Transgenerational Epigenetic Inheritance TOP ARTICLES SUPPLEMENT

Home , Epigenome, Germline, Midblastula, Somatic (biology)

CONTENTS EDITORIAL:: Epigenetics: where environment, society and genetics meet Epigenomics Vol. 6 Issue 1 REVIEW: Transgenerational epigenetic inheritance: adaption through the germline epigenome? Epigenomics Vol. 7 Issue 5 PERSPECTIVE: Transgenerational epigenetic inheritance: an open dicussion Epigenomics Vol. 7 Issue 5 Editorial

Editorial Epigenetics: where environment, society and genetics meet

“The goal is to use epigenetics to anticipate health in the individual and, more importantly, the population. Before this can be done, several challenges must be faced.”

Keywords: environment • epigenetics • noncoding RNA • socioeconomic status • transgenerational epigenetic inheritance

Understanding how environment, social fac- with environmental chemical exposure [3], tors and genetics combine to affect patterns race, gender [4] and income level [5]. How- of health is an urgent priority. Social factors ever, epigenetic gene regulation is bigger than that affect health patterns include race and methylation alone. socioeconomic status (SES), while biological Human and animal studies have begun factors include sex, genetics and epigenetics. to investigate environmental effects on epi- The consequence of biology and environ- genetics. While these publications are increas- ment can be observed in life expectancy ing, some factors need to be considered, these Amber V Majnik data. At 25 years old, life expectancy for include tissue specificity, epigenetic specific- Author for correspondence: white men is 4.4 years greater than for Afri- ity and transgenerational studies. Department of Pediatrics Medical can–American men. Females follow a similar College of Wisconsin, Milwaukee, pattern. Income also affects life expectancy Tissue specificity WI, USA with high income earners outliving their low Tissue-specific epigenetic profiles are well Tel.: +1 414 955 2374 [email protected] income counterparts, and white males and accepted, and so it must be noted that blood females outliving African–American males cells also have cell-specific profiles. Circulat- and females by 2–4 years at every income ing blood is a convenient source of material level [1] . Environment also affects perina- for epigenetic testing. It is minimally invasive tal mortality. US-born African–American to collect and can be collected multiple times women experience greater rates of perinatal over a period of time. However, develop- mortality than their foreign-born counter- ment and environment changes the cellular parts [2]. Similar patterns can be observed makeup of blood. with multiple health conditions, including Cellular heterogeneity may confound epi- cancer, depression, diabetes and heart disease genetic results, since epigenetics differs from highlighting the importance of this issue one blood cell to another [6,7]. For example, [1] . What is less clear is the mechanism by neutrophils predominate blood of a new- Robert H Lane which these disparities are caused. One way born, but, by 2–4 years, lymphocytes are Department of Pediatrics Medical in which biology and environment can both the majority population [8]. Furthermore, College of Wisconsin, Milwaukee, leave their fingerprint is through epigenetic cord blood also varies in its cellular make WI, USA changes. up compared to peripheral blood [9]. These Epigenetics refers to the regulation of gene two factors can confound the comparison of expression through modifications to DNA or methylation profiles from a newborn to an DNA-associated proteins. Epigenetic regula- adult, even in the same individual. The con- tion includes histone modifications, DNA stituents of the blood can also be changed if methylation and nonprotein coding RNA. the individual is experiencing inflammation, DNA methylation has been a particular focus an infection or high amounts of stress [10–12] . of environment-mediated changes in health. All of these factors can be influenced by an Methylation changes have been associated individual’s SES. These differences highlight part of

how developmental and environmental changes can established effect of environment on epigenetics, envi- alter epigenetic results. ronmental factors such as those effecting SES could The importance of different epigenetic patterns disrupt this process. While the role for noncoding between specific cell types and whole blood will vary RNA is still being investigated, many studies have depending on disease or condition. Subtle differences begun to investigate how the environment can change are likely more important for multifactorial diseases, the life course of multiple generations and the role of like obesity, where the environment, genetics and social epigenetics in this process. factors all contribute to pathology. Furthermore, while methylation patterns may appear similar between cell Transgenerational epigenetic inheritance types, changes in histone code or noncoding RNA may One of the most challenging and intriguing aspects of differ. epigenetics is the transmission of modifications to subsequent generations. Transgenerational studies indicate Epigenetic specificity that epigenetic modifications are not wiped clean for We believe social and environmental factors rely on each generation, as was originally thought [16,17]. This, epigenetics to effect biologic change. The study of how transgenerational epigenetic inheritance, coupled with the environment affects phenotype is environmental the environmental influences of epigenetics indicates epigenetics. A cell has many epigenetic tools to modify that the environment of one generation can modify its gene expression in response to the environment. the phenotype of the following generations. This far- However, to understand environmental epigenetics, we reaching epigenetic effect indicates the importance of must look beyond gene expression and protein coding understanding environmental influences, such as SES, regions and pay attention to regulatory and noncod- on epigenetics. ing regions. This includes studying methylation and The life course approach determines how the social histone modifications, as well as noncoding RNA. and environmental context of one generation can The majority of SES studies give a limited view of affect another. For example, a maternal grandmother’s the epigenetic landscape. The majority of these stud- exposure to neighborhood poverty during pregnancy ies investigate SES and changes in global methylation. predicts the risk for infant low birth weight and sub- Some studies have taken an extra step to investigate sequent mortality among African–American women, methylation in disease relevant genes. However, to independent of the mother’s status [18]. A detailed fully appreciate the epigenetic impact of the environ- understanding of transgenerational epigenetics is nec- ment, analysis must go beyond methylation. Animal essary to understand this complex inheritance. models allow for a more detailed investigation of the epigenetic profile. Pena et al. analyzed DNA methyla- “The role for noncoding RNA in epigenetics tion and histone modifications of the estrogen recep- is just beginning to be understood; however, tor alpha gene in the brains of dams and pups in a a number of studies indicate that RNA could be a powerful player in transgenerational maternal care model [13] . These types of tissue specific, epigenetics. detailed epigenetic studies provide a more rigorous ” demonstration of epigenetic consequences. Stress is a core component of socioeconomic disad- While methylation is the best studied epigenetic vantage. Both prenatal as well as postnatal stress has mark, followed by histone modifications, the composi- been shown to effect long-term health. Epigenetic tion of our DNA is telling us that RNA may be much changes in offspring exposed to maternal stress have more important. The discovery that <3% of the human been noted in multiple species. In Drosophila, heat genome encodes protein-coding genes was surprising stress induces phosphorylation of the transcription fac- to many people. But perhaps what is more astounding tor dATF-2 [19] . This leads to a loss of heterochroma- is that scientists have now found that approximately tin structures at several region of the genome, which is 76% of the human DNA is transcribed into some sort transmitted to the next generation. In mice, the early of RNA [14] . Much of the 76% is thought to have a life stress of chronic unpredictable maternal separation regulatory role, including a role in epigenetics. leads to changes in behavior and alters the methylation The role for noncoding RNA in epigenetics is just profile in the promoter of candidate genes in the germ- beginning to be understood; however, a number of line and brain of affected males, as well as in the germ- studies indicate that RNA could be a powerful player line of the offspring [20]. The licking and grooming in transgenerational epigenetics. The role for RNA as a behavior of rats has also been shown to be associated mechanism of transfer of transgenerational phenotype with epigenetic changes and inherited [13] . come from experiments of microinjections of oliogri- Several additional aspects of SES have been modeled bonucleotides into egg and sperm cells [15] . Given the in studies of transgenerational epigenetic inheritance;

2 Epigenomics (2014) 11(1) future science group Epigenetics: where environment, society & genetics meet Editorial one such aspect is prenatal diet. Areas of socioeconomic It is clear that numerous environmental factors can disadvantage are often part of food deserts, or areas contribute to an individual’s epigenetic inheritance of with limited access to healthy affordable food [21] . This health. The next step forward in advancing our under- can lead to high-fat, low-nutrient diets and contribute standing should be to integrate multiple environmen- to SES health disparity. This type of diet has negative tal influences. This will not only create a better life consequences for pregnant women, as well as subse- course model, but the increased complexity will help to quent generations. Animal studies have shown that discern the causality of health disparities. maternal exposure to high-fat diet results in increased Epigenetics exists at the intersection between genet- body size and reduced sensitivity to insulin through two ics and the environment. The goal is to use epigenetics generations via both the maternal and paternal lineages to anticipate health in the individual and, more impor- [22]. This effect continues into the third generation via tantly, the population. Before this can be done, several the paternal lineage. This transgenerational epigenetic challenges must be faced. These include investigating inheritance of diet is associated with imprinted gene the nonprotein coding region of the genome, studying patterns and changes in methylation. common diseases where the environment likely contrib- Transgenerational epigenetic studies have also begun utes to the gradation of severity, and developing new, to investigate the social and racial disparity in toxicant complex models that include multiple environmental burden. Economically disadvantaged areas have an influences and have transgenerational impacts. increased risk of environmental toxin exposure [23–25]. These toxins include endocrine disruptors, heavy met- Financial & competing interests disclosure als, tobacco smoke and air pollution, among others. The authors have no relevant affiliations or financial involve- Similar to diet studies, exposure of gestating females ment with any organization or entity with a financial inter- to environmental toxins leads to increased disease for est in or financial conflict with the subject matter or mate- multiple generations [3]. Also associated with the toxin rials discussed in the manuscript. This includes employment, derived phenotype are changes in epigenetics. Several consultancies, honoraria, stock ownership or options, expert studies have found changes in differential methylation testimony, grants or patents received or pending, or royalties. regions, which are proposed to act as biomarkers of No writing assistance was utilized in the production of this exposure [3]. manuscript.

References 7 Adalsteinsson BT, Gudnason H, Aspelund T et al. Heterogeneity in white blood cells has potential to confound 1 Williams DR, Mohammed SA, Leavell J, Collins C. Race, DNA methylation measurements. PloS ONE 7(10), e46705 socioeconomic status, and health: complexities, ongoing (2012). challenges, and research opportunities. Ann. NY Acad. Sci. 8 Orkin SH,Nathan DG, Ginsburg D, Look AT, Fisher 1186, 69–101 (2010). DE, Lux SE,Nathan and Oski’s Hematology of Infancy and 2 Collins JW Jr, Soskolne GR, Rankin KM, Bennett AC. Childhood. Saunders Elsevier, PA, USA (2009). Differing first year mortality rates of term births to white, 9 Lopez MC, Palmer BE, Lawrence DA. Phenotypic African-American, and Mexican-American US-born and differences between cord blood and adult peripheral blood. foreign-born mothers. Matern. Child Health J. 17(10), Cytometry B Clin. Cytom. 76(1), 37–46 (2009). 1776–1783 (2012). 10 Landmann RM, Muller FB, Perini C, Wesp M, Erne P, 3 Manikkam M, Tracey R, Guerrero-Bosagna C, Skinner MK. Buhler FR. Changes of immunoregulatory cells induced Plastics derived endocrine disruptors (BPA, DEHP and DBP) by psychological and physical stress: relationship to plasma induce epigenetic transgenerational inheritance of obesity, catecholamines. Clin. Exp. Immunol. 58(1), 127–135 (1984). reproductive disease and sperm epimutations. PLoS ONE 8(1), e55387 (2013). 11 Williams DM, Johnson NW. Alterations in peripheral blood leucocyte distribution in response to local inflammatory 4 Zhang FF, Cardarelli R, Carroll J et al. Significant stimuli in the rat. J. Pathol. 118(3), 129–141 (1976). differences in global genomic DNA methylation by gender and race/ethnicity in peripheral blood. Epigenetics 6(5), 12 Bonneau RH, Sheridan JF, Feng NG, Glaser R. Stress- 623–629 (2011). induced effects on cell-mediated innate and adaptive memory components of the murine immune response to herpes simplex 5 Tehranifar P, Wu HC, Fan X et al. Early life socioeconomic virus infection. Brain Behav. Immun. 5(3), 274–295 (1991). factors and genomic DNA methylation in mid-life. Epigenetics 8(1), 23–27 (2013). 13 Pena CJ, Neugut YD, Champagne FA. Developmental timing of the effects of maternal care on gene expression and 6 Wu HC, Wang Q, Delgado-Cruzata L, Santella RM, Terry epigenetic regulation of hormone receptor levels in female MB. Genomic methylation changes over time in peripheral rats. Endocrinology 154(11), 4340–4351 (2013). blood mononuclear cell DNA: differences by assay type and baseline values. Cancer Epidemiol. Biomarkers Prevent. 21(8), 14 Consortium TEP. An integrated encyclopedia of DNA 1314–1318 (2012). elements in the human genome. Nature 489(7414), 57–74 (2012).

future science group www.futuremedicine.com 3 Editorial Majnik & Lane

15 Kiani J, Grandjean V, Liebers R et al. RNA–mediated 21 Beaulac J, Kristjansson E, Cummins S. A systematic review epigenetic heredity requires the cytosine methyltransferase of food deserts, 1966–2007. Prevent. Chronic Dis. 6(3), A105 Dnmt2. PLoS Genet. 9(5), e1003498 (2013). (2009). 16 Lane N, Dean W, Erhardt S et al. Resistance of IAPs to 22 Dunn GA, Bale TL. Maternal high-fat diet effects on methylation reprogramming may provide a mechanism for third-generation female body size via the paternal lineage. epigenetic inheritance in the mouse. Genesis 35(2), 88–93 Endocrinology 152(6), 2228–2236 (2011). (2003). 23 Hajat A, Diez-Roux AV, Adar SD et al. Air pollution 17 Guerrero-Bosagna C, Skinner MK. Environmentally induced and individual and neighborhood socioeconomic status: epigenetic transgenerational inheritance of phenotype and evidence from the multi-ethnic study of atherosclerosis disease. Mol. Cell. Endocrinol. 354(1–2), 3–8 (2012). (MESA). Environ. Health Perspect. 121(11–12), 1325–1333 18 Collins JW Jr, David RJ, Rankin KM, Desireddi JR. (2013). Transgenerational effect of neighborhood poverty on low 24 Tyrrell J, Melzer D, Henley W, Galloway TS, Osborne birth weight among African Americans in Cook County, NJ. Associations between socioeconomic status and Illinois. Am. J. Epidemiol. 169(6), 712–717 (2009). environmental toxicant concentrations in adults in the 19 Seong K-H, Li D, Shimizu H, Nakamura R, Ishii S. USA: NHANES 2001-2010. Environ. Int. 59, 328–335 Inheritance of stress-induced, ATF-2-dependent epigenetic (2013). change. Cell 145(7), 1049–1061 (2011). 25 Vrijheid M, Martinez D, Aguilera I et al. Socioeconomic 20 Franklin TB, Russig H, Weiss IC et al. Epigenetic status and exposure to multiple environmental pollutants transmission of the impact of early stress across generations. during pregnancy: evidence for environmental inequity? J. Biol. Psychiatry 68(5), 408–415 (2010). Epidemiol. Commun. Health 66(2), 106–113 (2012).

