Chapter 26

Transgenerational Epigenetics: Current Controversies and Debates

David Crews*,† and Andrea C. Gore†,‡ *Department of Integrative Biology, †Institute for Cellular and Molecular Biology, ‡Division of Pharmacology & Toxicology; University of Texas at Austin, Austin, TX, USA

Chapter Outline Introduction 371 The Fallacy of Ecologically Relevant Dosages 381 Epigenetics and Behavioral Biology 374 Experimental Simulation of Exposures in the Real Epigenotypes Versus Epiphenotypes (Molecular World (Mixtures and Synergy) 382 vs. Molar Epigenetics) 375 Thresholds 383 The Epigenotype or Molecular Epigenetics 375 Species, Sex, and Individual Differences 384 The Epiphenotype or Molar Epigenetics 376 Species Differences in Hormone Abundance, Context-Dependent vs. Germline-Dependent Metabolism, and Receptor–Ligand Interactions 384 Epigenetic Modifications 376 Sex Differences in Responses 385 Context-Dependent Epigenetic Modification 377 Individual Differences 386 Germline-Dependent Epigenetic Modification 378 Conclusion 386 Transgenerational Inheritance 378 Glossary 387 Body Burdens Include Our Personal Exposures, and References 387 Those Inherited from Our Ancestors 380

Abbreviations Conrad H. Waddington (see,1 pp. 242), the “father” of modern developmental biology and the originator of the CRS Chronic restraint stress concept of epigenetics, sums it up as follows: E1 Estrone E2 Estradiol “Some years ago (e.g., 1947) I introduced the word ‘epigenetic’ E3 Estriol derived from the Aristotelian word ‘epigenesis,’ which had more EDC Endocrine-disrupting chemical or less passed into disuse, as a suitable name for the branch of GnRH Gonadotropin releasing hormone biology which studies the causal interaction between genes and GR Glucocorticoid receptor their products which bring the phenotype into being.” LH Luteinizing hormone PCB Polychlorinated biphenyls Today, the term epigenetics tends to be a catchall for any interaction of genes and the environment in the absence of overt mutation in the DNA sequence (Figure 26.1). From INTRODUCTION the molecular perspective, this includes DNA methylation Depending upon one’s age, scholarship, and/or discipline, and modifications to histones that change the availability of epigenetics can be viewed as either a very old problem in DNA for transcription. Other cellular processes such as post- biology or an exciting new breakthrough in molecular biol- transcriptional and/or post-translational modifications – ogy. Both, of course, are true, but the former perspective is e.g. non-coding RNAs that modify gene transcription or the more useful if the new methods are going to have any RNA stability, and transcription factors that can repress/ effect on current paradigms in evolutionary biology. activate gene expression – are sometimes considered to be

Transgenerational Epigenetics. http://dx.doi.org/10.1016/B978-0-12-405944-3.00026-X Copyright © 2014 Elsevier Inc. All rights reserved. 371 372 SECTION | XI Controversies and Debate of Generational Epigenetic Inheritance

BOX 26.1 History of Epigenetics Investigators in the field of epigenetics come from one of two distinct lineages, one derived from classic genetics and the other from evolutionary biology. Both molecular epigenetics and molar epigenetics share a common history, namely the 16th–17th century debates of preformationism versus epi- genesis. The central question at that time was how a fully integrated multicellular organism develops from a single cell (the fertilized egg); it is interesting that this basic question still remains unanswered by today’s reductionism or integra- tive emergence. Preformationists believed that adult features were present fully formed in the egg and simply unfolded during growth; August Weissman belonged to this group and asserted that the eggs contained all of the elements (later known as genes) to determine the phenotype that would develop. Those believing in epigenesis held that traits emerge as a conse- quence of the progressive interaction of the constituent parts of the zygote with the environment in which it develops. Although others such as Jean-Babtiste Lamarck and Charles FIGURE 26.1 The external environment interacts with the internal Darwin were believers in epigenesis, the pivotal role of the environment to influence fetal development with both immediate and environment in the developmental process was first dem- life-long consequences. Such environmentally induced changes can occur onstrated empirically by Oscar Hertwig (1894) and subse- at all levels of biological organization. Ultimately, these influences may quently by Richard Woltereck (1909). Woltereck’s work on be epigenetic in nature, inducing mitotically heritable alterations in gene expression without changing the DNA. Epigenetics can be studied in a Daphnia, an organism that usually reproduces asexually by reductionist manner (molecular) to understand the manner in which gene cloning, demonstrated that genetically identical individu- expression is altered. Alternatively, epigenetic modifications can be exam- als would develop very different morphs depending upon ined as consequences (molar) amplifying through higher levels of biologi- their environment. The modern rendition of this perspective cal organization. For example, these alterations can bring about functional is phenotypic plasticity. A telling difference between geneti- differences in brain and behavior that result in changes in the phenotype. cists and environmentalists was in the choice of study organ- These then modify how individuals respond to conspecifics and their envi- ism. The former chose species that were relatively immune ronment. Whether these eventually can have an evolutionary impact is still to environment during development, with little variance in open to question. What is known is that human society has changed the phenotype at hatch/birth. The latter group chose species ecosystem in a manner that has had demonstrable impact on the health of that illustrated the role of the environment in shaping the humans and wildlife (modified from 132). phenotype.

Hertwig, O. Zeit-und Streitfragen der Biologie I. Präformation oder epigenetic modifications. Returning to the idea of gene– Epigenese Grundzüge einer Entwicklungstheorie der Organismen. Gustav environment interactions, when the molecular mechanism Fischer, Jena. Translated as The Biological Problem of Today: Preformation by which an environmental factor may modify a gene or or Epigenesis? (P. C. Mitchell, transl.). Macmillan, New York. 1894. Weismann, A. The Germ-Plasm. New York, Scribner’s Sons. 1898. gene product is not known, scientists are quick to specu- Woltereck, R. Weitere experimentelle untersuchungen über artverände- late that “epigenetics” may be involved. Another issue that rung, speziell über das wesen quantitativer artunderscheide bei daph- is highly debated is that of the importance of heritability niden. Versuch. Deutsch. Zool. Ges. 1909:1909:110–172. in designating a process as epigenetic or otherwise. In the absence of DNA mutation, some of these cellular processes are potentially heritable if present in the germline, but How researchers have interpreted the meaning of epi- somatic epigenetic modifications may also be heritable by genetics reflects its dual origins4 and the nature of the ques- other, context-dependent mechanisms. Yet another contro- tion being posed. At a basic level, it is useful to differentiate versy is that of the permanence or transience of epigenetic two different forms of epigenetics, namely molecular and marks, especially in the case of DNA methylation. Until molar; in other words “bottom-up” versus “top-down” epi- recently, methylation of genes at specific nucleotides (CpG genetics (see Box 26.1, page ***). The former perspective sites) was thought to occur during embryogenesis and per- emerged within the last 25 years from modern genetics/ sist through life. This “dogma” has been contradicted sev- molecular biology and focuses on molecular levels of anal- eral times now,2,3 leading to increased recognition that this ysis. The latter perspective has a deeper history, preceding molecular property is much more plastic than previously the re-discovery of Mendel’s studies and focuses on ques- assumed. These and other controversial issues will be grap- tions of evolution and adaptive significance as evident in pled with in the course of this chapter. psychobiology and evolutionary biology. Thus, the object of Chapter | 26 Current Controversies and Debates 373

