 terminal.Theidentification againstHOScellswasfoundtobe40mg/L(Figure 1). Based of C/D box targets is quite different from miRNAs target on these results, sublethal concentrations (8 mg/L (1/5th) and screening. Identification of C/D box targets relies on the elu- 20 mg/L (1/2nd) of 24 h LC50) of NaF were chosen for further cidation of putative methylation or pseudouridylation sites experiments. at chromosome level. Total snoRNA probes with different The microarray experiment (Figure 2) was carried out lengths(48–237bp)wereanalyzedusinganoutlieranalysis using two different test concentrations, that is, 8 mg/L and (Supplementary data; S3A). The upper length threshold of 20 mg/L of NaF. Out of 5,639 probes, each test concentration snoRNA probe was fixed to l ≥ 112.87.Wehaveseparately displayed 128 differentially expressed miRNAs𝑃 ( < 0.05). analyzed the statistical outliers in order to investigate their Supplementary Table S2 displays representative miRNA sig- putative functions. Outlier analysis revealed 15 snoRNAs with natures which were differentially expressed in both the test different lengths ranging from (114–237 bps) (Supplementary concentrations (𝑃 < 0.05). Further screening with one and data; S3B). In order to check the homogeneity of the longer half fold change cut (FC ≥ 1.5) showed 82 and 62 differentially C/D box snoRNAs, the average two-parameter kimura’s expressed miRNAs in 8 mg/L and 20 mg/L test concentrations distance was estimated for the putative outliers set (⟨K2p⟩ = of NaF, respectively. In it, 56 and 21 miRNAs were found to 0.184) (Supplementary data; S3C). Further analysis revealed beupregulatedin8mg/Land20mg/Loftestconcentrations, differential distribution of C-box and D-box (Supplementary respectively. data; S3D). Each snoRNA sequence, except U22 and U97, mirTarBase and miRANDA software were used to under- was found to contain one C-box, whereas numbers of D-box stand the possible involvement of miRNAs in the regulation in the given set were found vary between 0 and 5. Majority of osteoclastogenic signaling pathway. In silico hybridization of the length outliers (l > 112.87) of snoRNAs with two or revealed that seed regions miR-124 and miR-155 can bind with more D-box sequences were found to be downregulated in the untranslated regions (UTRs) of the RUNX2 and RANKL NaF treated HOS cells compared to control. On the contrary, (Figure 3).OutofallthepossiblealignmentsofmiR-124/miR- snoRNAs with less than two D-box sequences were found to 155 with the RUNX2/RANKL, pairs depicting hybridization beupregulatedinNaFtreatedHOScellscomparedtocontrol with the conserved domains, good mirSVR and PhastCons (data not shown). Compared to CG dinucleotides, higher scores were selected using miRANDA software. Validation of numbers of UG dinucleotides were found in the sequences of miRNAs using qPCR demonstrated that sublethal concentra- all outliers. Single channel microarray exploratory protocol tions of NaF can significantly𝑃 ( < 0.05)upregulatemiR-124 analysis revealed up- and downregulation of C/D snoRNA and miR-155 expressions in HOS cells (Figure 4). Compared dataset, including length outliers and without outliers (Sup- to the higher dose level (20 mg/L) of NaF, expression levels plementary data; S3E). 14qII-13 snoRNA was found to be of miRNAs were more elevated in low dose level (8 mg/L). At downregulated at both the sublethal concentrations of NaF lowerdoselevel,foldchangeintheexpressionlevelsofmiR- when analyzed with and without length outliers. 14qII-9, 124andmiR-155wasobservedtobe3.24(±0.13) and 4.03 HBII-85-1,U42B,U53,U73b,andU84snoRNAswerefound (±0.3), respectively. At higher dose level, fold change in the to be upregulated at both the sublethal concentrations of expression levels of miR-124 and miR-155 was observed to be NaF when analyzed with length outliers. In the dataset 1.6 (±0.19) and 1.86 (±0.24), respectively. cleaned after outliers, the analyzed expression level of U58A Further validation through qPCR revealed that upregu- snoRNA was found to be downregulated at both the sublethal lation of miR-124 and miR-155 led to the downregulation of concentrations of NaF, whereas at the same condition the Runx2 (𝑃 < 0.05) in NaF treated HOS cells. At lower dose expression levels of 14qII-9, HBII-336, HBII-52-7,U36B, U43, BioMed Research International 5 NEE NEE NEE NEE 1 2 1 2

NEERI NEERI 1 2 15 15 12 12 C C

Figure 2: A Screen shot of Hcl clustering expression image of miRNAs. Red color shows overexpressed miRNAs (>0) and blue color shows underexpressed miRNAs (<0). The hcl heat map image has been generated on the basis of log2 normalized intensity value. Here, C1andC2 represent controls. 1 5and1 2 represent 8 mg/L and 20 mg/L NaF concentrations, respectively.

and U46 snoRNAs were found to be upregulated at both to find out epigenetic modifications in miRNAs and C/D the sublethal concentrations of NaF. Rest of the snoRNAs box snoRNAs that can regulate the key genes involved in the (Supplementary data; S3E) were specifically expressed at development of fluorosis. either of the sublethal doses of NaF. Noncoding RNAs are known to regulate majority of the genes involved in the cellular pathways [9, 13, 14, 19]. 4. Discussion Different xenobiotic compound may have different effects on the expression profiles of noncoding RNAs19 [ ]. In the At present, no successful therapeutics is available to treat flu- present study, sublethal concentrations of NaF were found orosis [2]. Available therapies against fluorosis mostly include to induce alterations in the expression profiles of noncoding supplementations (cofactor of enzymes, amino acids, and RNAs (Figure 2). Functional targets of selected miRNAs antioxidants) that can stimulate osteoblast/osteoclast medi- (miR-124 and miR-155) were identified using mirTarBase and ated bone formation and remodeling processes, macrophage in silico hybridization analysis. Seed regions of the miR-124 activation, and inhibition of reactive oxygen species (ROS) and miR-155 displayed in silico hybridization with untrans- [2, 3]. However, the therapeutic effects incurred due to these lated regions (UTRs) of the RUNX2. miR-124 displayed in 󸀠 therapies are not consistent. Reversals of the therapeutic silico hybridization with 3 untranslated regions of RANKL effects are often seen once the medication is discontinued (Figure 3). Further validation through quantitative RT-PCR [2]. It is because these medications do not have control showed that the expression profiles of the miR-124 and miR- on the epigenetic modifications that are responsible for 155 were inversely proportional to the expression profiles of posttranscriptional regulation of genes involved in the osteo- the RUNX2 and RANKL genes. Thus, it can be inferred that clastogenic pathway. Therefore, aim of the present study was miR-124/miR-155 could be involved in the posttranscriptional 6 BioMed Research International R A Y C W C 󳰀 G U R C S 3 U W C C A U C A A G A A A C C U U G U U R U C miR-155 U G G G G A U C G A A U U A G U A U K C W G U U 󳰀 A S G 5 C R U Seed region Y

c hsa-miR-155/RUNX2 alignment

󳰀 󳰀 3 uggggaUAGUGCUAAUCGUAAUu 5 hsa-miR-155 mirSVR score: −0.1710 󳰀 󳰀 3194: 5 cauggaAACACGA--AGCAUUAg 3 RUNX2 PhastCons score: 0.5879

C B B c hsa-miR-124/RUNX2 alignment Y S Y R Y R 󳰀 󳰀 mirSVR score: −0.0118 B G 3 ccguaaguggcgcACGGAAu 5 hsa-miR-124 K M 󳰀 󳰀 C G 715: 5 caucccuaaaaccUGCCUUc 3 RUNX2 PhastCons score: 0.6612 U A C G Y A Y A C C G C c hsa-miR-124/RUNX2 alignment U G G U U A U A C G 3󳰀 5󳰀 hsa-miR-124 mirSVR score: −0.0205 A U ccguaaGUGGCG-CACGGAAu C G 󳰀 󳰀 W G 1167: 5 3 RUNX2 PhastCons score: 0.7163 G C cucuaaCACAGCUUUGCCUUc Y G G C G A R C G C C G c hsa-miR-124/TNFSF11 alignment (RANKL) U A U A Seed region G U A W U U A A 3󳰀 5󳰀 hsa-miR-124 mirSVR score: −0.0976 U C ccguaaguggcgcACGGAAu A A 󳰀 󳰀 W U 754: 5 aaauuaaaauggaUGCCUUg 3 TNFSF11 PhastCons score: 0.6817 a A U C K U C miR-124

Figure 3: In silico hybridization analysis of mIR-124 and mIR-155 with the RUNX2 and RANKL.

regulation of the RUNX2 and RANKL in NaF exposed HOS Significant decrease in expression levels of the RANKL cells. was observed in NaF treated HOS cells (Figure 5). Simi- Yinetal.(2010)studiedroleofmiR-155inregulation lar results were reported in Xenopus laevis and B6 mice of BMP signaling pathway and demonstrated that miR- when exposed to sodium fluoride5 [ , 22, 23]. RANKL, a 󸀠 155 binds to 3 UTRs of target mRNAs and regulates the transmembrane protein, expresses on the osteoblast cells expression of key genes involved in BMP signaling pathway [6]. RANKL binds to its receptor (RANK) present on the (RUNX2,SMAD1,SMAD5,HIVEP2,CEBPB,andMYO10) progenitor osteoclast and stimulates osteoclastogenesis [6]. [10]. Further studies on knockout model demonstrated RANK-RANKL binding activate downstream target genes that miR-155 can directly inhibit BMP signalling pathway (TRAF6, NF-𝜅𝛽, and C-fos) in RANK signaling pathway [10]. [4, 7, 8]. RANK-RANKL signaling facilitates phagocyte RUNX2 is an osteoblastic transcription factor that medi- activation, osteoclast maturation, and bone resorption [4, 7, ates osteoblast and osteoclast signaling pathway during bone 8, 24]. Osteoclast at the end of remodeling step undergoes formation and bone resorption [20]. Otto et al. reported apoptosis via phagocytosis [4–8]. This results in the release that RUNX2 inhibition resulted in the death of new born of calcium in blood which assists the bone mineralization null mouse due to the disfunctioning of bone formation and process. Available literature entails that RUNX2 acts as a skeletal mineralization pathway [21]. Our results correspond transcription factor of RANKL [7],whereasourresultsbased with investigations on mice, which reported decrease in level on qPCR and in silico hybridization analysis revealed that of RUNX2 after NaF exposure5 [ , 22]. miR-124 can be involved in the direct regulation of RANKL BioMed Research International 7

4.5 the higher degree of differential expression of genes (RUNX2, 4 RANKL,BGLAP,andOPG)involvedintheosteoclastic differentiation. 3.5 Normal RANK-RANKL signaling provides proper struc- 3 ture and mechanical strength to bones [5, 6, 22]. However, 2.5 study on various animal models suggests that increase in RANKL-RANK interaction resulted in osteoporosis, exces- 2 sive loss of bone [24]. To regulate the interaction network Fold change Fold 1.5 between RANK-RANKL, osteoblast cells secret a soluble 1 decoy of RANKL called OPG [23, 24]. Thus, excessive loss of bonecanbepreventedwhenOPGbindstoRANKLandinter- 0.5 feres with osteoblast and osteoclast interactions [23]. OPG 0 and RANKL are synthesized by osteoblast cells. Decrease miR-124 miR-155 in RANKL or increase in OPG may result in osteopetrosis miRNAs [23, 24]. In our study, OPG was found to be significantly upregulated, whereas the RANKL was found to be downregu- 8 PPM 20 PPM latedafterNaFexposuretoHOScells(Figure 5). Thus, it can be seen that the sublethal concentrations of NaF can cause Figure 4: Validation of miRNAs by qPCR. The expression levels of aberrant changes in the RANKL/RANK/OPG system. qPCR miRNAs in NaF treated samples were normalized with the control. analysis revealed that the sublethal concentrations of NaF can downregulate the expression levels of BGLAP (osteocalcin) 3 (Figure 5). Osteocalcin plays an important role in the process of bone mineralization, energy metabolism, and bone remod- 2 eling [25].Carboxylatedosteocalcinisinvolvedinformation of calcium hydroxyapatite crystals which are deposited in the 1 triple helical fibrils space of bone matrix26 [ ]. Excess intake 0 of fluoride may result in the downregulation of osteocalcin. It results in the conversion of calcium hydroxyapatite crystals −1 to fluorapatite crystal [25, 26]. As a result, bones become Fold change Fold stiff and loss its mechanical strength. Low concentration of −2 osteocalcin can affect the expression of insulin and prolif- 𝛽 −3 eration of -cells [27]. This may limit leptin (secreted by adipocytes) mediated PTH (parathyroid hormone) release −4 RUNX2 RANKL BGLAP OPG from parathyroid glands [27]. Low levels of PTH blocks 1-hydroxylasemediatedupregulationofcalcitriol(vitamin Genes 8 PPM D) in the kidney [28]. As a result, process of absorption 20 PPM of dietary calcium in the small intestine and secretion of free calcium (via apoptosis of osteoclast) required for the Figure 5: Validation of gene expression by qPCR. The expression RANKL mediated bone formation and remodeling is severely levels of genes in NaF treated samples were normalized with the hampered (Figure 6)[27, 28]. control. Compared to CG dinucleotides, outlier analysis of C/D box snoRNAs revealed higher numbers of UG dinucleotides after NaF exposure to HOS cells. Spontaneous deamina- (Figures 3, 4,and5).Thus,itcanbeseenthatmiR-124/miR- tion of unmethylated cytosine residues of CG dinucleotides 155 can regulate the RANKL expression either through the may result in the conversion of cytosine to uracil residues direct binding of miR-124 with the RANKL untranslated [13]. Deamination reaction is caused due to hydrolysis of regions or via RUNX2 mediated transcriptional inactivation. cytosine [19]. Fluoride ion, being highly electronegative, Compared to the higher dose of NaF (20 mg/L), elevated can cause hydrolysis of water molecules inside the cellular levels of miR-124 and miR-155 expressions were seen at environment. In this regard, the possibility of spontaneous lower dose level of NaF (8 mg/L). It has been reported that deamination of unmethylated cytosine after the exposure of fluoride exposure may have biphasic effect on osteoblast and NaF cannot be denied. osteoclast cells [2]. Fluoride exposure at lower concentration Compared to the terminal C/D box, microarray analysis promotes bone development, whereas exposure to higher revealed more marked effect of fluoride exposure on the concentration of fluoride may lead to cellular stress2 [ ]. expression levels of internal C/D boxes of snoRNAs The C/D Thus, the higher expression levels of miR-124 and miR-155 at boxes snoRNAs are short ncRNAs that are mainly involved lower concentrations of NaF (8 mg/L) obtained in our study in RNA editing processes leaded by ADAR proteins. In could be attributed to the biphasic effects exerted by NaF. particular the C/D box snoRNAs are prominently involved Compared to the higher dose level, elevated expression levels in methylation processes, at 2-O-ribose position of different of miR-124 and miR-155 at lower dose of NaF resulted in cellular RNAs. Their name derives from the presence of 8 BioMed Research International

PTH release from Kidney parathyroid gland Calcitriol Intestine Ca++ Leptin Blood ++ Ca++ Ca

Apoptosis Osteoblast ++ 𝛽-cells Adipocytes Ca RANKL RANK

miR-155 miR-124 Preosteoclast Osteocalcin

mRNA BGLAP RUNX2 RANKL OPG

Altered RANKL-RANK signaling cascade

Preosteoclast Apoptosis

Ca++ RANK Mature osteoclast Mineralization

RANKL

Resorption Reversal Bone formation

Normal RANKL-RANK signaling cascade

Osteoblast cells Downregulation Osteocytes Inhibition Inhibition Mononuclear cells Activation Upregulation

Figure 6: miRNA mediated alterations in bone formation and remodeling pathway.

two conserved nucleotide stretches: the C-box (UGAUGA) reported that snoRNA displaying long ends (snoRNAL)were displaced near the 5 termini and the D-box (CUGA) located overexpressed in ovarian and breast cancer cell lines, whereas in proximity of 3 termini [29]. Many evidences support that snoRNA displaying short ends (snoRNASH)wereabundant one of the critical events, associated with environmental in normal cell lines [31]. Furthermore, snoRNAL were mostly stress, is the alteration of methylation processes at DNA or found to display canonical snoRNA features compared to RNA level. Recent studies have underlined the criticality of noncanonical features found in snoRNASH. C/D box snoRNA RNA methylation biological process such as cell stability [12]. mayhaveanadditionalinternalD-boxinthemiddleof Schroeder et al. reported that metal ions or binding the molecule. Addition of internal D-box motifs can affect 󸀠 󸀠 proteins are responsible for stabilizing the minor grooves of (1) hairpin structure formation of 5 and 3 motifs of C- the canonical and noncanonical stems of C-box and D-box box and D-box elements; (2) binding with contaminants; motifs [30]. In this regard, longer C/D box snoRNAs may (3) expression of other noncoding RNAs. Thus, it can be contain additional nucleotide motifs that can interact with stated that the conformational differences due to presence of the metals and other binding proteins and influence the func- varying numbers of C-box and D-box sequences may alter tional role of these molecules. Deschamps-Francoeur et al. the functional motifs of C/D box snoRNAs. BioMed Research International 9

Length outlier analysis showed high numbers of D- Conflict of Interests box containing snoRNAs in NaF treated HOS cells when compared to the numbers of C-box containing sequences The authors declare that there is no conflict of interests (Supplementary data; S3D). Downregulation of C/D box regarding the publication of this paper. length outliers with more numbers (≥2) of D-box sequences in NaF treated HOS cells suggest the possible interactions of Acknowledgments NaF with snoRNAs due to conformational alterations exerted by additional D-box sequences. Microarray analysis revealed Authors are thankful to Council of Scientific and Industrial that sublethal concentrations of NaF can induce mark dif- Research (CSIR), India, and Integrated NextGen approaches ferences in the expression profiles of snoRNAs in HOS in health, disease and environmental toxicity (INDEPTH) cells (Supplementary data; S3E). Structure of differentially Networking project for providing necessary facilities and expressed snoRNAs, their function, and reports on their pos- funding. This study was partially supported by the University sible interactions with osteoclastic pathway is given in Sup- of Genoa, Italy. plementary data S3F. Till date no reports are available on the possible interaction of snoRNAs in the osteoclastic pathway. References D-box has a vital role in the methylation of ribosomal RNA (rRNA) [32]. D-box recognizes the nucleotide of rRNAs [1] N. Deoki, T. Mathiyazhagan, P. Ambika, and T. Ravi, “An to be methylated (usually 5th nucleotide upstream from the overview of fluoride and fluorosis,” Newsletter National Institute D-box binding site) [32, 33]. These results suggest that NaF of Health and Family Welfare,vol.9,pp.1–8,2007. can affect snoRNA mediated methylation of 28 s ribosomal [2] E. T. Everett, “Fluoride’s effects on the formation of teeth and RNA. This may result in the activation of aberrant transcrip- bones, and the influence of genetics,” JournalofDentalResearch, tional cascades involved in osteoclastogenic pathway. The vol. 90, no. 5, pp. 552–560, 2011. present study was carried out only on a limited set of C/D box [3]S.ErdalandS.N.Buchanan,“Aquantitativelookatfluorosis, RNAs. This preliminary in silico analysis could be extended fluoride exposure, and intake in children using a health risk to whole set of currently identified C/D box RNAs for the assessment approach,” Environmental Health Perspectives,vol. 113, no. 1, pp. 111–117, 2005. further validation of the results obtained in this study. [4] T. Anna, “Bone development: overview of bone cells and 5. Conclusion signaling,” Current Osteoporosis Reports,vol.9,no.4,pp.264– 273, 2011. Effectoffluorideontheepigeneticmodificationsisnot [5] J. G. Carvalho, A. L. Leite, D. Yan, E. T. Everett, G. M. Whitford, yet reported. Herein, we report the direct/indirect possible andM.A.R.Buzalaf,“Influenceofgeneticbackgroundon interactions of mIR-124 and mIR-155 with the RUNX2 and fluoride metabolism in mice,” JournalofDentalResearch,vol. RANKL in NaF treated HOS cells. Obtained results suggest 88, no. 11, pp. 1054–1058, 2009. that mIR-124 and mIR-155 can be considered the new players [6] C. A. O’Brien, “Control of RANKL gene expression,” Bone,vol. involved in the development of fluorosis. In addition, our 46,no.4,pp.911–919,2010. resultsimplythepossibleroleofC/DboxsnoRNAsinthe [7]C.L.Joana,C.J.Helena,andF.Eurico,“Osteoblastsandbone development of fluorosis. However, these results need to be formation,” Acta Reumatologica´ Portuguesa,vol.32,no.2,pp. validatedwiththeotherinvitroandinvivomodels.Future 103–110, 2007. studies should be designed on the possible role of miRNAs [8] S. Avnet, R. Pallotta, F. Perut et al., “Osteoblasts from a (mIR-124 and mIR-155) and snoRNAs in the development of mandibuloacral dysplasia patient induce human blood pre- fluorosis and their implications as the possible therapeutic cursors to differentiate into active osteoclasts,” Biochimica et drug targets against fluorosis; in particular it is important Biophysica Acta—Molecular Basis of Disease,vol.1812,no.7,pp. to plan experiment to investigate the interactions between 711–718, 2011. fluoride and proteins involved in the methylation cascade [9]H.Jian,Z.Lan,X.Lianping,andC.Di,“MicroRNA-204reg- or in the miRNA dicer dependent and dicer independent ulates Runx2 protein expression and mesenchymal progenitor biogenesis pathways. cell differentiation,” Stem Cells,vol.28,no.2,pp.357–364,2010. [10] Q. Yin, X. Wang, C. Fewell et al., “MicroRNA miR-155 inhibits Abbreviations bone morphogenetic protein (BMP) signaling and BMP- mediated epstein-Barr Virus Reactivation,” Journal of Virology, NaF: Sodium fluoride vol.84,no.13,pp.6318–6327,2010. RUNX2: Runt-related transcription factor 2 [11] Z. Li, M. Q. Hassan, S. Volinia et al., “A microRNA signature RANKL: Receptor activator of nuclear factor for a BMP2-induced osteoblast lineage commitment program,” kappa-B ligand Proceedings of the National Academy of Sciences of the United BGLAP: Bone gamma-carboxyglutamic States of America, vol. 105, no. 37, pp. 13906–13911, 2008. acid-containing protein [12] G. A. Stepanov, D. V. Semenov, A. V. Savelyeva et al., “Artificial OPG: Osteoprotegerin box C/D RNAs affect pre-mRNA maturation in human cells,” RANK: Receptor activator of nuclear factor 𝜅-B BioMed Research International,vol.2013,ArticleID656158,9 miRNA: MicroRNAs pages, 2013. snoRNAs: Small nucleolar RNAs [13]M.S.Scott,M.Ono,K.Yamada,A.Endo,G.J.Barton,andA. PTH: Parathyroid hormone. I. Lamond, “Human box C/D snoRNA processing conservation 10 BioMed Research International

across multiple cell types,” Nucleic Acids Research,vol.40,no.8, [29] J. Lin, S. Lai, R. Jia et al., “Structural basis for site-specific ribose pp. 3676–3688, 2012. methylation by box C/D RNA protein complexes,” Nature,vol. [14] S. A. Nurul, L. H. Lee, M. S. Shiran, A. R. Sabariah, M. S. 469, no. 7331, pp. 559–564, 2011. Avatarsingh, and Y. K. Cheah, “miR-181a regulates multiple [30] K. T. Schroeder, S. A. Mcphee, J. Ouellet, and D. M. J. Lilley, “A pathways in hypopharyngeal squamous cell carcinoma,” African structural database for k-turn motifs in RNA,” RNA,vol.16,no. Journal of Biotechnology,vol.11,pp.6129–6137,2012. 8, pp. 1463–1468, 2010. [15] A. Nimura, T. Muneta, K. Otabe et al., “Analysis of human [31] G. Deschamps-Francoeur, D. Garneau, F. Dupuis-Sandoval et synovial and bone marrow mesenchymal stem cells in relation al., “Identification of discrete classes of small nucleolar RNA to heat-inactivation of autologous and fetal bovine serums,” featuring different ends and RNA binding protein dependency,” BMC Musculoskeletal Disorders, vol. 11, article 208, 2010. Nucleic Acids Research,vol.42,no.15,pp.10073–10085,2014. [16] X.-H. Luo, L.-J. Guo, H. Xie et al., “Adiponectin stimulates [32] A. Lapinaite, B. Simon, L. Skjaerven, M. Rakwalska-Bange, RANKL and inhibits OPG expression in human osteoblasts F. Gabel, and T. Carlomagno, “The structure of the box C/D through the MAPK signaling pathway,” Journal of Bone and enzyme reveals regulation of RNA methylation,” Nature,vol. Mineral Research,vol.21,no.10,pp.1648–1656,2006. 502, no. 7472, pp. 519–523, 2013. [17] X. Sun, L. Wei, Q. Chen, and R. M. Terek, “HDAC4 represses [33] M. Brameier, A. Herwig, R. Reinhardt, L. Walter, and J. vascular endothelial growth factor expression in chondrosar- Gruber, “Human box C/D snoRNAs with miRNA like func- coma by modulating RUNX2 activity,” The Journal of Biological tions: expanding the range of regulatory RNAs,” Nucleic Acids Chemistry,vol.284,no.33,pp.21881–21890,2009. Research,vol.39,no.2,pp.675–686,2011. [18] S. Sain, P. K. Naoghare, S. S. Devi et al., “Beta caryophyllene and caryophyllene oxide, isolated from Aegle marmelos, as the potent anti-inflammatory agents against lymphoma and neuroblastoma cells,” Anti-Inflammatory & Anti-Allergy Agents in Medicinal Chemistry,vol.13,no.1,pp.45–55,2014. [19] W. Valleron, E. Laprevotte, E. F. Gautier et al., “Specific small nucleolar RNA expression profiles in acute leukemia,” Leukemia,vol.26,no.9,pp.2052–2060,2012. [20] J. C. Liu, C. J. Lengner, T. Gaur et al., “Runx2 protein expression utilizes the Runx2 P1 promoter to establish osteoprogenitor cell number for normal bone formation,” Journal of Biological Chemistry, vol. 286, no. 34, pp. 30057–30070, 2011. [21]F.Otto,A.P.Thornell,T.Cromptonetal.,“Cbfa1,acandidate gene for cleidocranial dysplasia syndrome, is essential for osteoblast differentiation and bone development,” Cell,vol.89, no. 5, pp. 765–771, 1997. [22]D.Yan,A.Gurumurthy,M.Wright,T.W.Pfeiler,E.G.Loboa, and E. T. Everett, “Genetic background influences fluoride’s effects on osteoclastogenesis,” Bone,vol.41,no.6,pp.1036–1044, 2007. [23]M.Nair,Z.R.Belak,andN.Ovsenek,“Effectsoffluorideon expression of bone-specific genes in developing Xenopus laevis larvae,” BiochemistryandCellBiology,vol.89,no.4,pp.377–386, 2011. [24] D. H. Jones, Y. Y. Kong, and J. M. Penninger, “Role of RANKL and RANK in bone loss and arthritis,” Annals of the Rheumatic Diseases, vol. 61, no. 2, pp. ii32–ii39, 2002. [25] P. Roberto, A. Gloria, C. Fernando et al., “Bone-specific transcription factor Runx2 interacts with the1𝛼, 25-dihydro- xyvitamin D3 eceptor to up-regulate rat osteocalcin gene expression in osteoblastic cells,” Molecular and Cellular Biology, vol. 24, pp. 8847–8861, 2004. [26] T. M. Doherty, L. A. Fitzpatrick, D. Inoue et al., “Molecular, endocrine, and genetic mechanisms of arterial calcification,” Endocrine Reviews,vol.25,no.4,pp.629–672,2004. [27] P. Gilberto, M. Claudio, W.Ma-Li, and L. Julio, “Leptin therapy, insulin sensitivity, and glucose homeostasis,” Indian Journal of Endocrinology and Metabolism,vol.16,no.9,pp.S549–S555, 2012. [28]A.Patti,L.Gennari,D.Merlotti,F.Dotta,andR.Nuti, “Endocrine actions of osteocalcin,” International Journal of Endocrinology,vol.2013,ArticleID846480,10pages,2013. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 194031, 11 pages http://dx.doi.org/10.1155/2015/194031

Review Article Genotoxic Effect in Autoimmune Diseases Evaluated by the Micronucleus Test Assay: Our Experience and Literature Review

Olivia Torres-Bugarín,1 Nicole Macriz Romero,1 María Luisa Ramos Ibarra,2 Aurelio Flores-García,3 Penélope Valdez Aburto,1 and María Guadalupe Zavala-Cerna1

1 Facultad de Medicina, Universidad Autonoma´ de Guadalajara, 45129 Zapopan, JAL, Mexico 2CUCBA, Universidad de Guadalajara, 45221 Guadalajara, JAL, Mexico 3Unidad Academica´ de Medicina, Universidad Autonoma´ de Nayarit, 63155 Tepic, NAY, Mexico

Correspondence should be addressed to Mar´ıa Guadalupe Zavala-Cerna; g zavala [email protected]

Received 19 December 2014; Revised 17 March 2015; Accepted 8 April 2015

Academic Editor: Francesca Pacchierotti

Copyright © 2015 Olivia Torres-Bugar´ın et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Autoimmune diseases (AD) are classified into organ-specific, systemic, and mixed; all forms of AD share a high risk for cancer development. In AD a destructive immune response induced by autoreactive lymphocytes is started and continues with the production of autoantibodies against different targets; furthermore apoptosis failure and loss of balance in oxidative stress asa consequence of local or systemic inflammation are common features seen in AD as well. Micronucleus (MN) assay can be performed in order to evaluate loss of genetic material in a clear, accurate, fast, simple, and minimally invasive test. The MN formation in the cytoplasm of cells that have undergone proliferation is a consequence of DNA fragmentation during mitosis and the appearance of small additional nuclei during interphase. The MN test, widely accepted for in vitro and in vivo genotoxicity research, provides a sensitive marker of genomic damage associated to diverse conditions. In here, we present a review of our work and other published papers concerning genotoxic effect in AD, identified by means of the MN assay, with the aim of proposing this tool as a possible early biomarker for genotoxic damage, which is a consequence of disease progression. Additionally this biomarker could be used for follow-up, to asses genome damage associated to therapies.