4 Epigenomics (2014) 11(1) future science group Review

For reprint orders, please contact: [email protected]

7 Review 2015/07/28 Transgenerational epigenetic inheritance: adaptation through the germline epigenome?

Epigenomics Epigenetic modifications direct the way DNA is packaged into the nucleus, making Lexie Prokopuk1,2, Patrick genes more or less accessible to transcriptional machinery and influencing genomic S Western‡,1,2 & Jessica M stability. Environmental factors have the potential to alter the epigenome, allowing Stringer*,‡,1,2 1 genes that are silenced to be activated and vice versa. This ultimately influences disease Centre for Genetic Diseases, Hudson Institute of Medical Research, 27–31 susceptibility and health in an individual. Furthermore, altered chromatin states can Wright Street, Clayton, Victoria 3168, be transmitted to subsequent generations, thus epigenetic modifications may provide Australia evolutionary mechanisms that impact on adaptation to changed environments. 2Molecular & Translational Science, However, the mechanisms involved in establishing and maintaining these epigenetic Monash University, Clayton, Victoria 3168, modifications during development remain unclear. This review discusses current Australia *Author for correspondence: evidence for transgenerational epigenetic inheritance, confounding issues associated Tel.: +61 3 9594 7009 with its study, and the biological relevance of altered epigenetic states for subsequent [email protected] generations. ‡Authors contributed equally 846 Keywords: epigenetic inheritance • epigenetics • evolution • germline • transgenerational epigenetic inheritance

Multicellular organisms are composed of embryonic development, while retaining lin- hundreds of highly specialized cell types eage information and guiding organization that are ultimately derived from a single of the body plan. cell, the zygote, which forms at fertilization. After fertilization, the zygote has the The diverse range of cell types found in each capacity to make all cell types in the body, a individual is accomplished through speci- quality known as totipotency. As the embryo fication and differentiation of cells during develops, cells lose totipotency and become embryonic and fetal development. Changes restricted to specialized lineages. However, 5 in cell specialization are achieved through to ensure reproductive capacity, sexually complex interactions involving cell signaling, reproducing species must maintain a popu- transcriptional control and establishment lation of germ cells that have the capability of mitotically heritable epigenetic modifi- of regaining totipotency after fertilization. 2015 cations, which regulate global expression Thus, germ cells facilitate transmission of the patterns and restrict cells to their defined genome and epigenome between parent and lineages. Epigenetic modifications include offspring in a manner that supports embryo- DNA methylation and histone modifica- genesis. In mammals, this is achieved in part tions, which establish specialized chromatin through the ability of germ cells to remove states that restrict or permit gene expression. existing epigenetic information and estab- Although epigenetic modifications provide lish an epigenetic program that is compatible long-term cell lineage information, they are with totipotency. also dynamic and can respond to influences Epigenetic reprogramming includes the such as developmental cues or environmen- widespread removal and re-establishment of tal factors. This flexibility permits changes epigenetic information, and occurs exten- in cell differentiation and function during sively in germ cells and preimplantation part of

embryos during normal development (Figure 1). These Epigenetic reprogramming in the germline two reprogramming events remove existing epigenetic Epigenetic reprogramming in germ cells is most nota- information and establish new epigenetic modifica- bly defined by the genome-wide erasure and re-estab- tions throughout the genome. The first reprogram- lishment of DNA methylation in primordial germ cells ming event occurs specifically in the germline and (PGCs) (Figure 2). PGCs are the precursor cells that is required to reset the epigenome prior to the estab- develop into mature oocytes and sperm. During mouse lishment of sex-specific epigenetic information in embryonic development, PGCs are specified during the gametes [1] . The second wave of reprogramming gastrulation at embryonic day (E)6–7, before prolif- (preimplantation reprogramming) occurs after fertil- erating and migrating from the allantois to enter the ization and is required for the establishment of cell developing gonads by E10.5 [6,7]. During migration, differentiation pathways that underpin embryogenesis the PGCs undergo significant epigenetic reprogram- (reviewed in [2]). However, not all of the epigenetic ming, including the removal of DNA methylation information transmitted to the zygote by the sperm and remodeling of histone modifications [8–12]. The and oocyte is removed in the second wave of repro- PGCs then enter the developing gonads, and undergo gramming. This parent-specific epigenetic informa- further epigenetic reprogramming with DNA meth- tion is required for development of the embryo and ylation ultimately reduced to 14 and 7% by E13.5, in adult life, and is considered to mediate inherited epi- male and female germ cells, respectively [13] . Addition- genetic effects in the offspring. While studies inves- ally, reprogramming of histone modifications [14], and tigating epigenetic reprogramming in mammals have exchange of histone variants occurs [12], although the focused on the mouse, reprogramming has also been dynamic nature of these histone marks is partly attrib- reported in other species including the human (Homo uted to fluctuations in chromatin modifications dur- sapiens), rabbit (Oryctolagus cuniculus), ovine (Ovis ing cell cycle progression [15,16]. These changes demon- aries), bovine (Bos taurus), Caenorhabditis elegans and strate extensive reprogramming in the germline, and Arabidopsis thaliana) [3–5]. the molecular processes involved have been covered in Accumulating evidence indicates that altered pro- greater detail in recent reviews [2,17–18]. gramming of epigenetic information in the germline At the onset of sex determination, the level of DNA may affect offspring development and fitness. These methylation is at its lowest and provides an epigen- changes may result in epigenome–genome interac- etic ‘clean slate’ for the establishment of sex-specific tions that alter gene expression patterns that favor a epigenetic programming in the germline, includ- heritable genetic trait. On this basis it has been specu- ing the establishment of genomic imprints. Genomic lated that inherited epigenetic information can signifi- imprinting involves the epigenetic silencing, of either cantly contribute to an organism’s ability to adapt to the paternal or maternal copy of each imprinted gene. changing environments and hence influence evolution. This results in expression of alleles in a parent-of-origin In this review, we will discuss current evidence for specific manner [19–23] . Imprinted genes are marked transgenerational epigenetic inheritance and how this by parental allele-specific DNA methylation that is information may influence evolution. established in the germline after sex determination.

Germline reprogramming Preimplantation reprogramming Fertilization PGCs

PGCs specification PGC migration Gonad Mature Zygote ~E6-7 gametes

Figure 1. Germline and preimplantation reprogramming during mouse embryo development. PGCs are specified at around E7, before proliferating and migrating to the developing gonad (the genital ridge) at ∼E10.5. During migration and population of the developing gonad, PGCs undergo extensive epigenetic reprogramming that removes existing epigenetic information. The germ cells then commit to, and differentiate along the spermatogenic (male) and oogenic (female) lineages and sex-specific epigenetic information is established in the male and female germlines. After birth, germ cells complete gametogenesis, producing mature oocytes and sperm which are capable of supporting formation of a totipotent zygote at fertilization. After fertilization a second round of epigenetic reprogramming (preimplantation reprogramming) facilitates early development and lineage specification in the peri- and post-implantation embryo. PGC: Primordial germ cell.

830 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: adaptation through the germline epigenome? Review

Spermatogenesis

Histone-protamine exchange

Prolif and migration vels

eration

TRIM28 Maternal factor deposition STELLA (proteins and RNA) Global 5mc le Mitotic arrest

Oogenesis Entry into meiosis

E7 E12.5 Birth Puberty

Figure 2. Germline epigenetic reprogramming in the mouse. Primordial germ cells are specified at ∼E7 and DNA methylation is removed during primordial germ cell proliferation and migration. Germ cells commit to male and female development and enter mitotic arrest and meiosis, respectively. Sex-specific epigenetic information is established from E13.5. In male germ cells de novo DNA methylation increases rapidly with the activation of the de novo DNA methyltransferases, while DNA methylation levels remain low throughout oogenesis. In males, spermatogenesis occurs at sexual maturity and involves the exchange of histones for protamines, and compaction of the nucleus into the sperm head. In females, oocyte maturation involves the deposition of maternal factors that are required for supporting early development in the offspring. 5mC: 5-methylcytosine.

In male germ cells, DNA methylation at imprinted factors resulting in reduced male fertility and/or an loci is established from E14.5. This coincides with a altered ability to transition through meiosis (reviewed rapid increase in global DNA methylation levels from in [35]). In addition, epigenetic modifiers maintain approximately 10% to around 50% by E16.5 through epigenetic information in the gametes for utilization the action of DNMT3A, DNMT3B and the co-factor in the subsequent generation. This appears to be of DNMT3L [13,24–30]. In females, although global DNA significant importance during spermiogenesis, when methylation levels remain low throughout oogen- histones (modified and unmodified) are exchanged esis, DNMT activity is required for establishment of for protamines to facilitate the packaging of chroma- maternal imprints in the female germline [24,26,28,31–33]. tin into a sperm head, the smallest cell in the body While germ cell epigenetic reprogramming is vital (Figure 3). Since most histones are lost in this process, for resetting the epigenome in preparation for fertiliza- it may appear unlikely that transmission of histone tion and embryonic development, it may also be impor- modifications through the paternal germline would tant for germ cell development. A recent study identi- contribute to the offspring. However, a small percent- fied 513 and 727 genes in male and female fetal germ age of histones are retained at specific genes in sperm cells, respectively, that are bivalently enriched for both of mice (∼1%) and humans (∼10%), indicating that active trimethylated lysine 4 on histone 3 (H3K4me3) transmission of some histone modifications could and repressive trimethylated lysine 27 on histone 3 influence early development in the offspring [36,37]. (H3K27me3) histone modifications [34]. Out of these Erkek et al. demonstrated that methylated DNA is genes, 147 remained ‘poised’ between E12.5 until at packaged by protamines (Figure 3C), but CpG rich least E14.5 in both sexes, while other genes resolved to unmethylated transcriptional start sites (TSSs) retain either a repressed or active state. None of the genes that nucleosomes (Figure 3) [38]. Moreover, canonical his- resolved toward an active state were shared between tones (H3.1 and H3.2) are exchanged for the histone the sexes, indicating a role for these epigenetic modi- variant H3.3 at CpG rich active TSS which carry fications in controlling expression of genes involved in H3K4me3 (Figure 3A). By contrast, CpG rich TSSs germ cell differentiation and identity [34]. that are not methylated, are not transcribed and are marked by H3K27me3, do not exchange H3.1/H3.2 Spermatogenesis for H3.3 (Figure 3B). Interestingly, these genes that A number of epigenetic modifiers are required for male retain H3.1/H3.2 and H3K27me3 marks are primar- germline development and function, with loss of these ily associated with developmental processes. Reten-

future science group www.futuremedicine.com 831 Review Prokopuk, Western & Stringer

tion of H3K27me3 at these loci in sperm results in with sperm (∼25%) methylation levels within CGIs gene repression during preimplantation development. increase as oocytes mature. Approximately 1600 CGIs Polycomb repressive complex 2 (PRC2) functions to are completely methylated in mature oocytes and only catalyze the addition of methyl groups to H3K27. 185 CGIs are methylated in mature sperm. One hun- Therefore, the establishment of H3K27me3 by PRC2 dred of these completely methylated CGIs are shared by in the developing germline may provide a mecha- both mature oocytes and sperm and occur at intragenic nism for inherited epigenetic information that affects loci. Interestingly, in GV stage oocytes of Dnmt3a-/- and development in a parent’s offspring. Dnmt3l-/- mice, genome-wide CpG methylation is pro- foundly reduced, but CGIs remain highly methylated. Oogenesis As no new methylation can be established in the absence Female germ cell epigenetic programming continues of both Dnmt3a and Dnmt3l, these data suggest that after birth with de novo methylation occurring as oocytes CGIs are protected during PGC reprogramming. Fur- mature. Smallwood et al. used high-throughput reduced thermore, 15% methylated CGIs in wild-type oocytes representation bisulfite sequencing (RRBS) to investi- retain DNA methylation levels ≥40%, through to blas- gate CpG methylation in mature oocytes [39]. This tech- tocyst stage. This provides evidence that CGIs retain nique provides quantitative data specifically enriched methylation during germline reprogramming, and that for genomic regions with a high density of potential some of this methylation is also retained through the CpG methylation sites and enables investigation of preimplantation reprogramming period. DNA methylation at single-nucleotide resolution across Other factors that may influence maintenance and the genome [40]. Therefore, this analysis is restricted to inheritance of epigenetic information are mRNAs the majority of promoters and repeated sequences with and proteins that are deposited during oocyte CpG islands (CGIs). During oocyte maturation, CGI growth and maturation. These factors are essential methylation increases from 0.5% in day 5 oocytes to for oocyte function and early development of the approximately 11% in day 20 germinal vesicle (GV) embryo. Factors such as STELLA (DPPA3), ZFP57 stage oocytes and approximately 15% in ovulated and TRIM28 are required for the maintenance of metaphase II stage oocytes [39]. Although the overall genomic imprinting during preimplantation epigenetic CGI methylation level is lower in oocytes compared reprogramming [41–43].

A K4me3 B K27me3 C

H2A H2A H2A H4 H3.1/2 H4 H3.1/2 H4 H3.1/2 H2B H2B H2B

H3.1/2 H3.1/2

Histone exchange Retention of histones Histone–protamine exchange

H3.3 H2A H2A K4me3 K27me3 H4 H3.3 H4 H3.1/2 H2B H2B

Unmethylated CpG Methylated CpG Protamine

Figure 3. Retention and exchange of nucleosomes during spermatogenesis. During spermiogenesis histones are replaced by protamines, but this process is incomplete. (A & B) Unmethylated CpG-rich regions retain histones, (C) methylated CpG rich regions exchange all histones for protamines. (A) When an unmethylated CpG-rich transcriptional start site is associated with H3K4me3 and gene transcription, H3.1 and H3.2 are replaced by H3.3. (B) When an unmethylated CpG rich transcriptional start site is associated with H3K27me3 and gene repression, H3.1 and H3.2 are retained. These epigenetic conformations are transmitted to the offspring, where they may influence developmental programs and provide mechanisms for epigenetic inheritance.

832 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: adaptation through the germline epigenome? Review

Preimplantation reprogramming TET enzymes catalyze the conversion of 5-methyl- Preimplantation epigenetic reprogramming was first cytosine (5mC) to 5-hydroxymethylcytosine (5hmC) reported in the mouse as DNA methylation levels were and depletion of TET1 results in decreased levels of observed to be lower in the preimplantation embryos, 5hmC and the retention of 5mC levels (Figure 5A) [58]. compared with the levels in mature sperm or oocytes In the mature oocyte and zygote, Tet3 is present at and those found after implantation [44,45]. More recent high levels, before declining after the first zygotic divi- studies have determined the extent and timing of this sion. This coincides with the majority of hydroxylation reprogramming event in a variety of species [3,4]. In the of 5mC to 5hmC in the paternal genome, which results mouse the paternal genome releases protamines and in loss of DNA methylation after DNA replication [57]. is re-packaged by maternal nucleosomal histones to Critically, during preimplantation epigenetic repro- form the paternal pronucleus 1 h postfertilization [46]. gramming some epigenetic modifications, including By approximately 7–8 h postfertilization, the paternal those involved in the regulation of imprinted genes, genome appears to lose a substantial amount of DNA escape demethylation and maintain their methylated methylation, whereas low levels are retained in the state into adulthood [55,59]. Maternal factors such as maternal genome until E2.5 [47–50]. DNA methylation STELLA and the components of the TRIM28 complex levels in the blastocyst inner cell mass (ICM) are simi- are thought to be vital for maintaining this epigenetic lar to those observed in the maternal pronucleus, after information during reprogramming [60,61] . which DNA methylation rapidly increases as cells of Levels of the maternal factor STELLA are high in the ICM undergo differentiation and begin to establish the early embryo and are required for preimplanta- somatic lineages of the embryo. tion development [41] . STELLA binds to H3K9me2, A study by Li et al. challenged the paradigm of which is found at maternally imprinted loci and some active global demethylation during preimplantation paternally imprinted loci, and protects imprinted genes reprogramming using immunolocalization in preim- from epigenetic reprogramming (Figure 5B) [60,62]. plantation mouse embryos [51] . In this study, trypsin- Similarly another maternal factor, TRIM28, is a digestion resulted in detection of DNA methylation in scaffolding protein that recruits epigenetic machin- one-cell and two-cell stage embryos, indicating that ery involved in gene silencing, such as SETDB1 and DNA demethylation may be less pronounced at these UHRF1, which methylate H3K9me3 and recruit stages than previously thought. In other species such as DNMT1 and DNMT1o, respectively, maintaining Xenopus laevis (African clawed frogs) [52] and bovine [53], methylation within imprinted differentially methyl- very little DNA demethylation occurs in zygotes, while ated regions (DMRs) (Figure 5C) [43,63–67]. ZFP57 is a there is no reduction in DNA methylation in Danio KRAB zinc finger protein, which recognizes and binds rerio (zebra fish) and rabbit zygotes [53,54]. With the KRAB domains, and recruits TRIM28 (Figure 5C). caveat that analytical differences may exist, these stud- KRAB domains are located within the majority of ies suggest that preimplantation reprogramming may imprinting DMRs as well as other regions within the not occur to the same extent in all species. genome [42]. Deletion of either Trim28 or Zfp57 results Recently, Smith and colleagues used RRBS to in hypomethylation of some, but not all imprinted provide a more detailed analysis of CGI methylation DMRs [42,68]. Therefore, other factors and complexes throughout preimplantation development in mice [55]. may be involved in maintaining epigenetic information The authors identified dynamic changes in DNA during epigenetic reprogramming in preimplantation methylation during two developmentally important embryos. stages (Figure 4). Upon fertilization CGIs in the paternal genome are rapidly demethylated. This is followed Evidence of epigenetic inheritance by a more gradual demethylation, during cleavage, Epigenetic inheritance refers to the germline transmis- such that the level of DNA methylation is at its low- sion of epigenetic information to the somatic cells of est in the ICM, before increasing postimplantation. the resulting offspring. To achieve this, the epigenetic These global DNA methylation dynamics were also information transmitted by the sperm and/or oocyte identified in human embryos [56]. These genome-wide must escape preimplantation reprogramming and epigenetic data reveal epigenetic processes through be maintained during fetal, and later development. which developmental programs are likely to be set for There is clear evidence that epigenetic information is regulating differentiation of somatic lineages during inherited from parents to offspring (intergenerational postimplantation development. inheritance), with the strongest demonstration pro- The process of demethylation in both the PGCs and vided by imprinted genes. However, whether epigen- the zygote is thought to involve both passive, and active etic information is transmitted over multiple genera- mechanisms involving TET enzymes (Figure 5) [57]. tions without reprogramming in the germline is less

future science group www.futuremedicine.com 833 Review Prokopuk, Western & Stringer