study in molecular epigenetics is transcriptional and trans- Epigenetics is more an issue of one’s perspective or lational control, while in molar epigenetics, it is the individ- question, and less about the tools and techniques to be ual’s interactions with its biotic and physical environment implemented. For this reason virtually any aspect of biol- through time. More will be said about this distinction later. ogy and psychology is relevant to the investigator interested Another distinction that must be made is that between in epigenetic causes and outcomes: tissue differentiation, environmentally induced epigenetic modifications and developmental psychobiology, cognitive development, psy- parental genomic imprinting, a form of epigenetic marks chopathology, life history strategies, and phenotypic plas- that are carried in the male and female germline.5,6 Paren- ticity are just a few examples. tal genomic imprints refer to genes that are expressed in a In studies on environmental epigenetics, the goal is to parent-of-origin fashion, that is, from the mother (maternal identify a phenotypic outcome – but which one(s)? The imprint) or from the father (paternal imprint). Both envi- term “phenotype” is not meant to convey a unitary physi- ronmentally induced epigenetic modifications and parental cal feature but, in most instances, a consolidation of mul- genomic imprinting involve methylation and histone modi- tiple traits. Traditionally a trait is defined as any measurable fications, but genomic imprints are sex-specific, although aspect of the individual. In general, a deeper understanding most of these epigenetic control regions are matrilineal. of a particular phenotype increases proportionally with the Regardless of the parental origin, the imprinted gene is number of traits that are measured in the same individual. expressed or silenced in the same way in both male and Selection of particular genetic, morphological, physiologi- female offspring. Some imprints are tissue-specific but cal, behavioral, and brain traits should be predicated on the always show monoallelic expression (only one allele of a pertinent literature and demonstrated to be important for gene is actively transcribed). At this stage there is no evi- the question at hand. It is important to keep in mind that dence that the imprint is individually specific, or that each a gene has no greater meaning than body mass, circulating father (or mother) has an individual signature, but this is concentration of a hormone, etc. Indeed, because expres- due more to the species studied to date (e.g., inbred labora- sion of individual genes only has meaning in the context of tory rodents) than demonstrated as a general principle. It is other genes within and outside their functional categories, known that if strains are crossed, the imprint changes, but and because higher order traits are compounded and trans- as yet there is no information on naturally occurring species formed from lower levels (e.g., emergent properties of the and whether it is possible that the imprint may vary depend- combination of traits at lower levels of biological organiza- ing upon the population or perhaps lineage. This is a par- tion), the expression of any particular gene has relatively ticularly important question in the light of the role of sexual little importance due to epistasis and redundancy. selection in the evolution of traits in outbred (vs. inbred) With this in mind it should now be clear that when refer- species. For example, if a male mates with two females, ring to epigenetic outcomes or mechanisms we are discuss- would the paternal imprint of the offspring of the litters ing traits that are not determined by traditional genetic bases (or singleton) produced from those matings be different or of inheritance. A more precise definition would be that epi- the same? The converse question would apply to a female genetic effects are changes in the phenotype and/or specific that produces two litters by different males. Is her mater- traits that result from the environmental modification of the nal imprint identical in the respective litters? While DNA molecular factors and processes around DNA that regulate methylation is clearly involved in genomic imprinting, the genome activity yet are independent of the DNA sequence. signal for the imprint is not yet known. Sexual selection is Note that, in this instance, environment is inclusive of all not easily studied in conventional animal models (excep- stimuli, both internal and external to the individual, that tions being 7–9). Sexual selection sensu stricto refers to the may impinge on the organism during its life cycle. mutual choice exhibited by both partners, not the more typi- Finally, it is important to consider the issue of life cal experimental scenarios of mate preference where one stages. Individuals are particularly sensitive and vulner- individual is presented with a choice of two individuals able to environmentally induced epigenetic modifications usually of the opposite sex. Typically, such studies never during early life stages or in periods of transition from one proceed beyond the choice itself, leaving unanswered the stage to another. The time of maximal neuronal plasticity issues of mating and reproductive compatibility, both of is in the earliest stages of life, beginning before birth and which are all-important in biological evolution. lasting through adolescence. Although the individual’s This chapter will not deal with genomic (parent-of-origin) capacity to respond to environmental change or insult with imprinting but the authors refer the reader to excellent reviews heritable phenotypic variation at a later stage is possible, relating genomic imprinting to brain and behavior that are it is during this early period that hormones and genotype available (see 5,6,10). Instead, we will focus hereafter entirely predispose an individual’s responses to future experiences on environmentally induced epigenetic modifications, and use throughout the life cycle as well as its susceptibility to this term synonymously with context-dependent epigenetic developing disorders.11,12 Although most research has modifications. focused on the earliest life stages (fetus and neonate), 374 SECTION | XI Controversies and Debate of Generational Epigenetic Inheritance

research is now exploring the period surrounding adre- linked to profound sex differences in reproductive behav- narche (the increase in activity of the adrenal glands just iors.18–21 before puberty) and pubarche (the onset of puberty). It Behavioral neuroscience is most typically con- is during adolescence that the body, including the brain, ducted on conventional animal models such as labora- is reshaped by hormones, and the individual graduates tory rodents. However, non-conventional organisms have from dependence to independence, assuming the proper- provided unique insights that we will discuss briefly. In ties of maturity. Stressors experienced during this period designing an experiment, it is important to consider the also have enduring effects, including neural remodeling, advantages and disadvantages of the choice of species. We impaired learning and memory, and altered emotional highlight this point here, because later on we will discuss behaviors in adulthood. Prenatal, early postnatal, and ado- how environmental chemical insults cause behavioral and lescent (pubertal) life stages are often referred to as “criti- physiological changes, including molecular epigenetic cal” due to their unique sensitivity and responsiveness to modifications, that fundamentally affect functional out- stimuli such as endogenous hormones, or exogenous envi- comes. This field emerged in the 1950s based on concerns ronmental factors, possibly because of the plasticity of brought to public attention by Rachel Carson’s landmark molecular epigenetic processes during these periods. book Silent Spring,22 so the discipline is rooted in observa- tions on “real” animals in the wild. By studying diversity EPIGENETICS AND BEHAVIORAL BIOLOGY (naturally occurring species) we gain insights into evolu- tionary and ecological principles that can then be applied Behavioral neuroscience is a prime field for studying to other vertebrates, including mammals.23 However, nat- the interface between the environment and the organism urally occurring species have certain drawbacks, not the because the nervous system is the body’s “rapid response” least of which is that they require special environmentally system for adapting to external stimuli. In fact, of the three relevant cues not easily simulated in the laboratory; they physiological systems that have evolved in all vertebrates may need to be caught each year; and they may only breed to interpret environmental cues – nervous, endocrine, and once a season, or not at all, in captivity. Another important immune – it is the nervous system that enables almost disadvantage is the fact that usually only adult individuals instantaneous responses to external challenges. Further- are available for study. This is important because, unlike more, considering that all motivated behaviors occur in in animals bred for research, these individuals are the ones response to a stimulus, be it a cognizant thought (“move that have managed to survive to adulthood, but the experi- my left arm”), an environmental insult (withdrawing in ences they may have encountered as they grew are often response to pain), a sexual stimulus, availability of food, lost to us. Also, many of the molecular tools that are rou- etc., the nervous system must constantly evaluate and tine to those working with rats and mice are not readily summate all available stimuli and decide how to respond, usable in unconventional animals whose genomes are not whether consciously or unconsciously. Massive amounts sequenced, and whose proteins differ such that immuno- of research conducted on synaptic plasticity, learning and histochemical tools cannot be applied. memory, and many other areas of neurobiology, are predi- Thus, those interested in behavior and epigenetics, par- cated upon the fundamental ability of the nervous system ticularly how events early in life influence later behavior, to adapt on a second-to-second (or even faster) basis to the or how the epigenetic changes that occur from particular environment. experiences may alter future behavior, may find it neces- In epigenetic terms, it is becoming clear that neural sary to use conventional animal models such as laboratory plasticity includes the molecular events of DNA methyla- rodents. These animals are well studied and have been the tion and histone modifications, the consequences of which template on which molecular tools were forged. There are alter neurobiological functions both in the short and the limitations of such organisms, not the least of which are long term. As discussed further below, in rats the quality of that through inbreeding they have been stripped of their maternal care shapes the adult behavior of their pups, with ecologically relevant traits.24,25 In general, model systems those receiving high quality maternal being less anxious, are strains inbred to maximize fitness (reproduction and having attenuated corticosterone responses to stress, and growth rate) in an artificial and basically barren environ- increased expression of glucocorticoid receptor mRNA and ment. Thus, conventional animal model organisms can be protein in the hippocampus.13,14 Stressing mothers leads considered as a gonad, guided by a brain, with the “goal” to diminished maternal care that, in turn, alters Bdnf gene of attaining reproductive maturity as fast as possible. There expression in the young; as adults these individuals are fear- is no question that these laboratory models have been enor- ful and anxious with altered DNA methylation and histone mously useful in discerning molecular pathways activated modification at the BDNF gene.15–17 In the field of neuro- by environmental stimuli, but it is important to keep in mind endocrinology, sex differences in the molecular epigenetic these caveats when interpreting data and applying results to states of hypothalamic genes controlling reproduction are “outbred” species such as humans. Chapter | 26 Current Controversies and Debates 375