1. Introduction this loss of tolerance are unknown, but several studies suggest the participation of environmental, genetic and infectious The immune system functions in order to maintain an organ- factors. AD can be classified as organ-specific or systemic, ism survival against a threat through recognition of pathogen depending on whether the autoimmune response is directed derived antigens and the conduction of an immune response against a particular tissue as has been described in chronic to eliminate them; for that purpose the innate and adaptive lymphocytic thyroiditis (Hashimoto’s thyroiditis) where cir- immunity conduct organized and effective responses. Never- culating antibodies against thyroid peroxidase and thyroglob- theless, these mechanisms are not always accurate, sometimes ulin can be found [2], or against widespread antigens such a poor response may develop as seen in chronic infections, the as anti-nuclear antibodies in systemic lupus erythematosus recognition of self-proteins initiates autoimmunity, and the (SLE) [3]. presence of an exaggerated Th2 type of response is associa- Autoantibodies production then has a key role in AD dev- ted with allergic conditions [1]. From a clinical point of view, elopmentandisalsousedasadiseasemarkerforclassifi- autoimmune diseases (AD) are characterized by the pres- cation, evaluation, and monitoring disease activity. One of ence of autoantibodies that can lead to tissue damage and the most studied is antinuclear antibodies (ANA) detected destruction. Mechanisms in autoantibody generation can be mainly in patients with SLE but also found in patients with explained through a loss of tolerance, but how does tolerance other AD. Antinuclear antibodies are immunoglobulins that brake up occurs? In most instances mechanisms that initiate react against different nuclear (ssDNA, dsDNA, SSA/Ro, etc.) 2 BioMed Research International

(a) (b)

Figure 1: (a) Buccal mucosa cell with normal nuclei and (b) micronucleated cell from buccal mucosa. Photomicrographs stained with acridine orange viewed at 1000 magnification under fluorescence with an IVFL filter (450–490 nm). Binocular Microscope Carl Zeiss (Axiostar Plus). Sample was collected after a gentle swab of both cheeks oral mucosa with a polished slide; then the samples were spread directly in two separated slides previously identified, fixed with 80% ethanol, and stained for analysis; the sample must contain between 500 and 4000 cells to proceed with MN quantification16 [ , 17].

and cytoplasmic components (aminoacil tRNA synthetase, presence can reflect genotoxic damage occurred up to 14 days Jo-1, mitochondria, etc.). Autoantibodies and other clasto- prior to sample collection [7, 15]. For all of these reasons, MN genic factors (CF) have been identified in AD, capable of frequency can be considered as a biomarker for mutagenic inducing chromosomal breakage [4]. For example, the for- and genotoxic damage [14]. To further support this notion mation and scavenging activity of free radicals in biological we aimed to perform a review of the literature that included systems have been closely linked to a number of pathological information about MN frequencies in different AD. conditions, and there is increasing evidence that ROS and the resulting prooxidant/antioxidant imbalance play a major role 2. Methods in AD [5, 6]. Genomic damage is probably the most important funda- We reviewed data from the available medical literature. A mental cause of degenerative disease. It is also well established systematic search of papers was performed from data bases that genomic damage can be produced by environmental PUBMED,ClinicalKeyElsevier,andProQuest.Weusedthe exposure to genotoxic factors, medical procedures (e.g., radi- following search terms: “Autoimmune diseases” and “Micro- ation and chemicals), micronutrient deficiency (e.g., folate), nucleus.” We included all papers in English and Spanish, lifestyle (e.g., alcohol, smoking, drugs, and stress), and obtained with dates from 1995 until 2014. genetic defects, such as inherited defects in DNA metabolism and DNA repair [7]. It is essential to have a reliable and 3. Results and Discussion minimally invasive biomarker to improve monitoring, diagnostics, and treatment in AD. The micronucleus (MN) assay We identified 25 research papers: 17 were performed in lym- is an excellent biomarker candidate, since it is able to detect phocytes, 7 were performed in buccal mucosa, and 1 was per- chromosome breakage or malfunction of mitotic spindle formed in fibroblasts. MN frequency was associated with dif- caused by aneugenic mechanisms [8]. MN are formed during ferent pathological conditions listed in Table 1. mitosis when lagging chromosomes are left out the metaphase plate due to mitotic defects (aneuploid effect) or when 3.1. Organ Specific Autoimmune Disease. We found 3 publica- chromosome fragments lacking a centromere are not cap- tions about Graves’ disease, 2 studies related to type 1 diabetes tured by spindle microtubules (clastogenic effect); in both mellitus (T1DM), 2 studies in multiple sclerosis, and 2 with cases the genetic material was unable to be incorporated into affectionsoftheskin,oneinvitiligo,andonemoreinpsoriasis the nucleus of daughter cells [9]. The aneuploid effect can be patients. distinguished from the clastogenic damage by differences in Graves’ hyperthyroidism is an autoimmune thyroid dis- MN size [10] or by the presence of a centromere [11]. Both ease in which autoimmunity against thyroid antigens devel- events can develop spontaneously; nevertheless their fre- ops. The major antigen in Graves’ disease (GD) is the thy- quency can be increased under exposition to certain endoge- roid-stimulating hormone receptor (TSHR), but antibodies nous or exogenous agents [4, 12, 13]. Quantification of MN against thyroid peroxidase (TPO) and thyroglobulin (Tg) can frequency is easily performed in any tissue that is undergoing also be identified in up to 70% of patients with GD36 [ ]. constant proliferation [14], like oral mucosa epithelia [7, 15, Radioiodine therapy is the treatment of choice for recurrent 16](Figure 1), or in lymphocytes by two different assays [17]. hyperthyroidism; in radiation exposed patients, chromoso- With respect to oral mucosa, it has been described that MN mal damage may occur due to an increase in superoxide are observed in the basal layer of epithelial tissue and their radicals (clastogenic factors). After literature search, we did BioMed Research International 3 ] ] ] ] ] ] ] ] ] ] ] 18 23 22 25 19 21 12 26 24 27 20 2013 [ Year/reference 2011 [ 2008 [ 2007 [ 2011 [ 2013 [ 2004 [ 2006 [ 2009 [ c c a a a a a 13 (30 d) 2007 [ 16 2.8 2002 [ 19.4 15.9 22.8 c 1.4% ± 0.2% a 1.6 (7 d) ± 2.2 (7 d) ± ± 1.9 (21 d) ± 2.3 (21 d) ± 1.7 (7 d) NS NS ± NA 3.6 (21 d) ± 2.2 (180 d) 0.6 ± ± ± ± ± ± 3.1 347 ± ± 381 2.4 1.9 96.7 86.0 86.6 12.5 14.3 12.4 13.7 10.3 3.9 0.7 24.7 MN a b a a 3.3 6.1 6.5 4.4 1.1 1.6 1.8 1.5 1.0 1.4 3.2 1.2 3.1 0.8 1.2 2.9 0.6 3.5 1.1 1.3 1.4 0.3% ± ± 0.6% ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± NA NA NA NA NA NA ± Basal Posttreatment 1.9 3.1 1.2 27 1.8 3.5 3.1 5.2 8.7 2.5 1.5 2.6 4.0 0.8 3.2 14.2 15.2 3.3 12.7 45.3 2.4 0.6 0.9 I 131 I after 131 I I I 131 131 131 Systemic autoimmune diseases Ginkgo biloba Organ specific autoimmune diseases GD+vit.E,after T1DM T2DM DM + folic acid Healthy subjects MS + radiation Healthy subjects Vitiligo without treatmentHealthy subjects P + acitretin (oral) alone P + acitretin + NBUVB Acitretin + PUVA Healthy subjects RA without MTX RA with MTX RA+MTX+folicacid Healthy subjects GD 30 days after RRMS + radiation SPMS + radiation Healthy subjects + radiation T1DM Healthy subjects Study groups GD + GD + placebo after GD 7 days after Active RA Inactive RA Healthy subjects 55 23 30 81 23 20 21 21 5 6 5 5 29 50 40 39 18 18 20 35 15 𝑁 10 15 25 25 12 31 30 Table 1: Evaluation of micronucleus (MN) frequencies in organ specific and systemic autoimmune diseases. Ly 2000 cells 11 BM 2000 cells Ly 1000 cells BM 2000 cells Ly 1000 cells Ly 2000 cells Ly 2000 cells Ly NA Ly 1000 cells Disease Cells analyzed Graves’ disease (GD) Diabetes mellitus (DM) Multiple sclerosis (MS) VitiligoPsoriasis Ly 1000 cells LyRheumatoid 1000 arthritis cells (RA) 4 BioMed Research International ] ] ] ] ] ] ] ] 29 13 31 4 4 32 30 28 Year/reference 2014 [ 1999 [ 2003 [ 2010 [ 2004 [ 1995 [ 2000 [ 1999 [ c a a a b 13.1 2.1 1.3 1.6 17.9 26.1 2.3 2.8 ± ± ± ± ± ± ± ± 2.2 (14 d) 1.3 1.5 2.9 ± 7. 0 4.5 20.5 38.5 38.0 3.4 MN b b c b 1.7 2.7 2.1 1.1 3.3 12.0 2.2 13.6 2.4 1.2 2.2 1.4 9.5 4.6 4.7 3.9% 19.9 24.9 4.6% ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± NA ± Basal Posttreatment 1.3 1.3 0.9 9.1 5.5 4.7 9.4 19.1 6.2 25.9 23.5 9.1 15.4 14.0 10.1 18.7 10.6 27.4 38.0 Median cell counts of MN were significantly higheramong patients SLE ) − Table 1: Continued. /Scl70+) − SLE without treatment SLE with CYC Healthy subjects SSc SSc (ACA+/Scl70 SSc (ACA Healthy subjects SSc Idiopathic RP Secondary RP Healthy subjects SSc (nonaffected skin) + bleomycin SSc (affected skin) bleomycin + Healthy subjects + bleomycin SLE patients Healthy controls SLE without CYC SLE with CYC Healthy subjects Study groups SLE without treatment Healthy subjects SLE before first CYC bolus SLE 14 days after 1stSLE CYC before bolus second CYC bolus SLE 14 after the 2ndHealthy CYC bolus subjects 11 5 6 18 43 13 16 25 10 10 9 58 58 25 24 43 𝑁 15 18 10 10 7 7 43 Ly 2000 cells NA Ly 2000 cells Ly 1000 cells Fibroblasts 1000 BM 1000 cells BM 1000 cells Ly 2000 cells BM 1000 cells Disease Cells analyzed Systemic lupus erythematosus (SLE) Systemic sclerosis (SSc) BioMed Research International 5 ] ] ] 34 33 35 Year/reference 2005 [ 2013 [ 2009 [ MN c c a a a b 1.1 0.7 0.7 0.7 1.0 0.5 0.3 0.9 1.1 1.2 1.2 0.7 0.7 1.3 ± ± ± ± ± ± ± ± ± ± ± ± ± ± Basal Posttreatment te, SLE: systemic lupus erythematosus, CYC: cyclophosphamide, 3.4 1.7 1.7 1.9 2.3 2.2 2.2 0.8 3.4 3.1 3.2 3.6 3.5 2.6 ¸et’s disease, NA: not available, NS: not significant. Table 1: Continued. ; Ly: lymphocytes; BM: buccal mucosa; T1DM: type 1 diabetes mellitus; T2DM: type 2 diabetes mellitus; P: psoralen; PUVA: psoralen Patients with BD BD treated with colchicine BD without colchicines Healthy subjects Patients with BD BD without colchicine BD with colchicine Healthy subjects BD inactive BD active Healthy subjects BD negative HLA-B51 BD positive HLA-B51 Healthy subjects Study groups 𝑝 < 0.001 c tibodies against topoisomerase I, RP: Raynaud’s phenomenon, BD: Behc ,and 10 6 4 9 30 10 20 20 14 16 20 18 22 30 𝑁 𝑝 < 0.01 b , 𝑝 < 0.05 a BM 1000 cells Ly 1000 cells Ly 1000 cells Ly 1000 cells ¸et’s disease (BD) : sample size; d: days; statistical differences: Disease Cells analyzed Behc 𝑛 + UV A ultraviolet A irradiation; NBUVB:SSc: narrow-band ultraviolet systemic B sclerosis, irradiation; ACA: MS: anti-cytoplasmic multiple sclerosis, antibodies, RA: Scl-70: rheumatoid an arthritis, MTX: methotrexa 6 BioMed Research International not identified publications that compare MN frequencies (T2DM) when compared to type 1 (T1DM). Interestingly, the between untreated GD patients and healthy controls; it would frequency of MN was significantly decreased after 30 days have been desirable to state such differences for further ana- of folic acid supplementation (𝑝 < 0.001)[21]. T1DM is a lyzing the effect of specific treatments. The study publications complex autoimmune disorder caused by absolute deficiency that we found are related to the evaluation of micronuclei of insulin due to destruction of pancreatic 𝛽-cells. Cinkilic et 131 formation after I treatment in GD patients [18, 19]and al. in 2009 evaluated the frequency of MN in T1DM patients hyperthyroid patients (GD, toxic multinodular goiter and and compared to healthy subjects, they found no differences autonomous toxic nodule) [20]. In the study performed by in chromosomal aberrations or MN frequency. However Ballardin et al. [18], it was found that, in the hours imme- they were able to demonstrate a significantly higher sister diately following the administration of radioiodine, a mod- chromatids exchange frequency in T1DM when compared to ification in the oxidative equilibrium occurs, as documented healthy subjects, which may be due to generation of oxygen- by reduction in vitamin E levels, followed by the appearance derived free radicals occurring in this condition [22]. of genetic damage, shown by a twofold to threefold increase Multiplesclerosis(MS)isachronicautoimmunedisease, in MN frequency in observed lymphocytes. Authors con- in which myelin-reactive lymphocytes can be found in the cluded that MN formation in peripheral lymphocytes after central nervous system. Four types of MS have been char- radioiodine therapy correlated with the release of clastogenic acterized: relapsing-remitting (RRMS), secondary progres- factors in the plasma and negatively correlated to vitamin sive (SPMS), primary progressive (PPMS), and progressive E level. Nevertheless, in this study the greatest source of relapsing (PRMS). The RRMS form is present in 85% of pati- variation in MN induction was not attributed to absorbed ents at disease onset, but later during the course most of dose, radioiodine uptake, basal thyroid volume, age, or sex, them will develop the SPMS form [24]. We found two studies suggesting that other factors might influence MN induction aboutMNinMSpatients;theywereperformedtoevaluate and CF formation, such as polymorphic genes encoding the resistance to radiation of lymphocytes from MS patients; enzymes involved in reactive oxygen species detoxification resistance to radiation is defined as the capacity of lympho- and DNA repair [18]. In a later study Dardano et al. [19] cytes to prevent the formation of MN after radiation. The demonstrated a possible protective effect in genotoxic dam- first study was performed by Petcu et al. in 2006 [23]. They 131 age after I therapy with the administration of Ginkgo biloba found that the spontaneous MN frequency was significantly extract (EGb 761). Intriguingly basal frequencies of MN were higher in SPMS patients compared to healthy controls; such increased in both groups (placebo and EGb 761) when com- differences were attributed mainly to the increased level in pared to the previous study, as well as the result at the 7th day oxidative stress and decreased antioxidant capacity in serum but not at 21th day in the group receiving EGb 761. In both from MS patients. They also evaluated the radio sensitivity𝛾 ( - groups after 120 days MN frequency returned to basal values. irradiation) by mean of MN formation in peripheral blood 131 In conclusion I therapy induces a significant, although total lymphocytes, CD4+, and CD8+ T cells from SPMS transient, genetic damage characterized by an increase in patients compared to healthy controls; when analyzing total MN frequencies. Authors suggest that the administration of lymphocytes and CD8+ subset of cells, no differences were EGb761candecreasethistransientgeneticdamage[19]. found, but a decreased radiosensitive response was reported A third study performed by Shanmuga-Sundaram et al. in in CD4+ T cells from MS patients. When patients and 2011 [20] reported an analysis of peripheral blood lympho- healthy subjects were analyzed before and after irradiation the cytes cytotoxicity, in 32 hyperthyroid patients after the induction of MN was evident, although frequencies tended to administration of radioactive iodine. This study corrobo- be higher in healthy subjects after radiation; this fact could be rated previous findings, with a positive correlation between the consequence of a defective proliferation rate in patients 131 I dose and MN frequency, a significant increase in MN with MS or can be the reflection of MS therapies that render frequency at day 7th after therapy, and a decline after the CD4+ T cells resistant to MN formation after radiation when 30th day after therapy. The authors concluded that since compared to cells derived from healthy subjects. Lately in the cytotoxicity is transient and reversible, patients can be 2013, Milenkova et al. [24] reported that immunomodula- 131 torytherapywithIFN-𝛽 or glatiramer acetate in RRMS motivated to undertake I as an easy and safe treatment for patients reduces the number of MN in binucleated cells hyperthyroidism [20]. Nevertheless in this study the authors induced by in vitro 𝛾-irradiation. This data suggests that the did not provide information about baseline values of MN; it resistance to radiation by T lymphocytes in MS patients is wouldhavebeeninterestingtoknowhowmuchtheautoim- mainly due to treatment with immunomodulatory agents. mune pathology itself can contribute to the generation of MN The characterization of particular sets of cells derived from when compared to other inflammatory conditions. MS patients is needed to further clarify their cytophatic Diabetes mellitus (DM) is characterized by an elevation effects, since lymphocytes in MS and actually in any AD in blood glucose that eventually is associated with excessive have different characteristics that may influence cellular concentration in free radicals and oxidative damage to differ- responses, such as higher number of cells with the Th1 profile ent biomolecules, and importantly the DNA causing strand with increased production of proinflammatory cytokines, breaks [37]. Zu´niga-Gonz˜ alez´ et al. in 2007 found an incr- impaired apoptotic deletion (increased production of pro- eased frequency of MN formation in patients with DM; this liferative cytokines and growth factors, such as IL-2 and frequency was approximately twofold when compared to upregulation of survivin), and modified T helper/T suppres- healthy subjects (𝑝 < 0.001) and was increased in type 2 DM sor ratio [24].FinallythemechanismofactionofIFN-𝛽 BioMed Research International 7 is downregulation of vascular endothelial cells, decreasing Rheumatoid arthritis (RA) is a chronic systemic autoim- the production of adhesion molecules and metalloproteases mune condition that often presents with joint inflamma- which influences migration of T cells, and inhibition of the tion and destruction; furthermore, several studies have esti- generation of proinflammatory cytokines (TNF-𝛼,IFN-𝛾, mated the risk of malignancy in RA patients, with potential and IL-12). All together, these data suggest that therapy with attributable risk factors such as therapies, mainly disease- IFN-betacoulddecreasetheoxidativedamage,makingcells modifying antirheumatic drugs (DMARDs), and the influ- less sensitive to radiation in terms of MN formation. ence of RA itself. Ramos-Remus et al. in 2002 [12]evaluated Generalized vitiligo is an autoimmune disorder charac- thepresenceofMNinRApatientsandcomparedthemto terized by acquired white patches of skin and overlying hair, healthy controls. They also performed a subanalysis within as a result of loss in melanocytes function. Several theories the RA population and search for an increased frequency have been proposed to explain the pathogenesis of vitiligo, of MN within subjects taking methotrexate (MTX) as part including self-destructive, biochemical, neural, autoimmune, of their treatment. Authors found an increased baseline and genetic hypotheses, but each of them is only applicable frequencyofMNinRApatientswhencomparedtohealthy to a small proportion of cases [25]. High frequencies of skin- subjects, but the frequency was not significantly higher in homing melanocyte-specific cytotoxic T lymphocytes in the RA patients taking MTX; and although it was previously peripheral blood of patients with vitiligo have been reported, demonstrated in an in vitro study that folic acid could suggesting melanocyte death in vitiligo. The autocytotoxic inhibit the induction of MN by MTX, Ramos-Remus et al. hypothesis suggests that melanocyte impairment could be found that the MN number did not decrease after folic acid related to increased oxidative stress. Additionally patients supplementation, at least at the dose of 5 mg p.o. per day for 6 that suffer from vitiligo have decreased concentrations of hol- weeks. Authors concluded that the increase in MN frequency, otranscobalamin, vitamin B12, and folates [38]. It has been observed in this study, is mainly due to RA itself rather than to described previously that these micronutrients are important MTX use. A subsequent study performed in 2011 by Karaman for the maintenance of genome integrity [39]; therefore, we and coworkers [27] with the objective to determine if genetic believe that oxidative stress along with folate deficiency could lesions and DNA damage have an effect on the pathogenesis explain genomic instability in patients with vitiligo. This of RA tested the MN frequency in peripheral lymphocyte conclusion was demonstrated by Donmez-Altuntas et al. in cultures; authors found that the number of MN/1000 BNCs 2008 [25], since they found significantly higher frequencies of (binucleated cells) was significantly higher in RA compared MN in lymphocyte cultures of patients with vitiligo compared with controls. with control subjects, which points towards the use of MN as Systemic lupus erythematosus (SLE) is a chronic autoim- biomarkers during autoimmune conditions. mune disorder that may affect diverse organs, characterized Psoriasis vulgaris is a common, chronic, proliferative, and by the presence of antibodies against nuclear antigens (ANA) recurrent inflammatory skin disease, with lesions that appear with different specificities (ssDNA, dsDNA, RNA, and snRNP as a result of hyper proliferation, abnormal differentiation among others); treatment for this condition includes diverse regimens of high dose steroids and immunosuppressive of keratinocytes, and epidermis infiltration with inflamma- therapy. We found four studies related to MN frequency in tory cells including T cells and neutrophils. The etiology SLE. The first study included in this review was performed of psoriasis involves genetic, immune, and environmental bySeverinin1995[28]. She analyzed the frequency of MN factors [26, 40]. To evaluate cytotoxic and genotoxic damage cells in peripheral blood lymphocytes and found the highest in psoriasis patients after treatment, Silva et al.26 [ ]per- frequencies among SLE patients who had received cyclophos- formed a study in culture lymphocytes from psoriasis patients phamide for treatment, in comparison to patients with SLE who received different treatment regimens: acitretin alone, without medical treatment and control subjects. This higher acitretin + NBUVB (Narrow-band ultraviolet B irradiation), frequency was observed after 48 hours from the beginning of and acitretin + PUVA (psoralen + ultraviolet A irradiation). treatment. Later in 1999 Migliore et al. [4]performedastudy Results showed that at T0 all patients and controls pre- in SLE and systemic sclerosis (SSc) patients, with the aim to sented similar MN frequencies, at T12 (12 weeks after the assess the spontaneous frequency of cytogenetic damage in administration of treatments), both “acitretin + NBUVB” and lymphocytes, using the MN assay; they found that the MN “acitretin + PUVA”showed a trend towards an increase in MN frequency did not differ between SLE and controls. A third frequency; however, only “acitretin + PUVA”reached a signi- study performed by Rodr´ıguez-Vazquez´ et al. in 2000 [13]had ficant higher MN frequency. Authors state that the presence the objective to evaluate MN frequencies in SLE patients tak- of psoriasis induced an increase in MN frequencies, and these ing cyclophosphamide (CYC) treatment and compare them frequencies were increased after 13 weeks of treatment with to healthy controls; for this purpose they obtained buccal acitretin + PUVA [26]. mucosa samples from 10 SLE patients and 43 healthy subjects; samples were obtained before treatment and fourteen days after the administration of CYC. They found that the 3.2. Systemic Autoimmune Diseases. With respect to systemic frequency was similar in both groups at baseline, but in SLE autoimmune diseases, we found 2 published papers related 14 days after CYC administration the frequency in MN was to genotoxicity in rheumatoid arthritis (RA), 5 in systemic significantly increased and maintained elevated before the lupus erythematosus (SLE), 5 for systemic sclerosis (SSc), and second administration of CYC, with the highest frequency 4relatedtoBehc¸et’s disease. of MN observed 14 days after the second administration of 8 BioMed Research International