Sperm vels Global 5mc le

Oocyte

Zygote 2-cell 4-cell 8-cell Blastocyst

Figure 4. Regulation of DNA methylation levels throughout preimplantation reprogramming based on reduced representation bisulfite sequencing data from Smith et al. and Smallwood et al. CpG islands in the paternal genome undergo rapid demethylation, while maternal CpG islands retain methylation. DNA methylation levels are gradually reduced during preimplantation development, resulting in low methylation levels in the inner cell mass of the blastocyst. 5mC: 5-methylcytosine. Data taken from [39,55]. clear. For the purpose of this discussion it is important experiments led to the conclusion that the genomes in to distinguish between inter- and transgenerational the oocyte and sperm are not identical, and the pres- inheritance (Figure 6). Intergenerational transmission ence of both a male and female pronuclei is essential refers to the transmission of information from parent for normal development [69–71] . Genetic complementa- to offspring via the gametes and the maintenance of tion studies (via chromosomal translocations) demon- this information in somatic lineages of the offspring. strated that specific maternally and paternally derived Transgenerational transmission involves the transmis- regions function differently during embryogenesis [72]. sion of germline epigenetic information over multiple It was proposed that sex-specific modification of the generations, with epigenetic alterations being main- maternal and paternal genomes occurs during game- tained and affecting transcription and/or phenotype togenesis, a process that was referred to as genomic in multiple subsequent generations. For intergenera- imprinting [69]. Indeed, the vast majority of imprinted tional inheritance to occur epigenetic modifications genes in the mouse are regulated by DNA methyla- must be perpetuated only through preimplantation tion that is established in a sex-specific manner dur- reprogramming, either by ‘escaping’ reprogramming, ing germline development [73]. This methylation is or potentially, by being erased and re-written in the maintained through zygotic reprogramming resulting same context (Figure 6B). However, transgenerational in the retention of DMRs at imprinted genes in the inheritance requires epigenetic modifications to be postimplantation embryo. perpetuated through both germline and preimplanta- As PGCs are derived from cells that would otherwise tion reprogramming to achieve transmission between take on a somatic fate, the nascent germ cells carry the multiple generations (Figure 6C). same DMRs and other epigenetic information as the The first evidence that intergenerational epigenetic surrounding somatic cells of the embryo [1,74]. There- inheritance could significantly effect offspring devel- fore, germline epigenetic reprogramming ensures that opment came with the discovery of genomic imprint- each generation receives the correctly methylated allele ing [69,70]. Imprinting was discovered in the early from each parent, by erasing pre-existing DMRs and 1980s through a series of zygotic pronuclear trans- establishing new sex-specific DMRs during develop- plant experiments using genetically inbred mice [69,70]. ment of spermatogenic and oogenic lineages. Inter- When two paternal, or two maternal genetically iden- estingly, the remethylation of germline DMRs occurs tical genomes were used to initiate development in asynchronously. For example, the demethylated enucleated oocytes, the embryos failed to develop. In paternal allele of the H19 imprinting control region contrast, complementation of a maternal and paternal (ICR) becomes remethylated prior to the previously copy of the genome supported development. These unmethylated maternal allele in male germ cells [75–77].

834 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: adaptation through the germline epigenome? Review

Similarly, in developing oocytes the maternal Snrpn these sites is important for stability [84]. This is sup- allele is methylated before the paternal allele [78]. ported by recent evidence demonstrating that CGIs This suggests that these loci were either incompletely located close to IAPs maintain their methylation state demethylated during epigenetic reprogramming or that through germ cell reprogramming [13,85–86]. Combined, other marks such as histone modifications, chromatin these findings suggest that the chromatin environment binding proteins or non-coding RNAs (ncRNAs) are surrounding IAPs confer resistance to reprogramming retained in the absence of methylation to direct ordered at neighboring loci, potentially providing mechanisms remethylation at the appropriate sites in the develop- for transgenerational inheritance [13,85–86]. ing germline. Interestingly, transcriptional activation Other classes of repetitive elements include long of genes targeted for remethylation may be required interspersed elements (LINEs), which are generally to initiate de novo methylation during male germline hypomethylated in oocytes but hypermethylated in development [79] and oocyte maturation [80,81] . sperm; and nonautonomous short interspersed ele- DNA methylation must also be regulated at many ments (SINEs), which are hypomethylated in sperm, nonimprinted sequences. These include repeat ele- but hypermethylated in oocytes [84,87–88]. Given the ments and transposons, which make up a large pro- similarities between DNA methylation mechanisms portion of the mammalian genome (over 45% in the used to silence repetitive elements and imprinted human genome) [82]. It is imperative that these ele- genes, it has been suggested that genomic imprint- ments are silenced, as some of these repetitive elements ing may have adapted from host defense mechanisms cause genomic instability and mutations as a result of used to silence invading elements [89–94]. In support of their mobilization and random re-integration. This is this, most ICRs at imprinted loci are associated with particularly detrimental when transposons insert into direct repeats [95,96]. This also provides evidence for the protein-coding genes, regulatory regions and/or cause evolution of epigenetic mechanisms that regulate the abnormal chromosomal recombination [83]. To pro- genome in response to outside influences, particularly tect the integrity of the genome, epigenetic silencing genomic invasion by foreign sequences. mechanisms such as piwi-interacting RNAs (piRNAs), This continual shaping of the genome may also histone modifications and DNA methylation, have provide mechanisms for adaptation to environmental evolved to repress repetitive elements and their detri- pressures. Indeed, evidence suggests that the germline mental activity. Repeat elements such as intracisternal epigenome is subject to alteration by environmental A particles (IAPs) show resistance to demethylation, influences, such as diet, chemical exposure and even indicating that maintenance of DNA methylation at stress [97]. Typically these effects are measured by

A B C

TET

TET TET

K9me2 UHRF1 K9me2 ZFP57 STELLA TRIM28 DNMT1 Cell replication K9me2

SETDB1 K9me3

5mC

5hmC K9me2

Figure 5. Demethylation in the preimplantation embryo and the role of maternal factors in protecting specific sites from epigenetic reprogramming. (A) During preimplantation reprogramming TET enzymes catalyze the conversion of 5mC to 5hmC, and demethylation of 5hmC subsequently occurs as a result of cell replication. (B) Maternally supplied STELLA binds to H3K9me2 and protects CpG sites from TET demethylation. (C) ZFP57 binds to KRAB domain and recruits co-factor TRIM28. ZFP57/TRIM28 recruits SETDB1 to methylate H3K9me3 and co-factor UHRF1 which recruits DNMT1 and DMNT1o maintaining 5mC. 5mC: 5-methylcytosine; 5hmc: 5-hydroxymethylcytosine.

future science group www.futuremedicine.com 835 Review Prokopuk, Western & Stringer

Environmental effect F0 on germline

A B C

Not heritable Intergenerational Transgenerational inheritance inheritance

Figure 6. Modes of epigenetic inheritance. Exposure of the F0 germline to an environmental insult can alter epigenetic state, causing epimutations that may or may not be inherited. (A) No inheritance: epimutations in the parental F0 germline do not affect the offspring (F1). Such epimutations are presumably corrected during F1 preimplantation reprogramming. (B) Intergenerational inheritance: epimutations transmitted through the F0 germline escape preimplantation reprogramming and alter development in the F1 generation. However, these epimutations are corrected in the germline of F1 animals and are not transmitted to the F2 generation. (C) Transgenerational inheritance: epimutations escape both preimplantation and germline reprogramming in F1 and subsequent generations and affect development over multiple generations.

the appearance of variable phenotypes in subsequent that were differentially regulated in the F1 animals, generations, but in many cases the molecular mecha- methylation at the ICRs of these genes was maintained nisms underlying these effects remain unknown. and there were no changes in the methylation profile As imprinted DMRs are known to be inherited and of these genes in sperm. On this basis the authors survive preimplantation reprogramming, these loci concluded that epigenetic programming of ICRs in have been targets for many studies investigating epi- the germline is not susceptible to change as a result genetic inheritance [42,86,98]. However, it has been sug- of caloric restriction. However, as gene expression was gested that due to the targeted maintenance of the affected by under nutrition, it remains possible these epigenetic information during embryonic develop- imprinted genes may be targeted for dosage modula- ment, imprinted loci may be more resistant to envi- tion in the fetus as an adaptive response to adverse ronmental insult. Radford et al. challenged mice with in utero environment [98]. caloric restriction and investigated the susceptibility of Recently studies using C. elegans have provided imprinted genes to expression changes in a number of evidence that microRNAs regulate epigenetic inheri- different tissues over two generations [98]. Interestingly, tance, possibly by directing epigenetic machinery dur- imprinted genes, as a class, were neither more, nor less, ing reprogramming [99,100]. Gene-specific silencing was susceptible than nonimprinted genes to expression per- induced by feeding C. elegans bacteria expressing dou- turbation [98]. Although a number of imprinted genes ble-stranded RNAs (dsRNAs). When animals carry- were differentially regulated in a tissue-specific manner ing a histone-GFP transgene reporter were exposed to in the F1 generation, these alterations diminished in dsRNA corresponding to the GFP mRNA, the trans- the second (F2) generation. Out of the imprinted genes gene was epigenetically silenced. This silencing was

836 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: adaptation through the germline epigenome? Review maintained after removal of the dsRNA and was inher- Similarly, a single stimulus may cause a cascade of gene ited in over 60% of individuals for at least four gen- expression changes that are relayed through networks erations, suggesting that this silencing was maintained of genes to result in the observed phenotype. Clearly, through epigenetic programming in the germline [99]. in some cases the indirect nature of these mechanisms The PIWI argonaute proteins, and their associated can make it difficult to identify the causative modifica- piRNAs are distinct from small-interfering RNAs tion that initiated the gene expression cascade. These (siRNAs) as they have a specific and evolutionarily considerations provide profound challenges for iden- conserved role in transposon silencing in the germline tifying phenotypic changes that can be definitively in many invertebrate [101] and vertebrate [102] species linked to transgenerational inheritance over multiple (reviewed in [103]). It has been suggested that piR- generations. For these reasons, some studies identi- NAs have the ability to initiate an epigenetic memory fying transgenerational epigenetic inheritance may of ‘self’ and ‘nonself’ RNA [100] . When ‘nonself’ for- be confounded by influences of the parent in each eign RNA is detected, targeted silencing occurs [104] . subsequent generation [105] . While a piRNA trigger initially induced the targeted silencing, both small RNA factors and chromatin Maternal transmission regulators were required to complete and maintain Parents can influence their offspring in numerous ways, multigenerational transgene silencing, including the including mechanisms mediated directly through uter- putative H3K9me3 methyltransferases, HPL-2 and ine environment, or more indirectly through mecha- SET-25/32 [99]. Although the piRNA and siRNA sys- nisms such as behavioral cues. For example, studies tems vary between C. elegans and vertebrate species, have suggested that maternal care in mice is affected mechanisms regulating piRNAs and histone modifi- by stress, with stressed mothers more likely to produce cations are highly conserved and may produce similar stressed offspring [106,107]. However, what was origi- downstream silencing. nally thought of as a case of transgenerational epigenetic inheritance was subsequently attributed to altera- Confounding issues in the study of tions in the hypothalamic–pituitary–adrenal (HPA) epigenetic inheritance axis in the offspring [107] . Francis et al. carried out a Tracing the inheritance of DNA is relatively straight- series of experiments in rats, which involved cross- forward as mutations are rare and it is easy to track fostering pups from stressed mothers to control moth- the heritage of specific mutations through each gen- ers [106] . Naturally occurring variations in maternal eration by DNA sequencing. While the majority of care, such as licking and grooming (LG) behavior were epigenetic information, including DNA methylation, assessed to observe the effect on offspring stress [106] . is mitotically stable and important for cellular iden- They showed that female offspring of high LG moth- tity, some aspects are more plastic and can be altered ers showed significantly higher LG behavior (toward in response to different environmental stimuli. This their own pups) in comparison to the offspring of low enables cells to change, and potentially provides a LG mothers. However, cross fostering pups from low mechanism for individuals to adapt to altered environ- LG mothers to high LG mothers resulted in those pups ments. For transgenerational epigenetic inheritance having higher LG behavior in motherhood, which is to occur, altered epigenetic states need to be resistant presumably a behavior learned from the high LG foster to reprogramming and maintained in both germline mothers or due to factors and stress hormones absorbed and somatic lineages (Figure 6C). Therefore, to identify during lactation. Although this model provides inter- and trace specific epigenetic mutations across genera- esting insights into the influence of maternal behavior tions, the epigenome in both the germline and affected on offspring behavior, it does not provide evidence of somatic cells must be examined in each generation. In transgenerational epigenetic inheritance. addition, to test the effect of a specific environmen- During pregnancy a mother, her embryo and the tal factor on the epigenome in a founding individual, embryo’s germ cells may all be exposed to the same the environment that each subsequent generation is environmental influences (via the placenta). This expo- exposed to must be free of the stimulus of interest, and sure may alter the epigenome in the germ cells of the other confounding factors. For example, to examine embryo, and therefore, individuals derived from the the effect of maternal diet, it is necessary to distinguish embryo’s germline (i.e., the mother’s grandchildren). between the effect a mother’s diet has on the germline In mice, the methylation-sensitive metastable agouti of the fetus, and the effect on the somatic cells of the viable yellow (Avy) allele can influence coat color of offspring that may induce confounding factors in the genetically identical offspring (Figure 7). Mothers fed a following generation. These confounding factors may diet rich in methyl donors produce a reduced percent- in turn, affect outcomes in the following generation. age of yellow-coated pups [108] . The increased avail-

future science group www.futuremedicine.com 837 Review Prokopuk, Western & Stringer

effect that may alter methylation at the Avy allele in the offspring. However, the results from embryo-transfer experiments and specific breeding schemes indicate that the presence of the IAP in the Avy allele is more likely to retain epigenetic marks when inherited from the oocyte rather than through the sperm [113] . Inter- estingly, Blewitt et al. demonstrated that the Avy allele Percentage methylation of LTR at agouti locus is completely demethylated in the blastocyst before implantation suggesting that DNA methylation may Figure 7. Epigenetic control and inheritance at the not be the inherited epigenetic mark in this case [115] . vy A allele. Coat color variation ranging from 100% Although it is possible that DNA methylation leaves a yellow, mottled (∼50% yellow), to agouti (0% yellow) in isogenic mice carrying the methylation-sensitive footprint for remethylation after erasure in the epiblast, metastable Avy allele. Increased methylation at the the observations of Blewitt et al. indicate that epigen- long terminal repeat at the agouti locus correlates with etic effects observed using the Avy may be controlled by loss of transcription at the Avy allele and a consequent other epigenetic mechanisms. increase in agouti coat color. To avoid the effects of maternal environment, embryo LTR: Long terminal repeat. transfer methods have been utilized to study transgen- ability of methyl donors shifts the coat color toward erational epigenetic effects. Padmanabhan et al. found pseudoagouti due to higher DNA methylation and epi- that a hypomorphic mutation in the mouse Mtrr gene genetic silencing at the Avy allele [109] . In experiments resulted in intrauterine growth restriction, develop- by Cropley et al. pregnant females were fed a methyl mental delay and congenital malformations (neural donor rich diet, whereas their offspring were not and tube, heart and placental defects) [116] . The MTRR the coat color of F2 animals also shifted significantly enzyme is necessary for the metabolism of folate, a toward pseudoagouti, compared with controls [110] . As methyl-rich B group vitamin required for successful the F1 generation was only exposed to a methyl donor pregnancy and development of the fetus. Wild-type F2 diet in utero, this indicates that the diet altered Avy pro- and F3 offspring showed abnormal developmental phe- gramming in the germ cells of the F1 animals (i.e., the notypes consistent with their maternal Mtrr heterozy- founding germline for the F2 generation). However, gous grandparents. Global DNA methylation was sig- the phenotypic effect was heritable for only one gen- nificantly reduced in maternalMtrr deficient mice, and eration [111]. Therefore, it is likely that the phenotype their F2 and F3 wild-type offspring. To demonstrate was lost through reprogramming in the germ cells of that these effects were independent of maternal uterine the F2 animals (founders of the F3 generation) [112] . environment embryo transfers were performed. Wild- Importantly, this indicates that the effect of feeding type preimplantation embryos derived from pedigrees methyl donors was maintained only as an intergenera- with one heterozygous maternal grandparent were col- tional effect (i.e., only in the offspring derived from the lected at E3.25 [116], prior to DNA remethylation dur- affected PGCs), not as a transgenerational effect. ing preimplantation epigenetic reprogramming. These The Avy allele is controlled by an IAP inserted embryos were transferred to pseudo-pregnant wild-type upstream of the agouti allele. Interestingly, the spec- females and allowed to develop until E10.5. Growth trum of phenotypes in offspring from maternally inher- defects (intrauterine growth restriction and develop- ited Avy skews toward the maternal phenotype with a mental delay) were not observed in the transferred lit- higher proportion of pseudoagouti pups produced by ters. However, congenital malformations (neural tube, Avy females. However, when the Avy allele is inherited heart and placental defects) persisted in animals up to from the father, a higher percentage of pseudoagouti the F3 generation, indicating that the effects were due offspring are produced regardless of the father’s coat to transgenerational epigenetic inheritance, although color [108,113]. One possible explanation for this is that the mechanisms were not identified. IAPs are more stringently targeted for silencing by the PIWI pathway in the male germline (reviewed in [114]), Paternal transmission leading to greater silencing of the IAP-linked Avy locus. Rather than tracing epigenetic effects through the The observed maternal inheritance of the Avy allele was maternal line, it can be simpler to trace epigenetic initially presumed to be due to the maternal environ- transmission via the paternal germline. Paternal trans- ment that the pups were exposed to in utero. Yellow Avy mission of environmentally induced phenotypes pro- mice are obese and hyperinsulinemic, whereas pseu- vides a robust model for studying epigenetic inheri- doagouti Avy mice are lean and healthy. These pheno- tance, as the male contribution to the next generation typic differences provide a confounding environmental is mediated only through the sperm, without contribu-