EPIGENOTYPES VERSUS EPIPHENOTYPES BOX 26.2 Evolutionary Biology and Saltation (MOLECULAR VS. MOLAR EPIGENETICS) Étienne Geoffroy Saint-Hilaire endorsed a theory of salta- The Epigenotype or Molecular Epigenetics tional (jumps) evolution – that “monstrosities could become the founding fathers (or mothers) of new species by instan- The epigenotype was defined by Waddington in his epony- taneous transition from one form to the next.” Subsequently mous paper.26 Prior to the 1940s, the gene as the unit of Richard Goldschmidt (1940) suggested “the change from heritable material was a theoretical concept without a species to species is not a change involving more and physical identity. Waddington’s term, “epigenetics,” was more additional atomistic changes, but a complete change forwarded as a conceptual model to illustrate how genes of the primary pattern or reaction system into a new one, might interact with their environment and give rise to the which afterwards may again produce intraspecific varia- phenotype. It is in this sense that the term “epigenetics” is tion by micromutation” (Goldschmidt, 1940, pg. xxviii). commonly used in molecular and developmental genetics Goldschmidt was aware that the vast majority of macromu- tations have disastrous effects on the fitness of organisms – today, namely, “the study of the mechanisms of temporal these he called “monsters.” But in Goldschmidt’s view, every and spatial control of gene activity during the develop- once in a while a “hopeful monster” is generated that is 27 ment of complex organisms” (see , p. 329). This relatively adapted to a new mode of life. According to Goldschmidt, recent area of research focuses on processes such as DNA macroevolution proceeds by the rare success of these hope- methylation (the addition of a methyl group at specific posi- ful monsters rather than by an accumulation of small changes tions on two of the four DNA bases in a particular order within populations. – a cytosine (C) followed by a guanine (G)) and histone Theoretically it is possible that if novel phenotypes such modifications (changes to the proteins that package DNA), as hopeful monsters arise in the right environmental circum- that do not change the DNA sequence or create mutations. stances, the genotype/phenotype could become fixed, and Because the artificial manipulation of methylation patterns the population will found a new species. While this idea is often lethal, or at the least results in maladaptive traits was discounted during the Modern Synthesis, aspects of the hopeful monster hypothesis have been substantiated in or “monsters” (see Box 26.2, page ***), this method of recent years. This concept has had periodic re-awakenings, research illuminates normal development by comparison to with various levels of support; e.g., Søren Løvtrup (1974) abnormalities or anomalies in development. It also helps us suggested that macromutations interfere with various epi- to understand the processes that occur when environmental genetic processes and affect the casual processes in biologi- factors affect methylation of DNA during the normal devel- cal development. There is no question that dramatic changes opment of organisms. in phenotype can occur from few mutations in key develop- mental genes. Further, phenotypic differences among spe- “Genetic changes are stable and rarely reversed, whereas cies often map to relatively few genetic factors. epigenetic changes are often reversed. A good example of that is This idea is fundamentally different from the study of genomic imprinting, where the changes imposed on DNA sequences small genetic mutations that underlie human pathology and may be lost during development, or if they persist, are erased disease. Rather, such research bears more on the mecha- and re-set during gametogenesis. Environmental influences do nisms of adaptive evolution as hopeful monsters would be not change the genotype (leaving aside mutagens), and there is shaped by natural selection. no inheritance of acquired characteristics. Epigenetics is quite Goldschmidt, R. The Material Basis of Evolution. New Haven: Yale different, because normal development depends on communication University Press; 1940. Løvtrup, S. Epigenetics: A Treatise on Theoretical Biology. New York: John between cells. Thus, a hormone, morphogen or growth factor may Wiley and Sons; 1974. induce an epigenetic change that may be heritable. This means that the environment of a cell may be all-important in determining its properties or its fate in the developing organism. In this sense, continuous outcomes, these aspects have now been refuted. epigenetics encompasses Lamarckian inheritance.” This same conceptual framework is useful when thinking From 28, pp. 78–79 of the integrated organism and how it is shaped by its biotic environment to develop different phenotypes, to adopt dif- Waddington continues to dominate the way we think of epi- ferent life history trajectories, particularly at times of transi- genetics, and his image of an epigenetic landscape as an tions, such as from embryo to birth, the immediate postnatal emergent process is the defining concept of how epigenetics period, and adolescence. operates. His original structural depiction envisions how a Attendant concepts advocated by Waddington29 were cell is modified, taking different trajectories as development canalization and genetic assimilation. Canalization connotes proceeds, emphasizing the importance of timing as well the differentiation of the gene/cell type/embryo as develop- as interactions of genes and environment. Although Wad- ment progresses and pathways becomes more entrenched, dington’s formulation conveys the idea that development thereby making it harder for the canalized development to is irreversible and results in discrete outcomes rather than be dislodged and moved into another pathway. The concept 376 SECTION | XI Controversies and Debate of Generational Epigenetic Inheritance

of genetic assimilation emerged, in part, from his work with behaviors focused on the interaction of the genetic to the Drosophila. Waddington speculated that environmentally environmental levels of biological organization). Examples induced changes in phenotype could become incorporated of this integrative approach are now numerous, but two into the genome, as evidenced by the persistence of the phe- classic efforts were those of Daniel S. Lehrman and Jay S. notype even after the original selection pressure is relaxed. Rosenblatt. Lehrman conducted elegant work on the elabo- It is in this manner that natural selection acts on develop- rate interaction of parent and offspring in ring doves (Strepto- mental pathways leading to adaptive change in the genome pelia risoria) learning to care for their young, and Rosenblatt rather than relying on genetic mutation. This dynamic view carried out exquisite research on the physiological and of development incorporates both homeostasis (the stability behavioral events that underlie the development of maternal of a final steady state) and homeorhesis (the stability of the behavior in cats and later rats. Both were students of Schneirla process of development itself). and emphasized the dynamic nature of a process that involves the interaction of the internal milieu and the organism and the The Epiphenotype or Molar Epigenetics interaction of the organism and its environment. Moreover, they defined the environment broadly to include the behavior The concept of the epiphenotype was first advanced by E.O. and physiology of socially important species members. In so 30 Wiley as a term to evaluate the characters of an individual at doing they laid the foundation for psychobiology, a vibrant any point in time in its life history. Trait analysis is essential field that focuses on how experiences accumulate throughout in classifying organisms, with the total number of characters life to shape the way in which the individual interacts with its observed increasing the accuracy of the diagnostic. The same social and physical environment.35 principle holds for experimental scientists working in the It is not our purpose to venture into the relatively unex- laboratory with conventional animals. That is, analysis of a plored frontier that lies in uniting the two subdisciplines of single character (gene, morphology, behavior etc.) is of little molar epigenetics (namely that of evolutionary and devel- use in understanding the organism. It is only when multiple opmental biology, and psychobiology) and behavioral neu- traits are quantified that an accurate assessment of the organ- roendocrinology. However, it is useful to be reminded of ism and the effect of any manipulation is achieved. Ernst Mayr’s constant refrain that behavior is at the lead- There are two types of molar epigenetics. The first arose ing edge of evolution, and the observation of Michel and from early evolutionists who asked how different environ- Moore (see 36, p. 178) that “mechanisms that underlie much ments shaped the variations observed in phenotypes within of behavioral evolution may reside in the processes studied a species. This area of study fell out of favor for about 60– by developmental psychobiologists.” 70 years in European and American science. Interestingly, However, it is necessary before going further to empha- it continued as a major field of study in Russia, and was size that the individual, not the gene, is the unit of selection represented in small part in this country in the work of The- and that an approach that integrates both molecular and odosius Dobzhansky and his students, most notably Richard molar epigenetics will be necessary to reveal the mechanisms C. Lewontin.31 Today, it has re-emerged as a vigorous area that underlie behavioral evolution.12 That is, the continuity of research among evolutionary biologists and behavioral between molecular and molar epigenetics is revealed as the ecologists. New research on the origins of polymorphisms constituent elements interact both positively and negatively (multiple phenotypes in a single species) and polyphenisms in a temporal, spatial, and conditional (internal as well as (multiple phenotypes from a single genotype) has led to a the social and physical environments) context.37 As adap- concept now commonly referred to as “phenotypic plastic- tive responses emerge they, in turn, set the stage for future ity,” which is considered one of the driving forces in the variation. Thus, evolution is a tandem process involving first relatively new union of developmental biologists with evo- development, with its built-in flexible responsiveness to both lutionary biologists (“Evo–Devo”). gene products and environment, followed by selection, which The other type of molar epigenetics has an equally long dictates which variants are spread and maintained.31,38 In this history. In psychology there has long been an interest in sense the “genome learns from its experience.”39 behavioral development or behavioral organization. Zing- 32 Yang Kuo, who worked principally in the 1920s–1930s, CONTEXT-DEPENDENT VS. GERMLINE- created much of the theory. Unfortunately, Kuo returned to China where the political strife interrupted his research and, DEPENDENT EPIGENETIC MODIFICATIONS as a consequence, his contributions were marginalized.33,34 Next we distinguish between mitotic versus meiotic epigene- The other major figure in the field was Karl S. Lashley and tic modifications, or what has been termed context-dependent his students, most notable of whom being, for the purposes of versus germline-dependent epigenetic modifications.4,40–42 this review, Frank A. Beach (regarded as one of the founders The scope of environmental effects that influence patterns of neuroendocrinology) and Theodore C. Schneirla (whose of gene expression in the brain and subsequently behavior approach to the development and display of species-typical is virtually limitless. The extent to which environmentally Chapter | 26 Current Controversies and Debates 377

induced epigenetic modifications can become an inherited can be reversed by a different environmental factor. This trait depends both upon the nature of the stimulus and the mitotically based effect can be termed “context-dependent” mechanism of its action. At a molecular level, CpG sites epigenetic change.4 (regions of DNA where a cytosine nucleotide and a gua- Waddington’s structural depiction of epigenetics envi- nine nucleotide are adjacent) are often associated with 5′ sions how the environment shapes cell development, promoter regions of genes and have a higher probability stressing the importance of timing as well of genes and of undergoing mutation than other regions of the genome. environment. The same ideas can be applied to any phe- Consequent changes in DNA methylation patterns at CpG notypic outcome, not simply the fate of a cell. That is, islands (DNA regions that contain a high frequency of CpG with time and experience, the phenotype as such under- sites) would persist and, if imprinted in the germline, have goes modifications resulting from both internal and exter- the potential to become heritable. nal forces. His famous image of an epigenetic landscape The terms “heritable” and “inherited” are often used as an emergent process is the defining concept of how synonymously. They are, however, quite distinct and should epigenetics operates. Central to this depiction is his concept be used with discrimination. Heritable refers to the poten- of canalization, or the process that the differentiating gene/ tial for a trait to be passed to the subsequent generation, cell type/embryo undergoes as development progresses and while inherited refers to a trait that has been passed to pathways becomes more entrenched. Hence, the deepening the subsequent generation. The former does not imply the furrows make it harder for the canalized development to be mechanism (genetic or epigenetic), whereas the latter is dislodged and move into another pathway. usually reserved for classic genetic mechanisms. In the con- An example of how the environment, genes, and social text of this review, heritable will pertain to context-depen- experience can serve as agents of epigenetic modification dent epigenetic modifications while inherited will relate to can be found in the leopard gecko (Figure 26.2). In many germline epigenetic modifications, particularly to the trans- reptiles it is the egg’s incubation temperature during the generational properties of these modifications. As the next mid-trimester of development that determines gonadal sex, sections will discuss, the examples of a stress-induced trait not sex chromosomes.49 In the leopard gecko, low (26 °C) in offspring due to the parent’s experience is heritable in the and high (34 °C) incubation temperatures produce females, sense that it has the potential to affect the offspring.43 An while intermediate incubation temperatures produce both inherited trait is a phenotypic alteration involving germline sexes, but with different ratios: 30 °C produces a female- transmission. Although there are examples of such epigen- biased sex ratio (25:75 or Tf), and 32.5 °C a male-biased etic inheritance in mammals,44,45 examples in wild species sex ratio (75:25 or Tm). are to date restricted to plants (e.g., 46).