CYC. The last study about MN frequency in SLE included in significantly higher MN frequencies than controls. Further- this revision was performed by Aceves-Avila et al. in 2004 more, ACA positive patients showed the highest MN frequen- [29]. They analyzed the effect of CYC and the influence of cies, which suggest a possible role of ACA in determining CYP2D6 polymorphisms on MN formation in patients with cytogenetic anomalies [31].Confirmingtheresultsofthispre- SLE; they found that the genotoxicity is increased in patients vious study, Majone and coworkers described in 2007 [43] with SLE after CYC boluses without an association with the a significantly higher frequency of MN in SSc patients who CYP2D6 allele expression, which contributes to the idea that were positive for ACA compared to SSc patients who were irrespective from treatments patients with AD exhibit higher positive for antitopoisomerase or anti-RNA polymerase III as MNformationwhencomparedtohealthysubjects.Themost well as healthy controls. More importantly, they examined the recent study in this field was published in 2014 by Al-Rawi nature of chromosome damage within analyzed lymphocytes et al. [30]; after assessing nuclear morphology in 58 SLE for the presence of MN. Following the use of the in situ patients and 58 healthy age- and sex-matched controls by DIG-dUTP incorporation method using TdT enzyme, which the use of buccal micronucleus cytome assay, they proposed indicates deep genetic instability [44], they were able to show 󸀠 ascorewithacutoffvalueof≥4 that exhibited a high that the frequency of unstable fragments of DNA (free 3 - accuracy (93.1%), higher positive predictive value (98.1% at OH ends) was significantly higher in SSc patients who were 50% pretest probability and 99.8% at 90% pretest probability), positive for ACA and antitopoisomerase, with the highest sensitivity of 87.9%, and specificity of 98.3% for the diagnosis frequencies observed in ACA positive patients. Importantly, of SLE. After these findings, they concluded that the buccal this genetic abnormality was not found in SSc patients with micronucleus cytome assay was a valid and easy way for anti-RNA polymerase III antibody (𝑛=8), who had a low clinicians to assess individuals at high risk for developing SLE, frequencyofMN,similartothatinnormalsubjects[43]. The since a negative result in the total micronuclei count excluded mostrecentstudyincludedinthisreviewwithinsightsinto a possible diagnosis of SLE with 98.9% confidence in a clinical MN frequency in SSc patients was published by Martins et setting in which the differential diagnosis of SLE was of low al. in 2010 [32]; the objective of their study was to evaluate probability (10% pretest probability). Importantly this is the chromosome damage, by means of MN frequency, in dermal first report that analyzes MN frequency in an AD and proves fibroblasts from affected, nonaffected skin from SSc patients the potential that this test has to be used as a biomarker of (𝑛=10) and from controls (plastic surgery remnants; 𝑛= DNA damage in SLE. 9); all of the cultures were analyzed before and after the Systemic sclerosis (SSc) is an autoimmune condition administration of bleomycin. They found that the frequency where the presence of antibodies against the centromere or of MN was higher in cells from SSc patients, irrespective of directed to topoisomerase I (Scl70 antigen) can be found the area they were derived from, when compared to con- [4]. This systemic connective tissue disease is characterized trols. In bleomycin treated cultures, the MN frequency was by progressive fibrosis of the skin and internal organs; the increased in all groups but patterns were kept, with the high- presence of antinuclear antibodies (ANA) is frequently found est frequencies in cells derived from SSc patients. The authors in the serum with different specificities for nuclear targets conclude that increased chromosomal damage can affect [41]. Based on the extension of cutaneous sclerosis, SSc can be dermal fibroblast from clinically affected and nonaffected distinguished as a limited or diffuse; in most cases Raynaud’s skin in SSc patients, which may be related to oxidative stress derived from repetitive post-ischemic reperfusion episodes phenomenon (RP) is an early symptom of SSc that can pre- occurring in the tissues of SSc patients [32]. cedetheappearanceofSScbyseveralyears[31]. In 1999 Behçet’s disease (BD) is a chronic, multisystem, inflam- Migliore et al. [4] performed an analysis of MN frequency matory disorder characterized mainly by recurrent oral and in SSc patients and compared them to SLE and control sub- genital aphtha’s ulcerations, uveitis, and erythema nodosum; jects. Noteworthy, MN frequency was significantly higher although it can occur worldwide, it is most frequently seen in SSc patients when compared to both controls and SLE in Turkey and Japan [35]. The first study included in this patients, especially in patients who were anti-centromere review was published in 2005 by Hamurcu et al. [33], with antibodies (ACA) positive and anti-Scl70 negative, indicating an evaluation of MN frequency in peripheral lymphocytes that the presence of autoantibodies against nuclear targets from 30 patients with BD and 20 healthy controls and on could account for chromosome anomalies which have been buccal mucosa cells in 10 patients and 9 healthy controls. They reported since 1960 in association with SSc. Porciello et al. in found significantly higher MN rates in lymphocytes derived 2002 [42] described a first study in SSc patients; nevertheless from patients compared to controls and no difference in MN this study was replicated in 2003 with a higher number of frequency of buccal mucosa cells, even when a subgroup of patients and with similar results [31]; for this reason we only patients was exposed to the treatment colchicine. Later in included results from the latter study in the present review. 2009 Karaman et al. [34] performed a study to determine They evaluated the prevalence of spontaneous chromosome MN and sister chromatid exchange (SCE) frequencies in damage, by means of MN assay, in cultured peripheral lym- BD patients and to evaluate possible associations between phocytes of patients with SSc (𝑛=43), idiopathic RP (𝑛= oxidative stress and chromosome instability in BD. For this 13), suspected secondary RP (𝑛=16), and healthy cont- purpose lymphocyte cultures derived from 30 patients with rols (𝑛=25). Patients were also classified in terms of auto- BD and 20 healthy subjects were evaluated. Main results in antibody presence (ANA, ACA, or Scl70). They found that this study include a significantly higher frequency of SCE in patients with SSc and subjects with secondary RP showed BD patients when compared to healthy subjects, and even BioMed Research International 9 a significantly higher frequency was observed in active pati- a consequence of the disease. Another important observation ents (having at least three of the four major symptoms: oral is the fact that MN frequencies tend to be higher in systemic aphthae, genital ulceration, cutaneous lesions, and uveitis), AD when compared to local affections; in fact RA and SSc had when compared to inactive patients. Different from SCE, the highest MN frequencies even without the exposition to the MN rate was significantly higher in BD patients when treatments. Finally, more studies that involve different popu- compared to healthy subjects, but there was no significant lations of patients with other AD are needed to further clarify difference between active and inactive BD patients. Further- the mechanism of MN production in association with specific more, the presence of oxidative stress was confirmed mainly autoimmune conditions, since in some reports associations in patients with active BD. These differences are important, with specific autoantibodies could be detected. The influence since genetic abnormalities related to SCE are indicative of an of therapies on MN induction should be also clarified to give a earlybiologicaleffectthatmaynotbepermanentandmaynot correct interpretation of MN as early biomarkers of genomic have further consequences, while MN presence represents damage with possible prognostic significance. early but irreversible biological effects. Authors conclude that the elevated rates of SCE and MN in BD patients can be Conflict of Interests explained by increased oxidative stress. Several studies have shown a significant association of human leukocyte antigen- The authors declare that there is no conflict of interests (HLA-) B51 with BD, and for this reason Binici et al. in 2013 regarding the publication of this paper. [35] aimed to investigate whether patients with the HLA- B51 allele had higher MN frequencies; after analyzing MN frequencies in lymphocytes from 40 patients with BD and References 30 healthy controls, they found that the MN frequency was [1] P. Matzinger, “The danger model: a renewed sense of self,” significantly increased in BD patients and this frequency was Science,vol.296,no.5566,pp.301–305,2002. higher in HLA-B51 positive patients when compared to BD [2] A. P. Farwell and L. E. Braverman, “Inflammatory thyroid patients without this leukocyte antigen molecule. disorders,” Otolaryngologic Clinics of North America,vol.29,no. 4, pp. 541–556, 1996. 4. Conclusion [3] J. Cabiedes and C. A. Nuńez-˜ Alvarez,´ “Antinuclear antibodies,” Reumatologia Clinica,vol.6,no.4,pp.224–230,2010. It has been documented that AD with local or systemic affec- [4] L. Migliore, C. Bevilacqua, and R. Scarpato, “Cytogenetic tion are associated with genetic material loss; the presence of study and FISH analysis in lymphocytes of systemic lupus inflammatory cytokines, reactive oxygen species, and nitric erythematosus (SLE) and systemic sclerosis (SS) patients,” oxide is involved in the pathology of AD as well as in the Mutagenesis,vol.14,no.2,pp.227–231,1999. generation of MN, since the excess amount of these sub- [5] U.¨ Tursen,¨ “Pathophysiology of the Behçet’s disease,” Pathology stances may show cytotoxic and genotoxic effects. This review Research International,vol.2012,ArticleID493015,11pages, provides strong information that sustains the observation 2012. thatincreasedMNfrequenciescanbeobservedinassociation [6] A. Karaman, D. N. Binici, and M. A. Melikoglu,ˆ “Comet assay with the presence of any AD condition. On the other hand, and analysis of micronucleus formation in patients with rheu- treatments frequently used in AD can induce chromosomal matoid arthritis,” Mutation Research—Genetic Toxicology and damage as appeared evident in this review; such damage Environmental Mutagenesis,vol.721,no.1,pp.1–5,2011. can be identified with an increased frequency of MN, since [7] S. Bonassi, E. Coskun, M. Ceppi et al., “The HUman MicroNu- their presence provides an indirect and sensitive measure of cleus project on eXfoLiated buccal cells (HUMN XL): the role chromosomal breakage or missegregation. Therefore the MN of life-style, host factors, occupational exposures, health status, assay can be used to identify potential biomarkers that will and assay protocol,” Mutation Research,vol.728,no.3,pp.88– allow the identification of early genotoxic damage in a simple 97, 2011. and effective manner, with importance in early diagnosis, as [8]B.KashyapandP.S.Reddy,“Micronucleiassayofexfoliated well as an important tool for the differentiation of patients oral buccal cells: means to assess the nuclear abnormalities in with different AD from healthy controls in suspected cases. different diseases,” Journal of Cancer Research and Therapeutics, vol. 8, no. 2, pp. 184–191, 2012. Of consideration, after analyzing the information provided here we were able to observe that MN frequencies tend [9] W. Schmid, “The micronucleus test,” Mutation Research,vol.31, no.1,pp.9–15,1975. to be higher when lymphocytes are examined instead of buccal mucosa cells, irrespective of the disease evaluated. [10] L. Migliore, L. Cocchi, and R. Scarpato, “Detection of the centromere in micronuclei by fluorescence in situ hybridization: Furthermore, lymphocytes MN frequencies have higher vari- its application to the human lymphocyte micronucleus assay ability that is independent of the presence of an AD; just after treatment with four suspected aneugens,” Mutagenesis,vol. by examining healthy subjects different frequencies were 11, no. 3, pp. 285–290, 1996. ± ± obtained (from 0.6 0.3 to 27 3.5). These discrepancies [11] A. J. Afshari, P. W. McGregor, J. W. Allen, and J. C. Fus- should be minimized in order to be able to extrapolate results coe, “Centromere analysis of micronuclei induced by 2- in different populations and diseases. More important is the aminoanthraquinone in cultured mouse splenocytes using both evidence that in all AD that were included in this review the a gamma-satellite DNA probe and anti-kinetochore antibody,” presence of genomic instability is frequent when compared to Environmental and Molecular Mutagenesis,vol.24,no.2,pp.96– healthy subjects, even though it is not known if it is a cause or 102, 1994. 10 BioMed Research International

[12] C. Ramos-Remus, G. Dorazco-Barragan, F.J. Aceves-Avila et al., psoralen-ultraviolet A or narrow-band ultraviolet B-therapy in “Genotoxicity assessment using micronuclei assay in rheuma- psoriatic patients,” Mutation Research,vol.753,no.1,pp.42–47, toid arthritis patients,” Clinical and Experimental Rheumatol- 2013. ogy,vol.20,no.2,pp.208–212,2002. [27] A. Karaman, D. N. Binici, and M. A. Melikoglu,ˆ “Comet [13] M. S.-O. Rodr´ıguez-Vazquez,´ A. Ramos-Remus, C. Zuńiga,˜ assay and analysis of micronucleus formation in patients with and O. Torres-Bugar´ın, “Evaluacion´ de la genotoxicidad de la rheumatoid arthritis,” Mutation Research,vol.721,no.1,pp.1–5, ciclofosfamida mediante prueba de micronucleos´ en pacientes 2011. con lupus eritematoso sistemico,”´ Revista Mexicana de Reuma- [28] E. Severin, “Micronucleus frequency in peripheral blood lym- tolog´ıa,vol.15,no.2,pp.41–45,2000. phocytes from patients with systemic lupus erythematosus,” [14] J. A. Heddle, M. C. Cimino, M. Hayashi et al., “Micronuclei Romanian Journal of Morphology and Embryology,vol.41,no. as an index of cytogenetic damage: past, present, and future,” 1-2, pp. 55–58, 1995. Environmental and Molecular Mutagenesis,vol.18,no.4,pp. [29] F.J. Aceves-Avila, G. A. Esquivel Nava, M. P.Gallegos Arreola, B. 277–291, 1991. C. GomezMeda,G.M.Z´ uńiga˜ Gonzalez,´ and C. Ramos-Remus, [15] O. Torres-Bugar´ın, A. Ventura-Aguilar, A. Zamora-Perez et “Cyclophosphamide boluses induce micronuclei expression in al., “Evaluation of cisplatin + 5-FU, carboplatin + 5-FU, and buccal mucosa cells of patients with systemic lupus erythemato- ifosfamide + epirubicine regimens using the micronuclei test sus independent of cytochrome P450 2D6 status,” Journal of and nuclear abnormalities in the buccal mucosa,” Mutation Rheumatology,vol.31,no.7,pp.1335–1339,2004. Research—Genetic Toxicology and Environmental Mutagenesis, [30] Z. S. Al-Rawi, F. I. Gorial, R. F. Tawfiq et al., “Brief report: vol. 539, no. 1-2, pp. 177–186, 2003. a novel application of buccal micronucleus cytome assay in [16] N. Holland, C. Bolognesi, M. Kirsch-Volders et al., “The systemic lupus erythematosus: a case-control study,” Arthritis & micronucleus assay in human buccal cells as a tool for biomoni- Rheumatology, vol. 66, no. 10, pp. 2837–2841, 2014. toringDNAdamage:theHUMNprojectperspectiveoncurrent [31] G. Porciello, R. Scarpato, C. Ferri et al., “Spontaneous chromo- status and knowledge gaps,” Mutation Research,vol.659,no.1-2, some damage (micronuclei) in systemic sclerosis and Raynaud’s pp. 93–108, 2008. phenomenon,” JournalofRheumatology,vol.30,no.6,pp.1244– [17] M. Ceppi, F. Gallo, and S. Bonassi, “Study design and statistical 1247, 2003. analysis of data in human population studies with the micronu- [32] E. P. Martins, H. T. Fuzzi, C. Kayser et al., “Increased chromo- cleus assay,” Mutagenesis,vol.26,no.1,pp.247–252,2011. some damage in systemic sclerosis skin fibroblasts,” Scandina- [18] M. Ballardin, R. Barsacchi, L. Bodei et al., “Oxidative and vian Journal of Rheumatology,vol.39,no.5,pp.398–401,2010. genotoxic damage after radio-iodine therapy of Graves’ hyper- [33] Z. Hamurcu, H. Donmez-Altuntas,¨ M. Borlu, H. Demirtas, and thyroidism,” International Journal of Radiation Biology,vol.80, O.¨ Asçioslu, “Micronucleus frequency in the oral mucosa and no. 3, pp. 209–216, 2004. lymphocytes of patients with Behçet’s disease,” Clinical and [19] A. Dardano, M. Ballardin, M. Ferdeghini et al., “Anticlastogenic Experimental Dermatology,vol.30,no.5,pp.565–569,2005. effect of Ginkgo biloba extract in Graves’ disease patients [34] A. Karaman, M. Kadi, and F. Kara, “Sister chromatid exchange receiving radioiodine therapy,” Journal of Clinical Endocrinology and micronucleus studies in patients with Behçet’s disease,” and Metabolism,vol.92,no.11,pp.4286–4289,2007. Journal of Cutaneous Pathology,vol.36,no.8,pp.831–837,2009. [20] P. Shanmuga-Sundaram, S. Padma, S. Sudha, and K. Sasikala, [35] D. N. Binici, A. Karaman, and M. Kadi, “Micronucleus analysis 131 “Transient cytotoxicity of I beta radiation in hyperthyroid in behçet’s disease with and without HLA-B51,” Turkish Journal patients treated with radioactive iodine,” Indian Journal of of Physical Medicine and Rehabilitation,vol.59,no.1,pp.36–41, Medical Research,vol.133,no.4,pp.401–406,2011. 2013. [21] G. M. Zuńiga-Gonz˜ alez,C.M.Batista-Gonz´ alez,B.C.G´ omez-´ [36] G. Effraimidis and W. M. Wiersinga, “Mechanisms in endo- Meda et al., “Micronuclei in diabetes: folate supplementation crinology: autoimmune thyroid disease: old and new players,” diminishes micronuclei in diabetic patients but not in an animal European Journal of Endocrinology,vol.170,no.6,pp.R241– model,” Mutation Research,vol.634,no.1-2,pp.126–134,2007. R252, 2014. [22] N. Cinkilic, S. Kiyici, S. Celikler et al., “Evaluation of chromo- [37] P.Dandona,K.Thusu,S.Cooketal.,“OxidativedamagetoDNA some aberrations, sister chromatid exchange and micronuclei in in diabetes mellitus,” The Lancet,vol.347,no.8999,pp.444–445, patients with type-1 diabetes mellitus,” Mutation Research,vol. 1996. 676, no. 1, pp. 1–4, 2009. [38]L.F.Montes,M.L.Diaz,J.Lajous,andN.J.Garcia,“Folicacid [23]I.Petcu,D.Savu,H.Thierens,G.Nagels,andA.Vral,“Invitro and vitamin B12 in vitiligo: a nutritional approach,” Cutis,vol. radiosensitivity of peripheral blood lymphocytes in multiple 50,no.1,pp.39–42,1992. sclerosis patients,” International Journal of Radiation Biology, [39] M. Fenech, “Genomic stability: a new paradigm for recom- vol.82,no.11,pp.793–803,2006. mended dietary allowances (RDAs),” Forum of Nutrition,vol. [24] M. Milenkova, I. Milanov, K. Kmetska et al., “Chromosomal 56,pp.97–100,2003. radiosensitivity in patients with multiple sclerosis,” Mutation [40] I. Rucevic,´ V. Barisic-Drusko,´ L. Glavas-Obrovac, and M. Research,vol.749,no.1-2,pp.3–8,2013. Stefanic,´ “Vitamin D endocrine system and psoriasis vulgaris— [25] H. Donmez-Altuntas, Z. Sut, A. Ferahbas, Z. Hamurcu, review of the literature,” Acta Dermatovenerologica Croatica, and H. Demirtas, “Increased micronucleus frequency in vol.17,no.3,pp.187–192,2009. phytohaemagglutinin-stimulated blood cells of patients with [41] F. van den Hoogen, D. Khanna, J. Fransen et al., “2013 classi- vitiligo,” JournaloftheEuropeanAcademyofDermatologyand fication criteria for systemic sclerosis: an American College of Venereology,vol.22,no.2,pp.162–167,2008. Rheumatology/European League against Rheumatism collabo- [26] F. S. G. Silva, H. Oliveira, A. Moreiras et al., “Cytotoxic and rative initiative,” Arthritis and Rheumatism,vol.65,no.11,pp. genotoxic effects of acitretin, alone or in combination with 2737–2747, 2013. BioMed Research International 11

[42] G. Porciello, R. Scarpato, F.Storino et al., “The high frequency of spontaneous micronuclei observed in lymphocytes of systemic sclerosis patients: preliminary results,” Reumatismo,vol.54,no. 1,pp.36–39,2002. [43] F. Majone, F. Cozzi, M. Tonello et al., “Unstabilized DNA breaks in lymphocytes of patients with different subsets of systemic sclerosis,” Annals of the New York Academy of Sciences,vol.1108, pp.240–248,2007. [44] F. Majone, R. Luisetto, D. Zamboni, Y. Iwanaga, and K.-T. Jeang, “Ku protein as a potential human T-cell leukemia virus type 1 (HTLV-1) Tax target in clastogenic chromosomal instability of mammalian cells,” Retrovirology,vol.2,article45,2005. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 237183, 10 pages http://dx.doi.org/10.1155/2015/237183

Research Article Effect of 50 Hz Extremely Low-Frequency Electromagnetic Fields on the DNA Methylation and DNA Methyltransferases in Mouse Spermatocyte-Derived Cell Line GC-2

Yong Liu,1,2 Wen-bin Liu,2 Kai-jun Liu,2 Lin Ao,2 JuliaLiZhong,1 Jia Cao,2 and Jin-yi Liu2

1 College of Bioengineering, Chongqing University, Chongqing 400044, China 2InstituteofToxicology,CollegeofPreventiveMedicine,ThirdMilitaryMedicalUniversity,Chongqing400038,China

Correspondence should be addressed to Jin-yi Liu; [email protected]

Received 11 December 2014; Accepted 1 April 2015

Academic Editor: Davor Zeljezic

Copyright © 2015 Yong Liu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Previous studies have shown that the male reproductive system is one of the most sensitive organs to electromagnetic radiation. However, the biological effects and molecular mechanism are largely unclear. Our study was designed to elucidate the epigenetic effects of 50 Hz ELF-EMF in vitro. Mouse spermatocyte-derived GC-2 cell line was exposed to 50 Hz ELF-EMF (5 min on and 10 min off) at magnetic field intensity of 1 mT, 2 mT, and 3 mT with an intermittent exposure for 72 h. We found that 50 Hz ELF- EMF exposure decreased genome-wide methylation at 1 mT, but global methylation was higher at 3 mT compared with the controls. The expression of DNMT1 and DNMT3b was decreased at 1 mT, and 50 Hz ELF-EMF can increase the expression of DNMT1and DNMT3b of GC-2 cells at 3 mT. However, 50 Hz ELF-EMF had little influence on the expression of DNMT3a. Then, we established DNA methylation and gene expression profiling and validated some genes with aberrant DNA methylation and expression at different intensity of 50 Hz ELF-EMF.These results suggest that the alterations of genome-wide methylation and DNMTs expression may play an important role in the biological effects of 50 Hz ELF-EMF exposure.

1. Introduction the biological effects and its mechanism of ELF-EMF remain unclear. With the application of a number of electrical devices, living ELF-EMF exposure was considered genotoxicity and working environments are subject to different levels of although the energy transferred into the cells was not high extremely low-frequency electromagnetic fields (ELF-EMFs). enoughtocausedirectdamagetoDNA.However,itmay In recent years, the growing professionals and the public have have indirect effects on the DNA structure. Epigenetics is the dramatically concerned about the potential hazardous effects heritable and stable changes of gene expression or cellular of ELF-EMF exposure on human being. Many studies have phenotype without DNA sequence alterations, including reportedthathumanhealthisassociatedwiththeexposure DNA methylation, noncoding RNAs, and histone modifica- to ELF-EMF, especially focusing on a variety of diseases, tions [15, 16]. In recent years, many researches showed that including children leukemia, brain cancer, Lou Gehrig’s epigenetics is closely related to embryonic development, disease, and Alzheimer’s disease [1–8]. The male reproductive autoimmune diseases, cancer, and central nervous system system is one of the most sensitive organs to electromagnetic diseases [17, 18]. radiation. Although adverse effects of ELF-EMF on human DNA methylation occurs at the cytosine residues of CpG male reproduction are still controversial, substantial evidence dinucleotides by an enzymatic reaction that produces 5- indicates that exposure to ELF-EMF can decrease human methycytosine (5-mC), which is catalyzed by DNA methyl- andanimalssemenqualitybyreducingthemotilityand transferases (DNMTs). It is an extensively characterized increasing oxidative stress and morphometric abnormali- mechanism for epigenetic regulation and plays an important ties of spermatozoa [9–14]. In spite of numerous attempts, role in the regulation of gene expression [19]. Aberrant 2 BioMed Research International hypermethylation of CpG islands in gene promoter regions exposure time. After starvation (cultured with serum-free frequently occurs in human cancers. In higher eukaryotes, medium overnight), GC-2 cells were exposed to 50 Hz ELF- three active DNMTs have been identified. DNMT1 is involved EMF at magnetic field intensity of 1mT, 2mT, and 3mT with in the maintenance of methylation pattern that copies the an intermittent exposure (5 min on and 10 min off) for 72 h. preexisting methylation patterns onto the newly replicated The cultured cells were designed to two groups: one is sham strand during DNA replication, while DNMT3 is involved control; the other is 50 Hz ELF-EMF exposure. Cells were in de novo methylation for methylating previously unmethy- collected after 72 h of exposure. lated CpG sites and is encoded by 2 distinct genes, DNMT3a andDNMT3b.AsthemainenzymeforDNAmethylation, DNMTsisnotonlyassociatedwithDNAmethylation,but 2.3. Analysis of Global DNA Methylation. The total genomic also links to many important biological activities, including DNA was extracted from the GC-2 cells using the DNA cell proliferation, senescence, and cancer development. isolation Kit (Promega, Madison, WI, USA) according to the Previous studies have reported that DNA methylation is manufacturer’s instructions. The genome-wide methylation an important regulator of gene transcription and a large body level was detected by DNA Methylation Quantification Kit of evidence has demonstrated that aberrant DNA methyla- (Epigentek, New York, NY, USA) following the manufac- tion is associated with unscheduled gene silencing, and the turer’s protocol. The analysis provides the levels of global genes with high levels of 5-methylcytosine in their promoter DNA methylation, and it is not specific to any particular region are transcriptionally silent [20, 21]. Therefore, there is gene. The data were presented in terms of percent of control the possibility that ELF-EMF exposure could cause epigenetic (sham-exposed GC-2 cells). Experiments were carried out in modification changes in cells, and this may account forthe triplicate. adverse effects of ELF-EMF on the male reproductive system. In order to investigate a plausible mechanism, we exposed 2.4. RNA Extraction and Real-Time PCR. Total RNA was mouse spermatocyte-derived GC-2 cells to 50 Hz ELF-EMF extracted from the GC-2 cells using Trizol Reagent Kit exposure at an intensity of 1 mT, 2 mT, and 3 mT for 72 h. We (Invitrogen, Carlsbad, CA, USA). Complementary DNA was observed whether 50 Hz ELF-EMF exposure can induce the synthesized using 2 𝜇gRNAthroughthereversetranscription alterations of genome-wide methylation and the expression reaction by PrimeScript RT reagent kit with gDNA Eraser of DNMTs in GC-2 cells. (Takara, Otsu, Japan). The mRNA expression of DNMT1, DNMT3a, and DNMT3b was measured by real-time PCR. 2. Materials and Methods The real-time PCR was performed with the IQ5 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA), using 2.1. Cell Culture. Mouse spermatocyte-derived GC-2 cell line the SYBR Green I detection method. Each real-time PCR was obtained from the American Tissue Culture Collection reaction (20 𝜇Ltotalvolume)contained2𝜇LofcDNA,10𝜇L (ATCC, Rockville, MD, USA) and cultured in DMEM high of 2 × SYBR Green Real-time PCR Master (Takara), 0.5 𝜇Lof glucose medium (HyClone, Logan, UT, USA) containing 10% ∘ each of the forward and reverse primers, and 7 𝜇Lultrapure ∘ fetal bovine serum at 37 C in a humidified atmosphere with water. Samples were denatured by heating at 95 Cfor30sec, ∘ 5% CO2. followed by 40 cycles of amplification (95 Cfor5secand ∘ 60 C for 30 sec). The mRNA expression levels of DNMTs 2.2. Exposure Procedure. GC-2 cells were plated onto 90 mm were normalized to the expression level of the 𝛽-actin mRNA × 5 Petridishesatadensityof4 10 cells/5 mL. The ELF-EMF in each sample using the cycle threshold (Ct) method and −Δ exposure system used in our experiments is the EXF system the 2 Ct formula. Each measurement was repeated for three (Zurich, Switzerland). This EXF exposure system is designed times and was normalized against the control group. forthetestingofELF-EMFexposureinthe50Hzfrequency range such as alternating current transmission facilities. The exposure system was composed of a power frequency 2.5. Western Blot Analysis. Lysates of GC-2 cells were pre- generator, an arbitrary function generator, a narrow band pared for Western blot analysis. Proteins (40–60 𝜇g) were amplifier, and two rectangular waveguides. Exposed and resolved onto 10% sodium polyacrylamide gel electrophoresis sham-exposed cell dishes were simultaneously placed into and transferred onto nitrocellulose membranes (Bio-Rad, an incubator in which the environmental conditions were USA). The membranes were incubated in blocking buffer for ∘ constant (37 C, 5% CO2). The setup generated a vertical 2 hours at room temperature and then incubated with the pri- EMF, which is composed of two four-coil systems (two coils mary antibodies of DNMT1, DNMT3a, and DNMT3b (Santa with 56 windings, two coils with 50 windings), and was Cruz Biotechnology Inc., CA, USA) in antibody dilution ∘ placed inside a metal chamber. During the exposure, the buffer overnight at 4 C. The membrane was then washed with value of magnetic field was monitored and the temperature phosphate buffered saline for 3–5 times and incubated with of the medium was assessed as described previously [22]. HRP-conjugated secondary antibody (Beyotime, Shanghai, The temperature difference between sham and ELF-EMF China). The enhanced chemiluminescence kit (Beyotime) ∘ exposure never exceeded 0.3 C. The entire equipment was was used for the detection of signals. Equal loading of controlled by a computer, which can automatically control proteins on the gel was verified by reprobing the membrane the exposure parameters including exposure intensity and with 𝛽-actin antibody. BioMed Research International 3

2.6. Genome-Wide DNA Methylation Analysis and Validation. 140 ∗ ∗ The potential mechanism was investigated in detail using Affymetrix microarrays analysis. In brief, the methylation 120 of genes promoter were analyzed using Affymetrix Mouse Promoter 1.0R Array. Gene expression was analyzed by the Affymetrix Mouse Gene 2.0 ST microarrays. Microarray 100 experiments and data analyses were carried out at the Gminix ∗ company (Shanghai, China) as reported previously [23]. 80 Some genes were selected according to the highest fold change and verified with methylation-specific PCR (MSP) and real-time PCR. 60

2.7. Bisulfite Conversion and MSP. The DNA was extracted methylation Global DNA 40 from the GC-2 cells using the DNA isolation kit following the manufacturer’s instructions. DNA was bisulfite converted 20 EZ DNA Methylation-Gold Kit (Zymo Research, Orange, CA, USA) according to the manufacturer’s instruction. 0 Primer pairs that specifically amplified either methylated or 1 mT 2 mT 3 mT unmethylated sequences spanning the CpG island of these genes were used for MSP as detailed in Table S1 (see Supple- Sham mentaryMaterialavailableonlineathttp://dx.doi.org/10.1155/ Exposure 2015/237183). PCR conditions for MSP have been standard- Figure 1: Quantitative analysis of global DNA methylation in GC- ized in our laboratory [24, 25]. 2 cells. DNA methylation in GC-2 cells was lower than the sham- exposure group at magnetic field intensity of 1 mT and was higher 2.8. Statistical Analysis. All the experimental data were than the sham-exposure group at 2 mT and 3 mT exposure to 50 Hz expressed as the means ± SD. The statistical significance ELF-EMF. of difference between the values of sham and exposure was determined using analysis of variance. 𝑃 < 0.05 was considered as statistically significant. 1mTand2mTandweresignificantlyhigherthanthesham- exposure group at 3 mT. However, the expression of DNMT3a decreased exposed to 50 Hz ELF-EMF exposure compared 3. Results with the sham-exposure group at 2 mT (Figure 2(b)). 3.1. Effect of 50 Hz ELF-EMF on Genome Methylation Level in the GC-2 Cells. As male reproductive system is one of 3.3. Effect of 50 Hz ELF-EMF on the Protein Expression of the most sensitive organs to electromagnetic radiation, we DNMTs. We observed the differential mRNA expression of sought to determine whether there is a disruption of the DNA the DNMTs in the GC-2 cells exposed to ELF-EMF, and methylation patterns in GC-2 cells exposed to 50Hz ELF- then we examined the protein expression of DNMTs in EMF exposure for 72 h compared with the sham-exposure the GC-2 cells exposed to 50 Hz ELF-EMF exposure. We group. Quantitative analysis of the global DNA methylation found that the protein expression of DNMT1 decreased levels showed that the DNA methylation in GC-2 cells at magnetic field intensity of 1 mT and increased at 3mT were lower than the sham-exposure group at magnetic field electromagnetic field, while there were no obvious changes intensity of 1 mT and were higher than the sham-exposure in DNMT3a expression between the exposure and sham group at magnetic intensity of 2 mT and 3 mT (Figure 1). group (Figure 2(d)). Unfortunately, DNMT3b expression was These data showed that the DNA in GC-2 cells acquired under the detection level (data not shown). These results aberrant methylation pattern exposed to 50 Hz ELF-EMF were consistent with the quantitative analysis of the mRNA exposure. It suggested that global methylation alterations expression of the DNMTs using real-time PCR. might play a vital role in the biological effects and damage mechanismof50HzELF-EMFexposure. 3.4. Differentially Methylated DNA Analysis in 50 Hz ELF- EMF Exposure. In order to investigate the potential mech- 3.2. Effect of 50 Hz ELF-EMF on the mRNA Expression of anism, we screened differential methylated sites in detail DNMTs. As DNMTs play a crucial role in the regulation using Affymetrix Mouse promoter 1.0R Array. Genome- of DNA methylation pattern, we determined the expression wide methylation profiling across 50 Hz ELF-EMF exposure profiles of DNMT1, DNMT3a, and DNMT3b. The mRNA and control group were shown in Figure 3.Arrayanalysis expression of DNMT1, DNMT3a, and DNMT3b was mea- revealed that thousands of gene acquired aberrant methy- sured by real-time PCR. As shown in Figures 2(a) and 2(c), lation exposed to 50 Hz ELF-EMF exposure (fold change > the expression of DNMT1 and DNMT3b were significantly 2). Through DNA methylation chip analysis, there were a lower than the sham-exposure group at magnetic intensity of total of 296 differentially methylated sites (including 166 4 BioMed Research International

DNMT1 DNMT3a 1.2 1.2

1.0 1.0 ∗ ∗ ∗ 0.8 0.8

0.6 0.6

0.4 0.4 Relative mRNA expression mRNA Relative Relative mRNA expression mRNA Relative 0.2 0.2

0.0 0.0 1 mT 2 mT 3 mT 1 mT 2 mT 3 mT

Sham Sham Exposure Exposure (a) (b)

2.5 DNMT3b ∗

2.0

1.5 1 mT 2 mT 3 mT Sham Exposure Sham Exposure Sham Exposure 1.0 1 ∗ DNMT ∗

Relative mRNA expression mRNA Relative 0.5 DNMT3a

𝛽-actin 0.0 1 mT 2 mT 3 mT

Sham Exposure (c) (d)

Figure 2: Effect of ELF-EMF electromagnetic field exposures on the mRNA and protein of DNMT1, DNMT3a, and DNMT3b in GC-2cells. (a) The expression of DNMT1 was significantly lower than the sham-exposure group at magnetic field intensity of 1 mT and 2mTwas significantly higher than the sham-exposure group at 3 mT. (b) The expression of DNMT3a decreased exposure to 50 Hz ELF-EMF exposure compared with the sham-exposure group at magnetic field intensity of 2 mT. (c) DNMT3b expression was significantly lower than the sham- exposure group at magnetic intensity of 1 mT and 2 mT and significantly higher than the sham-exposure group at 3 mT. (d) The protein expression of DNMT1 decreased at magnetic field intensity of 1 mT and increased at 3 mT.

hypermethylation and 130 hypomethylation) in the 1 mT of those three genes was downregulated at 1.0 mT as shown exposed group compared with the control group. In 3 mT in Figure 3(b). At the same time, Fut11 was hypermethylated, group, there were only 70 differentially methylated sites, whereas Olfr969B and Lrrc9 were hypomethylated in GC-2 including 11 hypermethylated and 59 hypomethylated sites. cells at magnetic intensity of 3.0 mT as shown in Figure 3(c). The methylation status of differential methylation sites was The mRNA expression of Fut11 showed downregulation and confirmed by methylation-specific PCR and the mRNA the expression level of Olfr969B and Lrrc9 was upregulated in expression of genes was validated by real-time PCR in GC-2 cells at magnetic field intensity of 3.0 mTFigure ( 3(d)). Figure 3.AsshowninFigure 3(a),wefoundthatFut11, These results suggest that the methylation status of these Olfr969A,andTagln showed hypermethylation in GC-2 cells genes was inversely correlated with the mRNA expression in at magnetic field intensity of 1.0 mT. The mRNA expression GC-2 cells exposed to 50 Hz ELF-EMF. BioMed Research International 5

1 mT 1 mT 1.2 Sham Exposure M UMUMarker 1.0

Fut11 0.8

0.6 ∗ Olfr969A 0.4 ∗ Relative expression Relative ∗ 0.2 Tagln 0.0 Fut11 Olfr969A Tagln

Sham Exposure (a) (b) 3 mT 3 mT 2.5 ∗ Sham Exposure ∗ M UMUMarker 2.0

Fut11 1.5

∗ Olfr969B 1.0 Relative expression Relative 0.5 Lrrc9 0.0 Fut11 Olfr969B Lrrc9

Sham Exposure (c) (d)

Figure 3: Differentially methylated genes verified with methylation-specific PCR and real-time PCR in 50 Hz ELF-EMF exposure. (a) Representative MSP results of the three genes (Fut11, Olfr969A,andTagln) methylation in GC-2 cells at magnetic field intensity of 1.0 mT. M: methylated primers; U: unmethylated primers. (b) Validation of mRNA expression of the three genes (Fut11, Olfr969A,andTagln)by real-time PCR. (c) Representative MSP results of the three genes (Fut11, Olfr969B,andLrrc9) methylation in GC-2 cells at magnetic intensity of 3.0 mT. (d) Validation of mRNA expression of the three genes (Fut11, Olfr969B,andLrrc9)byreal-timePCR.