838 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: adaptation through the germline epigenome? Review tions from the uterine environment. Females not only notype in the second-generation, but further studies have the potential to influence the next generation via are required to determine whether this is the case. the oocyte, but also prenatally via the placenta and Male mice exposed to folate-deficient diets (FD) postnatally via lactation and maternal care. A recent in utero and throughout life, sire offspring with mul- study of paternal epigenetic inheritance demonstrated tiple musculoskeletal defects and placental abnor- that mice exposed to early life trauma sired progeny malities [128] . Methylated DNA immunoprecipitation with reduced fear and insulin hypersensitivity [117] . promoter array (me-DIP-chip) of FD male sperm com- Deep sequencing revealed that several microRNAs pared with control sperm identified DMRs associated were upregulated in the sperm of the traumatized mice with functions in the central nervous system and mus- compared with control mice. Reduced fear and insulin cle tissue. In the placentas of FD sired offspring (F2), hypersensitivity were observed in the F1, F2 and F3 only 2 out of approximately 400 differentially expressed animals compared with control animals. Changes in loci were correlated to DMRs in the FD sperm. How- miRNA expression were observed in blood serum and ever, no differences in DNA methylation were observed brain (hippocampus), but not in the sperm of the F2 at these two loci in the F2 placenta. Therefore, the dif- mice. Injection of the mutant sperm RNAs into wild- ferences in DNA methylation observed in the sperm type oocytes resulted in offspring with similar behav- are unlikely to have directly caused the observed phe- ioral and metabolic abnormalities as those observed notypes. In a similar study, Radford et al. exposed in traumatized males [117] . This study provides novel pregnant females to 50% caloric restriction during late evidence that an RNA-dependent mechanism affects gestation, coinciding with the time of epigenetic pro- paternal transmission of environmentally triggered gramming in developing male germ cells [126] . Caloric traits in mammals. restriction decreased DNA methylation at 111 loci, and The rapid global rise in diabetes, obesity and car- increased methylation at a further 55 loci in the sperm diovascular disease suggests that nongenetic factors of F1 male mice. When F1 males (control and restricted significantly contribute to disease risk. This has driven calorie groups) were mated to control females and fed debate over the past decade around the possibility that a normal calorie diet, differential methylation was lost environment can influence disease inheritance through in the F2 tissues examined (brain and liver of E16.5 epigenetic mechanisms [118]. While the underlying embryos). However, differences in gene expression asso- mechanisms are yet to be determined, environmentally ciated with glucose tolerance and metabolism persisted induced inter- and trans-generational effects have been in the absence of differential DNA methylation. In both reported [119–126] . Metabolic effects may be modulated of these studies [126,128] it is speculated that gene expres- through epigenetic mechanisms induced by high-fat sion differences might be perpetuated by mechanisms diet. Dunn et al. exposed female mice to high-fat diet other than DNA methylation, such as histone modi- before, throughout and after pregnancy [119] . First- fications. Interestingly, global H3K4 and K9 mono- and second-generation offspring of both maternal and methylation and H3K9 trimethylation were reduced in paternal lineages showed reduced insulin sensitivity the FD sperm compared with controls [128], indicating and increased body length. In the third generation that histone methylation levels may be sensitive to folate only increased body length was detected and this phe- availability. However, it should also be noted that folate notype was paternally, but not maternally, transmitted deficiency has been linked to increased DNA damage to female offspring. While these phenotypes were ini- and therefore genetic mutations could contribute in tially transmitted through both lineages, supporting a these models [129] . Indeed, a marker of DNA damage role for epigenetic inheritance, the mechanisms of this (γ-H2AX) was detected at higher levels in FD sper- inheritance remain unclear. Furthermore, DNA meth- matocytes compared with controls. This damage was ylation in the brain (arcuate nucleus) of second-gener- not observed in mature sperm [128] . Therefore, although ation offspring was reduced at the promoter region of it is possible that the observed phenotypes were due to Ghsr, which correlated with an increase in transcription alterations in the sperm epigenome, it is also plausible and a larger body size. However, it was not determined that aberrant DNA repair mechanisms may also con- whether this difference in DNA methylation origi- tribute. Therefore, it would be of particular interest to nated in the germline of the first-generation animals. identify specific regions within the genome that are par- Therefore, although these data provide evidence that ticularly sensitive to folate-driven DNA damage and to glucose homeostasis and body length are modulated determine if these genomic regions are associated with by maternal diet/caloric availability in utero, the role the observed phenotypes. These studies [126,128] high- for germline epigenetic inheritance is less clear [127] . It light the potential for confounding effects that can remains possible that the in utero environment affects influence the interpretation of inherited phenotypes, programming of the fetal germ cells, resulting in a phe- even when using a paternal inheritance model.

future science group www.futuremedicine.com 839 Review Prokopuk, Western & Stringer

Epigenetic inheritance may also occur in an indirect, ual level with minimal contribution of group selection. trans-acting manner. Disrupting epigenetic modifiers This results in gradual evolution through the selection can result in multiple errors during epigenetic program- of individuals with more advantageous phenotypes. ming in the germline that may subsequently alter pre- It is now clear that soft inheritance [136], which refers implantation reprogramming in the next generation. to nongenetic inheritance, such as epigenetic modifi- Using yellow (Avy/a) mothers, Chong et al. examined the cations, niche construction and possibly learning and coat colors of wild-type offspring sired by agouti (a/a) cultural influences, provides a potentially important males that were either heterozygous or wild-type for a influence on evolution. For example, mother–offspring point mutation in either Dnmt1 or Smarca5; epigenetic interactions in utero or postnatally affect the develop- modifiers that alter DNA methylation and histone orga- ment of an individual. These interactions could alter nization, respectively [130] . Mutations in either of these both survival and reproductive fitness. If these interac- genes caused a shift in the penetrance at the Avy allele, tions alter heritable epigenetic marks or are genetically such that wild-type animals from heterozygous fathers linked, these factors could influence evolution. Herita- were more likely to be yellow than offspring sired from ble epialleles are common in plants, fungi, invertebrates a wild-type father. Significantly this work indicates that and vertebrates and have been proposed to be particu- mutating epigenetic modifiers in the paternal germ- larly important for producing variation and promoting line can affect gene expression in the next generation adaptation in the absence of genetic variation [133] . Epi- through nongenetic information carried by the sperm genetic information in the germline, such as genomic to the offspring. Moreover, the Avy allele was inherited imprinting, may also initiate reproductive isolation from the mother, indicating that the effects resulting with variations resulting in incompatibilities within from the germline epimutations in the male caused a hybrid individuals [137] . Therefore, the modern synthe- trans-acting effect on gene expression in the offspring. sis view of heredity and evolution has been expanded to As this type of trans-acting effect is presumably indi- incorporate these soft inheritance mechanisms [138–142] . rect, Chong et al. speculate that RNA or chromatin Epigenetic processes may have evolved to calibrate modifying proteins in the sperm may be responsible organisms to stochastic environments. However, it has for altering expression of maternal alleles in the off- also been speculated that extreme adverse environmen- spring [130] . Indeed, recent studies using C. elegans have tal influences can promote programming defects in the associated small RNAs and histone modifications to germline that impair the individual to adapt to environ- multigenerational epigenetic inheritance [99–100,131–132]. mental stresses [143]. This may be particularly relevant These studies provide important insights into the rela- for human societies that have altered diets contain- tionship between germline epigenetic patterning and ing high-sugar and fat, and a greater exposure to toxic mechanisms facilitating the transmission of epigenetic chemicals, which may alter epigenetic processes in the information from one generation to the next. germline. A study by Li et al. used a mouse model with natural-onset maternal obesity and diabetes to demon- Adaptive evolution & natural selection strate that induced epigenetic changes have an adaptive It is possible that with each generation the epigenetic function [144] . Mice carrying the active Avy allele are pre- state can change to benefit survival of the offspring, disposed to becoming obese and developing type-2 dia- however due to confounding influences in most stud- betes [145] . When a/a offspring from obese yellow Avy/a ies, the significance of epigenetic mechanisms in inheri- mothers were compared with isogenic a/a offspring from tance remain to be clarified [133,134]. Although epigen- lean pseudoagouti Avy/a mothers after being fed con- etic modifications have mostly been studied in the trol and high-fat diets, offspring of yellow obese mice context of development and disease, they may also have showed a latent defect in glucose and lipid metabolism a role in evolutionary processes. Adaptive evolution, when fed a high-fat diet. Remarkably, this phenotype first described by Charles Darwin and Alfred Wallace, can be evaded by avoiding high-fat diets. This subop- involves the gradual process whereby environmental timal metabolic response is likely the result of damage pressure results in the selection of phenotypes within a induced by an abnormal intrauterine environment. population that benefit reproductive fitness (survival of Moreover, the molecular changes that occur due to the fittest). Over 70 years ago this Darwinian view of the in utero exposure do not directly indicate disrupted evolution was combined with the theory of Mendelian metabolic pathways, but rather developmental ontolo- genetics to produce the modern synthesis model for evo- gies [144] . Therefore, it is possible that programming lution [135] . Under this model, transmission of germline that occurred in utero facilitated survival of the devel- genes is equated to variations in DNA sequence and oping embryo, at the expense of increased disease risk importantly is not affected by the developmental his- later in life. Although, adult function may be negatively tory of the individual. Selection occurs at the individ- affected, if a mutation promotes early survival and repro-

840 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: adaptation through the germline epigenome? Review ductive success, then this adaptation, albeit outwardly of resulting offspring and therefore the species. Fur- suboptimal, will disseminate through the population ther developing our understanding of the mechanisms and influence the evolution of the species [146] . involved in epigenetic programming and mainte- Genomic imprinting provides another example of nance of epigenetic modifications in the germline will what might appear as ‘suboptimal’ epigenetic inheri- enhance our understanding of inherited epigenetic tance. For imprinting to have evolved, the selective information and its possible role in evolution. advantage of monoallelic expression must have out- weighed the cost of increased mortality due to the Future perspective increased rate of exposing recessive deleterious alleles. While evidence for transgenerational epigenetic inheri- Genomic imprinting affects both the growth and func- tance is strengthening, it remains challenging to isolate tion of organs involved in nutrient supply (e.g., pla- its specific effects on offspring from confounding factors. centa), and thereby modulates fetal nutrient supply and New technologies including genome-wide epigenetic growth [147] . Epigenetic regulation of these factors is studies are beginning to improve our understanding of vital for both embryonic and maternal survival during epigenetic programming in the germline and during pre- pregnancy and lactation [19–21,23,125]. This may provide implantation development. Expansion of this work will a balance between nutritional supply that ensures off- provide a more complete scaffold for understanding epi- spring success, and sufficient nutrient restriction that genetic programming and the transmission of epigenetic prevents a mother from being disadvantaged to the information between parent and offspring. These new point of compromised survival or future reproductive technologies are now being complemented with experi- success [148–151] . Genomic imprinting involves com- mental systems that directly test the function of epi- plex and strictly regulated mechanisms of epigenetic genetic modifiers in the germline by using models such programming that is essential for development. It is as C. elegans and mice. These models will identify how unlikely that such a mechanism of epigenetic inheri- changes in epigenetic modifications in the germline affect tance has evolved without evolutionary relevance, and development of the offspring and the epigenetic modifiers is likely an active mechanism that promotes survival that mediate these effects. This is essential to understand

Executive summary Epigenetic reprogramming in the germline • Germline epigenetic reprogramming involves the genome-wide erasure and re-establishment of DNA methylation in germ cells. • This mechanism is vital for the establishment of sex-specific epigenetic information and correct development in the next generation. • Histone marks are dynamic during germ cell epigenetic reprogramming, with some histone modifications remaining in a ‘poised’ state. Preimplantation reprogramming • Preimplantation reprogramming is essential for embryonic development, allowing cells to proliferate and differentiate into all cell types in the developing embryo. • Maternal factors such as STELLA and TRIM28 protect imprinted genes from epigenetic reprogramming in preimplantation embryos. Evidence of epigenetic inheritance • Epigenetic modifications need to ‘escape’ both germline and preimplantation epigenetic reprogramming to be inherited in a transgenerational manner. • Mechanisms maintaining retrotransposon silencing may have been adapted to enable epigenetic inheritance. Confounding issues of the study of epigenetic inheritance • Transgenerational epigenetic inheritance is difficult to study, due to limitations in isolating cause and effect, such as separating germline epigenetic changes from confounding in utero effects in the developing fetus. Adaptive evolution & natural selection • Epigenetic inheritance may provide evolutionary mechanisms that facilitate adaptation of an organism to adverse environments. Future perspective • Further study of how imprinted genes and IAPs escape reprogramming will advance our understanding of how loci are resistant to epigenetic reprogramming. • Genome-wide analyses of histone modifications throughout early development (germline and preimplantation reprogramming) will provide a greater understanding about epigenetic reprogramming mechanisms.

future science group www.futuremedicine.com 841 Review Prokopuk, Western & Stringer

the molecular mechanisms underlying intergenerational granted to PS Western and the Victorian Government’s Op- and transgenerational epigenetic inheritance and will erational Infrastructure Support Program. L Prokopuk is sup- have wide-ranging implications for human health. ported by an Australian Postgraduate Award. PS Western and JM Stringer are supported by National Health and Medical Re- Acknowledgements search Grants 1043939 and 1051223 awarded to PS Western. The authors would like to thank K Hogg for manuscript review The authors have no other relevant affiliations or financial in- and comments. volvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials Financial & competing interests disclosure discussed in the manuscript apart from those disclosed. This work was supported by funding from the Monash Univer- No writing assistance was utilized in the production of this sity Faculty of Medicine, Nursing and Health Sciences funding manuscript.