Context-Dependent Epigenetic Modification Best studied are the epigenetic modifications that either have an effect early in life, such as exposure to endocrine- disrupting chemicals (EDCs) in utero, or smoking during childhood and adolescence. In the first instance, the onset of disease manifests later during the individual’s lifetime while in the latter instance, the deleterious effects of smok- ing decline with time only if the individual is no longer exposed to the stimulus. Similarly, an epigenetic modifica- tion can be perpetuated across generations by simple per- sistence of the causal environmental factor such that each generation is exposed to the same conditions. For example, if the diet47,48 or an environmental toxicant such as lead con- tinues to be present in the environment, then the molecular epigenetic modification will be induced and manifested in each generation. This type of epigenetic modification lends FIGURE 26.2 Using the Waddingtonian epigenetic landscape as a itself to relatively straightforward therapeutic venues such template, the development of gonadal sex (first choice point) and sexu- as providing methyl donors in the diet,48 and removing the ality (second choice point) are depicted in the life of the leopard gecko. environmental toxicant, whether smoking or lead. Hence, Incubation temperature acts on the embryo to modify the expression of a suite of conserved genes to determine gonadal sex. Males are produced at the environmental exposure will induce epialleles (genes both temperatures (30 and 32.5 °C), but diverge in phenotypic traits at the that differ in the extent of methylation but otherwise are molecular through behavioral levels. During adulthood, sexual experience identical), but this environmentally induced epigenetic state has further modifying effects of the phenotype. 378 SECTION | XI Controversies and Debate of Generational Epigenetic Inheritance

Incubation temperature not only establishes the gonadal binds to the first exon of the GR gene, resulting in increased sex of the individual, but also accounts for much of the expression of GR. Such maternal care during this critical within-sex variation observed in the morphology, growth, period demethylates NGFI-A and acetylates histones. Just endocrine physiology, and aggressive and sexual behavior as cross-fostering can reverse these molecular and behav- of the adult.50 Male leopard geckos grow more rapidly and ioral changes, infusion of methionine, a histone deacety- attain larger adult size than females from the same incuba- lase inhibitor, into the hippocampus can also reverse these tion temperature. Within males, those from Tm incubation events. It is important to point out, however, that the effect temperature grow more rapidly and to a larger size than do of high and low quality mothering cannot be selected for males from a Tf incubation temperature. These two types of and eventually disappears after five generations. That is, it is males do not differ in circulating concentrations of andro- neither possible to selectively breed for quality of maternal gens as hatchlings, juveniles, or adults. Estrogen levels do behavior or to pass the effect non-genomically indefinitely. differ significantly, however, with Tf males having higher This underscores the idea that context-dependent epigenetic levels than do Tm males. Despite this similarity in circulat- modifications are not contained in the germline but require ing androgen levels in adulthood, males from the two tem- continued exposure to a stimulus, in this case, variations in perature morphs differ significantly in their scent-marking natural maternal behaviors. response to exogenous hormones in adulthood, indicating neuroendocrine differences between the Tf and Tm males. There are also between-sex as well as within-sex dif- Germline-Dependent Epigenetic Modification ferences in glucocorticoid levels in response to stress. In Germline-dependent epigenetic modifications are fun- rats, females have higher circulating levels of corticosterone damentally different from context-dependent epigenetic than males, but, for both females and males, Tm individuals modifications in that the epigenetic imprint has become have significantly lower levels than Tf individuals. Brain independent of the original causative agent. That is, the neurochemistry is also influenced by incubation tempera- epigenetic modification is transferred to subsequent gen- ture. For example, a significantly higher number of tyrosine erations because it has been incorporated into the germline. hydroxylase immunoreactive cells are found in the ventral Thus, the effect is manifested each generation, even in the tegmental area of sexually inexperienced Tf versus Tm absence of continued exposure to the causative agent. While males that had been castrated and androgen-implanted, sug- the mechanisms of heritability are still not fully under- gesting that embryonic temperature plays a role in differen- stood, they relate to the fact that during germline develop- tially organizing dopaminergic systems of the temperature ment, DNA methylation patterns of most genes (other than morphs. This is supported by the finding of significantly “parent-of-origin” imprinted genes) are first erased, and then higher dopamine levels in the nucleus accumbens of Tf re-established during embryogenesis. Interestingly, specific males compared to Tm males that have interacted with a methylation patterns of heritable epialleles are maintained receptive female across a barrier. Finally, sexually experi- both in the individual, and passed through to subsequent gen- enced Tf and Tm males both show strong preferences in a erations. It is important to note that because such epigenetic Y-maze apparatus to females or their odors, but the type of modifications are mediated through the germline, it is likely female they choose depends upon their incubation history. that they would be differentially expressed in males and Given the simultaneous choice between two females from females, whose germ cells develop into sperm and ova. Based different incubation temperatures, Tf males prefer females on sex differences in the timing of re-methylation in the male from eggs incubated at high temperatures (34 °C), while Tm versus the female germline in embryos (F1), re-methylation males prefer the Tf females. Among females, Tm females appears to occur earlier in male and later in female germ are less attractive to males than are Tf females and will even cells.51 In this regard, we have recently noted some differ- attack males, a typically male pattern of aggression. ences in maternal versus paternal lineages in polychlorinated The best example of context-dependent epigenetic biphenyl (PCB) effects (A.C. Gore, unpublished) which may modification and behavior is provided by Szyf, Meaney and explain the greater vulnerability of females to later embry- colleagues.13,14 With over a decade of elegant studies, this onic EDC exposure. In our pilot PCB work, the exposure was group has demonstrated that the nature and amount of care a in the last trimester of gestation, during a time when male pup receives from the mother modulates its reaction to stress germline re-methylation is near completion, but when female later in life, largely through effects on the glucocorticoid germline re-methylation is still ongoing, possibly explaining receptor (GR) in the hippocampus. This maternal effect can our maternal lineage effects. cross generations, but its heritability depends upon the pup’s experience in the first week of life. Recently this group has documented that being reared by a more highly attentive Transgenerational Inheritance mother results in the expression of the transcription factor The defining distinction between context- and germline- A (NGFI-A), a nerve growth factor-inducible protein, that dependent epigenetic modifications lies in the timing and Chapter | 26 Current Controversies and Debates 379

persistence of the exposure. To further illuminate this point, is induced and transmitted transgenerationally. Specifically, we will use the example of exposure to an EDC in the envi- in each generation, males of the treated lineage show accel- ronment. In the exposed individuals, EDCs will induce cer- erated onset of adult diseases such as cancer, prostate dis- tain physiological changes. If this happens in somatic (i.e., ease, kidney disease, and immune defects. This same group non-germ cells), the effect will be manifested as long as the demonstrated heritable DNA methylation changes transmit- environment stays contaminated. In addition, further genera- ted across generations.44 By contrast, Stouder and Paolini58 tions will also exhibit the modification if they grow up in the found effects of vinclozolin on methylation of imprinted same contaminated environment. Examples of such adapta- genes, an effect that declined over generations. tions were originally described in mice and chickens,52,53 The F3 descendants (great-grandchildren) show altera- and recent research on various fish species54–56 has docu- tions in their normal sociosexual behavior, and in learning and mented these changes at the molecular level. In this scenario, memory. For example, females discriminate and prefer male without molecular effects on the germline, the modification descendants of the line that was not exposed to the chemical; will be lost if the environmental insult is removed. On the similarly, epigenetically modified males do not exhibit such other hand, germline-dependent epigenetic modifications a preference59 (Figure 26.3). Partner preference tests reveal can be transmitted to future generations without the require- that females of both the ancestrally exposed lineage and con- ment of additional exposure. In such instances, removal of trol lineage discriminate and prefer unmanipulated males, the contaminant will not restore the original, non-modified that is, males who do not have a history of exposure. This state. Thus, only germline-dependent epigenetic modifica- preference is likely mediated by olfactory cues. Studies of tions are truly transgenerational in nature. anxiety-like behavior indicate that vinclozolin-lineage males Michael Skinner and colleagues have developed a rat spend more time in the light compartment and shuttle from model in which the male germline bears a permanent epi- the light to dark compartments more than do control males.60 genetic modification, thereby creating an epigenetic trans- Such studies indicate that the epigenetic modification is per- generational phenotype that is not context-dependent.44,57 manent and transmitted across generations. By exposing gestating female rats to vinclozolin during the Some might argue that using the term “epigenetics” with- period of fetal sex determination, an epigenetic modification out referring to a specific molecular epigenetic mechanism