3.5. Gene Expression Profiling in 50 Hz ELF-EMF Exposure. cells at magnetic field intensity of 1 mT and 3 mT to test and To further explore the effects of 50 Hz ELF-EMF, we used verify using real-time PCR. These results showed that the Affymetrix microarrays analysis to establish the gene expres- mRNAexpressionofselectedgenewasconsistentwiththe sion profiles. Clustering of differentially expressed genes Affymetrix Array in GC-2 cellsFigure ( 4(b)). across 50 Hz ELF-EMF exposure and control group were Through pathway analysis, we found that 30 genes in shown in Figure 4(a). Toinvestigate the potential mechanism, Olfactory transduction pathway, 6 genes in Ribosome path- we choose the fold change that was greater than 1.5 as way, and 4 genes in Steroid hormone biosynthesis pathway significantly differentially regulated genes at magnetic field upregulated, while 30 genes in Olfactory transduction path- intensity of 1 mT and 3 mT. Through gene expression chip way downregulated in the 1 mT exposed group compared analysis, there were a total of 84 differentially expression with the control group. However, more genes and more genes (including 44 genes upregulation and 40 downreg- pathways altered in 3 mT exposed group, including 21 genes ulation) in the 1 mT exposed group compared with the in Ribosome pathway, 51 genes in Olfactory transduction control group. In 3 mT, the altered genes are 324, including pathway, 10 genes in Systemic lupus erythematosus pathway, 235 increase and 89 decrease. According to the results of 7 genes in p53 signaling pathway, 5 genes in Circadian Affymetrix Array, we chose several differentially regulated rhythm pathway, 8 genes in Cytosolic DNA-sensing pathway, genes which were downregulated or upregulated in GC-2 9 genes in Drug metabolism pathway, 10 genes in Antigen 6 BioMed Research International Control 1 mT 3 mT

0.50

0.33

0.17

1 mT 1.8 ∗ 1.6 1.4 1.2 1.0 0.00 0.8 ∗ ∗ ∗ ∗ 0.6 ∗

Relative expression Relative 0.4 0.2 0.0 Prss3 fut11 Maoa Cypt3 Mpeg1 Zfp759 Rhox2f Olfr765 Olfr747 −0.17 3 mT 1.4 1.2

1.0 ∗ ∗ 0.8 ∗ 0.6 ∗ ∗ ∗ −0.33 ∗ ∗ 0.4 ∗ ∗ Relative expression Relative ∗ 0.2 0.0 Cfh Ifit1 ifi44 Mx2 irf73 ligp1 adh1 Gbp7 OAS3 OAS2 ddx60 OAS1g Ifi27l2a Sham −0.50 Exposure (a) (b) Figure 4: Clustering of differentially expressed genes across 50 Hz ELF-EMF exposure and control group (a) and the differentially regulated genesinGC-2cellsatmagneticfieldintensityof1mTand3mTwerevalidatedbyreal-timePCR(b).Geneexpressiondataarepresentedina matrix format. Each row represents an individual gene, and each column corresponds to an exposure group, with red indicating upregulation and green indicating downregulation. Black and gray indicate unchanged expression and missing value, respectively. BioMed Research International 7 processing and presentation pathway, 8 genes in Retinol report on the status of global DNA methylation in GC- metabolism pathway, and 17 genes in Cytokine-cytokine 2 cells subjected to 50 Hz ELF-EMF exposure. We found receptor interaction pathway. that genome-wide methylation in GC-2 cells decreased at magnetic intensity of 1 mT and increased at 3 mT exposure to 3.6. Network Analyses. To better elucidate the molecular 50 Hz ELF-EMF compared with the control. Taken together, mechanisms and biological pathways implicated in 50 Hz these data suggest that DNA hypomethylation may be an ELF-EMF exposure in GC-2 cells, network analyses were important epigenetic event in the process of 50 Hz ELF-EMF used to find the differentially expressed and target key genes exposure on GC-2 cells. Although our results need to be in GC-2 cells between the exposure group and the sham- confirmed in animal and human study, it appears that there exposure group. Network analyses were generated through is a direct relationship between DNA hypomethylation and the use of the software confirmed the major functionally 50 Hz ELF-EMF exposure. related gene groups. As seen from Figure 5(a), Maoa, Bhmt, Then, we applied real-time PCR and Western bolt Gng8, Ugt2b34, Dgat1,andAdh1 may be target genes in GC-2 to detect the mRNA and protein expression of DNMT1, cells at magnetic intensity of 1 mT. Cyp3a11, Ugt2b34, Adcy5, DNMT3a, and DNMT3b, respectively. We observed that Ptgs1,andCyp2b23 may be closely related with the epigenetic DNMT1 and DNMT3b decreased at magnetic field intensity of 50 Hz ELF-EMF exposure at magnetic intensity of 3 mT of 1 mT and increased at 3 mT. However, the expression of (Figure 5(b)). DNMT3a had no obvious change exposure to 50 Hz ELF- EMF exposure. DNMT1 is involved in the maintenance of 4. Discussion methylation, while DNMT3 is involved in de novo methylation. Thus, the alterative expression of DNMTs might In our study, we demonstrated for the first time that contribute to the alterations of genome-wide methylation in 50 Hz ELF-EMF exposure can induce the alterations of GC-2 cells. genome-wide methylation and the expression of DNMTs in In order to further investigate the potential mechanism, spermatocyte-derived GC-2 cells. That is, the epigenetics Affymetrix microarray was used to screen differential methy- might play an important role in the biological effects of 50 Hz lated sites in detail. The results of microarrays were confirmed ELF-EMF exposure. using real-time PCR. Through differential signal pathway analysis, we found that the most obvious change in expression Several studies have found reproductive toxicity of EMFs is olfactory transduction pathway. Olfactory receptor (OR) in male [26, 27]. In male rats, 50 Hz sinusoidal magnetic field genes, mainly expressed in the surface of neurons in the at approximately 25 mT for 18 consecutive weeks did not have olfactory epithelium, were also found in other tissues such effects on the weight of the testes but reduced significantly astestis,heart,lung,kidney,andbrain[32]. Genome analysis the weights of seminal vesicles and preputial glands and showed that about 5% to 10% of the olfactory receptor gene sperm count [26]. Continuous exposure of 50 Hz ELF-EMF were expressed in late spermatocytes and early and late of0.1or0.5mTfor8weekssignificantlyincreasedthe round sperm cells of testes in rodents [33, 34]. Fukuda et incidence of testicular germ cell death [28]. The similar results al. found that the sperm chemotaxis changed significantly in were repeated in mice by 16-week continuous exposure to mice with higher expression of human OR17-4 [35]. These 60 Hz MF of 14 mT [10]. In addition, 50 Hz ELF EMF induced studies suggest that olfactory receptor gene may be looking alterations of spermatozoa motility and kindling rate in to provide the guiding role of the egg cell or sperm selection. rabbits, therefore influencing fertility13 [ ]. At the cellular Thus, olfactory transduction pathway may play an important level, Li et al. found that short-term but not long-term role in the biological effect of 50 Hz ELF-EMF. ELF-EMF exposure may decrease the reproductive ability of Previous studies showed that immune cell activation may male Drosophila melanogaster through the caspase pathway have important role in biological effect of 50 Hz ELF-EMF mediated spermatogenesis [29]. The increase of free radicals 2+ [36–39]. In this study, we also found that the expression of and Ca levels may lead to the ELF-EMF induced toxicity 󸀠 󸀠 several immune-related gene has changed significantly. 2 ,5 - [11, 30]. In addition, ROS may be involved in cell growth Oligoadenylate synthetize genes (OAS1, OAS2, and OAS3) inhibition by apoptosis and arrested the cell cycle in 60 Hz encode a family of enzymes pivotal to innate antiviral defense. ELF-EMF exposed prostate cancer cells [31]. However, the Oligoadenylate synthetases (OAS)playanimportantrolein molecular mechanism is still unclear. the immune response against dengue virus. SNPs in the OAS Recently, DNA methylation via the regulation of chro- genes are known to affect OAS activity and are associated matin structure modifications and the expression of genes with outcome of viral infections [40, 41]. In addition, the involved in cell cycle checkpoints, apoptosis, and DNA repair gene Mx2 encodes a member of the Mx protein family of is closely related to embryonic development, autoimmune large GTPases and functions in the innate immunity system. diseases, cancer, and central nervous system diseases [31]. Interferon alpha/beta treatment or viral infection induces Therefore, we speculated that DNA methylation might be expression of this protein, which subsequently accumulates associated with 50 Hz ELF-EMF exposure. First, we detected in the cytoplasm and inhibits viral replication [42–44]. Ddx58 genome-wide methylation status of GC-2 cell line exposed to gene, also known as RIG-I, can sense viral RNAs and triggers 50 Hz ELF-EMF at magnetic field intensity of 1 mT, 2 mT, and innate antiviral responses through induction of type I IFNs 3 mT with an intermittent exposure for 72 h by Methylated and inflammatory cytokines. It has been reported that Rig-I DNA Quantification Kit. To our knowledge, this is the first functions as a positive regulator for NF-𝜅B signaling and is 8 BioMed Research International

Cyp1a1 Ctgf

Cyp2b9 c ex ex c Gsta1 c c Smad2 Ces2a c Smad3 c c c Chrm1 c c Ugt1a1 a(ind)(+p) Gpr119 c c Cyp3a11 c a(ind) c c Glp1r c c Gsta2 a Gm4906 c Ugt2a1 Grin2a Ugt2b34 c c c Tgfb2 a c a b c c a(ind) Gm14475 Cyp2c29 c c c b Mmp1b Cbr1 c c Gria2 Hdac1 a(ind) Gnas c c ex c c Adh1c Aldh1a3 ex b c c c a(+p) ex Cyp2d9 c Fosb b Jun 1700055N04Rik a(+p) Mapk8 c c disso + rep c a( p) Aox1 c c a Mapk11 Fos a a(ind) Hsd3b5 Comt c Maoa ex(ind) Gng8 disso Gnai1 Prkaca a(+p) a(ind) Ly96 a(ind)(+p) c a(+p) inh(+p) b a(ind)(+p) c inh a(ind) Tlr4 Aldh7a1 c Cacng6 Ifna12 c Atf3 a(inh) ind(+p) Il4i1 Pkia a Ifna1 Mapk1 ex(ind) c 4930438A08Rik ind c Irs1 ex(ind) c Lao1 Adcy1 inh c c Myd88 Socs2 Pde6h Bhmt a a(+p) ex(ind) ex(ind) c c inh(−p) c c inh(u) c Itpa Cxcl3 Ifna5 Dmgdh Igf1r lnsr c Polr2l Cbs Cdo1 Csf2rb c c a aa c c cc Entpd1 Pkm Sardh Pklr Gnaz Angpt2 Cth c c Gm5506 (a)

Anpep Oplah c c Ggt5 Cyp2d22 c c Fmo1 c Gpx2 c c c c Maoa c c Cyp3a44 c c c c c c Cyp3a11 c c c c c c c Gstm7c c c Gstk1 c c c c c c c c Cyp2c50 Aldh3b1 c c c c c c c c c c c c c Aldh3a1 c c c c Gng8 cGstt1 c c Aldh1l2 c c c c c c c c c c c Cyp2a4 c c c a(inh) c c c Adh1 c c Ptgs1 Cyp2b23 c c inh(ind) Cyp26b1 Entpd2 Mthfd2 c c c Adcy5 c c c Mthfd1 Rdh8 c c c Dgka c c c Ephx2 c c c Dgat1 c Dhrs9 c Polr1b c Aldh1a7 c c c c c c c c c Rdh16 c Polr1a c c c c c c c c Atic Lpin1c c c c c c c Lipg Pde5a Npr1 c c c c c c c c Nme1 Dgat2 c Egr1 Pde3b c Polr2h c Ccl7 Cxcl16 c c 1700080E11Rik c Pde4a c c Ppap2c c ex c a(b) a(b) Pde7b Nme6 Nrg1 Fosb Pde10a cc c Ccl5 rep a(b) Ccr1 Cmpk2 a ex c c Stat2 b Gp6 c a(b) c b Fos a(b) Erbb3 inh(ind) a(b) a(b) Arrb1 c Uprt a a Pold4 c Irf9 Cxcl10 a(b) a Dctpp1 ex Myd88 Cxcl12 c inh(ind) Cxcl11 Col6a1 Col5a3 Dctd ex ex ex(ind) Cxcl13 b Col5a2 ex ex Bdh2 Myc Cxcl3 a a a ex ex(ind) a a c Isg15 ex ex a(ind) a c ex Stat1 ex(ind) Itga6 Oxct2a Cdk6 ex ex Itga11 Oxct2b b Tfap4 Ccnd2 a a(b) c ex a c Gm12597 a b inh(+p)(a) Vdr Ifna12 a a(b) Lamb2 b b Hmgcs1 Ddx58 inh inh inh(+p) b inh Vegfa ex ex Bcar1 a Actg1 Lama1 c inh a(inh)(u) Cdkn1a a Irs1 b a Rb1 a Irf7 Actn3 Sv2c Hadhb Dhx58 Fgfr1 b a(b) Trim25 b c a a Myl9 a a Myh4 a a Zbp1 b(disso) Aldh6a1 Ngf Ifih1 Pdgfc Acta2 Hgf Efna2 Csf1 H2-T22 H2-Q6 s Hist1h2bc s Pycard s s s H2-M9 Tap2 Casp7 Tgfb1 Dtx3l Hist1h2ah Enpp3 s Nol6 Gdf5 Jag2 Hist1h2bn Ppp1r3c H2-K1 a b a(b) a a a b a(ind) c inh(−p) s b b b inh inh Gm14482 b s s b a(ind) a inh b b Hist3h2a Pygl Klrc3 s s s s Casp1 b s s s s Tapbp Chrd 1700012L04RiK b + s a as s a(ind) a(ind) b a( p) a s LOC100047957 Inhba Hdac4 b Hist2h2aa1 a a sa b a Mrc2 Notch3 b Calm5 H2-Q7 b b a(b)(ind) H2afv a s b b inh a(b)(ind) a b b s H2-T23 p inh c c s s H2-T10 s Hist1h2ao Hist1h2bg s a Tap1 C1qbp Bmp2 Naip6 Naip2 Acvr1b Grin2a a Jag1 Itpr2 Kir3df1 a H2-M10.1 (b)

Figure 5: Network analysis of dynamic gene expression in GC-2 cells at magnetic field intensity of 1 mT (a) and 3 mT (b). The red dot stands for upregulated genes, the blue dot is downregulated genes, and the lilac dot stands for the connection gene. BioMed Research International 9 involved in multiple biological processes in addition to host electromagnetic field exposure,” Neurology,vol.47,no.6,pp. antivirus immunity [45, 46]. 1477–1481, 1996. In conclusion, we demonstrated that 50 Hz ELF-EMF [9]J.S.Lee,S.S.Ahn,K.C.Jung,Y.-W.Kim,andS.K.Lee,“Effects exposure can induce the alterations of genome-wide methy- of 60 Hz electromagnetic field exposure on testicular germ cell lation and the expression of DNMTs. These results suggest apoptosis in mice,” Asian Journal of Andrology,vol.6,no.1,pp. that epigenetic regulation might play an important role in 29–34, 2004. the biological effects of 50 Hz ELF-EMF exposure. Although [10] Y.-W. Kim, H.-S. Kim, J.-S. Lee et al., “Effects of 60 Hz 14 𝜇T we could not confirm whether the long-term exposure of magnetic field on the apoptosis of testicular germ cell in mice,” alternating current does harm to our health based on the Bioelectromagnetics,vol.30,no.1,pp.66–72,2009. present study, we recommend that charge power supply and [11] N. Bernabo,` E. Tettamanti, M. G. Pistilli et al., “Effects of 50 Hz highvoltagewireshouldbekeptasfaraspossiblefromour extremely low frequency magnetic field on the morphology and body to reduce the absorption of radiation by cells. function of boar spermatozoa capacitated in vitro,” Theriogenol- ogy,vol.67,no.4,pp.801–815,2007. [12] N. Bernabo,` E. Tettamanti, V. Russo et al., “Extremely low Conflict of Interests frequency electromagnetic field exposure affects fertilization outcome in swine animal model,” Theriogenology,vol.73,no. The authors declare no conflict of interests. 9, pp. 1293–1305, 2010. [13]S.Roychoudhury,J.Jedlicka,V.Parkanyietal.,“Influenceof Authors’ Contribution a 50 Hz extra low frequency electromagnetic field on spermatozoa motility and fertilization rates in rabbits,” Journal of Yong Liu and Wen-bin Liu contributed equally to this work. Environmental Science and Health. Part A Toxic/Hazardous Substances and Environmental Engineering,vol.44,no.10,pp. Acknowledgment 1041–1047, 2009. [14] R. Iorio, R. Scrimaglio, E. Rantucci et al., “A preliminary This work was supported by the National Basic Research study of oscillating electromagnetic field effects on human Program of China (National 973 Program) (Grant no. spermatozoon motility,” Bioelectromagnetics,vol.28,no.1,pp. 72–75, 2007. 2011CB503700). [15] M. M. Suzuki and A. Bird, “DNA methylation landscapes: provocative insights from epigenomics,” Nature Reviews Genet- References ics,vol.9,no.6,pp.465–476,2008. [16] M. Monk, “Epigenetic programming of differential gene expres- [1] T. Koeman, P. A. van den Brandt, P. Slottje et al., “Occupational sion in development and evolution,” Developmental Genetics, extremely low-frequency magnetic field exposure and selected vol.17,no.3,pp.188–197,1995. cancer outcomes in a prospective Dutch cohort,” Cancer Causes and Control,vol.25,no.2,pp.203–214,2014. [17] E. Li, “Chromatin modification and epigenetic reprogramming in mammalian development,” Nature Reviews Genetics,vol.3, [2]C.Pedersen,O.Raaschou-Nielsen,N.H.Rodetal.,“Distance no.9,pp.662–673,2002. from residence to power line and risk of childhood leukemia: a population-based case-control study in Denmark,” Cancer [18] P. B. Talbert and S. Henikoff, “Spreading of silent chromatin: Causes and Control,vol.25,no.2,pp.171–177,2014. inaction at a distance,” Nature Reviews Genetics,vol.7,no.10, pp. 793–803, 2006. [3] J. Grellier, P.Ravazzani, and E. Cardis, “Potential health impacts of residential exposures to extremely low frequency magnetic [19] R. J. Klose and A. P.Bird, “Genomic DNA methylation: the mark fields in Europe,” Environment International,vol.62,pp.55–63, and its mediators,” Trends in Biochemical Sciences,vol.31,no.2, 2014. pp. 89–97, 2006. [4] S.J.London,D.C.Thomas,J.D.Bowman,E.Sobel,T.C.Cheng, [20] E. J. Richards, “Inherited epigenetic variatio—revisiting soft and J. M. Peters, “Exposure to residential electric and magnetic inheritance,” Nature Reviews Genetics,vol.7,no.5,pp.395–401, fields and risk of childhood leukemia,” American Journal of 2006. Epidemiology,vol.134,no.9,pp.923–937,1991. [21] P. A. Jones and G. Liang, “Rethinking how DNA methylation [5]J.M.Harrington,D.I.McBride,T.Sorahan,G.M.Paddle,and patterns are maintained,” Nature Reviews Genetics,vol.10,no. M. Van Tongeren, “Occupational exposure to magnetic fields 11, pp. 805–811, 2009. in relation to mortality from brain cancer among electricity [22]W.Duan,C.Liu,H.Wuetal.,“Effectsofexposuretoextremely generation and transmission workers,” Occupational and Envi- low frequency magnetic fields on spermatogenesis in adult rats,” ronmental Medicine,vol.54,no.1,pp.7–13,1997. Bioelectromagnetics,vol.35,no.1,pp.58–69,2014. [6] C. Johansen and J. H. Olsen, “Mortality from amyotrophic [23] M. Ebisuya, T. Yamamoto, M. Nakajima, and E. Nishida, lateral sclerosis, other chronic disorders, and electric shocks “Ripples from neighbouring transcription,” Nature Cell Biology, among utility workers,” American Journal of Epidemiology,vol. vol. 10, no. 9, pp. 1106–1113, 2008. 148, no. 4, pp. 362–368, 1998. [24] W. B. Liu, F. Han, X. H. Du et al., “Epigenetic silencing of [7]P.K.Verkasalo,J.Kaprio,J.Varjonen,K.Romanov,K.Heikkila,¨ Aristaless-like homeobox-4, a potential tumor suppressor gene and M. Koskenvuo, “Magnetic fields of transmission lines and associated with lung cancer,” International Journal of Cancer, depression,” American Journal of Epidemiology,vol.146,no.12, vol. 134, no. 6, pp. 1311–1322, 2014. pp.1037–1045,1997. [25]W.-B.Liu,X.Jiang,F.Hanetal.,“LHX6actsasanovelpotential [8] E. Sobel, M. Dunn, Z. Davanipour, Z. Qian, and H. C. Chui, tumour suppressor with epigenetic inactivation in lung cancer,” “Elevated risk of Alzheimer’s disease among workers with likely Cell Death & Disease,vol.4,no.10,articlee882,2013. 10 BioMed Research International

[26] M.-A. Al-Akhras, H. Darmani, and A. Elbetieha, “Influence [41] M. C. Tessier, H.-Q. Qu, R. Frechette´ et al., “Type 1 diabetes and of 50 Hz magnetic field on sex hormones and other fertility the OAS gene cluster: association with splicing polymorphism parameters of adult male rats,” Bioelectromagnetics,vol.27,no. or haplotype?” Journal of Medical Genetics,vol.43,no.2,pp. 2, pp. 127–131, 2006. 129–132, 2006. [27] E. K. Koh, B. K. Ryu, D. Y. Jeong, I. S. Bang, M. H. Nam, [42] O. Haller and G. Kochs, “Interferon-induced mx proteins: and K. S. Chae, “A 60-Hz sinusoidal magnetic field induces dynamin-like GTPases with antiviral activity,” Traffic,vol.3,no. apoptosis of prostate cancer cells through reactive oxygen 10, pp. 710–717, 2002. species,” International Journal of Radiation Biology,vol.84,no. [43] F. Flohr, S. Schneider-Schaulies, O. Haller, and G. Kochs, “The 11, pp. 945–955, 2008. central interactive region of human MxA GTPase is involved in [28] A. Elbetieha, M.-A. Al-Akhras, and H. Darmanl, “Long-term GTPase activation and interaction with viral target structures,” exposureofmaleandfemalemiceto50Hzmagneticfield: FEBS Letters,vol.463,no.1-2,pp.24–28,1999. effects on fertility,” Bioelectromagnetics,vol.23,no.2,pp.168– [44] T. Zurcher, J. Pavlovic, and P. Staeheli, “Mechanism of human 172, 2002. MxA protein action: variants with changed antiviral properties,” [29] S.-S. Li, Z.-Y. Zhang, C.-J. Yang, H.-Y. Lian, and P. Cai, The EMBO Journal, vol. 11, no. 4, pp. 1657–1661, 1992. “Gene expression and reproductive abilities of male Drosophila [45] M. Yoneyama, M. Kikuchi, T. Natsukawa et al., “The RNA melanogaster subjected to ELF-EMF exposure,” Mutation helicase RIG-I has an essential function in double-stranded Research—Genetic Toxicology and Environmental Mutagenesis, RNA-induced innate antiviral responses,” Nature Immunology, vol.758,no.1-2,pp.95–103,2013. vol. 5, no. 7, pp. 730–737, 2004. [30] R. P.Liburdy, “Calcium signaling in lymphocytes and ELF fields. [46] T. Kawai, K. Takahashi, S. Sato et al., “IPS-1, an adaptor trig- Evidence for an electric field metric and a site of interaction gering RIG-I- and Mda5-mediated type I interferon induction,” involving the calcium ion channel,” FEBS Letters,vol.301,no. Nature Immunology, vol. 6, no. 10, pp. 981–988, 2005. 1,pp.53–59,1992. [31] E. K. Koh, B.-K. Ryu, D.-Y. Jeong, I.-S. Bang, M. H. Nam, and K.-S. Chae, “A 60-Hz sinusoidal magnetic field induces apoptosis of prostate cancer cells through reactive oxygen species,” International Journal of Radiation Biology,vol.84,no. 11, pp. 945–955, 2008. [32] N. Kang and J. Koo, “Olfactory receptors in non-chemosensory tissues,” BMB Reports,vol.45,no.11,pp.612–622,2012. [33] X. Zhang, M. Rogers, H. Tian et al., “High-throughput microarray detection of olfactory receptor gene expression in the mouse,” Proceedings of the National Academy of Sciences of the United States of America,vol.101,no.39,pp.14168–14173,2004. [34] N.FukudaandK.Touhara,“Developmentalexpressionpatterns of testicular olfactory receptor genes during mouse spermatogenesis,” Genes to Cells, vol. 11, no. 1, pp. 71–81, 2006. [35]N.Fukuda,K.Yomogida,M.Okabe,andK.Touhara,“Func- tional characterization of a mouse testicular olfactory receptor and its role in chemosensing and in regulation of sperm motility,” JournalofCellScience,vol.117,no.24,pp.5835–5845, 2004. [36] S. H. Hefeneider, S. L. McCoy, F. A. Hausman et al., “Long-term effects of 60-Hz electric vs. magnetic fields on IL-1 and IL-2 activity in sheep,” Bioelectromagnetics,vol.22,no.3,pp.170–177, 2001. [37] M. Simko´ and M.-O. Mattsson, “Extremely low frequency electromagnetic fields as effectors of cellular responses in vitro: possible immune cell activation,” Journal of Cellular Biochem- istry,vol.93,no.1,pp.83–92,2004. [38]J.Frahm,M.Lantow,M.Lupke,D.G.Weiss,andM.Simko,´ “Alteration in cellular functions in mouse macrophages after exposure to 50 Hz magnetic fields,” Journal of Cellular Biochem- istry,vol.99,no.1,pp.168–177,2006. [39] Z. Akan, B. Aksu, A. Tulunay, S. Bilsel, and A. Inhan-Garip, “Extremely low-frequency electromagnetic fields affect the immune response of monocyte-derived macrophages to patho- gens,” Bioelectromagnetics,vol.31,no.8,pp.603–612,2010. [40] K. Alagarasu, T. Honap, I. M. Damle, A. P.Mulay, P.S. Shah, and D. Cecilia, “Polymorphisms in the oligoadenylate synthetase gene cluster and its association with clinical outcomes of dengue virus infection,” Infection, Genetics and Evolution,vol.14,no.1, pp. 390–395, 2013. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 587983, 13 pages http://dx.doi.org/10.1155/2015/587983

Review Article Environmental Epigenetics: Crossroad between Public Health, Lifestyle, and Cancer Prevention

Massimo Romani, Maria Pia Pistillo, and Barbara Banelli

Laboratory of Tumor Epigenetics, IRCCS AOU San Martino-IST, Largo Rosanna Benzi 10, 16132 Genova, Italy

Correspondence should be addressed to Massimo Romani; [email protected]

Received 14 January 2015; Accepted 1 April 2015

Academic Editor: Jia Cao

Copyright © 2015 Massimo Romani et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Epigenetics provides the key to transform the genetic information into phenotype and because of its reversibility it is considered an ideal target for therapeutic interventions. This paper reviews the basic mechanisms of epigenetic control: DNA methylation, histone modifications, chromatin remodeling, and ncRNA expression and their role in disease development. We describe also the influence of the environment, lifestyle, nutritional habits, and the psychological influence on epigenetic marks and how these factors are related to cancer and other diseases development. Finally we discuss the potential use of natural epigenetic modifiers in the chemoprevention of cancer to link together public health, environment, and lifestyle.