References 12 Hajkova P, Ancelin K, Waldmann T et al. Chromatin Papers of special note have been highlighted as: dynamics during epigenetic reprogramming in the mouse germ • of interest; •• of considerable interest line. Nature 452(7189), 877–881 (2008). 1 Hajkova P, Erhardt S, Lane N et al. Epigenetic 13 Seisenberger S, Andrews S, Krueger F et al. The dynamics of reprogramming in mouse primordial germ cells. Mech. Dev. genome-wide DNA methylation reprogramming in mouse 117(1–2), 15–23 (2002). primordial germ cells. Mol. Cell 48(6), 849–862(2012). • Demonstrates epigenetic reprogramming in primordial •• Used whole genome bisulfite sequencing and RNA-seq to germ cells (PGCs). identify dynamics of DNA methylation reprogramming in mouse PGCs. 2 Saitou M, Kagiwada S, Kurimoto K. Epigenetic reprogramming in mouse pre-implantation development and 14 Ancelin K, Lange UC, Hajkova P et al. Blimp1 associates with primordial germ cells. Development 139(1), 15–31 (2012). Prmt5 and directs histone arginine methylation in mouse germ cells. Nat. Cell. Biol. 8(6), 623–630 (2006). 3 Morgan HD, Santos F, Green K, Dean W, Reik W. Epigenetic reprogramming in mammals. Hum. Mol. Genet. 15 Kagiwada S, Kurimoto K, Hirota T, Yamaji M, Saitou M. 14(Suppl. 1), R47–R58 (2005). Replication-coupled passive DNA demethylation for the erasure of genome imprints in mice. EMBO J. 32(3), 340–353 (2013). 4 Feng S, Jacobsen SE, Reik W. Epigenetic reprogramming in plant and animal development. Science 330(6004), 622–627 16 Aoto T, Saitoh N, Sakamoto Y, Watanabe S, Nakao M. (2010). Polycomb group protein-associated chromatin is reproduced in post-mitotic G1 phase and is required for S phase progression. 5 Kelly WG. Transgenerational epigenetics in the germline J. Biol. Chem. 283(27), 18905–18915 (2008). cycle of Caenorhabditis elegans. Epigenetics Chromatin 7(1), 6 (2014). 17 Hackett JA, Zylicz JJ, Surani MA. Parallel mechanisms of epigenetic reprogramming in the germline. Trends Genet. 6 Mintz B, Russell ES. Gene-induced embryological 28(4), 164–174 (2012). modifications of primordial germ cells in the mouse.J. Exp. Zool. 134(2), 207–237 (1957). 18 Stringer J, Barrand S, Western P. Fine-tuning evolution: germ-line epigenetics and inheritance. Reproduction 146(1), 7 Tam PPL, Snow MHL. Proliferation and migration of R37–R48 (2013). primordial germ-cells during compensatory growth in mouse embryos. J. Embryol. Exp. Morph. 64, 133–147 (1981). 19 Stringer JM, Pask AJ, Shaw G, Renfree MB. Post-natal imprinting: evidence from marsupials. Heredity 113(2), 8 Tachibana M, Ueda J, Fukuda M et al. Histone 145–155 (2014). methyltransferases G9a and GLP form heteromeric complexes and are both crucial for methylation of 20 Constancia M, Kelsey G, Reik W. Resourceful imprinting. euchromatin at H3-K9. Genes Dev. 19(7), 815–826 (2005). Nature 432(7013), 53–57 (2004). 9 Yabuta Y, Kurimoto K, Ohinata Y, Seki Y, Saitou M. Gene 21 Li Y, Sasaki H. Genomic imprinting in mammals: its life cycle, expression dynamics during germline specification in molecular mechanisms and reprogramming. Cell Res. 21(3), mice identified by quantitative single-cell gene expression 466–473 (2011). profiling.Biol. Repro. 75(5), 705–716 (2006). 22 Bartolomei MS, Ferguson-Smith AC. Mammalian genomic 10 Seki Y, Yamaji M, Yabuta Y et al. Cellular dynamics imprinting. Cold Spring Harb. Perspect. Biol. 3(7), pii:a002592 associated with the genome-wide epigenetic reprogramming (2011). in migrating primordial germ cells in mice. Development 23 Charalambous M, Da Rocha ST, Ferguson-Smith AC. 134(14), 2627–2638 (2007). Genomic imprinting, growth control and the allocation of 11 Seki Y, Hayashi K, Itoh K, Mizugaki M, Saitou M, Matsui nutritional resources: consequences for postnatal life. Curr. Y. Extensive and orderly reprogramming of genome-wide Opin. Endocrinol. Diabetes Obes. 14(1), 3–12 (2007). chromatin modifications associated with specification and 24 Kaneda M, Okano M, Hata K et al. Essential role for de novo early development of germ cells in mice. Dev. Biol. 278(2), DNA methyltransferase Dnmt3a in paternal and maternal 440–458 (2005). imprinting. Nature 429(6994), 900–903 (2004).

842 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: adaptation through the germline epigenome? Review

25 Kato Y, Kaneda M, Hata K et al. Role of the Dnmt3 family 39 Smallwood SA, Tomizawa S, Krueger F et al. Dynamic in de novo methylation of imprinted and repetitive sequences CpG island methylation landscape in oocytes and during male germ cell development in the mouse. Hum. Mol. preimplantation embryos. Nat. Genet. 43(8), 811–814 Genet. 16(19), 2272–2280 (2007). (2011). 26 La Salle S, Mertineit C, Taketo T, Moens PB, Bestor TH, •• High-throughput bisulfite analysis of mouse oocytes, Trasler JM. Windows for sex-specific methylation marked by sperm and preimplantation embryos. DNA methyltransferase expression profiles in mouse germ 40 Chatterjee A, Rodger EJ, Stockwell PA, Weeks RJ, Morison cells. Dev. Biol. 268(2), 403–415 (2004). IM. Technical considerations for reduced representation 27 Sakai Y, Suetake I, Shinozaki F, Yamashina S, Tajima S. Co- bisulfite sequencing with multiplexed libraries.J. Biomed. expression of de novo DNA methyltransferases Dnmt3a2 and Biotechnol. 2012, 741542 (2012). Dnmt3L in gonocytes of mouse embryos. Gene Expr. Patterns 41 Payer B, Saitou M, Barton SC et al. Stella is a maternal 5(2), 231–237 (2004). effect gene required for normal early development in mice. 28 Shovlin TC, Bourc’his D, La Salle S et al. Sex-specific Curr. Biol. 13(23), 2110–2117 (2003). promoters regulate Dnmt3L expression in mouse germ cells. 42 Li X, Ito M, Zhou F et al. A maternal-zygotic effect gene, Hum. Reprod. 22(2), 457–467 (2007). Zfp57, maintains both maternal and paternal imprints. Dev. 29 Webster KE, O’bryan MK, Fletcher S et al. Meiotic Cell 15(4), 547–557 (2008). and epigenetic defects in Dnmt3L-knockout mouse 43 Messerschmidt DM, De Vries W, Ito M, Solter D, Ferguson- spermatogenesis. Proc. Natl Acad. Sci. USA 102(11), Smith A, Knowles BB. Trim28 is required for epigenetic 4068–4073 (2005). stability during mouse oocyte to embryo transition. Science 30 Western PS, Van Den Bergen JA, Miles DC, Sinclair AH. 335(6075), 1499–1502 (2012). Male germ cell differentiation involves complex repression of 44 Monk M, Boubelik M, Lehnert S. Temporal and the regulatory network controlling pluripotency. FASEB J. regional changes in DNA methylation in the embryonic, 24(8), 3026–3035 (2010). extraembryonic and germ cell lineages during mouse 31 Bourc’his D, Xu GL, Lin CS, Bollman B, Bestor TH. embryo development. Development 99(3), 371–382 (1987). Dnmt3L and the establishment of maternal genomic 45 Howlett SK, Reik W. Methylation levels of maternal and imprints. Science 294(5551), 2536–2539 (2001). paternal genomes during preimplantation development. 32 Hata K, Okano M, Lei H, Li E. Dnmt3L cooperates with Development 113(1), 119–127 (1991). the Dnmt3 family of de novo DNA methyltransferases to 46 Van Der Heijden GW, Dieker JW, Derijck AA et al. establish maternal imprints in mice. Development 129(8), Asymmetry in histone H3 variants and lysine methylation 1983–1993 (2002). between paternal and maternal chromatin of the early 33 Lucifero D, La Salle S, Bourc’his D, Martel J, Bestor TH, mouse zygote. Mech. Dev. 122(9), 1008–1022 (2005). Trasler JM. Coordinate regulation of DNA methyltransferase 47 Mayer W, Niveleau A, Walter J, Fundele R, Haaf T. expression during oogenesis. BMC Dev. Biol. 7, 36 (2007). Demethylation of the zygotic paternal genome. Nature 34 Lesch BJ, Dokshin GA, Young RA, Mccarrey JR, Page DC. 403(6769), 501–502 (2000). A set of genes critical to development is epigenetically poised 48 Oswald J, Engemann S, Lane N et al. Active demethylation in mouse germ cells from fetal stages through completion of of the paternal genome in the mouse zygote. Curr. Biol. meiosis. Proc. Natl Acad. Sci. USA 110(40), 16061–16066 10(8), 475–478 (2000). (2013). 49 Santos F, Hendrich B, Reik W, Dean W. Dynamic •• Performed ChIP-seq and RNA-seq analysis on male and reprogramming of DNA methylation in the early mouse female germ cells during embryogenesis and identified a set embryo. Dev. Biol. 241(1), 172–182 (2002). of developmentally critical epigenetically poised genes. 50 Wossidlo M, Nakamura T, Lepikhov K et al. 35 Kota SK, Feil R. Epigenetic transitions in germ cell 5-hydroxymethylcytosine in the mammalian zygote is development and meiosis. Dev. Cell 19(5), 675–686 (2010). linked with epigenetic reprogramming. Nat. Commun. 2, 36 Hammoud SS, Nix DA, Zhang H, Purwar J, Carrell DT, 241 (2011). Cairns BR. Distinctive chromatin in human sperm packages 51 Li Y, O’Neill C. Persistence of cytosine methylation of genes for embryo development. Nature 460(7254), 473–478 DNA following fertilisation in the mouse. PLoS ONE 7(1), (2009). e30687 (2012). 37 Brykczynska U, Hisano M, Erkek S et al. Repressive and 52 Stancheva I, Meehan RR. Transient depletion of xDnmt1 active histone methylation mark distinct promoters in human leads to premature gene activation in Xenopus embryos. and mouse spermatozoa. Nat. Struct. Mol. Biol. 17(6), Genes Dev. 14(3), 313–327 (2000). 679–687 (2010). 53 Beaujean N, Hartshorne G, Cavilla J et al. Non- 38 Erkek S, Hisano M, Liang CY et al. Molecular determinants conservation of mammalian preimplantation methylation of nucleosome retention at CpG-rich sequences in mouse dynamics. Curr. Biol. 14(7), R266–R267 (2004). spermatozoa. Nat. Struct. Mol. Biol. 20(7), 868–875 (2013). 54 Macleod D, Clark VH, Bird A. Absence of genome-wide •• Analyzed the location of nucleosome retention in mouse changes in DNA methylation during development of the spermatozoa, supporting the hypothesis for nucleosomal zebrafish. Nat. Genet. 23(2), 139–140 (1999). epigenetic inheritance.

future science group www.futuremedicine.com 843 Review Prokopuk, Western & Stringer

55 Smith ZD, Chan MM, Mikkelsen TS et al. A unique 70 McGrath J, Solter D. Completion of mouse embryogenesis regulatory phase of DNA methylation in the early requires both the maternal and paternal genomes. Cell mammalian embryo. Nature 484(7394), 339–344 (2012). 37(1), 179–183 (1984). •• High-throughput DNA methylation analysis of mouse 71 Mann JR, Lovell-Badge RH. Inviability of parthenogenones gametes and from zygote to postimplantation. is determined by pronuclei, not egg cytoplasm. Nature 310(5972), 66–67 (1984). 56 Smith ZD, Chan MM, Humm KC et al. DNA methylation dynamics of the human preimplantation embryo. Nature 72 Searle AG, Beechey CV. Noncomplementation phenomena 511(7511), 611–615 (2014). and their bearing on nondisjunctional effects. Basic Life Sci. 36, 363–376 (1985). 57 Inoue A, Zhang Y. Replication-dependent loss of 5-hydroxymethylcytosine in mouse preimplantation 73 Li E, Beard C, Jaenisch R. Role for DNA methylation in embryos. Science 334(6053), 194 (2011). genomic imprinting. Nature 366(6453), 362–365 (1993). 58 Tahiliani M, Koh KP, Shen Y et al. Conversion of 74 Saitou M, Payer B, Lange UC, Erhardt S, Barton SC, 5-methylcytosine to 5-hydroxymethylcytosine in mammalian Surani MA. Specification of germ cell fate in mice.Philos. DNA by MLL partner TET1. Science 324(5929), 930–935 Trans. R. Soc. Lond. B. Biol. Sci. 358(1436), 1363–1370 (2009). (2003). 59 Borgel J, Guibert S, Li Y et al. Targets and dynamics 75 Davis TL, Trasler JM, Moss SB, Yang GJ, Bartolomei of promoter DNA methylation during early mouse MS. Acquisition of the H19 methylation imprint occurs development. Nat. Genet. 42(12), 1093–1100 (2010). differentially on the parental alleles during spermatogenesis. Genomics 58(1), 18–28 (1999). 60 Nakamura T, Liu YJ, Nakashima H et al. PGC7 binds histone H3K9me2 to protect against conversion of 5mC to 76 Davis TL, Yang GJ, McCarrey JR, Bartolomei MS. The 5hmC in early embryos. Nature 486(7403), 415–419 (2012). H19 methylation imprint is erased and re-established differentially on the parental alleles during male germ cell • Demonstrated that STELLA binds H3K9me2 to protect development. Hum. Mol. Genet. 9(19), 2885–2894 (2000). CpG sites against TET demethylation. 77 Li JY, Lees-Murdock DJ, Xu GL, Walsh CP. Timing of 61 Messerschmidt DM. Should I stay or should I go: protection establishment of paternal methylation imprints in the and maintenance of DNA methylation at imprinted genes. mouse. Genomics 84(6), 952–960 (2004). Epigenetics 7(9), 969–975 (2012). 78 Lucifero D, Mann MR, Bartolomei MS, Trasler JM. 62 Nakamura T, Arai Y, Umehara H et al. PGC7/Stella protects Gene-specific timing and epigenetic memory in oocyte against DNA demethylation in early embryogenesis. Nat. imprinting. Hum. Mol. Genet. 13(8), 839–849 (2004). Cell Biol. 9(1), 64–71 (2007). 79 Henckel A, Chebli K, Kota SK, Arnaud P, Feil R. 63 Ratnam S, Mertineit C, Ding F et al. Dynamics of Dnmt1 Transcription and histone methylation changes correlate methyltransferase expression and intracellular localization with imprint acquisition in male germ cells. EMBO J. 31(3), during oogenesis and preimplantation development. Dev. 606 – 615 (2012). Biol. 245(2), 304–314 (2002). 80 Chotalia M, Smallwood SA, Ruf N et al. Transcription is 64 Quenneville S, Verde G, Corsinotti A et al. In embryonic required for establishment of germline methylation marks at stem cells, ZFP57/KAP1 recognize a methylated imprinted genes. Genes Dev. 23(1), 105–117 (2009). hexanucleotide to affect chromatin and DNA methylation of imprinting control regions. Mol. Cell 44(3), 361–372 (2011). 81 Smith EY, Futtner CR, Chamberlain SJ, Johnstone KA, Resnick JL. Transcription is required to establish maternal 65 Zuo X, Sheng J, Lau HT et al. Zinc finger protein ZFP57 imprinting at the Prader-Willi syndrome and Angelman requires its co-factor to recruit DNA methyltransferases and syndrome locus. PLoS Genet. 7(12), e1002422 (2011). maintains DNA methylation imprint in embryonic stem cells via its transcriptional repression domain. J. Biol. Chem. 82 Lander ES, Linton LM, Birren B et al. Initial sequencing 287(3), 2107–2118 (2012). and analysis of the human genome. Nature 409(6822), 860–921 (2001). 66 Kurihara Y, Kawamura Y, Uchijima Y et al. Maintenance of genomic methylation patterns during preimplantation 83 Slotkin RK, Martienssen R. Transposable elements and the development requires the somatic form of DNA epigenetic regulation of the genome. Nat. Rev. Genet. 8(4), methyltransferase 1. Dev. Biol. 313(1), 335–346 (2008). 272–285 (2007). 67 Cirio MC, Ratnam S, Ding F, Reinhart B, Navara C, 84 Lane N, Dean W, Erhardt S et al. Resistance of IAPs to Chaillet JR. Preimplantation expression of the somatic form methylation reprogramming may provide a mechanism for of Dnmt1 suggests a role in the inheritance of genomic epigenetic inheritance in the mouse. Genesis 35(2), 88–93 imprints. BMC Dev. Biol. 8, 9 (2008). (2003). 68 Mackay DJ, Callaway JL, Marks SM et al. Hypomethylation 85 Hackett JA, Sengupta R, Zylicz JJ et al. Germline DNA of multiple imprinted loci in individuals with transient demethylation dynamics and imprint erasure through neonatal diabetes is associated with mutations in ZFP57. 5-hydroxymethylcytosine. Science 339(6118), 448–452 Nat. Genet. 40(8), 949–951 (2008). (2013). 69 Surani MA, Barton SC, Norris ML. Development of 86 Guibert S, Forne T, Weber M. Global profiling of DNA reconstituted mouse eggs suggests imprinting of the genome methylation erasure in mouse primordial germ cells. during gametogenesis. Nature 308(5959), 548–550 (1984). Genome Res. 22(4), 633–641 (2012).