FIGURE 26.3 Experimental design for deriving F3 male and female rats from progenitor pregnant dams treated during pregnancy with vehicle or an endocrine-disrupting chemical (EDC) (vinclozolin). The Top Panel is a view of the experimental chamber, with a female in the center, separated on either side from a chamber holding a male F3-EDC on one side, and a male F3-vehicle rat on the other side. The chambers are divided by a wire mesh surrounded by Plexiglas. The Bottom Panel demonstrates that ancestral exposure to vinclozolin causes changes in attractiveness of male F3 descendants, even prior to manifestation of a disease phenotype. Right Panel: Females prefer males from the unexposed-vehicle lineage compared to the EDC lineage. Shown are data for the behaviors: Glass (time spent around the glass walls on the sides of the chamber); Still (standing still); Standing/sniffing, Walking, Grooming, Plexiglas (time spent sniffing a Plexiglas wall around the wire mesh), Facial Investigation (time in nose-to-nose contact across the wire mesh); and time at the Wire Mesh. Bars show the difference between time spent with F3-vehicle vs. F3-EDC lineage, with a negative value a preference for F3-EDC and a positive value a preference for F3-vehicle. A score of 0 on the X-axis indicates no preference. Females of either lineage (green: EDC, orange: vehicle) show a clear and significant preference for vehicle-F3 males. Left Panel: By contrast, males, whether from EDC lineage (green bars) or control lineage (orange bars), do not show any significant preferences. (Redrawn from 59.) 380 SECTION | XI Controversies and Debate of Generational Epigenetic Inheritance

is unacceptable. In fact, cancer biologists tend to be quite may decline with time if there is no further exposure. How- strict in their definition of an epigenetic modification and ever, body burden exposure need not include germline only apply the terminology to instances of heritability. epigenetic modifications. Thus, whether in a controlled However, as detailed above, a proper, historically accurate laboratory environment or in nature as a consequence of definition of epigenetics extends beyond this, incorporat- disaster, any study must be specific about the nature of the ing, in addition, functional outcomes. Thus, transcription exposure(s), as well as whether the exposure is sustained changes, physiological, brain, and behavior changes all fall through the generations (as opposed to a single exposure). within the proper definition of epigenetics. This is somewhat obvious for large-scale disasters such as Soveso in Italy (dioxin),62 Yusho in Japan and Yucheng in 63–65 BODY BURDENS INCLUDE OUR PERSONAL Taiwan (PCB and PCDF), or even smaller scale acci- dents as documented in ref 64 and recently in school lunches EXPOSURES, AND THOSE INHERITED FROM in Bihar, India (methylcyclopropene),66 in which the expo- OUR ANCESTORS sure was limited in time and scope. However, other ecologi- In the real world, exposure is not a single event, but lifelong. cal disasters in which contamination has been permanent, That is, animals living in contaminated environments have such as Lake Apopka in USA (dichlorodiphenyldichloro- 67 personal body burdens (from direct exposure) accumulated ethylene), means that multiple generations are exposed. throughout their lives. In addition, each individual experi- Another example to consider is that of periodic exposures enced the prenatal environment first as germ cells in their to seasonal spraying with insecticides and herbicides. For grandparents, and (in mammals) as fetuses in the mother’s the annual flora and fauna, this is a single exposure that uterus. In other words, each of us has a body burden of will affect their life history. For the longer-lived life forms, exposures that spans multiple generations (Figure 26.4). this becomes a body burden that may affect life history. For Any exception would only occur in situations where the those organisms exposed each year, the body burden may, contamination was transient with no lasting contaminants, in time, become a germline epigenetic transgenerational an unrealistic event. In most cases, though, if prenatal expo- inheritance borne by all descendants. sure occurred throughout the in utero period, then the dam It should be obvious that in the real world, germline- (F0 generation), that fetus (F1 generation), and its germ dependent and context-dependent epigenetic modifica- cells (F2 generation), have a personal exposure.44,61 There tions interact. Both heritable and experienced factors sculpt is also a second type of burden: the molecular epigenetic the phenotype and likely underlie the individual variation modifications occurring not only in immediate ancestors, observed in traits. The authors’ work has shown how the but also in predecessors even beyond the grandparents. In experience of past generations can influence how an indi- those cases, no trace of the original insult is no longer phys- vidual perceives and responds to events in their own life ically present, but the epigenetic mark remains. history. All germline epigenetic modifications induced by EDCs With this perspective, it is possible to create situations incorporate by definition body burden; this body burden in which normal life stressors, such as chronic restraint stress (CRS), that may occur during sensitive periods of development, might act on brain and behavior. The authors have used the vinclozolin model system described above to explore this question. CRS was chosen as a life stressor because of the extensive literature demonstrating how in rats it modifies an individual’s behavioral and neural phe- notypes as an adult.68 The authors’ specific hypothesis was that CRS during adolescence, a major life event that causes wholesale changes to an individual,69–71 would induce epi- genetic change in and of itself, and would also interact with the transgenerational epigenetic modification induced by vinclozolin three generations previously.72 At birth and at weaning, the body weight of vinclozolin- and control-lineage males are equivalent (Figure 26.5). However, with time, vinclozolin-lineage males become heavier. This dif- FIGURE 26.4 Theoretical model of exposure in the real world ference is greatest in non-stressed animals. As expected, CRS over years or decades, and the development of a chemical body bur- has an immediate effect on body weight regardless of lineage. den together with a “transgenerational” burden caused by germline changes in developing individuals. Modeled here are sequential responses Weight gain in stressed animals is half that of non-stressed of successive generations to three classes of EDCs and how these types of animals within two days of the onset of CRS. When CRS is body burdens accumulate. terminated the rate of change is increased in stressed animals Chapter | 26 Current Controversies and Debates 381

FIGURE 26.5 Effects of germline-dependent (Vinc-Lin) and context-dependent (Stress) epigenetic modifications on body weight. Animals are weighed at weaning (postnatal day (PND) 21) at Washington State University and again on receipt at University of Texas (PND 25). Illustrated is group mean body weight (g) ± SEM. Color code: purple – control lineage (Cont-Lin), non-stress condition; blue, vinclozolin lineage (Vinc-Lin), non-stress condition; yellow – Cont-Lin, stress condition; red – Vinc-Lin, stress condition. Vinc-Lin non-stress condition are significantly heavier than Cont-Lin non- stress condition. Within the non-stress groups, the Vinc-Lin males are heavier than Cont-Lin. There was no significant body weight difference between the Cont-Lin stress condition and Vinc-Lin stress condition males. Within a Cont-Lin, chronic restraint stress (CRS) has a significant effect on body weight. CRS resulted in a decrease in body weight in both lineages. This effect diminished after animals had recovered from the CRS treatment (PND 67) in the Cont-Lin males. However, body weight in stress condition continued to be lower than in the non-stress condition in the Vinc-Lin. Exposure to CRS does not seem to have differential effect between two lineages; however, it has a significant effect within a lineage, particularly in Vinc-Lin males. In these males, the decrease in body weight due to the CRS lasted longer than in the Cont-Lin males. Therefore, the transgenerational influence of vinclozolin does not affect body weight in a synergistic fashion with exposure to CRS; rather it changes sensitivity towards to external stress during this important life history stage. Inset graph shows the moving average (n = 3 consecutive weighings) of weight differences within each dyad for stress (red) and non-stress (blue) (Vinc-Lin minus Cont-Lin). (Modified from 72.) regardless of lineage. This “catch-up” in the rate of change ecologically relevant exposures in the laboratory, when occurs in both lineages and is consistent with other studies. everyone is different? Sometimes this terminology is used to imply an average THE FALLACY OF ECOLOGICALLY level, or range of levels, of exposure to a single compound based on body burden measurements. This soon loses all RELEVANT DOSAGES meaning, however, when it is realized that individuals vary, The term “ecologically relevant” comes up again and individuals of different ages vary, males and females vary, again in discussions of laboratory models of environmen- tissues vary, populations vary, sites and modes of expo- tal exposures. Papers or grant proposals are often criticized sure vary, times and duration of exposure vary, etc. The or rejected because the experimental model of exposure is most complete data set for these assertions is provided by perceived as being not ecologically relevant to humans or Schecter and colleagues.64,73–76 Indeed, the use of mean wildlife. We assert that the term “ecologically relevant” is values, even in studies in the laboratory with inbred spe- completely subjective and should be regarded as essentially cies, disguises the real point, namely how does the chemi- meaningless. What is relevant to you, the reader, and to us, cal exposure(s) affect the individual? Thus, terminology can the authors, depends upon our parents’ genes, behaviors, become downright dangerous when critics use it to dismiss geographical location, exposures to smoking, pollution, otherwise well-conducted science. They may argue that an chemicals, etc.; our more distant predecessors; and our own exposure level used in a laboratory experiment is not “rel- world, beginning in the unique maternal environment of the evant” and therefore nullifies the impact of that study alto- womb, and through infancy, childhood, adolescence, and gether. This is simply not correct. adulthood. Where we have lived, what we have eaten, and A related, and equally thorny, issue is how to measure our individual genetic make-ups result in an individual with ecological relevance. Are we talking about the levels in the a unique “exposome.” Thus, how is it possible to model air, water, or in/on the ground where an individual resides, 382 SECTION | XI Controversies and Debate of Generational Epigenetic Inheritance