1. Introduction the epigenetic asset of the live beings is now being increasingly appreciated and the role of epigenetics in nontumor The complete sequence of the human genome was released at pathologic conditions is actively investigated [6, 7]. Epige- the beginning of the XXI century and can be considered as netic modulation occurs during the entire lifespan from con- the master library where all the genetic information is stored. ception to adulthood. Maternal diet, alcohol consumption, All this information, to be used, must be properly read and and smoke habits can all influence the epigenetic landscape, interpreted. Even a well-known text like the Hamlet soliloquy as well as later in life the exposure to environmental chemicals ... “to be, or not to be ” would be hard to understand without can disrupt the epigenetic programming [8, 9]besideincreas- word interruptions and punctuation. Indeed, the genetic code ing the cancer risk. requires another code on top of it (from the Greek 𝜀𝜋𝜄)that, Epigenetic mechanisms are involved in drug resistance like the annotations on the side of a book, enables the com- [10–12] and are responsible, at least in part, for the interindi- prehension of the text [1]. The term “epigenetic landscape” vidual variation of the response to drugs [10]. In this respect, a wascoinedin1939byCHWaddington,beforeDNAwas particularly promising avenue for epigenetics is the develop- recognized as the molecule of inheritance, to describe the ment of new and effective therapies that could overcome drug mechanisms of transition of the cells from the totipotent resistance. The information contained in the DNA needs to be to the differentiated condition. In practice, epigenetics pro- timely used; in this respect epigenetic modifications, because vides the tools to translate the information (genotype) into of their reversibility and rapid change, confer phenotypic function (phenotype) [2].Thus,ifthesequenceoftheDNA plasticity in response to environmental or internal stimuli. storesallthedatanecessarytobuildalivingcellororganism, From a therapeutic point of view, the reversibility of the epi- epigenetic, like the operating system of a computer, decodes genetic variations makes them ideal drug targets [11]; indeed the information and determine when, how, and where a given in some experimental models drug resistance was found to set of instructions must be used. be reversible and mediated by epigenetic mechanisms [12]. Although epigenetic research has been traditionally Along this line some reports have shown the clinical utility of focusedondevelopmentalandcancer-relatedalterations[3– drug rechallenge and the possibility to resensitize the patients 5], the effect of the environment and of dietary factors on to first line chemotherapeutic agents intervening on the 2 BioMed Research International

Table 1: FDA approved epigenetic drugs. Common name Trade name Disease Route Mode of action SAHA, vorinostat Zolinza CTCL PO HDAC inhibitor Romidepsin Istodax CTCL IV HDAC inhibitor 5-Azacitidine Vidaza MDS IV DNMT inhibitor 󸀠 5-Aza-2 -deoxycytidine Dacogen MDS IV DNMT inhibitor CTCL: cutaneous T cell lymphoma. MDS: myelodisplastic syndromes.

Table 2: Epigenetic modifier drugs in noncancer clinical trials (partial list). Drug Disease NCT number Phase 5-Azacitidine Beta thalassemia NCT00005934 Phase 2 5-Azacytidine + Na phenylbutyrate Thalassemia major NCT00007072 Phase 2 Resveratrol Cardiovascular diseases NCT01449110 Phase 2 Resveratrol Trauma NCT01321151 Phase 1 Resveratrol Metabolic syndrome, obesity NCT01150955 Phase 1 Curcumin Irritable bowel syndrome NCT00779493 Phase 4 Curcumin Alzheimer disease NCT00164749 Phase 2 Curcumin Psoriasis NCT00235625 Phase 2

epigenome [13–18]. Only four drugs are currently utilized as Table 3: Enzymes involved in epigenetic modifications. single agents or in combination for the therapy of Myelodys- Enzyme 𝑁 Function plastic Syndrome and Cutaneous T Cell Leukemia and other hematological disorders (Table 1). However, at the end of DNA methyltransferase (DNMT) 5 2014, 195 open trials based on or including epigenetic drugs Histone Acethyltransferase (HAT) 19 Writer were listed in the database of the clinical trials of the NIH Histone methyltransferase (HMT) 41 Histone deacethylase (HDAC) 13 (https://clinicaltrials.gov/). Noticeably, several of these ongo- Eraser ing trials are not aimed at treating cancer diseases (Table 2) Histone demethylase (KDM) 26 indicating that the potential therapeutic use of “epigenetic Methyl binding proteins (MBD1) 5 Proteins that recognize and react to specific drugs” is extending beyond the boundaries of cancer. 𝑁 Readers modified histone residue In this review we will discuss the interaction between > theenvironmentandtheepigenomeandhownaturaland Total(April2012) 109 synthetic molecules that modulate epigenetic factors can have preventive properties against cancer. The number of epigenetic modifier proteins is steadily increasing: in 2009 they were “only” 91 and in 2012 more than 2. Mechanistic Aspects of 109 different proteins were identified. Epigenetic Regulation Epigenetic inheritance occurs through four basic layers 2.1. DNA Methylation and Hydroxymethylation. DNA methy- deeply interconnected: lation is the most widely studied epigenetic alteration and was the first one to be linked to cancer [19]. The only (1) DNA methylation and hydroxymethylation, biologically relevant C-methylation occurs at CpG doublets and is mediated by DNA methyltransferases (DNMTs) that (2) histone modification, catalyze the addition of a methyl group at C-5 of the cytosine. (3) chromatin remodeling, It is generally accepted that only DNMT1, DNMT3a, and DNMT3b are capable of C-methylation and that DNMT1 is (4) ncRNA. responsible, mainly but not exclusively, of the maintenance of DNA methylation throughout development and cell life, thus Epigenetic modifications are controlled by a set of enzymes preserving genomic integrity [20]. DNMT3a and DNMT3b whose functions can be summarized as follows: the “writers” are generally accepted as de novo DNMT for setting DNA are the enzymes that modify their target by adding residues methylation patterns. However, de novo and maintenance (i.e., methyl groups to DNA or histones); the “erasers” DNMTs are not tightly compartmentalized. The bulk of DNA remove the added residues and the “readers” are the proteins methylation is carried out by DNMT1 which is also capable that recognize and bind to the modified targets and act of de novo methylation. On the other hand, DNMT3a and as intermediate for subsequent protein-protein interactions DNMT3b can methylate the CpGs missed by DNMT1 at (Table 3). the replication fork thus serving as “maintenance DNMTs.” BioMed Research International 3

7 The human genome contains approximately 3 × 10 CpGs Table 4: Effect of histone modifications on transcription in mam- and although methylation at single doublets may, in princi- mals. ple, have functional consequences [21, 22], the biologically Modification Histone Site Effect relevant DNA methylation is occurring at CpG cluster (CpG H2A Lys5 Activation islands) [23, 24]. These clusters can be localized at the gene promoter and their methylation is inversely associated Lys5 Activation Lys12 Activation with transcription. CpG islands can be also intergenic or H2B intragenic; these clusters are generally hypermethylated to Lys15 Activation prevent spurious initiation particularly at internal promoters. Lys20 Activation An example of inappropriate transcriptional activation is that Lys4 Activation of ΔN-p73, oncogenic and antiapoptotic variants of the p73 Lys9 Activation gene from an internal promoter kept silenced by methylation Acetylation Lys14 Activation H3 of a small CpG island [25–27]. Lys18 Activation DNA methylation is deeply altered in cancer cells that Lys23 Activation present diffuse hypomethylation along with focal hyperme- Lys27 Activation thylation of selected genes or regions of the genome. The Lys5 Activation general idea is that hypomethylation contributes to genomic Lys8 Activation instability while hypermethylation inactivates tumor sup- H4 pressor genes. However, the picture is likely much more Lys12 Activation complex and methylation changes outside promoter regions Lys16 Activation mayhaveunexpectedeffectsongeneexpression[28, 29]. H1 Lys26 Repression Aberrant DNA methylation is not restricted to cancer and Lys4 Activation is present also in disorders of imprinting like the Beckwith- Arg8 Repression Wiedemann syndrome [30], the Prader-Willi syndrome, the Lys9 Repression ICF syndrome [31], and other neurodevelopmental disorders H3 Arg17 Activation [32]. Methylation Lys27 Repression 5-Hydroxymethylcytosine (5-hmC) is considered as the Lys36 Activation “sixth base” of the genome, with 5-methylcytosine (5-mC) being the “fifth,” and is an intermediate in the removal of Lys79 Activation Lys20 Repression methyl groups from cytosine by the TET 1 enzymes. 5-hmC H4 has the opposite function of 5-mC and is a transcriptional Lys59 Repression activator [33, 34]. It is not clear if the mechanism through H1 Ser27 Activation which 5-hmC activates transcription is the removal of the H2A Ser1 Repression Phosphorylation methyl group with the consequent displacement of the MDB Ser10 Activation H3 or if TET proteins act as “readers” preventing the binding of Ser28 Activation DNMT. Ubiquitylation H2B Lys120 Activation

2.2. Histone Modifications. DNA is tightly compacted around an octamer of histones forming a structure named “nucleosome”whichisthebasicunitofthechromatinandincludes The effects on transcription of some of these modifications are 146 bp of DNA wrapped around a disk-like structure com- reported in Table 4. Histone modifications occur at specific posedoftwomoleculeseachofH2A,H2B,H3,andH4.A sites and in various combinations and, along with the dis- 80 bp linker DNA and a fifth histone (H1) separate adjacent covery of specific functions of histone variants, generated the nucleosomes. The electron microscopy appearance of the hypothesis of the “Histone Code” that postulates that distinct chromatin is that of a “beads-on-a-string” whose packaging modifications, alone, or in combination or sequentially can is determined by the histones. bereadbyeffectorproteinstobringaboutdownstreamevents Histones for many years were considered simply as [38]. structural proteins whose function was to assemble DNA Initially, the transcriptional control mechanism of histone into chromosomes. Only in the 1960s Allfrey at al. [35, 36] modifications was considered to be primarily “mechanic.” showed that histones undergo postsynthetic modifications The transfer of acetyl groups mediated by Histone Acetyl- that are related to the control of transcription. Because of transferases (HAT) to the lysines of histone tails neutralizes his pioneeristic work, Allfrey must be considered one of the thepositivechargeoftheAAandresultsintheweakeningof fathers of epigenetics. Beside the five major histones, several the interaction with DNA [39, 40] and in a “relaxed” chro- variants with unique distribution patterns and functions not matin conformation. Deacetylation of histones by histone exerted by the “classic” histones have been described [36, 37]. deacethylases (HDAC) restores the positive charge of lysines The N-terminal tail of histones is subjected to various types of and the “closed” chromatin conformation. It is believed that modification including acetylation, methylation, phosphory- the opening of the chromatin is a key step for the recruitment lation, ubiquitination, ADP-ribosylation, and biotinylation. of the transcription machinery; thus, HAT and HDAC are 4 BioMed Research International

Table 5: Principal ncRNAs. ncRNA Length (bp) Function miRNA 21–23 mRNA targeting siRNA 20–25 Targeting of specific genes by sequence complementarity piRNA 27–30 Chromatin regulation and transposon silencing XiRNA 24–42 Control X chromosome methylation and inactivation Long ncRNA >200 Various, including targeting of specific genes transcriptional activators and silencers, respectively. Simi- processing and function of other RNAs thus intervening larly, the phosphorylation of serine, threonine, and tyrosine in complex pathways and cell mechanisms. The principal mediated by kinases and phosphatases changes the net charge ncRNA and their functions are reported in Table 5. of histones contributing to the changes of the chromatin In general ncRNAs interfere with the functionality of structure. other RNAs through a mechanism called RNA interference Also the methylation of histones occurring at lysine (RNAi). RNAi regulates gene expression in a sequence- and arginine was discovered by Allfrey et al. in 1964 [40]. specific way without altering the target sequence; accordingly Differently from acetylation and phosphorylation, methy- this mechanism is considered epigenetic. The role of micro- lation does not change the protein charge and does not RNAs as epigenetic controllers is well recognized and this modify the interaction between DNA and histones. For many class of ncRNA is the most studied in this respect. miRNAs years histone methylation was considered an irreversible inactivate transcription by base pairing of nucleotides 2– 󸀠 modification and only in 2004 the first histone demethylase 8(the“seed”regionofmiRNA)andthe3UTR of the (KDM), the LSD1 (KDM1) amine oxidase, was identified [41]. mRNA. This leads to premature degradation or to the stop of Since then many other KDM genes were identified and they translation through the formation of the “silencing complex” represent a complex and expanding family involved in many [53–55]. miRNAs are involved in a variety of biological pro- aspects of cell physiology and pathology [42–44]. Several cesses including cell growth and differentiation, apoptosis, KDM genes act as oncogenes or antioncogenes, are involved maintenance of cell identity and they are deregulated in in anticancer drug response, and have been proposed as cancer and other diseases [56]. A miRNA can have multiple therapeutic targets [12, 43, 45, 46]. The effect of methylation targets and a gene can be targeted by many miRNAs; thus, of histone tails on transcription is that of recruiting effector the regulatory pathway determined by these molecules can proteins with promoting or inhibiting properties. be extraordinary complex. Interestingly, miRNAs expression can be modulated by 2.3. Chromatin Remodeling. The tight packaging of chro- DNA methylation and histone modifications and they can matin prevents the binding of transcription factors and RNA modulate the DNA methylation machinery leading to a polymerases. Chromatin remodeling can be obtained not “loop” of epigenetic regulation [57–59]. only by histone acetylation/deacetylation but also through Importantly and interestingly, exogenous miRNA ATP-dependent protein complex formation, histone modi- sequences may enter the cells from the outside and exert their fication by polycomb proteins (PcG), and by interaction of biological effect [60] demonstrating that the environment noncoding RNA (ncRNA). canhaveprofoundeffectsontheepigenome. Nucleosomes can be repositioned by the SWI/SNF ATP- dependent complex through the formation of a DNA loop 3. Influence of the Environment on that moves repositioning the nucleosome and reduces the distance between adjacent nucleosomes [47, 48]. This process Epigenetic Marks may activate transcription in regions where nucleosomes are Nucleated somatic cells of the human body contain approx- relaxed and more distant and can silence the regions where imately 3 billion bp of DNA that, in length, correspond to chromatinismorecompact. a filament of roughly 3 meters. Although tightly packaged PcG proteins repress transcription compacting the chro- in chromatin and hidden in the nucleus, DNA is exposed matin through DNA methylation and multiple histone mod- to a variety of agents that may influence epigenetic regula- ifications following a multistep protein recruitment that tion. DNA methylation is considered a stable modification; prevents mRNA elongation [49, 50]. however, it is sensitive to multiple agents including those that ncRNA can influence transcription by targeting PcG reduce the bioavailability of S-adenosylmethionine, the major proteins to specific sets of genes [51]. The interaction between methyl donor involved in DNA methylation [61]. Intuitively, ncRNA and PcG is particularly important in chromosome X the highly dynamic histone modifications are very sensitive to inactivation. A ncRNA (Xist) is expressed from the noncod- the environmental changes since each of the many enzymes ing X chromosome, recruits a PcG complex denoted as PRC2, involved in these processes is a potential target of epigenetic and initiates silencing by binding the X chromosome [52]. regulation. Moreover, drugs and dietary compounds and the exposure to contaminants that pass the placenta can 2.4. ncRNA. ncRNA are a class of RNA that are not translated produce subtle alterations of the developmental pattern of into proteins, whose function is that of controlling the the fetus [62, 63]. According to the “developmental origin BioMed Research International 5 of adult health and disease” hypothesis [64, 65], exposure pathways by chemicals (denoted by “Endocrine Disrupting to chemicals or natural bioactive products during pregnancy Chemicals” (EDCs)) that mimic the action of the natural may have long-term effects on health also through epigenetic molecules [80]. mechanisms. Considering that these chemicals are utilized since the 1940s and that some of them, like biphenyls, are extremely resistant to degradation while the metabolites of others, like 3.1. Epigenetic and Lifestyle. The environmental agents that DDT, are biologically very active, it is not surprising that they can interfere with DNA methylation are widespread and are considered to constitute a health hazard. depend on lifestyles. Smoking, alcohol consumption, UV The effect of EDC has been considered to be transgenera- light exposure, or factors linked to oxidative stress are some tional and it is believed that these molecules could affect also of the most common and important lifestyle aspects that may the offspring through epigenetic mechanisms81 [ , 82]. alter the DNA methylation profile [66]. Epigenetic inheritance has been advocated for many TheeffectsofsmokeonDNAmethylationhavebeen pathologic conditions and many experimental evidences have extensively studied but the results obtained are contradic- been accumulated in support of this hypothesis [83–86]. The tory and no consensus has yet been reached. Indeed, if concept of transgenerational epigenetic inheritance is very Benzo[a]pyrene has no effect on the methylation of the DNA attractive and it has been demonstrated for plants and lower sequences known to be involved in lung cancer [67], smoke organisms like nematodes. However, recent surveys of the condensate containing the full spectrum of carcinogenic literature have raised several questions in particular on the substances found in cigarettes modifies the methylation role of the environment in epigenetic inheritance and on its patterns of tumor suppressor genes involved in the early extent in mammals [87, 88]. stages of lung carcinogenesis [68, 69]. One of the most common EDCs found in the environ- Experimental studies failed to demonstrate the direct ment is Bisphenol A (BPA) a compound used in the manu- carcinogenic activity of alcohol; however, the causal relation facture of polycarbonate and epoxy resin. The most common between alcohol intake and cancer has been established mode of assumption of BPA in humans is through canned several years ago [70]. In a Dutch study on diet and cancer it foodsinceepoxyresincoatingfoodcansandreleasesBPA.It has been demonstrated that the methylation pattern of tumor hasbeenestimatedthattheintakefromfoodofthischemical suppressor and DNA repair genes was altered in colorectal is approximately 7 𝜇g/person/day. BPA and its derivatives can cancer patients with low folate/high alcohol intake [71]. How- be easily detected in bodily fluids at bioactive concentrations ever, a subsequent study failed to show the significant associa- [80]. Several reports describe the effect of BPA on epigenetic tion between folate intake, alcohol and methylation of MLH1, programming and demonstrate that this chemical can alter a gene hypermethylated early during colon carcinogenesis the expression of selected genes through histone methylation, [72]. It must be pointed out that these studies were conducted DNA methylation changes, and miRNA expression. The using qualitative methylation analysis techniques that may exposuretoBPA,indevelopmentalin vivo models, increases notbesufficientlyaccuratetodisclosesubtle,butclinically the susceptibility to prostate cancer and it is likely involved relevant differences of methylation. In this respect it has been in breast carcinogenesis [89–91]. In another experimental demonstrated that quantitative methylation analysis offers model, the continuous exposure of mice before mating and significant advantages over purely qualitative techniques and during gestation and lactation produced hypomethylation of enables to identify methylation cut-off values that define the DNA, obesity, and diabetes and increased tumorigenicity disease-specific methylation patterns73 [ –78]. in the offspring92 [ ]. Interestingly, the epigenetic alterations Alcohol-related epigenetic changes influence neuronal inducedbyBPAcouldbecounteractedbydietarysupplemen- growth and development and affect memory and learning tation with methyl donors [92]. possibly modifying the methylation status of many genes. Diethylstilbestrol (DES), a synthetic estrogen antagonist In particular, a genome-wide methylation analysis, fol- widely utilized between 1938 and 1971, induces vaginal cancer lowed by stringent validation by quantitative methodologies, in the offspring of the treated mothers93 [ ]. Persistent showed remarkable differences in 84 genes involved in brain epigenetic changes that can be passed to the next generation metabolism and differentiation79 [ ]. This study indicated can be induced by DES treatment [94, 95]. Among the genes that exposure to alcohol of early mouse embryo changed the whose expression and/or methylation is altered by in utero methylation pattern and the expression of several genes and exposure to DES, two are particularly relevant for cancer these alterations were put in relation with neural tube defects. development: the EZH2 histone methyltransferase involved Interestingly, the human homologues of many of the genes inbreastcancer,inglioblastoma,andinothertumors[96, 97] epigenetically modified by acute alcohol administration in and the HOXA10 gene, involved in cell stemness and in mice are involved in neurologic pathologies like Alzheimer glioblastoma [75, 98]. disease (APP), ALS-Parkinson (TRPM7), myotonic dystro- Beside DTT, other insecticides and pesticides, like phy (DMPK), Angelman syndrome (UBE3A), and others. Methoxychlor, Vinclozolin, and the widely used Permethrin, are widespread environmental contaminants and EDC; some 3.2. Endocrine Disrupting Chemicals in the Environment of them have various effects on the epigenome since they can and Epigenetics. A particularly serious and often poorly alterDNAmethylationaswellasthelevelofexpressionof considered effect of the accumulation of chemicals in the the DNA methyltransferase DNMT3B [82, 99–101]. Several environment is the alteration of the endocrine signaling genes, including the paternally imprinted H19, Gtl2,and 6 BioMed Research International

Diet overcome nutritional deficiencies. As discussed earlier, the Dietary folate Dietary cathecol- polyphenols maternal exposure to epigenetic modifiers can affect the Vitamins Methionine transmission of epigenetically controlled traits. In principle, Dietary proteins Vitamin B6 B2, B6, B12 Zn, selenium the same effect can result from the administration with the Folate pathway SAM pathway diet of nutrients affecting the epigenome. Indeed, in mice, the addition of folic acid and choline, both involved in one- carbon metabolism, can increase the methylation of IGF2 Selenium [114], an imprinted gene whose loss of imprinting is involved genistein DNMT in colon cancer [115], and can alter histone modification polyphenols possibly through the increase of histone methyltransferase expression [116]. In humans, understanding the effect on the DNA methylation offspring’s epigenome of folic acid administration to pregnant women is complicated by the widespread utilization of folic Figure 1: One-carbon metabolism and DNA methylation schematic acid as food supplement. representation of the interaction of dietary components on folate In animals, the effect of maternal diet on epigenetic marks and SAM pathways and their effect on DNA methylation. has been widely studied and the results showed the complexity of the interplay between many different components [109]. In humans the impact of the diet on developmental repro- Meg3 and the maternally imprinted Peg1, Peg3,andSnrp gramming has never been formally proven; however, the genes, are affected through altered CpG methylation levels effect of famine on the imprinting of the IGF2 locus has been in exposed animals. Interestingly, the effect persisted through evaluated in 60 individuals of the “Hunger Winter Families three generations although diminished from F1 to F3 [102]. Study” [117] conceived in Holland and in their siblings and in matched controls. The results of this study showed that the methylation level of IGF2 was significantly lower in the 3.3. Nutrition and Epigenetics. The effect of the dietary habits individuals periconceptionally exposed to famine compared on health and disease prevention has been the subject of many to the controls [118] supporting the hypothesis that epigenetic investigations and a proper and balanced diet is one of the changes occurred in the early-life can be maintained during key points of a “good lifestyle.” Through food we are exposed adulthood. However, the phenotypic consequences of the daily to many toxic substances like, for example, BPA. At the periconceptional exposure to epigenetic-modifying condi- same time certain components of the diet can modify the tions are not known. epigenetic pattern through natural bioactive components that Folic acid and vitamins B2, B6, and B12 are not the only can act on DNA methylation or histone modification [103– dietary components acting on the epigenome. Polyphenols, 106] or, as in the case of exogenous miRNA, that can be direct found in green tea and vegetables, are a group of natural com- epigenetic actors [60]. In a recent survey on the effects of pounds that do not modify the availability of methyl groups dietary compounds on the basic mechanisms of epigenetics, but interfere with the activity of DNMTs, HATs, and HDACs it has been recognized that many vegetable components have [119, 120]. Polyphenols have gained a particular interest when detectable activities on HAT in human subjects [107, 108]. it was shown that they could revert malignancy-associated Several nutrients in the diet have a key role in the epigenetic alterations in cell lines [121–123] and that they methylation of all biological substrates and can influence could modulate DNA methylation in humans [124, 125]. DNA methylation either by changing the availability of Phytoestrogens are another group of naturally occurring methyl donors or by modulation of the DNMTs activity molecules found predominantly in soybeans that inhibit (Figure 1)[109]. In particular, folic acid and Vitamins B6 DNMT1, DNMT3a, and DNMT3b and HDAC that were and B12 are essential for the one-carbon metabolism and hypothesized to reduce the risk of hormone-related tumors their insufficient dietary intake (or an excess) can alter the [126–128]. availability of S-adenosylmethionine, the methyl donor for DNA methylation. Folic acid prevents neural tube defects Selenium is an essential component of the diet that andin1998hasledtheFDAtorequirethatcertainfoods can interfere with the epigenome by inhibiting DNMT and and dietary supplements are enriched with folates. The in HDAC (Figure 1). Importantly, selenium can reactivate the vivo effect of folic acid administration on DNA methylation expression of genes involved in the response to oxidative has been well documented and intake or deprivation of folic stress and to protection from carcinogen [129, 130]. acid in colon cancer and healthy individuals resulted in The effects of nutrition at the early stages of gestation increase or decrease of methylation in colon mucosa and and in the early postnatal life are part of the complex field lymphocytes, respectively [110–113]. However, the impact of of disease chemoprevention that will be discussed in another exogenous folate administration and of the modulation of part of this review. DNA methylation in humans has not been fully determined as it will be discussed in more details in another section of 3.4. Psycho-Epigenetics. The early-life experiences are con- this review. In particular, it is not clear how dosages and the sidered the building blocks for adulthood and their impor- length of the treatment affect DNA methylation and if there is tanceiswellrecognizedbypsychologists.Alterationofthe atime-window,in utero, when treatment could significantly methylation levels of neuronal glucocorticoid receptors in BioMed Research International 7 hippocampus of suicide victims was put in relation with Inhibitors of HDAC and of DNA synergize together and childhood abuse and similar results were obtained in rats are being utilized in combination in clinical trials. TSA exposed to stress [131, 132]. and SAHA cannot reactivate hypermethylated genes unless Although all these data support the concept that epige- a minimal demethylation is achieved by DNMT inhibition netic mechanisms are regulated, by dietary and environmen- [144]. tal factors, and may have an influence on health and behavior, thefunctionalconsequencesofthistypeofcontrolarelargely 4.2. Natural Bioactive Food Components. We have previ- incomplete and a task for future research will be to link, at ously seen that some diet components (folates, Vitamins B2, the genome-wide and population level, epigenetic changes, B6, and B12) are essential components of the one-carbon expression of the genes, and the resulting phenotype. metabolisms and are involved in global DNA methylation. The cancer preventive effects of folates are, likely, tissue specific and age dependent. In rats, maternal supplementation 4. Role of Epigenetics in Chemoprevention of folic acid increases the methylation in the colon of the offspring and reduced by 64% the risk of colorectal cancer [145]. The possibility of modulating the epigenome through the However,thesametreatmentreducedtheoverallmethylation diet opens the possibility of utilizing bioactive molecules for in mammary glands and doubled the risk of mammary the prevention of diseases characterized by the alteration carcinoma [146]. Folic acid can increase overall methylation of the epigenetic asset. Intuitively, the utilization of syn- in liver; however, this effect seems to be age-related occurring thetic or natural molecules with hypothetic chemopreventive in mice older than 18 months but not in those below 4 months properties was pioneered mainly for cancer where the effect of age [147].Trialsinhumansubjectswerecontradictory. of demethylating the genome is beneficial133 [ ]. Indeed, In a randomized trial, folic acid supplementation did not along with the generalized hypomethylation that, supposedly, changetherecurrencerateofcolorectaladenomainpatients activates genes with oncogenic properties, cancer cells have with a previous history of adenoma [148]. In another study a distinct hypermethylation profile that frequently involves conducted on healthy subjects, folic acid administration for 3 genes with antioncogenic properties. The rationale for a years resulted in the persistent methylation of cancer-related hypomethylating therapy is to reactivate these dormant genes genesinthecolon[149]. tocounteracttheeffectoftheactiveoncogenes.Thebasic Folates and related drugs interfere with global methyla- concept of chemoprevention is to prevent the initiation or tion, and other molecules present some specificity in their progression of premalignant lesions through the administra- action. Polyphenols, like epigallocatechin-3-gallate (EGCG), tion of synthetic compounds or through food additives. Ide- curcumin, and resveratrol are three molecules widely dis- ally, chemoprevention should be focused toward individual at tributed in vegetables and other dietary components that risk either familiar or because they are exposed to carcinogen. are being studied for their anticancer properties [120]. Epigenetic chemoprevention is particularly attractive because Polyphenols exert their activity through several mechanisms: epigenetic alterations are reversible and are an early event in carcinogen detoxification, DNA repair, cell cycle progres- cancer development [134, 135]. sion, activation of differentiation, and epigenetic modulation. The major drawback of epigenetic therapies is the lack EGCG is active against many types of cancer cells [150], of specificity and the global hypomethylation achieved by demethylates and reactivates several genes involved in cancer DNMT inhibitors might be highly detrimental. Indeed, repet- like p16, RAR𝛽, MGMT, hMLH1, and GSTP1.Although itive elements and cryptic internal promoter are normally EGCGseemstobeacompetitiveinhibitorofDNMT1[123], it kept hypermethylated but can be reactivated by epigenetic can also have apparent demethylating activities or be inactive drugs. This leads to genomic instability and to the illegitimate in some biological systems [151, 152]. transcription of genes that, at a certain stage of development, Genistein, a phytoestrogen extracted from soybean, can should not be functional [136–139]. interfere with multiple pathways through hormone receptors Thus, DNMT inhibitors, along with recognized anti- and has the same activities of EGCG on p16, RAR𝛽, and cancer activity, can promote oncogenic transformation that MGMT [153]. Many in vitro studies have documented the could be counteracted only with more specific drugs or by antiproliferative and anticancer activities of genistein par- combination therapies targeting other epigenetic determi- ticularly, but not exclusively, in prostate and breast cancer nants and pathways [140]. [154, 155]. In vivo model studies have shown that genistein can modify methylation levels in the prostate of mice [156]and 4.1. Synthetic Drugs. Synthetic inhibitors of DNMT and that, overall, epigenetic changes could be detected by genis- HDAC are being utilized mainly as therapeutic drugs rather tein treatment; however, the results on the anticancer activity than as chemopreventive agents in hematopoietic tumors of this compound were contradictory. Genistein seems to [141–143]. These molecules inhibit DNA methylases by have protective effects against induced carcinogenesis [157] incorporating into DNA (5-Azacytidine and 5-Aza-2-deox- but it was also observed that genistein could even stimulate ycytidine) or inhibit histone deacethylase (SAHA and TSA). tumor growth through hormone receptors [153]. Moreover, To be active, nucleoside analogues require their incorpo- the plasma concentration required for these effects is unlikely ration into DNA; for this reason, they are more effective on to be reached by dietary intake. It must be stressed that these rapidly proliferating cells rather than on quiescent or nearly studies were conducted in models using cell lines that may quiescent cells. notberepresentativeofthedisease. 8 BioMed Research International