844 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: adaptation through the germline epigenome? Review

87 Peaston AE, Evsikov AV, Graber JH et al. Retrotransposons 106 Francis D, Diorio J, Liu D, Meaney MJ. Nongenomic regulate host genes in mouse oocytes and preimplantation transmission across generations of maternal behavior and embryos. Dev. Cell 7(4), 597–606 (2004). stress responses in the rat. Science 286(5442), 1155–1158 88 Rubin CM, Vandevoort CA, Teplitz RL, Schmid CW. Alu (1999). repeated DNAs are differentially methylated in primate germ 107 Francis DD, Champagne FA, Liu D, Meaney MJ. Maternal cells. Nucleic Acids Res. 22(23), 5121–5127 (1994). care, gene expression, and the development of individual 89 Sapienza C. Sex-linked dosage-sensitive modifiers as differences in stress reactivity. Ann. NY Acad. Sci. 896, imprinting genes. Development 108(Suppl.), 107–113 (1990). 66–84 (1999). 90 Barlow DP. Methylation and imprinting: from host defense 108 Wolff GL, Kodell RL, Moore SR, Cooney CA. Maternal to gene regulation? Science 260(5106), 309–310 (1993). epigenetics and methyl supplements affect agouti gene expression in Avy/a mice. FASEB J. 12(11), 949–957 (1998). 91 McDonald JF. Transposable elements: possible catalysts of organismic evolution. Trends Ecol. Evol. 10(3), 123–126 109 Waterland RA, Jirtle RL. Transposable elements: targets for (1995). early nutritional effects on epigenetic gene regulation. Mol. Cell. Biol. 23(15), 5293–5300 (2003). 92 Surani MA. Imprinting and the initiation of gene silencing in the germ line. Cell 93(3), 309–312 (1998). 110 Cropley JE, Suter CM, Beckman KB, Martin DI. Germ- line epigenetic modification of the murine A vy allele by 93 Matzke MA, Mette MF, Matzke AJM. Transgene silencing nutritional supplementation. Proc. Natl Acad. Sci. USA by the host genome defense: implications for the evolution 103(46), 17308–17312 (2006). of epigenetic control mechanisms in plants and vertebrates. Plant Mol. Biol. 43(2), 401–415 (2000). 111 Cropley JE, Dang TH, Martin DI, Suter CM. The penetrance of an epigenetic trait in mice is progressively yet 94 Mcdonald JF, Matzke MA, Matzke AJ. Host defenses reversibly increased by selection and environment. Proc. Biol. to transposable elements and the evolution of genomic Sci. 279(1737), 2347–2353 (2012). imprinting. Cytogenet. Genome Res. 110(1–4), 242–249 (2005). 112 Waterland RA, Travisano M, Tahiliani KG. Diet-induced hypermethylation at agouti viable yellow is not inherited 95 Constancia M, Pickard B, Kelsey G, Reik W. Imprinting transgenerationally through the female. FASEB J. 21(12), mechanisms. Genome Res. 8(9), 881–900 (1998). 3380–3385 (2007). 96 Reik W, Walter J. Genomic imprinting: parental influence on 113 Morgan HD, Sutherland HG, Martin DI, Whitelaw E. the genome. Nat. Rev. Genet. 2(1), 21–32 (2001). Epigenetic inheritance at the agouti locus in the mouse. Nat. 97 Daxinger L, Whitelaw E. Understanding transgenerational Genet. 23(3), 314–318 (1999). epigenetic inheritance via the gametes in mammals. Nat. Rev. 114 Zamudio N, Bourc’his D. Transposable elements in the Genet. 13(3), 153–162 (2012). mammalian germline: a comfortable niche or a deadly trap? 98 Radford EJ, Isganaitis E, Jimenez-Chillaron J et al. An Heredity 105(1), 92–104 (2010). unbiased assessment of the role of imprinted genes in an 115 Blewitt ME, Vickaryous NK, Paldi A, Koseki H, Whitelaw intergenerational model of developmental programming. E. Dynamic reprogramming of DNA methylation at an PLoS Genet. 8(4), 41–53 (2012). epigenetically sensitive allele in mice. PLoS Genet. 2(4), e49 99 Ashe A, Sapetschnig A, Weick EM et al. piRNAs can trigger (2006). a multigenerational epigenetic memory in the germline of C. 116 Padmanabhan N, Jia D, Geary-Joo C et al. Mutation elegans. Cell 150(1), 88–99 (2012). in folate metabolism causes epigenetic instability and 100 Shirayama M, Seth M, Lee HC et al. piRNAs initiate transgenerational effects on development. Cell 155(1), 81–93 an epigenetic memory of nonself RNA in the C. elegans (2013). germline. Cell 150(1), 65–77 (2012). •• Showed transgenerational epigenetic effects using embryo 101 Bagijn MP, Goldstein LD, Sapetschnig A et al. Function, transfer experiments. targets, and evolution of Caenorhabditis elegans piRNAs. 117 Gapp K, Jawaid A, Sarkies P et al. Implication of sperm Science 337(6094), 574–578 (2012). RNAs in transgenerational inheritance of the effects of early 102 Lim SL, Tsend-Ayush E, Kortschak RD et al. Conservation trauma in mice. Nat. Neurosci. 17(5), 667–669 (2014). and expression of PIWI-interacting RNA pathway genes 118 Schübeler D. Function and information content of DNA in male and female adult gonad of amniotes. Biol. Reprod. methylation. Nature 517(7534), 321–326 (2015). 89(6), 136 (2013). 119 Dunn GA, Bale TL. Maternal high-fat diet promotes body 103 Malone CD, Hannon GJ. Small RNAs as guardians of the length increases and insulin insensitivity in second-generation genome. Cell 136(4), 656–668 (2009). mice. Endocrinology 150(11), 4999–5009 (2009). 104 Lee HC, Gu W, Shirayama M, Youngman E, Conte D Jr., 120 Carone BR, Fauquier L, Habib N et al. Paternally induced Mello CC. C. elegans piRNAs mediate the genome-wide transgenerational environmental reprogramming of metabolic surveillance of germline transcripts. Cell 150(1), 78–87 gene expression in mammals. Cell 143(7), 1084–1096 (2010). (2012). 121 Ng SF, Lin RC, Laybutt DR, Barres R, Owens JA, Morris 105 Grossniklaus U, Kelly B, Ferguson-Smith AC, Pembrey M, MJ. Chronic high-fat diet in fathers programs beta-cell Lindquist S. Transgenerational epigenetic inheritance: how dysfunction in female rat offspring. Nature 467(7318), important is it? Nat. Rev. Genet. 14(3), 228–235 (2013). 963–966 (2010).

future science group www.futuremedicine.com 845 Review Prokopuk, Western & Stringer

122 Pembrey ME. Male-line transgenerational responses in 136 Mayr E. The Growth of Biological Thought: Diversity, humans. Hum. Fertil. (Camb) 13(4), 268–271 (2010). Evolution, and Inheritance. Harvard University Press, USA 123 Jimenez-Chillaron JC, Isganaitis E, Charalambous M et al. (1982). Intergenerational transmission of glucose intolerance and 137 Wolf JB, Oakey RJ, Feil R. Imprinted gene expression obesity by in utero undernutrition in mice. Diabetes 58(2), in hybrids: perturbed mechanisms and evolutionary 460–468 (2009). implications. Heredity 113(2), 167–175 (2014). 124 Wei Y, Yang C-R, Wei Y-P et al. Paternally induced 138 Jablonka E, Lamb MJ. Epigenetic Inheritance and Evolution: transgenerational inheritance of susceptibility to diabetes The Lamarckian Dimension. Oxford University Press, UK, in mammals. Proc. Natl Acad. Sci. USA 111(5), 1873–1878 p360 (1995). (2014). 139 Jablonka E, Lamb MJ. The evolution of information in the 125 Guerrero-Bosagna C, Settles M, Lucker B, Skinner MK. major transitions. J. Theor. Biol. 239(2), 236–246 (2006). Epigenetic transgenerational actions of vinclozolin on 140 Jablonka E, Lamb MJ. Soft inheritance: challenging the promoter regions of the sperm epigenome. PLoS ONE 5(9), modern synthesis. Genet. Mol. Biol. 31(2), 389–395 (2008). e13100 (2010). • Summarizes the assumptions of the Modern Synthesis 126 Radford EJ, Ito M, Shi H et al. In utero effects. In utero model for evolution and discusses the need for expansion to undernourishment perturbs the adult sperm methylome and incorporate soft inheritance. intergenerational metabolism. Science 345(6198), 1255903 (2014). 141 Bossdorf O, Richards CL, Pigliucci M. Epigenetics for ecologists. Ecol. Lett. 11(2), 106–115 (2008). 127 Dunn GA, Bale TL. Maternal high-fat diet effects on third-generation female body size via the paternal lineage. 142 Pigliucci M, Müller GB. Evolution, the Extended Synthesis. Endocrinology 152(6), 2228–2236 (2011). MIT Press, Cambridge, USA (2010). 128 Lambrot R, Xu C, Saint-Phar S et al. Low paternal dietary 143 Suter CM, Boffelli D, Martin DI. A role for epigenetic folate alters the mouse sperm epigenome and is associated inheritance in modern evolutionary theory? A comment in with negative pregnancy outcomes. Nat. Commun. 4, 2889 response to Dickins and Rahman. Proc. Biol. Sci. 280(1771), (2013). 20130903; 20131820 (2013). 129 Blount BC, Mack MM, Wehr CM et al. Folate deficiency 144 Li CC, Young PE, Maloney CA et al. Maternal obesity and causes uracil misincorporation into human DNA and diabetes induces latent metabolic defects and widespread chromosome breakage: implications for cancer and neuronal epigenetic changes in isogenic mice. Epigenetics 8(6), damage. Proc. Natl Acad. Sci. USA 94(7), 3290–3295 (1997). 602–611 (2013). 130 Chong S, Vickaryous N, Ashe A et al. Modifiers of epigenetic 145 Duhl DM, Vrieling H, Miller KA, Wolff GL, Barsh GS. reprogramming show paternal effects in the mouse. Nat. Neomorphic agouti mutations in obese yellow mice. Nat. Genet. 39(5), 614–622 (2007). Genet. 8(1), 59–65 (1994). 131 Gu SG, Pak J, Guang S, Maniar JM, Kennedy S, Fire A. 146 Bateson P, Gluckman P. Plasticity and robustness in Amplification of siRNA in Caenorhabditis elegans generates development and evolution. Int. J. Epidemiol. 41(1), 219–223 a transgenerational sequence-targeted histone H3 lysine 9 (2012). methylation footprint. Nat. Genet. 44(2), 157–164 (2012). 147 Reik W, Constancia M, Fowden A et al. Regulation of 132 Buckley BA, Burkhart KB, Gu SG et al. A nuclear Argonaute supply and demand for maternal nutrients in mammals by promotes multigenerational epigenetic inheritance and imprinted genes. J. Physiol. 547(Pt 1), 35–44 (2003). germline immortality. Nature 489 (7416) 447 –451 (2012). 148 Haig D, Westoby M. Parent-specific gene expression and the 133 Youngson NA, Whitelaw E. Transgenerational epigenetic triploid endosperm. Am. Nat. 134(1), 147–155 (1989). effects. Annu. Rev. Genomics Hum. Genet. 9, 233–257 149 Haig D. Genomic imprinting and kinship: how good is the (2008). evidence? Annu. Rev. Genet. 38(1), 553–585 (2004). 134 Daxinger L, Whitelaw E. Transgenerational epigenetic 150 Wolf JB, Hager R. A maternal–offspring coadaptation theory inheritance: more questions than answers. Genome Res. for the evolution of genomic imprinting. PLoS Biol. 4(12), 20(12), 1623–1628 (2010). e380 (2006). 135 Huxley J. Evolution. The Modern Synthesis. Allen & Unwin, 151 Keverne EB, Curley JP. Epigenetics, brain evolution and London (1942). behaviour. Front. Neuroendocrinol. 29(3), 398–412 (2008).

846 Epigenomics (2015) 7(5) future science group Perspective

For reprint orders, please contact: [email protected]

7 Perspective 2015/07/28 Transgenerational epigenetic inheritance: an open discussion

Epigenomics Much controversy surrounds the idea of transgenerational epigenetics. Recent papers Corina Nagy1 argue that epigenetic marks acquired through experience are passed to offspring, & Gustavo Turecki*,1,2 but as in much of the field of epigenetics, there is lack of precision in the definitions 1McGill Group for Suicide Studies, Douglas Hospital University Institute, and perhaps too much eagerness to translate animal research to humans. Here, we 6875 Lasalle boul, Montreal, QC, Canada review operational definitions of transgenerational inheritance and the processes 2Department of Psychiatry, McGill of epigenetic programing during early development. Subsequently, based on this University, Montreal, QC, Canada background, we critically examine some recent findings of studies investigating *Author for correspondence: transgenerational inheritance. Finally, we discuss possible mechanisms that may explain [email protected] transgenerational inheritance, including transmission of an epigenetic blueprint, which may predispose offspring to specific epigenetic patterning. Taken together, we conclude that presently, the evidence suggesting that acquired epigenetic marks are passed to the subsequent generation remains limited.