is it the level in the individual itself, which may or may not EXPERIMENTAL SIMULATION OF have been acquired elsewhere, or both? Thus, how can we EXPOSURES IN THE REAL WORLD simulate exposure in order to study its effects? All modes (MIXTURES AND SYNERGY) of exposure to a single or a few chemicals have been used, including inhalation, transdermal, injection, diet, or even Clearly, present day risk-assessment paradigms must simu- gavage. The best of course is to simulate the actual nature late more closely what occurs in nature. Recently, studies of exposure. Again, how do we do this when everyone’s have shifted away from a single-exposure dose–response exposure is unique? This can be done in the case of a toxic approach using individual compounds to assaying the spill, an occupational hazard, or even in cases of household effects of combinations of these compounds. Lake Apopka application of certain chemicals. But it is not clear in every- in Florida has become a classic example of how mixtures day life. We also note that because stress, maternal behav- of environmental contaminants can affect reproduction of ior, and hormones cause epigenetic modifications, the stress animals in nature. The resemblance of gonadal and penile caused by a standard method such as injection and gavage, abnormalities of the American alligator in this lake to those how animals are housed (which affects maternal behavior), described in mice treated with the potent, synthetic estrogen and many other husbandry conditions can potentiate or even diethylstilbestrol77 led to detailed studies documenting that attenuate the effects of the exposure. chronic pollution by agricultural run-off – exacerbated by One way to relate body burden to disease is to consider a chemical spill of dicofol – was the most likely cause of cases where known exposures have occurred at the popu- the observed reproductive abnormalities.67,78 Dicofol and its lation level. For example, documented chemical exposure components bind the estrogen receptor (ER) of the alliga- disasters (see above) affected adults, children, and fetuses. tor, and, therefore, may mimic estrogens in this species. In For those individuals, the ecologically relevant dosage addition to DDE/DDT (dichlorodiphenyldichloroethylene/ was a short-term high-level exposure to a small number of dichlorodiphenyltrichloroethane) contamination, the PCB chemicals. Importantly, while exposed to the same event, mixture Aroclor 1242, and a variety of pesticides, have been individuals differ markedly in the body burden expressed detected in alligator eggs, including dieldrin, toxaphene, and, with time, these levels diminish differently accord- cis/trans nonachlor, chlordane, and pp’DDD. Exposure of ing to individuals (see above). The overemphasis on group alligator78 and red-eared slider turtle79 embryos to combina- mean values in particular hide the individual variation in tions and concentrations of these compounds typically found such changes in body burden. It is obvious that as the gen- in nature result in anomalous reproductive development. erations unfold, these exposure levels are diminished. Even As mixtures, EDCs may behave differently compared in the absence of further disastrous events, the body bur- to their behavior individually, and it becomes necessary to den of the descendants versus the “ecologically relevant” determine whether their actions are additive or synergistic. exposure of that particular population becomes even more There is much controversy as to whether EDCs exhibit syn- ambiguous. At the writing, an emerging case of toxic expo- ergistic activity. The heat of that debate stems from the fact sures to what is believed to be cooking oil contaminated that a number of EDCs have a lower potency than natural with pesticides and used in preparation of school lunches steroidal estrogens77,80,81 and, when considered individu- in India, may recapitulate some of the tragedies of the past. ally, these chemicals may exist in the environment in con- Dozens of children have died, and those who survive will centrations believed to be too low to be of concern. It is now carry a chemical body burden and its consequences for the authors’ opinion that much work in this area has been themselves and their future offspring. in realm of theoretical constructs and fantasy biology, and In the context of ecological relevance, these “what, not based on real-life physiology or hormone actions, or the where, when, why, and how” issues have great importance. real-world nature of exposures. What was the exposure? usually refers to the nature of the Synergism occurs when the effect of two factors together chemical(s) exposure. Where did the exposure occur? may is greater than the sum of their effects when given separately refer to geography as well as the nature of the exposure (open (i.e., greater-than-additive effects).82 This idea is the same air vs. confined space). When? is usually more restricted, usu- as the concept of cumulative effects as defined by Sonnen- ally relating to the age of the individual, but can also refer schein and Soto.83 Cumulative effects occur when com- to time of day since there is abundant evidence of circadian pounds are estrogenic in combination even though singly rhythms in sensitivity to environmental stimuli. Why? can be they may have no effect. Do dosages of certain compounds either active or passive involvement with the chemical(s) that in combination act synergistically to produce a strong estro- in turn can relate to the quantity, quality (pure vs. mixture). genic response? Low-dose synergy was first shown with How? can refer to direct versus indirect exposure. Hence, the polychlorinated biphenyls (PCBs) using an in vivo sex deter- concept of “ecologically relevant” dosages that is currently mination assay in the red-eared slider, a species in which driving much of research has little relevance to understanding sex is determined by temperature but which chemicals can the nature and consequences of exposure. override. In the case of estrogenic chemicals, the sex ratio Chapter | 26 Current Controversies and Debates 383

is skewed in a female-biased manner, and thus the ratio of activation. Most assume that the major circulating corticoid females to males is a read-out of an estrogenic exposure. is either corticosterone or cortisol, and this varies with spe- Using this bioassay, the effects of 11 common PCBs on cies. However, all species produce both corticoids as well as sex ratio were assayed by Bergeron and coworkers.84 Only dihyroepiandrostanedione, and the ratio of these hormones two of the compounds tested, 2′,4′,6′-trichloro-4-biphenylol is important. Indeed, Wynne-Edwards and colleagues92 (3-PCB) and 2′,3′,4′,5′-tetrachloro-4-biphenylol (4-PCB), have demonstrated in various species, including humans, both hydroxylated PCBs, were found to have estrogenic that what is traditionally considered to be the non-dominant activity as indicated by the production of female or intersex (the lower circulating concentration) glucocorticoid has an hatchlings from eggs incubated at a male-producing temper- independent role in stress responses. ature. These results of combined effects have been replicated Taken together the evidence indicates the need to not in further organismal and molecular studies using the mouse only examine combinations of compounds but also to assess uterus, where these same two compounds show an appre- the endocrine milieu of the animal, both at the time of expo- ciable affinity for ER, due, in part, to their conformational sure as well as at the endpoint, rather than focusing on sin- properties as hydroxybiphenyls.85,86 When eggs are treated gle hormone profiles. with various mixtures of PCBs except 3- and 4-PCB, there is no evidence of an effect on sex ratio. However, when 3- and 4-PCB are combined, they act synergistically, resulting in THRESHOLDS a significant increase in female and intersex hatchlings at a The traditional toxicological approach uses single doses dose of 10 µg or less than 1 ppm; when administered alone, of a single chemical at increasing concentrations to gener- 3-PCB and 4-PCB require at least a tenfold higher dose to ate a dose–response curve and to identify the lowest dose show sex reversal. This is a clear demonstration of syner- at which no adverse effects are observed (the no-observed- gism on an unambiguous endpoint. Additional studies were adverse-effect level or NOAEL). This approach, used in performed, evaluating the effects of various EDCs, selected standard risk assessment, is contingent on the presumption because they had been identified in alligator eggs from that each chemical has a threshold dose below which no Lake Apopka. The red-eared slider sex determination assay adverse effects occur. As applied today, the NOAEL derives showed that not only did several of these compounds reverse from work performed in rodents to establish the highest dose the sex of significantly greater numbers of individuals com- without toxicity and dividing by 100. The NOAEL then is pared to the controls, but the sex steroid hormone levels of used as a surrogate for a threshold. This concept was first the treated hatchlings were compromised in the same pattern challenged by Gaylord and colleagues,93 who suggested that that had been observed in the alligators.87 a threshold dose might not apply to EDCs. This hypothesis As elaborated below, the endocrine state of the individ- was verified using the red-eared slider turtle model system ual is a critical determinant of the eventual outcomes of an of temperature-dependent sex determination.94 In this sys- exposure. This extends beyond whether the individual is an tem, E2 can override effects of temperature depending upon embryo, newborn, juvenile, or sexually mature adult. Not dose of E2 and temperature of incubation. A retrospective only are the hormonal states in each of these life states differ- analysis of publications that varied doses of E2 at different ent but they also obviously vary according to sex. It is also a incubation temperatures in directing female sex determina- mistake to assume that only one hormone should be consid- tion was performed. The resulting dose–response curves fit a ered. Hormones constitute an internal milieu that provides a Michaelis–Menten model of a single protein–molecule inter- background against which the external agent will interact. action driving a reversible process. Using power analysis in Consider the three main estrogens: estrone (E1), estradiol this large (2400 egg) dose–response study, it was unambigu- (E2), and estriol (E3). In most scientific studies, it is usually ously shown that there is no threshold dosage for exogenous only E2 that is administered or measured. However, estro- estrogen. Development is an exquisitely calibrated process gens other than E2 are part of the natural milieu and they can that depends upon a precise amount and sequence of hor- exhibit synergy among themselves. In the red-eared slider88 monal action. Thus, because EDCs might mimic the actions E2 synergizes with E3, resulting in a rate of sex reversal that of endogenous hormones important to development, the addi- is twice than that shown by E2 alone (36% reversal for E2 tion of any exogenous chemicals and/or EDCs automatically alone vs. 69% reversal for 0.2 µg E2 + 0.01 µg E3). exceed thresholds for regulating for these developmental The interactions of natural estrogens have not been well changes.79 Since the time that those studies were published, studied in mammalian systems, but evidence of interac- there has been extensive research on low-dose effects of tions exists.89–91 This raises the issue of how EDCs, par- EDCs that, taken together, challenge the concept of a thresh- ticularly estrogenic EDCs, might interact not only with E2, old dosage for EDCs.95 That is, even very low concentra- but also with E1 and E3, to affect, possibly differentially, tions of endocrine-disrupting, receptor-dependent chemicals synthesis, degradation, and/or actions upon estrogen recep- presently found in the environment exceed the threshold for tors. A similar situation occurs with measures of adrenal endogenous hormones. The only conclusion feasible is that 384 SECTION | XI Controversies and Debate of Generational Epigenetic Inheritance