Retinoids interfere in the DNA methylation mechanisms APP: Amyloid beta (A4) precursor protein intervening in the one-carbon metabolism upregulating BPA: Bisphenol A the glycine-N-methyltransferase and their cancer preventive DES: Diethylstilbestrol activity has been extensively studied [158, 159]. Retinoids DNMT: DNA revert the cancer phenotype of breast and promyelocytic DMPK: Dystrophia myotonica-protein kinase leukemia by demethylation and reactivation of the RAR𝛽2 EDC: Endocrine disrupting chemicals receptor [160–162] and a recent epigenome-wide analysis EGCG: Epigallocatechin-3-gallate identified a subset of genes, including stem cell genes, HAT: Histone Acethyltransferase selectively modulated by retinoids [163]. ICF syndrome: Immunodeficiency, centromeric instability, Retinoids, polyphenols, and fatty acids can target the and Facial dysmorphism syndrome polycomb transcriptional repressor complexes that partic- KDM: Lysine histone demethylase ipate in the epigenetic silencing through methylation of MDB: Methyl DNA binding proteins Histone 3 (H3K27me3) [164]. Retinoic acid displaces PcG MLH1: mutL homolog 1 from its target genes including HOXA1 and RAR𝛽2 [165], miRNA: Micro-RNA whereas ECGC reduces the expression of the PcG compo- ncRNA: Noncoding RNA nents BMI1 and EZH2, slows the proliferation of cancer PcG: Polycomb proteins cells, and promotes apoptosis. A similar effect is exerted by RNAi: RNA interference curcumin [166, 167]. SAHA: Suberanilohydroxamic acid The inhibition of breast cancer cells growth was observed SWI/SNF: SWItch/sucrose nonfermentable complex by dietary omega-3 fatty acids treatment and was related to TET 1: Ten-eleven translocation methylcytosine the downmodulation of EZH2 [168]. dioxygenase 1 enzymes TRPM7: Transient receptor potential cation channel, member 7 5. Conclusions TSA: Trichostatin A UBE3A: Ubiquitin protein ligase E3A We have summarized the effects of environmental and of Xist: X-inactive specific transcript (ncRNA). some lifestyle factors on epigenetic regulatory processes and we have reported the results of studies showing that some dietary components may have a role in cancer prevention. Conflict of Interests Many environmental factors, including mother’s diet and chemical pollution, can induce epigenetic alterations at con- The authors declare that there is no conflict of interests ception or later in the uterine life. However, the phenotypic regarding the publication of this paper. consequences of these early-life modifications should be still disclosed in full. Moreover, although epigenetic marks are Acknowledgments considered inheritable, transgenerational epigenetic inheritance can be the result of confounding mechanisms such as This study was supported by the “Fondazione Compagnia di cryptic genomic variations, behavioral or microbial effects. San Paolo-Torino,” by the Italian Ministry of Health, Core Overall, many experimental evidences indicate that cer- Grant “Ricerca Corrente” to the IRCCS AOU San Martino- tain food components may interfere with cancer development IST, and by funds “5 × 1000.”Barbara Banelli is the recipient of and growth through epigenetic mechanisms. However, much the “Young Investigators” Grant GR-2008-1143408 from the work is needed to efficiently translate the promising results Italian Ministry of Health. obtained in vitro andinanimalmodelstothehealthsystem. In this respect, the antineoplastic properties of folates were References not confirmed in clinical trials [148]. Moreover, it remains to be precisely established if the oral administration of [1] R. G. Gosden and A. P. Feinberg, “Genetics and epigenetics— unsupplemented food is sufficient to gain clinically relevant nature’s pen-and-pencil set,” The New England Journal of dosages of bioactive compounds, if there is a preferred timing Medicine,vol.356,no.7,pp.731–733,2007. of the life when the addition of food components could be [2] C. M. Rivera and B. Ren, “Mapping human epigenomes,” Cell, effective and the extent and the duration of the biological vol.155,no.1,pp.39–55,2013. effect. In conclusion, the now standardized technologies of [3] J. D. Choi and J. S. Lee, “Interplay between epigenetics and epigenome analysis coupled with whole-genome expression genetics in cancer,” Genomics & Informatics,vol.11,no.4,pp. studies are required to determine the biological and clinical 164–173, 2013. impact of environmentally induced epigenetic modifications [4] H. Easwaran, H.-C. Tsai, and S. B. Baylin, “Cancer epigenetics: in humans. tumor heterogeneity, plasticity of stem-like states, and drug resistance,” Molecular Cell,vol.54,no.5,pp.716–727,2014. List of Abbreviations [5] R. M. Campbell and P. J. Tummino, “Cancer epigenetics drug discovery and development: the challenge of hitting the mark,” 5-hmC: 5-Hydroxymethylcytosine The Journal of Clinical Investigation,vol.124,no.1,pp.64–69, 5-mC: 5-Methylcytosine 2014. BioMed Research International 9

[6]R.R.Kanherkar,N.Bhatia-Dey,andA.B.Csoka,“Epigenetics [23] J. G. Herman and S. B. Baylin, “Gene silencing in cancer in across the human lifespan,” Frontiers in Cell and Developmental association with promoter hypermethylation,” The New England Biology,vol.2,article49,2014. Journal of Medicine,vol.349,no.21,pp.2042–2054,2003. [7] S. A. Tammen, S. Friso, and S.-W. Choi, “Epigenetics: the link [24] A. M. Deaton and A. Bird, “CpG islands and the regulation of between nature and nurture,” Molecular Aspects of Medicine,vol. transcription,” Genes and Development,vol.25,no.10,pp.1010– 34, no. 4, pp. 753–764, 2013. 1022, 2011. [8] A. J. Bernal and R. L. Jirtle, “Epigenomic disruption: the effects [25] I. Casciano, B. Banelli, M. Croce et al., “Role of methylation in of early developmental exposures,” Birth defects research Part A: the control of ΔNp73 expression in neuroblastoma,” Cell Death Clinical and molecular teratology,vol.88,no.10,pp.938–944, and Differentiation,vol.9,no.3,pp.343–345,2002. 2010. [26] I. Casciano, K. Mazzocco, L. Boni et al., “Expression of ΔNp73 [9]L.S.Rozek,D.C.Dolinoy,M.A.Sartor,andG.S.Omenn, is a molecular marker for adverse outcome in neuroblastoma “Epigenetics: relevance and implications for public health,” patients,” Cell Death and Differentiation,vol.9,no.3,pp.246– Annual Review of Public Health,vol.35,pp.105–122,2014. 251, 2002. [10] M. Ivanov, M. Kacevska, and M. Ingelman-Sundberg, “Epige- [27] M. Romani, G. P.Tonini, B. Banelli et al., “Biological and clinical nomics and interindividual differences in drug response,” Clin- role of p73 in neuroblastoma,” Cancer Letters,vol.197,no.1-2, ical Pharmacology and Therapeutics,vol.92,no.6,pp.727–736, pp. 111–117, 2003. 2012. [28] J. F. Smith, S. Mahmood, F. Song et al., “Identification of 󸀠 [11] M. Rius and F. Lyko, “Epigenetic cancer therapy: rationales, DNA methylation in 3 genomic regions that are associated targets and drugs,” Oncogene,vol.31,no.39,pp.4257–4265, withupregulationofgeneexpressionincolorectalcancer,” 2012. Epigenetics,vol.2,no.3,pp.161–172,2007. [29] M. Weber, J. J. Davies, D. Wittig et al., “Chromosome-wide and [12]S.V.Sharma,D.Y.Lee,B.Lietal.,“Achromatin-mediated promoter-specific analyses identify sites of differential DNA reversible drug-tolerant state in cancer cell subpopulations,” methylation in normal and transformed human cells,” Nature Cell,vol.141,no.1,pp.69–80,2010. Genetics,vol.37,no.8,pp.853–862,2005. [13] R.M.Glasspool,J.M.Teodoridis,andR.Brown,“Epigeneticsas [30] S. Azzi, W. A. Habib, and I. Netchine, “Beckwith-Wiedemann a mechanism driving polygenic clinical drug resistance,” British and Russell-Silver Syndromes: from new molecular insights to Journal of Cancer,vol.94,no.8,pp.1087–1092,2006. the comprehension of imprinting regulation,” Current Opinion [14] H. M. Strik, J.-H. Buhk, A. Wrede et al., “Rechallenge with in Endocrinology, Diabetes and Obesity,vol.21,no.1,pp.30–38, temozolomide with different scheduling is effective in recurrent 2014. malignant gliomas,” Molecular Medicine Reports,vol.1,no.6,pp. [31] E. Walton, C. Francastel, and G. Velasco, “Dnmt3b prefers 863–867, 2008. germ line genes and centromeric regions: lessons from the ICF [15] J.R.Perry,P.Rizek,R.Cashman,M.Morrison,andT.Morrison, syndrome and cancer and implications for diseases,” Biology, “Temozolomide rechallenge in recurrent malignant glioma vol. 3, no. 3, pp. 578–605, 2014. by using a continuous temozolomide schedule: the ‘rescue’ [32] S. Rangasamy, S. R. D’Mello, and V. Narayanan, “Epigenetics, approach,” Cancer, vol. 113, no. 8, pp. 2152–2157, 2008. autism spectrum, and neurodevelopmental disorders,” Neu- [16]X.Chen,Y.Pan,S.Zhangetal.,“Rechallengewithgefitinib rotherapeutics,vol.10,no.4,pp.742–756,2013. following severe drug-induced hepatotoxicity in a patient with [33] T. F. Kraus, V. Guibourt, and H. A. Kretzschmar, “5-hydro- advanced non-small cell lung cancer: a case report and literature xymethylcytosine, the ‘sixth base’, during brain development review,” Oncology Letters,vol.7,no.3,pp.878–880,2014. and ageing,” Journal of Neural Transmission,2014. [17]D.Gasparotto,G.Miolo,E.Torrisietal.,“Improvedout- [34] J. Wang, J. Tang, M. Lai, and H. Zhang, “5-Hydroxyme- come with multimodal treatment and imatinib rechallenge in thylcytosine and disease,” Mutation Research/Reviews in Muta- advanced GIST,” International Journal of Colorectal Disease,vol. tion Research,vol.762,pp.167–175,2014. 29,no.5,pp.639–640,2014. [35] V. G. Allfrey and A. E. Mirsky, “Structural modifications of [18] D. Matei, F. Fang, C. Shen et al., “Epigenetic resensitization to histones and their possible role in the regulation of RNA platinum in ovarian cancer,” Cancer Research,vol.72,no.9,pp. synthesis,” Science,vol.144,no.3618,p.559,1964. 2197–2205, 2012. [36] G. Vidali, E. L. Gershey, and V. G. Allfrey, “Chemical studies of [19] A. D. Riggs and P. A. Jones, “5-methylcytosine, gene regulation, histone acetylation. The distribution of epsilon-N-acetyllysine and cancer,” Advances in Cancer Research,vol.40,pp.1–30,1983. in calf thymus histones,” TheJournalofBiologicalChemistry,vol. [20] F. Xu, C. Mao, Y. Ding et al., “Molecular and enzymatic profiles 243,no.24,pp.6361–6366,1968. of mammalian DNA methyltransferases: structures and targets [37] S. E. Polo, “Reshaping chromatin after DNA damage: the for drugs,” Current Medicinal Chemistry,vol.17,no.33,pp. choreography of histone proteins,” Journal of Molecular Biology, 4052–4071, 2010. vol. 427, no. 3, pp. 626–636, 2014. [21]C.J.Nile,R.C.Read,M.Akil,G.W.Duff,andA.G.Wilson, [38] C. M. Weber and S. Henikoff, “Histone variants: dynamic “Methylation status of a single CpG site in the IL6 promoter is punctuationintranscription,”Genes & Development,vol.28,no. related to IL6 messenger RNA levels and rheumatoid arthritis,” 7, pp. 672–682, 2014. Arthritis and Rheumatism,vol.58,no.9,pp.2686–2693,2008. [39] T. Jenuwein and C. D. Allis, “Translating the histone code,” [22] I. P. Pogribny, M. Pogribna, J. K. Christman, and S. J. Science,vol.293,no.5532,pp.1074–1080,2001. James, “Single-site methylation within the p53 promoter region [40]V.G.Allfrey,R.Faulkner,andA.E.Mirsky,“Acetylationand reduces gene expression in a reporter gene construct: possible methylation of histones and their possible role in the regulation in vivo relevance during tumorigenesis,” Cancer Research,vol. of Rna synthesis,” Proceedings of the National Academy of 60, no. 3, pp. 588–594, 2000. Sciences of the United States,vol.51,no.5,pp.786–794,1964. 10 BioMed Research International

[41] Y.Shi, F.Lan, C. Matson et al., “Histone demethylation mediated methyltransferases 3A and 3B,” Proceedings of the National by the nuclear amine oxidase homolog LSD1,” Cell,vol.119,no. Academy of Sciences of the United States of America,vol.104,no. 7, pp. 941–953, 2004. 40, pp. 15805–15810, 2007. [42] D. Pasini, A. P. Bracken, K. Agger et al., “Regulation of stem [60] L. Zhang, D. Hou, X. Chen et al., “Exogenous plant MIR168a cell differentiation by histone methyltransferases and demethy- specifically targets mammalian LDLRAP1: evidence of cross- lases,” Cold Spring Harbor Symposia on Quantitative Biology,vol. kingdom regulation by microRNA,” Cell Research,vol.22,no. 73, pp. 253–263, 2008. 1,pp.107–126,2012. [43] P. A. C. Cloos, J. Christensen, K. Agger, and K. Helin, “Erasing [61] B. Richardson, “Primer: epigenetics of autoimmunity,” Nature the methyl mark: histone demethylases at the center of cellular Clinical Practice Rheumatology,vol.3,no.9,pp.521–527,2007. differentiation and disease,” Genes and Development,vol.22,no. [62] V.Bollati and A. Baccarelli, “Environmental epigenetics,” Hered- 9, pp. 1115–1140, 2008. ity,vol.105,no.1,pp.105–112,2010. [44] K. Agger, J. Christensen, P. A. Cloos, and K. Helin, “The [63] A. Baccarelli and V. Bollati, “Epigenetics and environmental emerging functions of histone demethylases,” Current Opinion chemicals,” Current Opinion in Pediatrics,vol.21,no.2,pp.243– in Genetics and Development, vol. 18, no. 2, pp. 159–168, 2008. 251, 2009. [45] K. Helin and D. Dhanak, “Chromatin proteins and modifica- [64] P. D. Gluckman, M. A. Hanson, and T. Buklijas, “A conceptual tionsasdrugtargets,”Nature,vol.502,no.7472,pp.480–488, framework for the developmental origins of health and disease,” 2013. JournalofDevelopmentalOriginsofHealthandDisease,vol.1, [46] D. Rotili and A. Mai, “Targeting histone demethylases: a new pp. 6–18, 2010. avenue for the fight against cancer,” Genes and Cancer,vol.2, [65] M. A. Hanson and P. D. Gluckman, “Developmental origins of no. 6, pp. 663–679, 2011. health and disease: new insights,” Basic & Clinical Pharmacology [47] K. Havas, A. Flaus, M. Phelan et al., “Generation of superhelical &Toxicology,vol.102,no.2,pp.90–93,2008. torsion by ATP-dependent chromatin remodeling activities,” [66] G. J. Gorelik, S. Yarlagadda, and B. C. Richardson, “Protein Cell,vol.103,no.7,pp.1133–1142,2000. kinase C𝛿 oxidation contributes to ERK inactivation in lupus T [48]M.L.Dechassa,A.Sabri,S.Pondugulaetal.,“SWI/SNFhas cells,” Arthritis and Rheumatism,vol.64,no.9,pp.2964–2974, intrinsic nucleosome disassembly activity that is dependent on 2012. adjacent nucleosomes,” Molecular Cell,vol.38,no.4,pp.590– [67] S. Tommasi, S.-I. Kim, X. Zhong, X. Wu, G. P. Pfeifer, and A. 602, 2010. Besaratinia, “Investigating the epigenetic effects of a prototype [49] T. H. Kwaks, P. Barnett, W. Hemrika et al., “Identification of smoke-derived carcinogen in human cells,” PLoS ONE,vol.5, anti-repressor elements that confer high and stable protein no.5,ArticleIDe10594,2010. production in mammalian cells,” Nature Biotechnology,vol.21, [68] J. A. Marwick, P. A. Kirkham, C. S. Stevenson et al., “Cigarette pp. 553–558, 2003. smoke alters chromatin remodeling and induces proinflamma- [50]W.Zhou,P.Zhu,J.Wangetal.,“HistoneH2Amonoubiquiti- tory genes in rat lungs,” American Journal of Respiratory Cell and nation represses transcription by inhibiting RNA polymerase II Molecular Biology,vol.31,no.6,pp.633–642,2004. transcriptional elongation,” Molecular Cell,vol.29,no.1,pp.69– [69] F. Liu, J. K. Killian, M. Yang et al., “Epigenomic alterations 80, 2008. and gene expression profiles in respiratory epithelia exposed [51] A. Kanhere, K. Viiri, C. C. Araujoetal.,“ShortRNAsare´ to cigarette smoke condensate,” Oncogene,vol.29,no.25,pp. transcribed from repressed polycomb target genes and interact 3650–3664, 2010. with polycomb repressive complex-2,” Molecular Cell,vol.38, [70] P. Boyle, P. Autier, H. Bartelink et al., “European code against no. 5, pp. 675–688, 2010. cancer and scientific justification: third version (2003),” Annals [52] J.Zhao,B.K.Sun,J.A.Erwin,J.-J.Song,andJ.T.Lee,“Polycomb of Oncology,vol.14,no.7,pp.973–1005,2003. proteins targeted by a short repeat RNA to the mouse X [71] M. Van Engeland, M. P. Weijenberg, G. M. J. M. Roemen et chromosome,” Science, vol. 322, no. 5902, pp. 750–756, 2008. al., “Effects of dietary folate and alcohol intake on promoter [53] K. Ranganathan and V. Sivasankar, “MicroRNAs—Biology and methylation in sporadic colorectal cancer: the Netherlands clinical applications,” JournalofOralandMaxillofacialPathol- cohort study on diet and cancer,” Cancer Research,vol.63,no. ogy,vol.18,no.2,pp.229–234,2014. 12, pp. 3133–3137, 2003. [54] D. P. Bartel, “MicroRNAs: genomics, biogenesis, mechanism, [72]S.deVogel,B.W.C.Bongaerts,K.A.D.Woutersetal., and function,” Cell,vol.116,no.2,pp.281–297,2004. “Associations of dietary methyl donor intake with MLH1 pro- [55] L. He and G. J. Hannon, “MicroRNAs: small RNAs with a big moter hypermethylation and related molecular phenotypes in role in gene regulation,” Nature Reviews Genetics,vol.5,no.7, sporadic colorectal cancer,” Carcinogenesis,vol.29,no.9,pp. pp.522–531,2004. 1765–1773, 2008. [56] G. A. Calin and C. M. Croce, “MicroRNA signatures in human [73] U. Lehmann, B. Hasemeier, R. Lilischkis, and H. Kreipe, cancers,” Nature Reviews Cancer,vol.6,no.11,pp.857–866, “Quantitative analysis of promoter hypermethylation in laser- 2006. microdissected archival specimens,” Laboratory Investigation, [57] A. Barski, R. Jothi, S. Cuddapah et al., “Chromatin poises vol.81,no.4,pp.635–638,2001. miRNA- and protein-coding genes for expression,” Genome [74] B. Banelli, S. Bonassi, I. Casciano et al., “Outcome prediction Research,vol.19,no.12,pp.1742–1751,2009. and risk assessment by quantitative pyrosequencing methyla- [58] F. Sato, S. Tsuchiya, S. J. Meltzer, and K. Shimizu, “MicroRNAs tion analysis of the SFN gene in advanced stage, high-risk, and epigenetics,” FEBS Journal,vol.278,no.10,pp.1598–1609, neuroblastic tumor patients,” International Journal of Cancer, 2011. vol. 126, no. 3, pp. 656–668, 2010. [59] M. Fabbri, R. Garzon, A. Cimmino et al., “MicroRNA-29 family [75] A. D. Vinci, I. Casciano, E. Marasco et al., “Quantitative reverts aberrant methylation in lung cancer by targeting DNA methylation analysis of HOXA3, 7, 9,and10 genes in glioma: BioMed Research International 11

association with tumor WHO grade and clinical outcome,” [90] A. Veiga-Lopez, L. J. Luense, L. K. Christenson, and V.Padman- Journal of Cancer Research and Clinical Oncology,vol.138,no.1, abhan, “Developmental programming: gestational bisphenol- pp. 35–47, 2012. A treatment alters trajectory of fetal ovarian gene expression,” [76] A. Di Vinci, C. Brigati, I. Casciano et al., “HOXA7, 9, and 10 Endocrinology, vol. 154, no. 5, pp. 1873–1884, 2013. are methylation targets associated with aggressive behavior in [91] S.-M. Ho, W.-Y. Tang, J. Belmonte de Frausto, and G. S. meningiomas,” Translational Research,vol.160,no.5,pp.355– Prins, “Developmental exposure to estradiol and bisphenol A 362, 2012. increases susceptibility to prostate carcinogenesis and epigenetically regulates phosphodiesterase type 4 variant 4,” Cancer [77]B.Banelli,D.F.Merlo,G.Allemanni,A.Forlani,andM. Research, vol. 66, no. 11, pp. 5624–5632, 2006. Romani, “Clinical potentials of methylator phenotype in stage 4 high-risk neuroblastoma: an open challenge,” PLoS ONE,vol. [92] D. C. Dolinoy, D. Huang, and R. L. Jirtle, “Maternal nutri- 8, no. 5, Article ID e63253, 2013. ent supplementation counteracts bisphenol A-induced DNA hypomethylation in early development,” Proceedings of the [78]K.Brakensiek,L.U.Wingen,F.Langer,¨ H. Kreipe, and U. National Academy of Sciences of the United States of America, Lehmann, “Quantitative high-resolution CpG island mapping vol.104,no.32,pp.13056–13061,2007. with pyrosequencing reveals disease-specific methylation patterns of the CDKN2B gene in myelodysplastic syndrome and [93]S.J.Robboy,R.E.Scully,W.R.Welch,andA.L.Herbst, myeloid leukemia,” Clinical Chemistry,vol.53,no.1,pp.17–23, “Intrauterine diethylstilbestrol exposure and its consequences: 2007. pathologic characteristics of vaginal adenosis, clear cell ade- nocarcinoma, and related lesions,” Archives of Pathology and [79]Y.Liu,Y.Balaraman,G.Wang,K.P.Nephew,andF.C.Zhou, Laboratory Medicine,vol.101,no.1,pp.1–5,1977. “Alcohol exposure alters DNA methylation profiles in mouse [94]S.Li,K.A.Washburn,R.Mooreetal.,“Developmental embryos at early neurulation,” Epigenetics,vol.4,no.7,pp.500– exposure to diethylstilbestrol elicits demethylation of estrogen- 511, 2009. responsive lactoferrin gene in mouse uterus,” Cancer Research, [80] E. Diamanti-Kandarakis, J.-P. Bourguignon, L. C. Giudice et vol. 57, no. 19, pp. 4356–4359, 1997. al., “Endocrine-disrupting chemicals: an Endocrine Society [95] M. K. Skinner, “Environmental epigenetic transgenerational scientific statement,” Endocrine Reviews,vol.30,no.4,pp.293– inheritance and somatic epigenetic mitotic stability,” Epigenet- 342, 2009. ics,vol.6,no.7,pp.838–842,2011. [81] M. D. Anway and M. K. Skinner, “Epigenetic transgenerational [96] S. Karsli-Ceppioglu, A. Dagdemir, G. Judes et al., “Epigenetic actions of endocrine disruptors,” Endocrinology,vol.147,pp. mechanisms of breast cancer: an update of the current knowl- S43–S49, 2006. edge,” Epigenomics,vol.6,no.6,pp.651–664,2014. [82] M. D. Anway, A. S. Cupp, N. Uzumcu, and M. K. Skinner, [97] K. Zhang, X. Sun, X. Zhou et al., “Long non-coding RNA “Toxicology: epigenetic transgenerational actions of endocrine HOTAIR promotes glioblastoma cell cycle progression in an disruptors and male fertility,” Science,vol.308,no.5727,pp. EZH2 dependent manner,” Oncotarget,vol.6,no.1,pp.537–546, 1466–1469, 2005. 2015. [83] M. K. Skinner, M. Manikkam, R. Tracey, C. Guerrero-Bosagna, [98] A. Murat, E. Migliavacca, T. Gorlia et al., “Stem cell-related ‘self- M.Haque,andE.E.Nilsson,“Ancestraldichlorodiphenyl- renewal’ signature and high epidermal growth factor receptor trichloroethane (DDT) exposure promotes epigenetic transgen- expression associated with resistance to concomitant chemora- erational inheritance of obesity,” BMC Medicine, vol. 11, no. 1, diotherapy in glioblastoma,” Journal of Clinical Oncology,vol. article 228, 2013. 26, no. 18, pp. 3015–3024, 2008. [84] R. Tracey, M. Manikkam, C. Guerrero-Bosagna, and M. K. [99] A. M. Zama and M. Uzumcu, “Fetal and neonatal exposure Skinner, “Hydrocarbons (jet fuel JP-8) induce epigenetic trans- to the endocrine disruptor methoxychlor causes epigenetic generational inheritance of obesity, reproductive disease and alterations in adult ovarian genes,” Endocrinology,vol.150,no. sperm epimutations,” Reproductive Toxicology,vol.36,pp.104– 10, pp. 4681–4691, 2009. 116, 2013. [100] C. Stouder and A. Paoloni-Giacobino, “Specific transgenerational imprinting effects of the endocrine disruptor methoxy- [85] M. K. Skinner, “Environmental stress and epigenetic transgen- chlor on male gametes,” Reproduction,vol.141,no.2,pp.207– erational inheritance,” BMC Medicine,vol.12,article153,2014. 216, 2011. [86] E. E. Nilsson and M. K. Skinner, “Environmentally induced epi- [101] C. Stouder and A. Paoloni-Giacobino, “Transgenerational genetic transgenerational inheritance of disease susceptibility,” effects of the endocrine disruptor vinclozolin on the methyla- Translational Research,vol.165,no.1,pp.12–17,2014. tion pattern of imprinted genes in the mouse sperm,” Reproduc- [87]U.Grossniklaus,B.Kelly,A.C.Ferguson-Smith,M.Pembrey, tion,vol.139,no.2,pp.373–379,2010. and S. Lindquist, “Transgenerational epigenetic inheritance: [102]A.E.Armenti,A.M.Zama,L.Passantino,andM.Uzumcu, how important is it?” Nature Reviews Genetics,vol.14,no.3,pp. “Developmental methoxychlor exposure affects multiple repro- 228–235, 2013. ductive parameters and ovarian folliculogenesis and gene [88] E. Heard and R. A. Martienssen, “Transgenerational epigenetic expression in adult rats,” Toxicology and Applied Pharmacology, inheritance: myths and mechanisms,” Cell,vol.157,no.1,pp.95– vol. 233, no. 2, pp. 286–296, 2008. 109, 2014. [103] H. Bartsch and J. Nair, “Oxidative stress and lipid peroxidation- [89] L.F.Doherty,J.G.Bromer,Y.Zhou,T.S.Aldad,andH.S.Taylor, derived DNA-lesions in inflammation driven carcinogenesis,” “In utero exposure to diethylstilbestrol (DES) or bisphenol-A Cancer Detection and Prevention,vol.28,no.6,pp.385–391, (BPA) increases EZH2 expression in the mammary gland: an 2004. epigenetic mechanism linking endocrine disruptors to breast [104]M.W.Lawless,K.J.O’Byrne,andS.G.Gray,“Oxidativestress cancer,” Hormones and Cancer,vol.1,no.3,pp.146–155,2010. induced lung cancer and COPD: opportunities for epigenetic 12 BioMed Research International