Keywords: animal models • early development • histone modifications • human • intergenerational • methylation • transgenerational

Conrad Waddington first used the term epi- germline transmission; epigenetic marks genetics over half a century ago and defined might also be propagated forward through it as ‘the branch of biology which studies the germline to offspring. This notion was the causal interactions between genes and fuelled by both research studies in plants their products which bring the phenotype and longitudinal epidemiological studies into being.’ A more modern definition of in humans, which have suggested that the epigenetics is: heritable chemical modifica- effects of environmental influences were tions to DNA capable of influencing tran- observed in subsequent generations. Stud- 5 scriptional activity with no direct alteration ies in the plant Arabidopsis thaliana showed to the DNA sequence itself. It should be that alterations in DNA methylation can be noted that heritable in this definition refers transmitted through several generations [3]. to ‘mitotic stability’ [1] or, in other words, to Famously in humans, epidemiological stud- 2015 the fact that epigenetic information is passed ies examining the Dutch famine found that to, and maintained, in the daughter cell upon those individuals who were exposed to the mitotic division. Certainly, this type of epi- famine while in utero, had increased sus- genetic heritability is of relevance to cancer ceptibility to diseases in adulthood such as, research [2]. However, the concept of epigen- diabetes, cardiovascular disease and obesity, etic heritability has broadened over the years. to name a few [4]. These findings begged the Studies have progressively suggested that not question of lastingness, expressly, whether only might epigenetic marks be inherited these predisposing factors, which seemingly through mitotic processes, but also through stemmed from in utero exposure to famine, meiosis, a concept akin to genetic inheri- could be transmitted to the next generation. tance. That is, similar to the way we inher- Though there is no direct molecular link ited genetic traits from our parents through to these major health concerns, second- part of

generation offspring was found to demonstrate the could be carried forward. In fact, some loci, includ- same lower birth weights, as seen in first-generation ing repeated elements such as retrotransposons, do affected individuals [4]. not undergo reprogramming during this phase [6,7]. In contrast to the mitotic stability seen in somatic Interestingly, Molaro and colleagues [8] found a strik- cells, transgenerational epigenetics requires germline ing difference in methylation of retrotransposon sub- transmission of acquired epigenetic information across families when examining the sperm of humans and generations. In cases of male exposure, the second chimps. Furthermore, when comparing these sperm generation may be sufficient to assess transgenera- methylomes to embryonic stem cells, which they con- tional effects. This is because exposure will affect the sider a mature germ cell having the product of the F0 male as well as his germline (F1), therefore only two reprogramming events in mammal, they found starting with the F2 generation can the effects of trans- distinct characteristic, suggesting that sperm and generational epigenetic inheritance be considered. In somatic cells have different features, which deter- females, however, studies require a third generation to mine DNA methylation patterns in each cell type. test whether any of these DNA modifications are truly Furthermore, approximately 100 nonimprinted, non- transmitted from one generation to the next. This is repetitive genes have been identified as maintaining because one needs to rule out in utero exposure of the promoter methylation levels throughout a range of second generation to an environmental effect associ- developmental stages, from mature gametes to blas- ated with epigenetic changes. As an illustration, expos- tocysts [9]. Nonmammalian organisms on the other ing a gestating female, her fetus and thus the fetal hand, appear to not undergo whole genome demeth- germline to a given environment, produces an F1 gen- ylation upon fertilization [10] . In a comprehensive eration with an associated epigenetic change reflecting study using zebrafish, Jiang and colleagues [10] were direct environmental effects through in utero exposure able to demonstrate that no global erasure occurs at (Figure 1). Prior to the F2 generation in males and the fertilization and by the midblastula stage the embryo’s F3 generation in female, the inheritance detected in methylome is almost identical to the of sperm, while the offspring can only be considered intergenerational, the oocyte’s methylome contribution lessens by the in other words, parental effect due to direct expo- 16-cell state. This suggests that the early reprogram- sure. This is important information to consider when ming mechanisms are different between mammalian reviewing some of the recent publications in this field. and non-mammalian organisms. As mentioned above, both the mammalian mater- Early developmental mechanisms: nal and paternal genomes undergo demethylation and epigenetic reset they also experience active chromatin remodeling. This Early epigenetic reprogramming has been extensively is particularly evident for the paternal genome. The studied in rodents. Though the same processes of early process of spermatogenesis in humans requires that developmental epigenetic erasure occur in humans, anywhere between 85 and 96% of histones be replaced the definitive characteristics and timeline of these by protamines, resulting in a tenfold compaction of events have still to be established [5]. Work in mice the DNA [12,13]. It is believed that this process protects has demonstrated that maternal and paternal cells the paternal genome from physical and chemical dam- not only undergo a global erasure of DNA methyla- age. Protamines, like histones, carry post-translational tion but the chromatin also undergoes active remod- modifications such as phosphorylation. The role of eling. Global erasure occurs at two points. First, as these modifications remains unknown, although they the primordial germ cells (PGC) or gamete precursors are believed to act similarly to histone modifications develop and become part of the embryo, the epigenetic altering the availability of the DNA sequencing to landscape is cleared to allow for cellular totipotency the cellular transcriptional machinery [12] . Thus, the and the development of future generations through majority of paternally derived histone marks are lost PGCs. This stage is followed by gametogenesis when during this process while the histones not replaced by the genomes undergo de novo methylation, a process protamines are thought to belong to genes expressed that occurs later in the maternal genome. The second early in development [13] . Interestingly, the repressive wave of demethylation occurs after fertilization as mark, histone H3 lysine 27 trimethylation, is retained the gametes fuse to form the zygote (Figure 2). Here, on some genes in the human and mouse spermatozoa. however, imprinted genes escape the second round of This is arguably another possible avenue for inheri- demethylation (Figure 2, dotted lines), to carry for- tance; conversely, it could be a methodological arti- ward parent-specific monoallelic expression in somatic fact, and simply be the result of a histone mark being cells. If epigenetic marks were to be maintained across re-established so quickly that its period of absence is generations, it is presumably at this point that they not detected. Gradually, studies are demonstrating

782 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: an open discussion Perspective

Female Male

F1 F0

F2 F1

F3 and subsequent generations F2 and subsequent generations

Figure 1. The difference between intergenerational and transgenerational epigenetic inheritance in females and males. F0 represents maternal or paternal exposure to environmental factors, which directly impacts the fetus (F1) and its already formed germline (F2) in females or the germline in males (F1). This demonstrates the importance of studying the F2 generation in males and the F3 generation in females to avoid parental effects, in other words, intergenerational epigenetics. Studying the F2 (in males) and F3 (in females) and/or their subsequent generations could provide information on transgenerational epigenetics as they would be free of direct environmental exposure and parental effects. For color images please see online www.futuremedicine.com/doi/full/10.2217/epi.15.46 the impressive dynamics of chromatin, for example, takes the form of nucleosome, whereas following the the protein involved in heterochromatin formation histone to protamine exchange during spermiogen- HP1 binds for just a few minutes [14] and the wrap- esis, the paternal genome incorporates histones from ping and unwrapping of nucleosomes, a process which the maternal genome [18]. Importantly, epigenetic pro- erases histone modifications, is so rapid that the turn- cesses in early mammalian development result in non- over is faster than a complete cell cycle [15] . Thus, the canonical forms of histones in both the maternal and highly dynamic nature of histone marks makes them paternal genomes. less likely to act as a mechanism of transgenerational Finally, there is interesting epigenetic asymmetry epigenetics. Equally, canonical forms of histones in between parental genomes, which remains poorly the maternal genome, such as H2A, are replaced with understood. For example, de novo methylation in H2A.Z during early embryonic development [16] . It gametes, and demethylation after fertilization occurs is thought that H2A.Z is responsible for establishing more rapidly in paternal genome, denoted respectively heterchromatin in early development [17], furthermore, by the blue and pink line graph in Figure 2. In males, its absence in the early embryonic development results during protamine removal, the paternal genome is in death shortly after implantation [16] . The maternal repackaged into new histones, as well as into mater- genome has also been shown to transfer components nally derived histones, whereas maternal chromatin of a repressive complex, PRC1, to the paternal genome maintains histone methylation throughout early cleav- after fertilization [18], resulting in direct silencing in age [19] . The timing of methylation could be directly the zygote. At gamete fusion, the maternal genome related to the presence or absence of histone marks

future science group www.futuremedicine.com 783 Perspective Nagy & Turecki

like H3K9me2, which are intimately related to DNA Escaping intergenerational reprogramming methylation, the absence of this mark may promote In light of early developmental reprogramming, the DNA demethylation [5]. For instance, PGCs show sex question researchers must address is whether certain differences in the imprinted Igf2r gene, in both the epigenomic transcriptional drivers, in the form of timing of demethylation in males and de novo methyl- DNA methylation or histone modifications, escape ation in females [6]. This asymmetry is likely an expla- this prototypical intergenerational reprogramming. nation of why some traits are only propagated through The classic experiment in the agouti viable yellow one parental line, examples of which are discussed in mouse (Avy) was the first study to report intergenera- later sections. tional inheritance in mammals. If a repeat element

De novo E10.5–13.5 methylation DNA demethylation DNA methylation

Gametes Zygote Blastocyst Post-implantation

PGC Imprints (gonads) maintained Epiblast Epiblast

PGC DNA demethylation precursor

Methylation cycle Male Imprinting ylation Meth

Female

PGC in Gamatogenesis Fertilization Cell division embryo E10.5–13.5

F1 F2

Figure 2. Very early embryonic development corresponds to epigenetic programing. Primordial germ cells in the embryo undergo global DNA methylation erasure, or ‘reprogramming’ from their epiblast state (red arrow). This first wave of demethylation is also denoted in the methylation cycle graph depicting both the male and female genomes devoid of DNA methylation, including imprinted genes. Gametes are then de novo methylated at different rates, with maternal methylation marks being established later (graph pink line) than paternal marks. A second round of ‘reprogramming’ occurs upon fusion of the gametes (sperm and oocyte) producing totipotent or pluripotent cell states. At this point, demethylation occurs more rapidly in the paternal genome (graph blue line), moreover, imprinted genes escape erasure (graph dotted lines) maintaining their methylation marks. Genome- wide remethylation occurs in both parental genomes at implantation (green arrow). The timeline denoted in this schematic refer to event in the mouse life cycle. The timeline in humans is not yet full defined, though the events are considered to occur in a similar manner. PGC: Primordial germ cell. Adapted with permission from [11].

784 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: an open discussion Perspective

called an intracisternal A particle (IAP), of which there tance of an epigenetic state [28–32]. In other words, in are thousand of copies throughout the mouse genome, some cases, transmission of epigenetic marks can be inserts itself upstream of the agouti locus, phenotypic secondary to inherited DNA sequence mutations. variations can arise including altered coat color and obesity [20]. DNA methylation at the promoter region Epigenetic inheritance: acquiring of this IAP, which in turn regulates the agouti gene, knowledge is inversely related to its transcriptional activity [20]. A number of recent studies have implied that epi- As indicated above, repeat elements may escape repro- genetic alteration induced by environmental factors gramming, which appears to be the case for this epial- affect the behavior of subsequent generations. Vas- lele that is passed along the maternal but not paternal soler and colleagues [33] found decreased cocaine self- germline [20]. Interesting as this may be, this is not administration in the first generation offspring of the result of a past environmental factor being propa- cocaine exposed male rats. The behavioral outcome gated forward via epigenetic inheritance. Though the suggests an almost protective effect of fathers consum- insertion of the IAP at the agouti locus may not be ing cocaine, which is not consistent with the clinical entirely random [21], it does not result from the effects literature [34,35]. Molecularly, this is associated with of environmental factors. However, once the IAP is increased Bdnf levels as well as increased H3 acetyla- inserted and susceptible to methylation changes, envi- tion, an epigenetic mechanism potentially responsible ronment can play a role. This has been demonstrated for this change. The authors of this intriguing study by altering the food source of pregnant dams. Diets correctly suggested that more generations are required high in methyl groups such as folic acid or betaine can to definitively conclude that cocaine exposure results vy effectively ‘turn off’ the A locus by methylating the in transgenerational epigenetic alterations [36]. Cer- promoter region of the IAP [22,23]. Though it is tempt- tainly, cocaine itself can alter chromatin states [37], but ing to look at this as transgenerational inheritance, it it is possible that cocaine directly influenced the gam- is rather the result of direct environmental exposure on etes of the parental generation, and thus the F1 prog- the germline, and is thus the result of parental influ- eny. In a similar vein, a recent study in Caenorhabditis ence, more appropriately termed, intergenerational elegans showed paternal transmission to the embryo of inheritance. the repressive mark H3K27me3, which is under the Early life environment is also known to alter meth- control of the Polycomb repression complex 2 [38]. ylation states as classically demonstrated through rat Interestingly, histones containing this mark are not maternal behavior paradigms [24]. Caution must be replaced by protamines during spermatogenesis in taken when conducting transgenerational experi- mammals [13] . This very interesting finding demon- ments to control for both in utero and early life expe- strates mechanistic properties for the propagation of riences to avoid confounding early-life environmental epigenetic marks through generations, but as pointed effects with transgenerational inheritance. To avoid out previously, there is a need to demonstrate trans- these potential confounding effects, both in utero and generational inheritance beyond the first generation of in early life, studies in rodent models have focused progeny. on paternal lineage as males are not involved in the A study examining early stress in mice addressed in utero environment and are less likely to be involved the question of whether transgenerational inheritance in the early-life postnatal environment. An interesting could be explained by the effect of parental environ- example of this was found using outbred rats exposed ment on the gametes or through the transmission of to the fungicide, vinclozolin in utero. As a result, males acquired epigenetic marks to the offspring, by observ- exhibited diminished fertility over three to four gen- ing the transmission of depressive-like behaviors up to erations of offspring, which was found to be transmit- the third generation [39,40]. The behavioral traits cose- ted through the male germline [25]. Interestingly, these gregated with altered DNA methylation in the male effects were not observed with another strain of rats. It germ line. The same genes were tested for methylation is possible that this discrepancy results from method- levels in both the F2 sperm and the F2 female brains. ological differences between studies but it also raises The methylation patterns in these cell types were simi- the possibility that interaction with genetic variation lar, though certainly not identical. However, of note might be responsible for the observed effect [26]. If this was a discrepancy in behavioral phenotype, where the is the case, it represents an example of secondary epi- F1 and F3 males showed depressive-like behaviors but mutations, whereby DNA sequence variants have an not the F2 males. The depressive-like phenotype was effect on epigenetic marks (in cis or trans) [27]. Indeed only seen in the F2 females and the molecular changes a number of studies have demonstrated that genetic- reported in this study were correlational, as there was based alterations can be responsible for the inheri- no experiment in this study showing a causal link

future science group www.futuremedicine.com 785 Perspective Nagy & Turecki

between methylation changes and phenotype. Further- tions may act in a probabilistic manner. In other words, more, the exposure of females to stressed males, even these changes are more or less likely to occur in broad for a short period, could have had subtle effects on the genomic regions, resulting in altered behavior without female behaviors and hormonal expression, which in producing a consistent molecular phenotype. A recent turn may have affected maternal care. In vitro fertiliza- paper demonstrating that paternal sugar consumption tion (IVF) of the sperm directly into control females in fruit flies influences the metabolic properties of the has been used to avoid these potential confounding F1 generation, leading to an obesity phenotype, sup- effects from fathers. When IVF was used in social ports this concept. Here, the authors demonstrate that defeat paradigms, which is an animal model of depres- a network of genes is required for proper intergenera- sion, the depressive-like phenotype that seemed to be tional metabolic reprogramming involving a number propagated forward under normal mating conditions of changes to chromatin structure. No specific sites was largely absent [40]. Together, these results suggest were identified but rather a pattern or signature of gene that broad behavioral conditioning can be inherited. dysfunction was identified as conferring susceptibility Although there is a possibility that these may cosegre- to obesity [44]. gate with specific epigenetic profiles, additional work is necessary to definitively make such conclusions. Epigenetic inheritance: from rodents to Another study using the same early stress paradigm humans in mice showed depressive-like behaviors associated The vast majority of studies examining transgenera- with an upregulation of several miRNA in sperm tional inheritance have been conducted in animals which appeared to affect serum and hippocampal other than humans. Though mice and rats provide miRNAs levels in the subsequent generation [41] . convenient models for human disorders and behav- Because small RNAs are highly present in sperm, they ioral traits, there are different physiological and bio- have become candidate vectors for conveying trans- logical processes between humans and rodents. For generational inheritance [42]. However, in this study, instance, a recent paper exploring the methylation although the behavioral phenotype was detected in landscape of early human embryos showed features the F3 generation, the miRNA levels of five candi- distinctive to humans, specifically levels of methyla- date miRNAs in F2 sperm were unchanged and the tion and timing of genome-wide demethylation [45], alterations seen in these candidate miRNAs from F2 while another study suggests the symmetry of epigen- were no longer detectable in F3. The authors suggest etic reprogramming cycles may differ between spe- that the initial change in miRNA levels resulting from cies [6]. Moreover, many transcriptionally relevant early stress may have been transferred to other epigen- epigenetic marks show considerable sequence diver- etic marks, but give no evidence of this effect. Another gence between mice and humans [46]. Furthermore, study from the same group set out to demonstrate a information about epigenetic erasure during early causal mechanism for the inherited stress responses in development comes primarily from mouse mod- the offspring of stressed sires [43]. In this study, the els, and it is possible to be different in humans, but authors were able to pharmacologically induce the remains, as yet, untested. Transgenerational effects same behavioral traits seen in the stressed offspring, have been observed in humans through longitudinal with corresponding alteration to gene expression levels studies [47–49], whereas gene–environment interplay is resulting from histone methyl- and acetyl-transferase suspected as a mediating factor in health outcome. inhibitors, suggesting a causal role for mineral corti- Some researchers have attempted to approach epigen- coid receptors in the altered stress response of pater- etic inheritance using in a systems biology approach, nally stressed offspring. Here, however, the epigenetic whereby acquired behavioural and physiological changes seen in the male germline differed from what changes, are considered through the framework of a was found in the adult hippocampus of the offspring, “two hit” model. This approach recognizes both, the specifically methylation changes where seen in the biology established in a previous generation and the sperm whereas altered histone modifications were seen organism’s present environment, as involved in the in the brains. Collectively, it appears that whatever development of these changes [50]. In a study explor- environmentally induced epigenetic alteration is tak- ing vinclozolin-lineage rats presumed to be carrying place in one generation may be propagated for- ing epigenetic alterations from previous generations, ward, but as noted above, although possible, it is still the authors identified gene network and behavioural premature to conclusively suggest a clear association alterations associated with the lineage specific rats between one acquired epigenetic mark and transgener- when subjected to a stress paradigm [51] . Though this ational behavioral phenotypes. An enticing possibility study does not answer the question of whether or not is that epigenetic changes transmitted across genera- these genetic and behavioural changes would have