the concept of a threshold is meaningless. To entertain other- factors contributing to this variation include the embryonic wise is to leave the realm of biology and enter into ideology. environment, which itself is dependent upon the psycholog- In fact, a recent letter from toxicologists96 challenging the ical and physiological condition of the mother, circulating endocrinological principles of low-dose EDC effects and the levels of adrenal and gonadal steroid hormones during and concept of no-thresholds is an example of frivolous and dan- subsequent to life stage transitions, age, and experience. gerous misinterpretation of an extant, peer-reviewed body of The latter aspect is often unappreciated, yet these are quite literature. That letter is based on what the authors call “com- separate; further, it is possible to age without experience, mon sense.” If scientific research were based on common but not possible to gain experience without aging. Taken sense instead of empirical experimentation, there would be together this is the essence of the sexuality of the organism. no need to perform experiments at all. A series of responses from the endocrine, pediatric, and reproductive community Species Differences in Hormone Abundance, ensued to set the record straight on the need to base decisions Metabolism, and Receptor–Ligand on high-quality, peer-reviewed science.97–101 Interactions SPECIES, SEX, AND INDIVIDUAL The circulating concentrations of steroid and other hormones vary, often by an order of magnitude, between species. We DIFFERENCES have already discussed the fact that multiple members of The diversity of vertebrate species is matched by the diver- a class of steroid can exist within individuals. We add to sity in sex determining mechanisms, processes of sexual that discussion by pointing out that there are also species differentiation, and the development of sexuality.8 In the differences in the dominant hormone (mentioned above for literature, sex differences in almost all cases refers to gono- glucocorticoids) but also for estrogens and androgens. It chorism, or separate sexes in separate individuals. Differ- may sound simple, but mice are not rats in most aspects ences between sexual stages in hermaphroditic species is of endocrine physiology and response to EDCs. Relevant a promising area of sex difference research, but relatively to this point is that transgenerational epigenetic modifica- limited at this stage. tions have been demonstrated in rats (see above) and, more To facilitate an understanding of sex differences, it is recently in mice.102,103 The effects are quite different, which important to understand the major stages of an emergent may be due to the chemical (vinclozolin vs. BPA), dosage process that is the central to the life history of all organisms. (pharmacological vs. low dose), method of testing between The first relates to the gonad: sex determination relates to laboratories, or species. Developmental stage or physiologi- the mechanisms by which the gonadal ridge is directed to cal state also profoundly affects the abundance and metabo- develop into ovary or testes. In species with genotypic sex lism of hormones. As examples, the dominant androgen in determination (GSD), sex is determined at fertilization by some species is testosterone, but in others it may be dihy- genetic constitution; the presence of the Y chromosome- drotestosterone and, in still others, 11-ketotestosterone. In linked SRY gene (mammals and male heterogamety) or females, E1–3 are commonly found, but their levels and double dosage of the W chromosome (birds and female ratios vary enormously depending upon life stage and phys- heterogamety). An alternative mechanism is environmental iological state. For example, during pregnancy, levels of E1 sex determination (ESD) in which some aspect of the envi- rival that of E2.104,105 The “dominant” steroid can also vary ronment (such as temperature in early embryonic develop- with life stage. In humans, cortisol is the dominant steroid, ment or behavioral dominance in adulthood) that acts after but in the fetus it is corticosterone.92 fertilization determines what type of gonad will develop. Furthermore, there is a relationship between circulating This is to be distinguished from gonadal differentiation, or steroid hormones and sensitivity to these hormones. Species the process by which the testis or the ovary develop from having low circulating levels of E2 are more sensitive to a bipotential germinal ridge. As would be expected, the both endogenous and exogenous estrogen than are species mechanisms and processes are fundamentally different in having relatively high estrogen titers.106 Earlier research sex determination and gonadal differentiation. suggests that species differences in hormone levels and sen- The sexual phenotype can also be divided into two lev- sitivity to the hormone are, in turn, related to the abundance els of organization. The primary level of organization is the of hormone receptor mRNA in neuroendocrine controlling process of sexual differentiation that follows gonadal deter- brain areas. For example, rats, guinea pigs, and hamsters mination. Here the primary and accessory sex structures, the have very different circulating levels of E2, with rats having gonads and the duct systems and associated structures that the lowest and hamsters the highest. Behavioral sensitivity convey the gametes, are organized sequentially by distinct and estrogen binding capacity in the brain to exogenous mechanisms. The second level of organization establishes estrogen is proportional to these levels in the same order; in individual variation in the morphological, physiological, other words, ER abundance in the brain is inversely propor- and behavioral aspects of the sexual phenotype. Some of the tional to estrogen levels. Chapter | 26 Current Controversies and Debates 385

The relative potencies of naturally occurring steroids relatively high estrogen titers.106 Site 2 is contaminated and can vary with species.79,107 In the red-eared slider turtle, depicts the situation whereby EDCs in the environment can estriol is far more potent than E2 in overcoming the effects lead to altered endogenous circulating concentrations of sex of a male-producing temperature. Similarly, DDE and its hormones and their actions on receptors. Several scenarios metabolites have been implicated in a variety of repro- are envisioned for Site 2. In Scenario A, EDCs and endog- ductive anomalies, but not in all species. In the red-eared enous hormones may form heterodimers that do not activate slider, DDE will cause female development when applied hormone response elements on the DNA as effectively. This to the egg in concentrations found in a contaminated envi- would diminish hormone action, feedback mechanisms, ronment,108 but has no detectable effect in the green sea and ultimately levels of sex hormones in the circulation. In turtle,109 and is anti-estrogenic in the tiger salamander.110 Scenario B, such heterodimers as described in Scenario A In the female rat, developmental exposure to methoxychlor occur, but in this case bind as effectively or even more effec- causes aberrant reproductive cycles in adults,111 but in ham- tively to the DNA response element, thereby increasing hor- sters is without apparent effect on estrous cycles.112 On the mone action. Scenario C is a case in which an exogenous other hand, the pesticide kepone, when given early in life, ligand may “hijack” the estrogen receptor, causing dimer- masculinizes the behavior of female hamsters,113 and in rats ization and gene activation even under very low endogenous seems to inhibit female sexual behavior when administered E2 conditions, a situation that has been reported for dioxin at proestrus,114 fails to induce lordosis in ovariectomized in female mice.117 In yet another scenario, D, binding of the females,115 and masculinizes female tissues.116 Thus, rather EDC to the hormone receptor, which typically happens at than concentrating on a few species to serve as standards for much lower affinity than binding of the endogenous ligand, defining a chemical as an EDC, a more instructive approach destroys the receptor, thereby preventing activation by the might be to document and then understand why species dif- hormone of its response element on target genes. fer in their sensitivity to specific compounds. Finally, physiological effects of EDCs may involve The “sensitivity compensation” model explains such neuroendocrine feedback mechanisms that lie upstream to differences.106 This model proposes that the sensitivity of gonadal steroid biology and include hypothalamic gonado- a species (or an individual) to a compound is related to the tropin-releasing hormone (GnRH) and pituitary gonadotro- typical circulating concentrations of steroid hormone and pin release and regulation. The hypothalamic–pituitary axis the physiological effects these hormones may exert. Evi- is exquisitely sensitive to endogenous feedback, as well as dence shows that this model can be applied to species dif- EDCs.42 EDC effects on levels of circulating concentrations ferences in sensitivity to different EDCs. In addition, the of reproductive hormones, in turn, influence receptor abun- sensitivity compensation model enables predictions as to dance. For example, luteinizing hormone (LH) levels are which EDCs an animal may be particularly sensitive. For lowered or delayed following exposure to chlordecone in example, species with relatively low endogenous circu- female rats,118 1,2-dibromo-3-chloropropane in male rats,119 lating levels of E2 are more sensitive to estrogen-mimics and chlordimeform in female hamsters.120 Other EDCs – and demonstrate equivalent physiological responses when including procymidone and vinclozolin121–123 – cause exposed to them. Thus, circulating hormone levels in ani- increased LH levels in rodents, which, in turn, affect levels of mals from uncontaminated areas could serve as sentinels of natural hormones and thus, the abundance of hormone recep- potential EDC effects. Because circulating levels of steroid tor.106 Finally, as mentioned, circulating endogenous hormone hormones affect the abundance of sex hormone receptors, levels in exposed animals may be an index of EDC exposure, receptor abundance may be useful as an index for mea- as indicated by the discovery of abnormally low levels of suring the effects of EDCs. Because levels of natural hor- circulating hormones in alligators and red-eared slider tur- mones can affect receptor abundance, the authors predict tles. Thus, these avenues for measuring endocrine effects of that EDCs may also affect hormone receptor levels in brain EDCs – circulating hormone levels, LH levels, and receptor areas important for the regulation of reproduction. Thus, abundance – can all serve as markers for sensitivity. receptor abundance could be useful as an index for measur- ing the effects of EDCs. We present hypothetical situations about possible actions Sex Differences in Responses of contaminants on hormone–receptor interactions. Con- Another obvious issue is sex differences, a topic we have sider Site 1, an uncontaminated location where animals have covered above in discussing differential effects of EDCs normal expected endogenous sex hormone concentrations, on hormones, gene expression in the brain, and behav- hormone–receptor interactions, and feedback mechanisms. iors. There is abundant evidence that males and females In such cases it is known that species having low circulat- vary in the production, secretion, and sensitivity to ste- ing concentrations of E2, for example, have a greater abun- roid and other hormones.124 Such sex differences extend dance of estrogen receptor (ER) in the brain and are more to EDC sensitivity and responsiveness.79,125 Again, much sensitive to exogenous estrogens than are species having of this topic hinges upon the realization that development 386 SECTION | XI Controversies and Debate of Generational Epigenetic Inheritance