therapy,” Journal of Cellular and Molecular Medicine,vol.13,pp. [120] A. Link, F. Balaguer, and A. Goel, “Cancer chemoprevention by 2800–2821, 2009. dietary polyphenols: promising role for epigenetics,” Biochemi- [105] N. Druesne, A. Pagniez, C. Mayeur et al., “Diallyl disulfide cal Pharmacology,vol.80,no.12,pp.1771–1792,2010. (DADS) increases histone acetylation and p21waf1/cip1 expression [121] J. Paluszczak, V. Krajka-Kuzniak,´ Z. Małecka et al., “Frequent in human colon tumor cell lines,” Carcinogenesis,vol.25,no.7, gene hypermethylation in laryngeal cancer cell lines and the pp.1227–1236,2004. resistance to demethylation induction by plant polyphenols,” [106] B. Kiec-Wilk, U. Razny, J. C. Mathers, and A. Dembinska-Kiec, Toxicology in Vitro, vol. 25, no. 1, pp. 213–221, 2011. “DNA methylation, induced by beta-carotene and arachidonic [122] M. Z. Fang, Y. Wang, N. Ai et al., “Tea polyphenol (-)- acid, plays a regulatory role in the pro-angiogenic VEGF- epigallocatechin-3-gallate inhibits DNA methyltransferase and receptor (KDR) gene expression in endothelial cells,” Journal of reactivates methylation-silenced genes in cancer cell lines,” Physiology and Pharmacology,vol.60,no.4,pp.49–53,2009. Cancer Research, vol. 63, no. 22, pp. 7563–7570, 2003. [107] J. Chen and X. Xu, “Diet, epigenetic, and cancer prevention,” [123] W. J. Lee, J.-Y. Shim, and B. T. Zhu, “Mechanisms for the Advances in Genetics,vol.71,pp.238–255,2010. inhibition of DNA methyltransferases by tea catechins and bioflavonoids,” Molecular Pharmacology,vol.68,no.4,pp.1018– [108] R. H. Dashwood and E. Ho, “Dietary histone deacetylase 1030, 2005. inhibitors: from cells to mice to man,” Seminars in Cancer Biology,vol.17,no.5,pp.363–369,2007. [124]Y.Yuasa,H.Nagasaki,Y.Akiyamaetal.,“DNAmethylation status is inversely correlated with green tea intake and physical [109]J.A.MckayandJ.C.Mathers,“Dietinducedepigeneticchanges activity in gastric cancer patients,” International Journal of and their implications for health,” Acta Physiologica,vol.202, Cancer,vol.124,no.11,pp.2677–2682,2009. no. 2, pp. 103–118, 2011. [125]Y.Yuasa,H.Nagasaki,Y.Akiyamaetal.,“Relationshipbetween [110]M.Cravo,P.Fidalgo,A.D.Pereiraetal.,“DNAmethylation CDX2 gene methylation and dietary factors in gastric cancer as an intermediate biomarker in colorectal cancer: modulation patients,” Carcinogenesis, vol. 26, no. 1, pp. 193–200, 2005. by folic acid supplementation,” European Journal of Cancer [126] W.Qin, W.Zhu, H. Shi et al., “Soy isoflavones have an antiestro- Prevention,vol.3,no.6,pp.473–479,1994. genic effect and alter mammary promoter hypermethylation in [111] Y.-I. Kim, H. W. Baik, K. Fawaz et al., “Effects of folate sup- healthy premenopausal women,” Nutrition and Cancer,vol.61, plementation on two provisional molecular markers of colon no. 2, pp. 238–244, 2009. cancer: a prospective, randomized trial,” American Journal of [127] Y. Li, L. Liu, L. G. Andrews, and T. O. Tollefsbol, “Genistein Gastroenterology,vol.96,no.1,pp.184–195,2001. depletes telomerase activity through cross-talk between genetic [112] M. Pufulete, R. Al-Ghnaniem, A. Khushal et al., “Effect of folic and epigenetic mechanisms,” International Journal of Cancer, acid supplementation on genomic DNA methylation in patients vol. 125, no. 2, pp. 286–296, 2009. with colorectal adenoma,” Gut,vol.54,no.5,pp.648–653,2005. [128]H.S.Cross,E.Kallay,´ D. Lechner, W. Gerdenitsch, H. Adler- [113] G. C. Rampersaud, G. P. A. Kauwell, A. D. Hutson, J. J. creutz, and H. J. Armbrecht, “Phytoestrogens and vitamin D Cerda, and L. B. Bailey, “Genomic DNA methylation decreases metabolism: a new concept for the prevention and therapy in response to moderate folate depletion in elderly women,” of colorectal, prostate, and mammary carcinomas,” Journal of American Journal of Clinical Nutrition,vol.72,no.4,pp.998– Nutrition,vol.134,no.5,pp.1207S–1212S,2004. 1003, 2000. [129]E.S.Fiala,M.E.Staretz,G.A.Pandya,K.El-Bayoumy,and [114] R. P. Steegers-Theunissen, S. A. Obermann-Borst, D. Kremer et S. R. Hamilton, “Inhibition of DNA cytosine methyltransferase al., “Periconceptional maternal folic acid use of 400 𝜇gperday by chemopreventive selenium compounds, determined by an is related to increased methylation of the IGF2 gene in the very improved assay for DNA cytosine methyltransferase and DNA young child,” PLoS ONE,vol.4,no.11,ArticleIDe7845,2009. cytosine methylation,” Carcinogenesis,vol.19,no.4,pp.597– [115] T. Sakatani, A. Kaneda, C. A. Iacobuzio-Donahue et al., “Loss of 604, 1998. imprinting of Igf2 alters intestinal maturation and tumorigene- [130]N.Xiang,R.Zhao,G.Song,andW.Zhong,“Selenitereactivates sis in mice,” Science, vol. 307, no. 5717, pp. 1976–1978, 2005. silenced genes by modifying DNA methylation and histones in [116] J. M. Davison, T. J. Mellott, V. P. Kovacheva, and J. K. Blusztajn, prostate cancer cells,” Carcinogenesis,vol.29,no.11,pp.2175– “Gestational choline supply regulates methylation of histone 2181, 2008. H3, expression of histone methyltransferases G9a (Kmt1c) and [131] P. O. McGowan, A. Sasaki, A. C. D’Alessio et al., “Epigenetic Suv39h1 (Kmt1a), and DNA methylation of their genes in rat regulation of the glucocorticoid receptor in human brain fetal liver and brain,” The Journal of Biological Chemistry,vol. associates with childhood abuse,” Nature Neuroscience,vol.12, 284, no. 4, pp. 1982–1989, 2009. no.3,pp.342–348,2009. [117] L. H. Lumey, A. D. Stein, H. S. Kahn et al., “Cohort profile: the [132] N. M. Tsankova, O. Berton, W. Renthal, A. Kumar, R. L. Neve, Dutch Hunger Winter families study,” International Journal of and E. J. Nestler, “Sustained hippocampal chromatin regulation Epidemiology, vol. 36, no. 6, pp. 1196–1204, 2007. in a mouse model of depression and antidepressant action,” [118] B. T. Heijmans, E. W. Tobi, A. D. Stein et al., “Persistent Nature Neuroscience,vol.9,no.4,pp.519–525,2006. epigenetic differences associated with prenatal exposure to [133] J.-P.J. Issa, “DNA methylation as a therapeutic target in cancer,” famine in humans,” Proceedings of the National Academy of Clinical Cancer Research,vol.13,no.6,pp.1634–1637,2007. Sciences of the United States of America,vol.105,no.44,pp. [134] J. Tost, “DNA methylation: an introduction to the biology 17046–17049, 2008. and the disease-associated changes of a promising biomarker,” [119] L. Fini, M. Selgrad, V. Fogliano et al., “Annurca apple polyphe- Molecular Biotechnology,vol.44,no.1,pp.71–81,2010. nols have potent demethylating activity and can reactivate [135] M. T. McCabe, J. C. Brandes, and P. M. Vertino, “Cancer DNA silenced tumor suppressor genes in colorectal cancer cells,” methylation: molecular mechanisms and clinical implications,” Journal of Nutrition,vol.137,no.12,pp.2622–2628,2007. Clinical Cancer Research,vol.15,no.12,pp.3927–3937,2009. BioMed Research International 13

[136] A. Eden, F. Gaudet, A. Waghmare, and R. Jaenisch, “Chromo- [153]C.K.Taylor,R.M.Levy,J.C.Elliott,andB.P.Burnett,“The somalinstabilityandtumorspromotedbyDNAhypomethyla- effect of genistein aglycone on cancer and cancer risk: a review tion,” Science,vol.300,no.5618,p.455,2003. of in vitro, preclinical, and clinical studies,” Nutrition Reviews, [137] B. Weber, S. Kimhi, G. Howard, A. Eden, and F. Lyko, vol. 67, no. 7, pp. 398–415, 2009. “Demethylation of a LINE-1 antisense promoter in the cMet [154] M. Z. Fang, D. Chen, Y. Sun, Z. Jin, J. K. Christman, and locus impairs Met signalling through induction of illegitimate C. S. Yang, “Reversal of hypermethylation and reactivation of transcription,” Oncogene,vol.29,no.43,pp.5775–5784,2010. p16INK4a, RAR𝛽, and MGMT genes by genistein and other [138] E. M. Wolff, H.-M. Byun, H. F. Han et al., “Hypomethylation of isoflavones from soy,” Clinical Cancer Research, vol. 11, no. 19, a LINE-1 promoter activates an alternate transcript of the MET pp. 7033–7041, 2005. oncogene in bladders with cancer,” PLoS Genetics,vol.6,no.4, [155] S.Majid,A.A.Dar,A.E.Ahmadetal.,“BTG3tumorsuppressor Article ID e1000917, 2010. gene promoter demethylation, histone modification and cell [139] P.M. Lizardi, “As we bring demethylating drugs to the clinic, we cycle arrest by genistein in renal cancer,” Carcinogenesis,vol.30, better know the DICE being cast,” Oncogene,vol.29,no.43,pp. no. 4, pp. 662–670, 2009. 5772–5774, 2010. [156] J. K. Day, A. M. Bauer, C. Desbordes et al., “Genistein alters [140]K.D.Hansen,W.Timp,H.C.Bravoetal.,“Increasedmethy- methylation patterns in mice,” JournalofNutrition,vol.132,no. lation variation in epigenetic domains across cancer types,” 8, pp. 2419S–2423S, 2002. Nature Genetics,vol.43,no.8,pp.768–775,2011. [157]A.I.Constantinou,D.Lantvit,M.Hawthorne,X.Xu,R.B. [141] J.-P. J. Issa, G. Garcia-Manero, F. J. Giles et al., “Phase 1 study of van Breemen, and J. M. Pezzuto, “Chemopreventive effects of low-dose prolonged exposure schedules of the hypomethylating 󸀠 soy protein and purified soy isoflavones on DMBA-induced agent 5-aza-2 -deoxycytidine (decitabine) in hematopoietic mammary tumors in female Sprague-Dawley rats,” Nutrition malignancies,” Blood,vol.103,no.5,pp.1635–1640,2004. and Cancer, vol. 41, no. 1-2, pp. 75–81, 2001. [142]E.Kaminskas,A.Farrell,S.Abrahametal.,“Approvalsum- [158] G. L. Johanning and C. J. Piyathilake, “Retinoids and epigenetic mary: azacitidine for treatment of myelodysplastic syndrome silencing in cancer,” Nutrition Reviews,vol.61,no.8,pp.284– subtypes,” Clinical Cancer Research,vol.11,no.10,pp.3604– 289, 2003. 3608, 2005. [159] M. J. Rowling, M. H. McMullen, and K. L. Schalinske, [143] J.-P.F. Issa, V.Gharibyan, J. Cortes et al., “Phase II study of low- “Vitamin A and its derivatives induce hepatic glycine N- dose decitabine in patients with chronic myelogenous leukemia methyltransferase and hypomethylation of DNA in rats,” Jour- resistant to imatinib mesylate,” Journal of Clinical Oncology,vol. nal of Nutrition,vol.132,no.3,pp.365–369,2002. 23,no.17,pp.3948–3956,2005. [160] M. Esteller, M. Guo, V. Moreno et al., “Hypermethylation- [144] E.E.Cameron,K.E.Bachman,S.Myoh¨ anen,¨ J. G. Herman, and associated inactivation of the cellular retinol-binding-protein S. B. Baylin, “Synergy of demethylation and histone deacetylase 1 gene in human cancer,” Cancer Research,vol.62,no.20,pp. inhibition in the re-expression of genes silenced in cancer,” 5902–5905, 2002. Nature Genetics,vol.21,no.1,pp.103–107,1999. [161]L.DiCroce,V.A.Raker,M.Corsaroetal.,“Methyltransferase [145] K.K.Y.Sie,A.Medline,J.VanWeeletal.,“Effectofmaternaland recruitment and DNA hypermethylation of target promoters by postweaning folic acid supplementation on colorectal cancer an oncogenic transcription factor,” Science,vol.295,no.5557,pp. risk in the offspring,” Gut,vol.60,no.12,pp.1687–1694,2011. 1079–1082, 2002. [146] A. Ly, H. Lee, J. Chen et al., “Effect of maternal and postweaning [162] F. Fazi, L. Travaglini, D. Carotti et al., “Retinoic acid targets folic acid supplementation on mammary tumor risk in the DNA-methyltransferases and histone deacetylases during APL offspring,” Cancer Research, vol. 71, no. 3, pp. 988–997, 2011. blast differentiation in vitro and in vivo,” Oncogene,vol.24,no. [147]M.K.Keyes,H.Jang,J.B.Masonetal.,“Olderageand 11,pp.1820–1830,2005. dietary folate are determinants of genomic and p16-specific [163] H. S. Cheong, H. C. Lee, B. L. Park et al., “Epigenetic modifica- DNA methylation in mouse colon,” Journal of Nutrition,vol.137, tion of retinoic acid-treated human embryonic stem cells,” BMB no.7,pp.1713–1717,2007. Reports,vol.43,no.12,pp.830–835,2010. [148] B. F. Cole, J. A. Baron, R. S. Sandler et al., “Folic acid for the [164] A. P. Bracken and K. Helin, “Polycomb group proteins: naviga- prevention of colorectal adenomas: a randomized clinical trial,” tors of lineage pathways led astray in cancer,” Nature Reviews Journal of the American Medical Association,vol.297,no.21,pp. Cancer,vol.9,no.11,pp.773–784,2009. 2351–2359, 2007. [149] K. Wallace, M. V. Grau, A. J. Levine et al., “Association between [165] L. J. Gudas and J. A. Wagner, “Retinoids regulate stem cell folate levels and CpG island hypermethylation in normal differentiation,” Journal of Cellular Physiology,vol.226,no.2, colorectal mucosa,” Cancer Prevention Research,vol.3,no.12, pp. 322–330, 2011. pp.1552–1564,2010. [166]S.Balasubramanian,G.Adhikary,andR.L.Eckert,“The [150] M. Fang, D. Chen, and C. S. Yang, “Dietary polyphenols may Bmi-1 polycomb protein antagonizes the (-)-epigallocatechin- affect DNA methylation,” Journal of Nutrition,vol.137,no.1, 3-gallate-dependent suppression of skin cancer cell survival,” supplement, pp. 223S–228S, 2007. Carcinogenesis,vol.31,no.3,pp.496–503,2010. [151] J. B. Berletch, C. Liu, W. K. Love, L. G. Andrews, S. K. Katiyar, [167] W.-F. Hua, Y.-S. Fu, Y.-J. Liao et al., “Curcumin induces down- and T. O. Tollefsbol, “Epigenetic and genetic mechanisms con- regulation of EZH2 expression through the MAPK pathway in tribute to telomerase inhibition by EGCG,” Journal of Cellular MDA-MB-435 human breast cancer cells,” European Journal of Biochemistry,vol.103,no.2,pp.509–519,2008. Pharmacology,vol.637,no.1–3,pp.16–21,2010. [152]J.C.Chuang,C.B.Yoo,J.M.Kwanetal.,“Comparisonofbio- [168]M.Dimri,P.V.Bommi,A.A.Sahasrabuddhe,J.D.Khandekar, logical effects of non-nucleoside DNA methylation inhibitors and G. P. Dimri, “Dietary omega-3 polyunsaturated fatty acids 󸀠 versus 5-aza-2 -deoxycytidine,” Molecular Cancer Therapeutics, suppress expression of EZH2 in breast cancer cells,” Carcinogen- vol.4,no.10,pp.1515–1520,2005. esis,vol.31,no.3,pp.489–495,2010. Hindawi Publishing Corporation BioMed Research International Volume 2015, Article ID 949514, 10 pages http://dx.doi.org/10.1155/2015/949514

Research Article Impact of Cadmium on Intracellular Zinc Levels in HepG2 Cells: Quantitative Evaluations and Molecular Effects

Chiara Urani,1 Pasquale Melchioretto,1 Maurizio Bruschi,1 Marco Fabbri,2 Maria Grazia Sacco,3 and Laura Gribaldo4

1 Department of Earth and Environmental Sciences, University of Milan Bicocca, Piazza della Scienza 1, 20126 Milan, Italy 2Department of Clinical and Experimental Medicine, University of Insubria, 21100 Varese, Italy 3Molecular Biology and Genomics Unit, Institute for Health and Consumer Protection, DG JRC, European Commission, Via Enrico Fermi 2749, 21027 Ispra, Italy 4Chemical Assessment and Testing Unit (CAT), Institute for Health and Consumer Protection, DG JRC, European Commission, Via Enrico Fermi 2749, 21027 Ispra, Italy

Correspondence should be addressed to Chiara Urani; [email protected] and Laura Gribaldo; [email protected]

Received 17 December 2014; Accepted 2 March 2015

Academic Editor: Luciana dos Reis Vasques

Copyright © 2015 Chiara Urani et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Cadmium is classified as a human carcinogen, and its disturbance in zinc homeostasis has been well established. However, its extent as well as molecular mechanisms involved in cadmium carcinogenesis has yet to be fully clarified. To this end, we used the zinc specific probe Zinquin to visualize and to quantitatively evaluate changes in the concentration of labile zinc, inan in vitro model of human hepatic cells (HepG2) exposed to cadmium. A very large increase (+93%) of intracellular labile zinc, displaced by cadmium from the zinc proteome, was measured when HepG2 were exposed to 10 𝜇M cadmium for 24 hrs. Microarray expression profiling showed that in cells, featuring an increase of labile zinc after cadmium exposure, one of the top regulated genes is Snail1 (+3.6), which is included in the adherens junction pathway and linked to cancer. In the same pathway MET, TGF-𝛽R,andtwomembers of the Rho-family GTPase, Rac,andcdc42 all implicated in the loss of adherence features and acquisition of migratory and cancer properties were regulated, as well. The microRNAs analysis showed a downregulation of miR-34a and miR-200a, both implicated in the epithelial-mesenchymal transition. These microRNAs results support the role played by zinc in affecting gene expression at the posttranscriptional level.

1. Introduction pathways have been suggested and recently reviewed [5]. In mammalian cells well established pathways involve Cadmium is a highly persistent pollutant harmful to humans receptor-mediated endocytosis (megalin/cubilin) of Cd- andanimals,listedasoneofthetoptenhazardoussubstances metallothionein complex and cotransport mechanisms by bytheAgencyforToxicSubstancesandDiseaseRegistry[1], divalent metal ion transporter 1 (DMT1) and ZIP proteins and classified as a human carcinogen, based on epidemio- (ZIP8 and ZIP14A/B). A recently discovered candidate path- 2+ logical studies and animal experiments, by the International way for Cd entry into cells is through the T-Type calcium Agency for Research on Cancer [2]. The widespread presence 3.1 2+ channels Cav [6]. Due to its molecular mimicry, cadmium of Cd makes it a severe environmental health problem follows a Trojan horse strategy leading to interference with that needs to be considered thoroughly. Apart from natural essential metals homeostasis. In mammalian cells cadmium sources, major fonts of exposure are cadmium-contaminated toxicity has been closely related to zinc homeostasis. Once food and water, providing around 30 𝜇g per day for adults, inside the cells, cadmium is expected to displace zinc in cigarette smoke [3], and cosmetic products [4]. proteins and enzymes in which zinc has a sulphur-dominated Cadmium enters the cells by utilizing transport path- coordination sphere, such as the metallothioneins, the zinc ways typically evolved for essential metals. Varieties of sensor MTF-1, and multiple zinc-finger proteins3 [ , 7, 8]. 2 BioMed Research International

However, even though the disturbance of cadmium in zinc incubated with Zinquin to a final concentration of 25 𝜇M ∘ homeostasis has already been recognized, its extent has not in PBS and left in the dark for 30 min at 37 Cforbinding been a matter of study yet. with zinc. Samples incubated with PBS alone were used for In this context, the aim of our work was (1) to demon- autofluorescence evaluations of HepG2 cells. The solution strate that the exposure of a human hepatic model system was removed and the coverslips were washed twice with PBS (HepG2 cells) to cadmium leads to an increase of the and once with distilled water. The dried coverslips were subse- intracellular pool of labile Zn(II) and (2) to quantitatively quently mounted on glass slides using glycerol and PBS (9 : 1), measure this increase. To this end, we have used a fluorescent containing 2.5% (v/v) 1,4-diazabicyclo[2.2.2]octane solution probe (Zinquin) specifically developed for the visualization (DABCO) to reduce fluorescence quenching. and measurement of intracellular labile zinc ions [9]. A Slides were examined with a fluorescent microscope further aim was to analyze the consequences at molecular (Zeiss Axioplan) utilizing a standard UV filter set. All images level of this increase, as zinc is a recognized second messenger were acquired by a digital camera (CoolSnap-ProColors and transcriptional regulator [10, 11]. Media Cybernetics, Bethesda, MA, USA) and stored by Image Proplus software (Media Cybernetics). 2. Materials and Methods 2.3. Spectrofluorimetry. Spectrofluorimetry experiments were 2.1. Cell Cultures and Treatments. Human hepatoblastoma performed essentially according to Coyle and coworkers cells (HepG2) represent a useful in vitro model for cadmium [22] with minor modifications. HepG2 cells were seeded in 2 6 accumulation and toxicity studies, as previously reported 165 cm flasks (3 × 10 cells/flasks) and left to recover for [12–14]. Furthermore HepG2 are, among others, a widely 24 hrs before treatments with zinc or cadmium, as described used alternative cell system to primary hepatocytes as they in par. 2.1. Treated and control cells were harvested by retain, under specific cultivation conditions, many metabolic tripsinization, and a small aliquot from the cell suspension functions of normal liver cells [15] The HepG2cells were was counted by a Coulter counter to have an estimate of cells routinely cultured in Opti-MEM medium (Life Technologies, number in each sample. All cell suspensions were centrifuged Monza, Italy) supplemented with 10% inactivated fetal bovine (200 ×g, 10 min) and the pellets resuspended in 2 mL of serum and 1% antibiotics (streptomycin/penicillin). Cells ∘ Hanks Balanced Salt Solution (HBSS: 137 mM NaCl, 5.4 mM were kept in incubator at constant 37 C under a humidified KCl, 0.3 mM KH2PO4,0.33mMNa2HPO4, 1.23 mM CaCl2× 5% CO2 atmosphere. 2H2O, 0.81 mM MgSO4× 7H2O, 4.2 mM NaHCO3,and For fluorescence microscopy visualization and spec- 5.6 mM D-Glucose). The cells in suspension were incubated trofluorimetric measurements the cells were cultured either with Zinquin (final concentration 25 𝜇M) in HBSS for 40 min ∘ in complete medium (control) or in complete medium at 37 C. Then the cells were washed three times with HBSS, 𝜇 containing CdCl2 (Cd, 0.1 and 10 M) or ZnSO4 (Zn, 10, 50, and fluorescence was measured at an excitation wavelength 𝜇 and 170 M) for 24 hrs. of 370 nm and emission wavelength of 490 nm (slit width The viability of HepG2 cells exposed to the highest 5–10 nm) in a Jasco FP-777 spectrofluorimeter. Fluorescence 𝜇 CdCl2 (Sigma-Aldrich, Milan, Italy) concentration (10 M) intensities are expressed as arbitrary units and are normalized for 24 hrs was between 80 and 90% and above 80% in ZnSO4 on a fixed number of cells (100.000 cells). (Sigma-Aldrich,Milan,Italy)treatedcells,aspreviously published [16, 17],andconfirmedherewith(datanotshown). 2.4. Microarray Expression Profiling. In this study we rean- alyze the gene expression data that we recently published 2.2. Fluorescent Visualization of Free Zinc. Cells were seeded [23] and submitted to NCBI repository (accession number on sterile glass coverslips in 35 mm culture plates (80.000 GSE3128). cells/plate) in complete culture medium. The medium was The whole gene expression of HepG2 cells treated with removed 24 hrs after seeding and substituted with Zn- or Cd- 𝜇 containing medium at concentrations indicated in par. 2.1. 10 M Cd was analyzed by Agilent microarray. In order to detect expression changes between treated and control The cells were left with the control or treatment media in 𝑡 𝑡 incubator for further 24 hrs. cells the moderated test was applied. Moderated statistics were generated by Limma Bioconductor package. Modulated At the end of treatment time, the cells were processed for 2 fluorescent experiments with Zinquin according to Sarwar geneswerechosenasthosewithlog fold change greater Nasir et al. [21] with some modifications. than 2 and a false discovery rate (Benjamini and Hochberg’s method) corrected 𝑃 value smaller than 0.05 [24]. All of the Zinquin [ethyl (2-methyl-8-p-toluenesulphonamido-6- 𝑅 quinolyloxy)acetate] is a fluorescent probe used to detect above computations were conducted using the statistics labile intracellular Zn and to reveal Zn in cultured cells. programming environment. Zinquinispermeableandisreadilytakenupbythecells A validation of ten genes regulated by 10 𝜇MCdwasdone and is essentially nonfluorescent until it complexes Zn with by a real-time quantitative PCR (qPCR) analysis on the same a high selectivity [9, 22], thus providing a means for labile RNA samples that were used for the microarray. The trends of zinc detection and visualization. At the end of all treatments, all validated genes were confirmed as presented in our recent themediumwasremovedandthecellswererinsedtwice paper [23]. with warm PBS and fixed in formaldehyde (3.7% in PBS) The regulated genes have been mapped on the KEGG ∘ for 30 min at 37 C.AfterrinsingwithPBS,thecellswere (Kyoto Encyclopedia of Genes and Genomes) PATHWAY, BioMed Research International 3 a collection of manually drawn pathway maps representing However, most of the zinc bound to proteins is unreactive molecular interactions and reaction networks [25]. toward Zinquin. The term “labile” will be used in the The heat map for visualization of expression profiles was following to indicate all of the Zn(II) sensed by Zinquin. The performed with TM4 system for microarray data analysis. displacement of Zn(II) from the Zn-proteome by Cd(II) and In the map, red and green colours means higher and lower its labilization has been recently demonstrated on the isolated expression levels [26]. proteome extracted by pig kidney LLC-PK1 cells by using Zinquin as a fluorescent probe8 [ ]. Here, we are investigating 2.5. MicroRNAs Expression Profiling. RNA was extracted theeffectofCdexposureontheZn-homeostasisinintact using the MIRVANA kit (AMBION) and was reverse tran- human cells. scribed with Taqman MicroRNA Reverse Transcription Kit HepG2 cells were exposed to noncytotoxic Cd concentra- using Megaplex RT Primers (Applied Biosystems). tions (0.1 and 10 𝜇M) and to Zn concentrations representing MicroRNA expression was measured and quantified an overload of the metal (50 and 170 𝜇M). Food and water are using microfluidic cards (Taqman ArrayMicroRNA Cards, major sources of Cd intake for nonoccupationally exposed set A, V2.2 and set B, V3, Applied Biosystems) allowing population. This metal cannot undergo metabolic degrada- the detection of about 754 unique assays specific and four tion, is poorly excreted, and is retained in the liver with a very candidate endogenous control assays (Applied Biosystems, long biological half life (20–30 years) [2], thus providing us FosterCity,CA).Samplesofcontroland10𝜇MCdwere with the fact that the maximum concentration used (10 𝜇M) tested. Analyses were carried out on the ABI Prism 7900HT maybecompatiblewithbioaccumulationintheliver.On Sequence Detection System (SDS) (Applied Biosystems). the other hand, since zinc is an essential element, the cells MicroRNA levels were normalized to endogenous control U6 can buffer high concentrations of this metal. Thus, with the (Applied Biosystems). MiRNAs with a threshold cycle <33 goal of inducing in our cells an observable effect, we exposed that showed a log 2 fold change greater than two in samples HepG2 to high nonphysiological zinc concentrations, up to treated with cadmium as compared to control samples were 170 𝜇M,thatwasshownnottobecytotoxic[16, 17]. Moreover, considered as induced. the exposure to 10 𝜇M Zn provided the results for a direct comparison to the equimolar concentration of Cd (10 𝜇M). 2.6. Statistical Analysis. All experiments were performed in Low physiological labile zinc in the control HepG2 (grown in triplicate, and two or three different technical replicates were complete medium) incubated with Zinquin was confirmed as analyzed in each experiment. Data are expressed as mean shown by fluorescence microscopy undetectable zinc levels ± SD. Student’s t-test was used for sample comparison, and (Figure 1(b)). Cells without Zinquin probing were also ana- the software package Statgraphics Plus version 5.0 was used lyzed for the evaluation of the background autofluorescence (Statistical Graphics Corp., Manugistic Inc., Rockville, MD, (Figure 1(a)). USA). The growth of HepG2 cells for 24 hrs in medium containing 10 or 50 𝜇M Zn (Figures 1(d) and 1(e), resp.) or containing 3. Results and Discussion 0.1 𝜇MCd(Figure 1(c))ledtoaverylowandalmostcom- parabletothatofcontrolsorslightlyincreasedfluorescence 3.1. Cd Exposure in HepG2 Cells Increases Intracellular Labile relatedtointracellularlabileZn.Onthecontrary,adramatic Zn: Qualitative Evaluations and Intracellular Distribution. To increase in fluorescence intensity was observed when the provide evidence that exposure of HepG2 cells to Cd induces cells were exposed for 24 hrs to the highest Zn concentration an increase of intracellular zinc, we have used the zinc- (Figure 1(g),170𝜇M Zn), as expected. In these samples, the specific fluorescent probe Zinquin to visualize by microscopy fluorescence was predominantly distributed in the cytoplasm and to measure by spectrofluorimetry the fluorescence inten- and in the perinuclear region and had a punctate appearance, sities. “Free” zinc concentration (where the term free is as previously reported in other cell types [22, 32]. Notably, referred to as freely available, zinc readily bound to chelating an intense and punctuate fluorescence was observed in agents) is maintained by a complex buffering system in the HepG2 cells grown in culture medium containing 10 𝜇MCd picomolartonanomolarrangeincellsgrownundernormal (Figure 1(f)), showing the high increase of intracellular labile conditions, while the total zinc in these cells is evaluated zinc when HepG2 cells are exposed to Cd. These punctuate, in the high 𝜇M[27–29]. In fact, the main pool of intracel- vescicular-like structures of high fluorescence intensity were lularzincisboundtoavastpopulationofmetalloproteins, previously designated “zincosomes,” possibly representing referred to as the zinc proteome of the cells, or its excess late endosomes, at least in certain cells type. These structures is stored in zinc-containing vesicles or organelles [30]. In may have a role in detoxifying excess zinc increased by toxic this respect, it should be noted that Zinquin not only binds agents [10, 30], as in our cells by Cd. free zinc (reaction (1)), but can also access Zn-protein with open coordination sites to form fluorescent ternary adducts 2+ 3.2. Quantification of Intracellular Labile Zinc in Cd- and Zn- (reaction (2)), or chelating Zn from the proteome (reaction Treated Cells. The assessment of the amount of intracellular (3))[31, 32]: 2+ labile Zn in HepG2 cells exposed to excess of Zn or to Cd was Zn +2ZQ 󳨀→ Zn (ZQ)2 (1) carried out by spectrofluorimetric measurements. The fluorescence enhancement after Zinquin binding to intracellular Zn-protein + ZQ 󳨀→ ZQ-Zn-protein (2) labileZnwasmeasuredasarbitraryunitsandexpressedas Zn-protein +2ZQ 󳨀→ Zn (ZQ)2 + apo-protein (3) a relative increase of the signal with respect to basal control 4 BioMed Research International

Control Control + Zinquin Cd 0.1 𝜇M

(a) (b) (c)

Zn 10 𝜇M Zn 50 𝜇M Cd 10 𝜇M (d) (e) (f)

Zn 170 𝜇M

(g)

Figure 1: Microscopy images of HepG2 cells loaded with 25 𝜇M of the zinc specific fluorescent probe Zinquin for intracellular visualization and distribution of labile zinc. HepG2 cells grown in control medium without Zinquin were used for autofluorescence visualization (a). Cells grown in control medium show undetectable basal levels of loosely bound zinc (b). HepG2 cells grown in medium containing, respectively, 0.1 (c) and 10 𝜇M (f) Cd for 24 hrs; (d, e, and g): cells exposed to increasing zinc concentrations (10, 50, and 170 𝜇M, resp.). Microscope magnification 400x.