786 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: an open discussion Perspective occurred in the absence of the induced environmental One should be particularly cautious when interpret- stress (context) nor does it confirm the presence of ing the results of studies testing transgenerational epi- epigenetically altered states, it nonetheless represents genetic inheritance in animals to the lay media or gen- the attractive notion of dynamic interactions between eral public. Though likely not the intention, as pointed context dependence and “predisposing factors”, in out previously [52], animal studies have been vulgarized this case, environment and inherited epigenetic state. and commonly anthropomorphized into a public mes- Intuitively, most people can recognize that their envi- sage that overemphasizes the molecular impact of the ronment plays a role in health and behavior, but we environment, including parents and grandparents onto still lack conclusive evidence to indicate that such their children. transgenerational effects are explained by acquired The field of epigenetics is fascinating and holds epigenetic mechanisms inherited from one generation great potential in medicine, both to uncover dis- to the next. ease biomarkers and therapeutic interventions. The results suggesting that acquired epigenetic factors Conclusion may be transmitted through generations and explain The studies reviewed above suggest that there is some acquired phenotypic traits are promising and intrigu- evidence that certain epigenetic marks escape erasure ing, but nevertheless, studies published thus far have in early development. There is also evidence that cer- limitations and prevent us from making definitive tain acquired behavioral phenotypes are transmitted conclusions. through subsequent generations. Although promising, the evidence suggesting that acquired epigenetic marks Future perspective cosegregate with acquired behavioral phenotypes An important missing link in the study of transgen- remains inconclusive and open to a number of potential erational epigenetic inheritance is a mechanism by alternative interpretations, including methodological which gene-regulatory information is transferred explanations. from somatic cells to germs cells. Efforts to uncover

Executive Summary Early developmental mechanisms: epigenetic reset • Epigenetic reprogramming occurs at two points during early development, after fertilization and in primordial germ cells. • Some genomic features, such as retrotransposons, as well as certain histone marks can carry forward epigenetic marks in spite of these erasure mechanisms, supporting the possibility that epigenetic marks can be maintained across generations. Escaping intergenerational reprogramming • Mice and rat models have been used to show that certain epigenetic marks can escape intergenerational reprogramming, affecting the phenotype of future generations. • Effects resulting from parental influences result in intergeneration inheritance, which is not to be confused with transgenerational inheritance (effects that survive across generations in the absence of direct exposure). • Secondary epimutations represent another factor that can influence possible transgenerational inheritance. In these cases, epigenetic marks are secondary to inherited DNA sequence mutations. Epigenetic inheritance: acquiring knowledge • A number of studies including recent articles providing mechanistic insight show that epigenetic marks can be inherited intergenerationally. • Data for transgenerational inheritance on the other hand show certain inconsistencies, which may be suggestive of probabilistic epigenetic changes capable of influencing phenotypic outcome in subsequent generations. Epigenetic inheritance: from rodents to humans • It is difficult to provide concrete evidence that transgenerational epigenetics occurs in humans owing to the lack of studies in humans and the physiological and biological differences between humans and rodents. Conclusion • There is evidence that epigenetic inheritance occurs intergenerationally by escaping erasure in early development. • However, evidence is lacking for the co-segregation of acquired epigenetic marks with acquired behavioral traits transgenerationally. • The field is very promising but more studies are required in order to provide definitive conclusions on the topic of transgenerational inheritance, particularly in humans.

future science group www.futuremedicine.com 787 Perspective Nagy & Turecki

these mechanisms have been made in nonmamma- phenotypic traits and acquired epigenetic marks, and lian animals. For example, a number of studies in C. their co-segregation, through three generations at elegans have implicated small RNAs in the process least. In addition to correlational evidence, studies of inherited epigenetic marks [53–55]. It was recently should also investigate causative links between the demonstrated in that dsRNA act as mobile elements molecular changes investigated and the phenotypes. that mediating intertissue transfer of regulatory infor- Finally, although more difficult to conduct, studies mation by entering the cytosol via a dsRNA-selective are also necessary in humans. importer [56]. Equally, it has been speculated, notably by a paper discussed here [41] and others [57,58], that Financial & competing interests disclosure small noncoding RNAs, like miRNA, mediate soma The authors have no relevant affiliations or financial in- to germline transfer of regulatory information in volvement with any organization or entity with a financial mammals, however, experimental evidence in mam- interest in or financial conflict with the subject matter or mals is lacking. Tackling this question in mammals materials discussed in the manuscript. This includes employ- and particularly in humans is an important next step ment, consultancies, honoraria, stock ownership or options, in the research of transgenerational epigenetics. expert testimony, grants or patents received or pending, or Moving forward in the field of transgenerational royalties. epigenetics requires more precision in experimental No writing assistance was utilized in the production of design. First, studies should investigate both acquired this manuscript.

References 11 Fresard L, Morisson M, Brun JM et al. Epigenetics and Papers of special note have been highlighted as: phenotypic variability: some interesting insights from birds. • of interest; •• of considerable interest Genet. Sel. Evol. 45, 16 (2013). 1 Skinner MK. Environmental epigenetic transgenerational 12 Brunner AM, Nanni P, Mansuy IM. Epigenetic marking inheritance and somatic epigenetic mitotic stability. of sperm by post-translational modification of histones and Epigenetics 6(7), 838–842 (2011). protamines. Epigenetics Chromatin 7(1), 2 (2014). 2 Dawson MA, Kouzarides T. Cancer epigenetics: from 13 Brykczynska U, Hisano M, Erkek S et al. Repressive and active mechanism to therapy. Cell 150(1), 12–27 (2012). histone methylation mark distinct promoters in human and mouse spermatozoa. Nat. Struct. Mol. Biol. 17(6), 679–687 3 Johannes F, Porcher E, Teixeira FK et al. Assessing the (2010). impact of transgenerational epigenetic variation on complex traits. PLoS Genet. 5(6), e1000530 (2009). 14 Festenstein R, Pagakis SN, Hiragami K et al. Modulation of heterochromatin protein 1 dynamics in primary mammalian 4 Painter RC, Osmond C, Gluckman P, Hanson M, Phillips cells. Science 299(5607), 719–721 (2003). DI, Roseboom TJ. Transgenerational effects of prenatal exposure to the dutch famine on neonatal adiposity and 15 Deal RB, Henikoff S. Histone variants and modifications in health in later life. BJOG 115(10), 1243–1249 (2008). plant gene regulation. Curr. Opin. Plant Biol. 14(2), 116–122 (2011). 5 Leitch HG, Tang WW, Surani MA. Primordial germ-cell development and epigenetic reprogramming in mammals. 16 Faast R, Thonglairoam V, Schulz TC et al. Histone variant Curr. Top. Dev. Biol. 104, 149–187 (2013). h2a.Z is required for early mammalian development. Curr. Biol. 11(15), 1183–1187 (2001). 6 Hyldig SM, Croxall N, Contreras DA, Thomsen PD, Alberio R. Epigenetic reprogramming in the porcine germ line. BMC 17 Banaszynski LA, Allis CD, Lewis PW. Histone variants in Dev. Biol. 11, 11 (2011). metazoan development. Dev. Cell 19(5), 662–674 (2010). 7 Law JA, Jacobsen SE. Establishing, maintaining and 18 Puschendorf M, Terranova R, Boutsma E et al. Prc1 modifying DNA methylation patterns in plants and animals. and suv39h specify parental asymmetry at constitutive Nat. Rev. Genet. 11(3), 204–220 (2010). heterochromatin in early mouse embryos. Nat. Genet. 40(4), 411–420 (2008). 8 Molaro A, Hodges E, Fang F et al. Sperm methylation profiles reveal features of epigenetic inheritance and 19 Zhou LQ, Dean J. Reprogramming the genome to totipotency evolution in primates. Cell 146(6), 1029–1041 (2011). in mouse embryos. Trends Cell. Biol. 25(2), 82–91 (2014). 9 Borgel J, Guibert S, Li Y et al. Targets and dynamics 20 Morgan HD, Sutherland HG, Martin DI, Whitelaw E. of promoter DNA methylation during early mouse Epigenetic inheritance at the agouti locus in the mouse. Nat. development. Nat. Genet. 42(12), 1093–1100 (2010). Genet. 23(3), 314–318 (1999). 10 Jiang L, Zhang J, Wang JJ et al. Sperm, but not oocyte, 21 Baillie JK, Barnett MW, Upton KR et al. Somatic DNA methylome is inherited by zebrafish early embryos.Cell retrotransposition alters the genetic landscape of the human 153(4), 773–784 (2013). brain. Nature 479(7374), 534–537 (2011). • In this study, the authors demonstrate that the timing and 22 Waterland RA, Jirtle RL. Transposable elements: targets for inheritance of epigenetic marks in zebrafish are driven by early nutritional effects on epigenetic gene regulation. Mol. the male germline. Cell. Biol. 23(15), 5293–5300 (2003).

788 Epigenomics (2015) 7(5) future science group Transgenerational epigenetic inheritance: an open discussion Perspective

• Evidences that transposable elements can escape early 39 Franklin TB, Russig H, Weiss IC et al. Epigenetic developmental reprogramming. transmission of the impact of early stress across generations. 23 Wolff GL, Kodell RL, Moore SR, Cooney CA. Maternal Biol. Psychiatry 68(5), 408–415 (2010). epigenetics and methyl supplements affect agouti gene 40 Dietz DM, Laplant Q, Watts EL et al. Paternal transmission expression in avy/a mice. FASEB J. 12(11), 949–957 (1998). of stress-induced pathologies. Biol. Psychiatry 70(5), 408–414 (2011). 24 Weaver IC, Cervoni N, Champagne FA et al. Epigenetic programming by maternal behavior. Nat. Neurosci. 7(8), 41 Gapp K, Jawaid A, Sarkies P et al. Implication of sperm rnas 847–854 (2004). in transgenerational inheritance of the effects of early trauma in mice. Nat. Neurosci. 17(5), 667–669 (2014). • Demonstrates how environment can have a relatively long- lasting effect on behavior through epigenetic alterations. 42 Kawano M, Kawaji H, Grandjean V, Kiani J, Rassoulzadegan M. Novel small noncoding rnas in mouse spermatozoa, 25 Anway MD, Cupp AS, Uzumcu M, Skinner MK. zygotes and early embryos. PLoS ONE 7(9), e44542 (2012). Epigenetic transgenerational actions of endocrine disruptors and male fertility. Science 308(5727), 1466–1469 (2005). 43 Gapp K, Soldado-Magraner S, Alvarez-Sanchez M et al. Early life stress in fathers improves behavioural flexibility in 26 Schneider S, Kaufmann W, Buesen R, Van Ravenzwaay B. their offspring. Nat. Commun. 5, 5466 (2014). Vinclozolin – the lack of a transgenerational effect after oral maternal exposure during organogenesis. Reprod. Toxicol. •• Provides evidence in mammals that certain histone marks 25(3), 352–360 (2008). can be mantained through developmental reprogramming. 27 Oey H, Whitelaw E. On the meaning of the word 44 Ost A, Lempradl A, Casas E et al. Paternal diet defines ‘epimutation’. Trends Genet. 30(12), 519–520 (2014). offspring chromatin state and intergenerational obesity. Cell 159(6), 1352–1364 (2014). 28 Bell JT, Pai AA, Pickrell JK et al. DNA methylation patterns associate with genetic and gene expression variation 45 Guo H, Zhu P, Yan L et al. The DNA methylation landscape in hapmap cell lines. Genome Biol. 12(1), R10 (2011). of human early embryos. Nature 511(7511), 606–610 (2014). 29 Gertz J, Varley KE, Reddy TE et al. Analysis of DNA 46 Yue F, Cheng Y, Breschi A et al. A comparative encyclopedia methylation in a three-generation family reveals widespread of DNA elements in the mouse genome. Nature 515(7527), genetic influence on epigenetic regulation.PLoS Genet. 355–364 (2014). 7(8), e1002228 (2011). 47 Bygren LO, Kaati G, Edvinsson S. Longevity determined by 30 Gutierrez-Arcelus M, Ongen H, Lappalainen T et al. paternal ancestors’ nutrition during their slow growth period. Tissue-specific effects of genetic and epigenetic variation on Acta Biotheor. 49(1), 53–59 (2001). gene regulation and splicing. PLoS Genet. 11(1), e1004958 48 Kaati G, Bygren LO, Edvinsson S. Cardiovascular and (2015). diabetes mortality determined by nutrition during parents’ 31 Zhang D, Cheng L, Badner JA et al. Genetic control of and grandparents’ slow growth period. Eur. J. Hum. Genet. individual differences in gene-specific methylation in 10(11), 682–688 (2002). human brain. Am. J. Hum. Genet. 86(3), 411–419 (2010). 49 Pembrey ME, Bygren LO, Kaati G et al. Sex-specific, male- 32 Cini G, Carnevali I, Quaia M et al. Concomitant mutation line transgenerational responses in humans. Eur. J. Hum. and epimutation of the mlh1 gene in a lynch syndrome Genet. 14(2), 159–166 (2006). family. Carcinogenesis 36(4), 452–458 (2015). 50 Crews D, Gore AC. Transgenerational epigenetics – current 33 Vassoler FM, White SL, Schmidt HD, Sadri-Vakili G, controversiesand debates. In: Transgenerational Epigenetics. Pierce RC. Epigenetic inheritance of a cocaine-resistance Tollefsbol T (Ed.). Elsevier, MA, USA, 371–387 (2014). phenotype. Nat. Neurosci. 16(1), 42–47 (2013). 51 Crews D, Gillette R, Scarpino SV, Manikkam M, Savenkova 34 Johnson JL, Leff M. Children of substance abusers: MI, Skinner MK. Epigenetic transgenerational inheritance of overview of research findings.Pediatrics 103(5 Pt. 2), altered stress responses. Proc. Natl Acad. Sci. USA. 109(23), 1085–1099 (1999). 9143–9148 (2012). 35 Delaney-Black V, Chiodo LM, Hannigan JH et al. Prenatal 52 Juengst ET, Fishman JR, McGowan ML, Settersten RA Jr. and postnatal cocaine exposure predict teen cocaine use. Serving epigenetics before its time. Trends Genet. 30(10), Neurotoxicol. Teratol. 33(1), 110–119 (2011). 427–429 (2014). 36 Vassoler FM, Sadri-Vakili G. Mechanisms of 53 Buckley BA, Burkhart KB, Gu SG et al. A nuclear argonaute transgenerational inheritance of addictive-like behaviors. promotes multigenerational epigenetic inheritance and Neuroscience 264, 198–206 (2014). germline immortality. Nature 489(7416), 447–451 (2012). 37 Maze I, Feng J, Wilkinson MB, Sun H, Shen L, Nestler 54 Ashe A, Sapetschnig A, Weick EM et al. Pirnas can trigger a EJ. Cocaine dynamically regulates heterochromatin and multigenerational epigenetic memory in the germline of C. repetitive element unsilencing in nucleus accumbens. elegans. Cell 150(1), 88–99 (2012). Proc. Natl Acad. Sci. USA 108(7), 3035–3040 (2011). 55 Burton NO, Burkhart KB, Kennedy S. Nuclear rnai 38 Gaydos LJ, Wang W, Strome S. Gene repression. H3K27me maintains heritable gene silencing in Caenorhabditis elegans. and PRC2 transmit a memory of repression across Proc. Natl Acad. Sci. USA 108(49), 19683–19688 (2011). generations and during development. Science 345(6203), 56 Devanapally S, Ravikumar S, Jose AM. Double-stranded 1515–1518 (2014). RNA made in C. elegans neurons can enter the germline and

future science group www.futuremedicine.com 789 Perspective Nagy & Turecki

cause transgenerational gene silencing. Proc. Natl Acad. Sci. 58 Sharma A. Novel transcriptome data analysis implicates USA 112(7), 2133–2138 (2015). circulating microRNAs in epigenetic inheritance in 57 Sharma A. Bioinformatic analysis revealing association of mammals. Gene 538(2), 366–372 (2014). exosomal mrnas and proteins in epigenetic inheritance. J. Theor. Biol. 357, 143–149 (2014).

790 Epigenomics (2015) 7(5) future science group