Table 1 Cd 0.1 𝜇MCd10𝜇MZn10𝜇MZn50𝜇MZn170𝜇M ∗ ∗ 23.17 ± 14.01 93.03 ± 6.55 12.57 ± 10.24 21.37 ± 4.24 225.22 ± 24.59 Fluorescence intensities (arbitrary units) of labile zinc in cadmium- or zinc-treated HepG2 cells. The results are expressed as increment of basalontrol) (c values ∗ ± SD of at least three independent biological replicates. P < 0.05 significantly different (Student’s 𝑡-test) from the corresponding Cd or Zn treatment at lower concentration. values (cells grown in complete medium). Indeed, it was nec- expressed as mean ± SD of Δ%ofbasalfluorescenceis essary to express the fluorescence intensities as a Δ%versus presented. controls due to fluctuations of basal conditions possibly due At low zinc concentrations (10 and 50 𝜇M), the inten- to variations in zinc content amongst different experiments. sity of Zinquin fluorescence is relatively low and features However, despite this basal difference, the increase was fairly mean values of +21% versus basal fluorescence intensities, in constant within a specific zinc or cadmium treatment, as agreement with a strict regulation of the cellular zinc content demonstrated by the relatively small standard deviations and distribution [10]. Zinquin fluorescence was significantly obtained in the different biological replicates. Figure 2 gives raised (+225% versus controls, 𝑃 < 0.05) in HepG2 grown an overview of the increment of intracellular labile Zn ions in the presence of the highest zinc concentration (170 𝜇M), in all treated samples, expressed as fluorescence intensities, used as a positive control for zinc excess. The most interesting while, in Table 1 a summary of all fluorescence intensities result is the fluorescence intensity of Zinquin bound to BioMed Research International 5

250 ∗ is the adherens junctions, represented in Figure 4.Snail1 belongs to a superfamily of zinc-finger transcription factors 200 involved, among others, in the acquisition of invasive and 150 migratory properties during tumor progression [34]. The ∗ 100 mechanism of Snail1 activation has been recently related in breast cancer cells to zinc influx and intracellular zinc 50

% versus basal levels % versus increase, either directly or indirectly [35], thus strongly Δ 0 supporting our data. Cd 0.1 𝜇MZnCd 10 𝜇MZn10 𝜇MZn50 𝜇M 170 𝜇M Other upregulated genes in the adherens junction pathway (Figure 4)areMET (1.4 fold change), TGF-𝛽R (1 fold Figure 2: Spectrofluorimetric quantification of intracellular labile change), and the two members of the Rho-family GTPase, zinc. Samples are expressed as Δ% ± SD versus controls (cells grown Rac (1.5 fold change), and cdc42 (1.5 fold change). The MET ∗ in control medium in the presence of 25 𝜇M Zinquin). 𝑃 < 0.05 tyrosine kinase receptor (known also as the HGF receptor) significantly different (Student’s 𝑡-test) from the corresponding Cd promotes, among others, cancer growth and metastasis by or Zn treatment at lower concentration. conveying proliferative, antiapoptotic, and promigratory signals (see for a review [36]). The transforming growth factor receptor (TGF-𝛽R) is a member of the TGF-𝛽 signalling intracellularlabileZnincellsgrowninthepresenceof pathway, involved in cell proliferation and cell migration 𝛽 10 𝜇M Cd: a mean value of +93% versus basal fluorescence functions. In addition, TGF- is a signalling molecule impli- intensities was obtained. To the best of our knowledge, this cated in Snail1 activation and in our samples exposed to Cd is is the first time that the amount of labile zinc is visualized upregulatedwitha1.4foldchange[37]. The Rac and the cdc42 and quantitatively evaluated in Cd-exposed mammalian genes promote lamellipodia and filopodia formation, thus cells. In our previous works [13, 14]thetotalintracellular regulating cell migration through cytoskeleton remodelling amount of zinc and cadmium in HepG2 cells exposed to (see for a review [38]). 170 𝜇MZnand10𝜇MCdfor24hrswasmeasuredbyan The role of elevated zinc concentrations was demon- analytical technique (ICP-AES). The increment in the total stratedtoaffectgeneexpressionincancercellsatposttran- concentrations with respect to the basal values was about scriptional level [39]. Thus, intriguing results linking zinc 6 0.075 ppm/10 cells for zinc in cells exposed for 24 hrs to levels and gene expression can emerge from microRNAs 6 170 𝜇MZnand0.050ppm/10 cells for cadmium in samples (miRNA) analysis in HepG2 cells exposed to cadmium. The exposed to 10 𝜇MCd.Thus,itcanbenotedthat,atthe miRNA are small noncoding RNAs initially transcribed and same exposure time, the uptake of cadmium is only slightly processedinthenucleusasprecursors.Thesemoleculesare smaller than that of zinc, although cells have grown in the then exported to the cytoplasm where they become mature presence of a concentration of cadmium which is about 20 miRNA of about 23 nucleotides whose main function is to times smaller than that of zinc. This result clearly indicates negatively regulate gene expression at the posttranscriptional that the uptake of cadmium, that uses aspecific transport level by repression of protein translation or promotion of pathways [5], occurs to a significantly larger extent to that mRNA degradation [40]. In addition, miRNAs by targeting of zinc, whose concentration is, on the contrary, strictly multiple transcripts play a crucial role in tumorigenesis and regulated.Thisresultisconsistentwiththelargeincrement cancer progression [41]. Even if the role played by zinc in cell of the fluorescence intensity of HepG2 cells exposed to 10 𝜇M growth and proliferation is well known [10], the involvement Cd as due to cadmium to zinc replacement in proteins and of this metal in the regulation of gene expression at post- mobilization of the latter metal ion, a mechanism previously transcriptional level is still to be explored. Recently it was proposed [7] and recently demonstrated in a prokaryotic demonstrated that elevated intracellular zinc concentrations zinc-finger domain [33], and proteome extracted from pig modulate the miRNA expression in mammalian cells, with kidney LLC-PK1 cells [8]. miR-223 among the top downregulated miRNA [39]. It is important to highlight that miR-223 is involved in cell cycle regulation, proliferation, and survival [42], thus suggesting 3.3. Biological Effects of Labile Zinc Increase in HepG2 Cells: the relevance of zinc concentrations in the epigenetic mech- Genes and MicroRNAs Regulation. Prompted by the results anisms of cancer, at least those known up to now. discussed above, we have reanalyzed the molecular data and Two major miRNAs were downregulated in our samples: pathways to better understand the mechanisms of cadmium amiR-34familymember(−1.1 fold change) and a miR-200 biologicaleffectsandtocorrelatetheroleofzincingeneand family member (−1.2 fold change). molecular regulation. Very interestingly, a decrease in miR-34, which normally The elevated concentrations of Zn(II) visualized and antagonizes Snail1, was recently described as part of the measured in HepG2 cells after 24 hrs of Cd(II) exposure p53/miRNA-34 axis. Namely, in the absence of a functional activated a number of zinc-related processes. One of the p53 and of a decrease of miRNA-34, Snail1 is upregulated, as top regulated genes in our samples, as identified by the we found in our samples. This axis promotes the epithelial- Microarray expression profiling (Figure 3), is Snail1,withan mesenchymal transition (EMT) and the invasion program upregulation of +3.6 fold change with respect to controls. in neoplastic cells [20]. Further supporting this mecha- Snail1 is included in many biological pathways, one of which nism are our results in HepG2 cells on miR-200 family. 6 BioMed Research International

−1.0 0.0 1.0 2 3 1 2 3 1 _R _R _R 10 10 10 Ctrl_R Cd Cd Cd Ctrl_R Ctrl_R Rac TGF-𝛽R MET Rac/cdc42 Snail

(a)

KEGG ID Gene symbol Description FC Cd10

Ras-related C3 botulinum toxin substrate 2 Rac Rac2 1.5 (rho family, small GTP binding protein Rac2)

Transforming growth factor, 𝛽 receptor II TGF-𝛽R TGF-𝛽R2 1.0 (70/80 kDa)

Met proto-oncogene (hepatocyte growth MET MET 1.4 factor receptor)

Ras-related C3 botulinum toxin substrate 2 Rac/cdc42 Rac2 1.5 (rho family, small GTP binding protein Rac2)

SnailSnail1 Snail homolog 1 (Drosophila) 3.6

(b)

Figure 3: The genes regulated by 10 𝜇M Cd treatment and belonging to adherens junction pathway represented in Figure 4 are displayed. (a) Heat map graphically representing the gene regulation. Each row represents a mRNA (green = low expression, red= high expression). (b) In the table the data are expressed as log 2 FC (fold change), comparing HepG2 treated with 10 𝜇MCdtocontrolcells.

The miR-200 family has been related to the suppression of labile zinc concentrations possibly due to the replacement of the epithelial-mesenchymal transition as well. A correlation Zn(II) from the Zn-proteome by Cd(II) and its correspond- between downregulation of the miR-200 family member ing labilization. As known, both MT and ZnT-1 are activated and induction of TGF-𝛽1 in renal proximal tubule epithelial by the metal transcription factor-1 (MTF-1) that functions as cells (NRK-52E) was observed and are responsible for the ametalsensor.MTF-1in vitro is robustly regulated by zinc epithelial-mesenchymal transition [43]. On the other hand, butnotbycadmiumions,asjustpublished[46],and further an overexpression of the family member miR-200b was confirming that labile zinc displaced by cadmium is the pri- demonstrated to significantly reduce cellular proliferation mary mediator in MTF-1 activation. The regulation of MTF-1 and inhibition of cell migration and metastasis in gastric by metals, such as Cd, could be a direct consequence of zinc to cadmium displacement from zinc-containing proteins. The carcinoma cells [44]. All these data strengthen the correlation Cd-ZnexchangeandthepossiblesubstitutionofCdinZn- between high zinc levels, Sanil1 upregulation, miR-34, and binding sites were previously described [7, 18]. The results miR-200 family members downregulation. presented here demonstrate the increment of labile zinc levels in the presence of cadmium, supporting the validity of the 4. Conclusions mechanism proposed above. Further work is underway by our research group to investigate the gene expression and the We have previously demonstrated [23, 45] that Cd exposure biological implications of the modulation of these genes, due in HepG2 cells modulates genes and proteins responsible for to the displacement of zinc with cadmium as a function of metal homeostasis, such as the metallothioneins (MT) and their concentrations. membrane transporters responsible for the elimination of Taking into account our present and previously published zinc excess (i.e., ZnT-1). However, we posed [45] a question on data, the following mechanisms emerge (Figure 5): (i) HepG2 ZnT-1 regulation by Cd exposure: how does Cd regulate ZnT- cells in the presence of Cd have a nonfunctional p53 (previ- 1 levels? We proposed a mechanism of activation by elevated ously published data [19]); (ii) Cd determines an increase of BioMed Research International 7

Adherens junction

Inside Outside Inside Tight junction Transinteractions +p of E-cadherins PAR3 Src FRG Regulation of Transinteractions GTP of nectins actin cytoskeleton GTP Nectin Nectin Afadin Cdc42 Rac (active) p120 ctn GTP 7 LMO ADIP Ponsin 1 IRSp53 Rac/cdc42 NWASP WASP IQGAP 𝛼-actinin Vinculin ZO-1 WAVE 1 Strong adhesion IQGAP Cadherin Cadherin 𝛽-catenin 𝛼-catenin Actin p120 ctn −p − + Actin Formation of A Js 2+ −p − p p polymerization Ca RPTPs p RhoA LMW-PTF CKII (active) GTP PTP𝜇 PTPs Trafficking VEPTP PTP1B LAR 1 ER and Golgi Gene SHP- Free𝛽-catenin TCF/LEF 1 expression DEP Recycling Wnt signaling 𝛽-catenin Cell growth pathway DNA differentiation Endocytosis Degradation +p GDP (inactive) Rac/cdc42 p120 ctn GDP (inactive) 1 IQGAP RhoA Weak adhesion 𝛽-catenin 𝛼-catenin Cadherin Cadherin 120 Actin p ctn + 𝛼 + p +p -actinin ZO-1 + p IQGAP1 p Vinculin Afadin Fer Src RPTKs IGF-1R + Rac Fyn p INSR + GDP p PTKs Yes External signals MET MAPK Cytokine-cytokine ErbB1/2 ERK Slug signaling pathway Cadherin receptor interaction 1 FGFR Snail DNA 𝛽 TGF- 3 DNA TGF-𝛽R signaling pathway Smad Smad4 CBP +p TAK1 NLK

Figure 4: Representation of the adherens junction pathway map from KEGG. Upregulated genes in HepG2 cells by 10 𝜇MCdtreatmentfor 24 hrs are colored in the signaling pathway map.

intracellular labile zinc level, which is a known proliferation open new intriguing interpretations on Cd-induced carcino- signal [10].ThisresponseleadstotheactivationofSnail1, that genicity. along with the upregulation of other genes responsible for loss of adherence, represents a signal promoting cell migration Highlights and metastasis (present work, Figure 4 adherens junction); (iii) miR-34 downregulation contributes to the abnormal (i) HepG2 cells exposed to Cd show an increase of expression of Snail1,whichisnormallyantagonizedbymiR- intracellular labile zinc. 34 and whose pathological expression has been linked to (ii) Labile zinc is visualized and measured by the zinc cancer cell epithelial-mesenchymal transition; (iv) miR-200 specific probe Zinquin. downregulation contributes to the epithelial-mesenchymal transition as well. (iii) Genes belonging to loss of adherence and cell migra- Therefore, our results provide new information on the tion pathways are activated. existing proposed mechanisms for Cd toxicity and biological (iv) Cd exposure affects the expression of miR-34a and effects and along with those already reported in the literature miR-200a. 8 BioMed Research International

2+ Cd

2+ 2+ Cd Zn

2+ Cd 2+ 2+ Zn Zn 2+ Zn MET TGF-𝛽R 2+ Zn Rac/cdc42

+93% Snail

p53

miRNAs 2+ Proliferation and Zn ∙ miR-34a EMT signals 2+ ∙ miR-200a Zn

Cancer

Figure 5: Overview of cellular and molecular effects of 10 𝜇M Cd in HepG2 cells. Cd enters the cells through aspecific sites. Intracellular Cd accumulation determines an increase of labile zinc, a known proliferation signal, and second messenger. The replacement of Zn with Cd in thezincproteome[18] is the hypothetical process underlying the above described mechanism. Genes (Snail, MET, TGF-𝛽R,andRac/cdc42) involved in the loss of cell adherence and also comprised in the mechanism of epithelial-mesenchymal transition are disregulated. MicroRNAs (miR-34a, miR-200a) with tumor-suppressor functions are downregulated. Cd-exposed HepG2 cells have a nonfunctional p53 [19]thatalong with miR-34 downregulation represents the disregulated axis in Snail1-dependent epithelial-mesenchymal transition [20].

Conflict of Interests [4]B.Bocca,A.Pino,A.Alimonti,andG.Forte,“Toxicmetals contained in cosmetics: a status report,” Regulatory Toxicology The authors declare that there is no conflict of interests and Pharmacology,vol.68,no.3,pp.447–467,2014. regarding the publication of this paper. [5] F. Thevenod,´ “Catch me if you can! Novel aspects of cadmium transport in mammalian cells,” BioMetals,vol.23,no.5,pp.857– Acknowledgments 875, 2010. 2+ [6] K. V.Lopin, F.Thevenod,J.C.Page,andS.W.Jones,“Cd´ block The authors are thankful to Dr. Pietro Fumagalli for his and permeation of CaV3.1 (𝛼1G) T-type calcium channels: can- 2 assistance in spectrofluorimetric experiments. Chiara Urani didate mechanism for Cd + influx,” Molecular Pharmacology, and Maurizio Bruschi thank the partial financial support vol.82,no.6,pp.1183–1193,2012. by Fondo di Ateneo, University of Milan Bicocca, and the [7] C. Meplan,´ K. Mann, and P. Hainaut, “Cadmium induces EuropeanCommission.M.FabbriisastudentofthePhD conformational modifications of wild-type p53 and suppresses program in Biotechnology, School of Biological and Medical p53 response to DNA damage in cultured cells,” The Journal of Sciences, University of Insubria (Italy). Biological Chemistry,vol.274,no.44,pp.31663–31670,1999. [8]M.A.Namdarghanbari,J.Bertling,S.Krezoski,andD.H. Petering, “Toxic metal proteomics: reaction of the mammalian References 2+ zinc proteome with Cd ,” Journal of Inorganic Biochemistry, [1] Agency for Toxic Substances and Disease Registry, Toxicological vol. 136, pp. 115–121, 2014. Profile for Cadmium,CentersforDiseaseControlandPreven- [9] P. D. Zalewski, I. J. Forbes, and W. H. Betts, “Correlation tion, Atlanta, Ga, USA, 2012, http://www.atsdr.cdc.gov/toxpro- of apoptosis with change in intracellular labile Zn(II) files/tp5.pdf. using Zinquin [(2-methyl-8-p-toluenesulphonamido-6- [2] IARC, A Review of Human Carcinogens; Part C: Arsenic, Metals, quinolyloxy)acetic acid], a new specific fluorescent probe for Fibers, and Dusts,IARC,Lyon,France,2012. Zn(II),” Biochemical Journal,vol.296,no.2,pp.403–408,1993. [3] A. Martelli, E. Rousselet, C. Dycke, A. Bouron, and J.-M. Moulis, [10] D. Beyersmann and H. Haase, “Functions of zinc in signaling, “Cadmium toxicity in animal cells by interference with essential proliferation and differentiation of mammalian cells,” BioMet- metals,” Biochimie,vol.88,no.11,pp.1807–1814,2006. als,vol.14,no.3-4,pp.331–341,2001. BioMed Research International 9

[11] S. Yamasaki, K. Sakata-Sogawa, A. Hasegawa et al., “Zinc is a [26] A. I. Saeed, V. Sharov, J. White et al., “TM4: a free, open- novel intracellular second messenger,” JournalofCellBiology, source system for microarray data management and analysis,” vol. 177, no. 4, pp. 637–645, 2007. BioTechniques, vol. 34, no. 2, pp. 374–378, 2003. [12] P. F. Dehn, C. M. White, D. E. Conners, G. Shipkey, and [27] A. Kręzel˙ and W.Maret, “Zinc-buffering capacity of a eukaryotic T.A.Cumbo,“Characterizationofthehumanhepatocellular cell at physiological pZn,” Journal of Biological Inorganic Chem- carcinoma (HEPG2) cell line as an in vitro model for cadmium istry,vol.11,no.8,pp.1049–1062,2006. toxicity studies,” In Vitro Cellular & Developmental Biology— [28] C. E. Outten and T. V. O’Halloran, “Femtomolar sensitivity Animal,vol.40,no.5-6,pp.172–182,2004. of metalloregulatory proteins controlling zinc homeostasis,” [13] C. Urani, V. Calini, P. Melchioretto, F. Morazzoni, C. Canevali, Science,vol.292,no.5526,pp.2488–2492,2001. and M. Camatini, “Different induction of metallothioneins and [29] J. L. Vinkenborg, T. J. Nicolson, E. A. Bellomo, M. S. Koay, G. Hsp70 and presence of the membrane transporter ZnT-1 in A. Rutter, and M. Merkx, “Genetically encoded FRET sensors to 2+ HepG2 cells exposed to copper and zinc,” Toxicology in Vitro, monitor intracellular Zn homeostasis,” Nature Methods,vol.6, vol. 17, no. 5-6, pp. 553–559, 2003. no.10,pp.737–740,2009. [14] C. Urani, P. Melchioretto, C. Canevali, F. Morazzoni, and L. [30] D. J. Eide, “Zinc transporters and the cellular trafficking of zinc,” Gribaldo, “Metallothionein and hsp70 expression in HepG2 Biochimica et Biophysica Acta,vol.1763,no.7,pp.711–722,2006. cells after prolonged cadmium exposure,” Toxicology in Vitro, [31] A. B. Nowakowski and D. H. Petering, “Reactions of the vol.21,no.2,pp.314–319,2007. fluorescent sensor, zinquin, with the zinc-proteome: adduct [15]J.G.Hengstler,S.Hammad,A.Ghallab,R.Reif,andP. formation and ligand substitution,” Inorganic Chemistry,vol.50, Godoy, “In vitro systems for hepatotoxicity testing,” in In Vitro no. 20, pp. 10124–10133, 2011. Toxicology Systems, A. Bal-Price and P. Jennings, Eds., chapter [32] J. A. L. Figueroa, K. S. Vignesh, G. S. Deepe Jr., and J. 2, pp. 27–44, 2014. Caruso, “Selectivity and specificity of small molecule fluores- 2+ 2+ [16] C. Urani, P. Melchioretto, C. Canevali, F. Morazzoni, and M. cent dyes/probes used for the detection of Zn and Ca in Camatini, “Copper and zinc uptake and hsp70 expression in cells,” Metallomics,vol.6,no.2,pp.301–315,2014. HepG2 cells,” Toxicology in Vitro,vol.15,no.4-5,pp.497–502, 2001. [33] G. Malgieri, M. Palmieri, S. Esposito et al., “Zinc to cadmium replacement in the prokaryotic zinc-finger domain,” Metal- [17] C. Urani, P. Melchioretto, C. Canevali, and G. F. Crosta, “Cyto- lomics,vol.6,no.1,pp.96–104,2014. toxicity and induction of protective mechanisms in HepG2 cells exposed to cadmium,” Toxicology in Vitro,vol.19,no.7,pp.887– [34] M. A. Nieto, “The Snail superfamily of zinc-finger transcription 892, 2005. factors,” Nature Reviews Molecular Cell Biology,vol.3,pp.155– 166, 2002. [18] L. Tang, R. Qiu, Y. Tang, and S. Wang, “Cadmium–zinc exchange and their binary relationship in the structure of Zn- [35] C. Hogstrand, P. Kille, M. L. Ackland, S. Hiscox, and K. M. related proteins: a mini review,” Metallomics,vol.6,no.8,p.1313, Taylor, “A mechanism for epithelial-mesenchymal transition 2014. and anoikis resistance in breast cancer triggered by zinc channel ZIP6 and STAT3 (signal transducer and activator of transcrip- [19] C. Urani, P. Melchioretto, M. Fabbri, G. Bowe, E. Maserati, and tion 3),” Biochemical Journal,vol.455,no.2,pp.229–237,2013. L. Gribaldo, “Cadmium Impairs p53 Activity in HepG2 Cells,” ISRN Toxicology,vol.2014,ArticleID976428,9pages,2014. [36] L. Trusolino, A. Bertotti, and P. M. Comoglio, “MET signalling: principles and functions in development, organ regeneration [20] N. H. Kim, H. S. Kim, X.-Y. Li et al., “A p53/miRNA-34 axis and cancer,” Nature Reviews Molecular Cell Biology, vol. 11, no. regulates Snail1-dependent cancer cell epithelial-mesenchymal 12, pp. 834–848, 2010. transition,” Journal of Cell Biology,vol.195,no.3,pp.417–433, 𝛽 2011. [37] J. Massague,´ “TGF in cancer,” Cell,vol.134,no.2,pp.215–230, 2008. [21] M. Sarwar Nasir, C. J. Fahrni, D. A. Suhy, K. J. Kolodsick, C. P. Singer, and T. V. O’Halloran, “The chemical cell biology [38] A. Sadok and C. J. Marshall, “Rho GTPases—masters of cell of zinc: structure and intracellular fluorescence of a zinc- migration,” Small GTPases,vol.5,no.4,ArticleIDe29710,pp. quinolinesulfonamide complex,” Journal of Biological Inorganic 1–7, 2014. Chemistry,vol.4,no.6,pp.775–783,1999. [39] J. Zheng, X.-X. Zhang, H. Yu, J. E. Taggart, and W.-Q. Ding, [22] P. Coyle, P. D. Zaleewski, J. C. Philcox et al., “Measurement of “Zinc at cytotoxic concentrations affects posttranscriptional zinc in hepatocytes by using a fluorescent probe, zinquin: rela- events of gene expression in cancer cells,” Cellular Physiology tionship to metallothioneins and intracellular zinc,” Biochemical and Biochemistry,vol.29,no.1-2,pp.181–188,2012. Journal,vol.303,no.3,pp.781–786,1994. [40] D. P. Bartel, “MicroRNAs: target recognition and regulatory [23] M. Fabbri, C. Urani, M. G. Sacco, C. Procaccianti, and L. functions,” Cell,vol.136,no.2,pp.215–233,2009. Gribaldo, “Whole genome analysis and microRNAs regulation [41] B. Adams, A. Kasinski, and F. Slack, “Aberrant regulation and in HepG2 cells exposed to cadmium,” ALTEX,vol.29,no.2,pp. function of microRNAs in cancer,” Current Biology,vol.24,no. 173–182, 2012. 16, pp. R762–R776, 2014. [24] G. K. Smyth, “Linear models and empirical Bayes methods for [42]M.Haneklaus,M.Gerlic,L.A.J.O’Neill,andS.L.Masters, assessing differential expression in microarray experiments,” “MiR-223: infection, inflammation and cancer,” Journal of Statistical Applications in Genetics and Molecular Biology,vol. Internal Medicine, vol. 274, no. 3, pp. 215–226, 2013. 3,no.1,article3,25pages,2004. [43] M. Xiong, L. Jiang, Y. Zhou et al., “The miR-200 family [25] M. Kanehisa, S. Goto, M. Furumichi, M. Tanabe, and M. regulates TGF-𝛽1-induced renal tubular epithelial to mesenchy- Hirakawa, “KEGG for representation and analysis of molecular mal transition through smad pathway by targeting ZEB1 and networks involving diseases and drugs,” Nucleic Acids Research, ZEB2 expression,” The American Journal of Physiology—Renal vol.38,no.1,pp.D355–D360,2009. Physiology,vol.302,no.3,pp.F369–F379,2012. 10 BioMed Research International

[44] J. Kurashige, H. Kamohara, M. Watanabe et al., “MicroRNA- 200b regulates cell proliferation, invasion, and migration by directly targeting ZEB2 in gastric carcinoma,” Annals of Surgical Oncology,vol.19,no.3,supplement,pp.S656–S664,2012. [45] C. Urani, P. Melchioretto, and L. Gribaldo, “Regulation of metallothioneins and ZnT-1 transporter expression in human hepatoma cells HepG2 exposed to zinc and cadmium,” Toxicol- ogy in Vitro,vol.24,no.2,pp.370–374,2010. [46] S. Choi and A. J. Bird, “Zinc’ing sensibly: controlling zinc homeostasis at the transcriptional level,” Metallomics,vol.6,no. 7, pp. 1198–1215, 2014.