(especially fetal and infant) is a highly sensitive period to differentiation and the potential role of EDCs in disturbing both endogenous and exogenous hormones or hormonally these normal processes is unknown in human twins. active chemicals. Once altered by EDC exposure during a Related to this are a myriad of environmental effects sensitive period, the remaining life trajectory is modified, on adult social and anxiety behaviors. Some are more again differentially in the sexes. These individuals have important than others. For example, it is common in inves- been permanently altered, such that they may be designated tigations using genetically modified mice to take animals as non-canonical males or females, and should be investi- as needed to fill groups without considering “family-of- gated more with respect to how they differ from “normal” origin” issues such as sex ratio, genotype ratio, size of males and females, both in laboratory environments as well litter, maternal care, unanticipated stresses that occur in as in the real world.40,59,72,102,103,126,127 any colony, etc. It is not valid to assume that Mendelian The mode of sex determination needs consideration, as genotype ratio and neutral sex ratio based on group means mammals, reptiles, fish, and other vertebrates vary enor- for a population (colony) will apply to the experimental mously in these processes,49 including genotypic (sex chro- animals actually used in the study. It is the distribution mosome) and environmental (temperature, humidity, social of individuals within the groups that is essential. Statisti- structure, etc.) patterns of sex determination, as well as their cians are beginning to recognize this and consider the litter intermediate forms. While we can speak of sex reversal, it is as the unit of analysis (the N) rather than the individual not appropriate to consider species with environmental sex within a litter. An N of 8 females from 8 litters is very dif- determination as being reversed, but rather altered. It is also ferent from an N of 8 females from 2 litters, and statistical clear that environmental sex determination alters the mode methods must be used appropriately. and tempo of EDC action.50 Thus, species and sex differ- Experiences during sensitive periods of development can ences are fundamental components of understanding when, alter behavioral and brain phenotypes. By keeping geno type and how, EDCs act. constant, Freund and colleagues131 have showed how indi- vidual differences among inbred mice become more pro- nounced over time in enriched environments. However, the Individual Differences genotype can also be a factor. This is particularly the case The effects of epigenetic modification by EDCs at the level in genetically modified animals. This concept needs to be of the individual is important. For example, the intrauterine applied to EDC research. position of the individual, in animals with litters, is a signif- icant factor in phenotypic differences in morphology, hor- mones, and behavior later in life. That is, a female rodent that CONCLUSION develops between two females in utero produces litters with We have discussed some of the most critical and unresolved a female-biased sex ratio, whereas a female that developed issues in epigenetics. In order for the field to move forward, between two males in utero produces litters with a male- it is necessary to reconcile some fundamental differences biased sex ratio. An individual’s response to EDCs as adults in vocabulary and definitions used in seemingly contra- can depend upon its particular embryonic environment.128 dictory ways by different groups. For example, molar and Previous work indicates that female offspring of rats fed molecular epigenetics (Box 26.1) are complementary, not bisphenol-A while pregnant exhibit a subtle, but significant, contradictory, perspectives; those who insist that epigenetic increase in postnatal growth and an advancement of the traits must be heritable through germline transmission must onset of puberty as a group. However, when the intrauterine appreciate the validity and history of other aspects of the position of the fetus was considered, the effect is restricted field. Only then can epigenetics progress in a meaningful to those individuals having the highest background levels way into translational applications. of endogenous E2 during fetal life, that is, females between Epigenetics is a perspective on how life histories may two female neighbors. In gerbils, during pregnancy the cir- unfold and be transmitted through time: that is, how evolu- culating concentrations of androgens in the pregnant female tion may work. Epigenetics will continue to grow as long bearing male-biased litters are higher than in females bear- as it borrows from other fields. Again, history matters ing female-biased litters.129,130 Thus, male-biased litters (Box 26.2). produce individuals can differ in sensitivity to exogenous Suites of genes underlie the fundamental plasticity of EDCs than individuals from female-biased litters. Because an organism, particularly during development or life stage the hormonal milieu differs during development depending transitions. Our understanding of the way that these gene on intrauterine position, such results indicate that the back- networks interact with the experiences that cumulate dur- ground hormonal profile influences an individual’s subse- ing an individual’s life history is just beginning. An impor- quent sensitivity to this EDC. While this concept may not tant interface between the environment (either internal or seem applicable to humans that do not have litters, twinning external) and the genotype is that of epigenetic modifica- is common (approximately 1:80 births). However, sexual tions. Exactly how these modifications come about is still Chapter | 26 Current Controversies and Debates 387

relatively unknown, but recent studies at both the molecu- Genetic assimilation A concept attributed to Waddington in which lar and molar levels indicate that the origin of such effects phenotypes that are environmentally malleable become genetically may occur in previous generations. That is, experiences of fixed and no longer require the original environmental stimulus in earlier generations can modify regulatory factors affecting order to be manifested. gene expression such that the DNA sequence itself is not Germline-dependent epigenetic modification Epigenetic modifica- tion arising from exposure to an environmental cue during germline changed but the individual’s physiology and behavior are development. By being incorporated into the germline, the effect is substantially influenced. By changing the individual’s per- manifested each generation, even in the absence of continued expo- ception and behavior, the nature of its interaction with the sure to the causative agent. environment is altered. Understanding how such modifica- Homeorhesis Stability of the process of development itself. tions actually occur will increase our understanding of how Homeostasis Stability of a final steady state. the environment influences the relationship between geno- type and behavior during sensitive developmental periods. Finally, not discussed in this chapter but something that REFERENCES should be recognized and appreciated, is ethnicity. When speaking of humans we tend to think of races, geography, 1. Waddington CH. The basic ideas of biology. In: Waddington CH, editor. Biological Theory.. Cambridge, MA: MIT Press; 1968. pp. and culture. When referring to nonhuman animals this can 238–53. be thought of as subspecies. In both, the issues are the same. 2. McCarthy MM. Sex and the Developing Brain: Morgan & Claypool Whether individuals are sedentary throughout their life Life Sciences; 2011. cycle or mobile is also important, since movement to new 3. Ho SM, Tang WY, Belmonte de Frausto J, Prins GS. Developmen- locations alters the nature of the environment to which the tal exposure to estradiol and bisphenol A increases susceptibility to individual is exposed; that is, behavioral physiology reflects prostate carcinogenesis and epigenetically regulates phosphodiester- geography and is shaped by elemental composition of the ase type 4 variant 4. Cancer Res 2006;66:5624–32. natal sites. 4. Crews D. Epigenetics and its implications for behavioral neuroendo- crinology. Frontiers Neuroendocrin 2008;29:344–57. 5. Keverne EB, Curley JP. Epigenetics, brain evolution and behaviour. Frontiers Neuroendocrin 2008;29:398–412. GLOSSARY 6. Keverne EB. Significance of epigenetics for understanding brain Canalization Developmental process proposed by Waddington by development, brain evolution and behaviour. pii:S0306-4522 Neuro- which the cell, tissue, organ, or organism initially is subject to inter- science 2012;(12):01141–4. nal and external forces but, through time, less subject to influences 7. Carson HL. Mate choice theory and the mode of selection in sexual from these forces, ultimately leading to reduced variability in natu- populations. Proc Natl Acad Sci 2003;100:6584–7. rally occurring species. 8. Crews D. Diversity and evolution of neuroendocrine mechanisms Context-dependent epigenetic modification Epigenetic modi- underlying reproductive behavior. In: Becker J, Breedlove SM, fication caused by experiences during sensitive periods of life Crews D, McCarthy MM, editors. Behavioral Endocrinology. (prenatal, early postnatal, adolescence) that modify the biologi- Cambridge, MA: M.I.T. Press; 2002. pp. 223–88. cal system but that are not intrinsically heritable in the absence 9. Drickamer LC, Gowaty PA, Holmes CM. Free female mate choice in of continued exposure. Such experiences can include passive or house mice affects reproductive success and offspring viability and active exposures and have long-lasting effects on the individual’s performance. Anim Behav 2000;59:371–8. subsequent life history. These modifications can be perpetuated 10. Keverne EB. Neuroscience: A mine of imprinted genes. Nature across generations due to persistence of the causal environmental 2010;466:823–4. factor such that each generation is exposed to the same conditions. 11. Gilbert SF, Epel D. Ecological Developmental Biology. Sunderland, Endocrine-disrupting chemical (EDC) Anthropogenic chemical(s) MA.: Sinauer: Integrating Epigenetics, Medicine, and Evolution; 2009. that interfere(s) with any aspect of hormone action. 12. Bateson P, Gluckman P. Plasticity, Robustness, Development and Epialleles A group of otherwise identical genes that differ in the degree Evolution. Cambridge: Cambridge University Press; 2011. of methylation and produce novel phenotypes that are heritable 13. Kappeler L, Meaney MJ. Epigenetics and parental effects. Bioessays across generations. 2010;32:818–27. Epigenotype A term coined by Waddington to suggest a process by 14. Szyf M, Weaver I, Meaney M. Maternal care, the epigenome and which a cell may differentiate as a consequence of changes in the phenotypic differences in behavior. Reprod Toxicol 2007;24:9–19. internal and external environments; a potential for change that is 15. Day JJ, Sweatt JD. Epigenetic mechanisms in cognition. Neuron heritable and the basic genetic program of that cell. Refers to a 2011;70:813–29. stable pattern of gene expression that is outside the actual base pair 16. Roth TL, Sweatt JD. Epigenetic marking of the BDNF gene by early- sequence of DNA. life adverse experiences. Horm Behav 2011;59:315–20. Epiphenotype A concept developed by E.O. Wiley as a means to eval- 17. Zovkic IB, Guzman-Karlsson MC, Sweatt JD. Epigenetic regulation of uate the total or essential characters of an individual at any point in memory formation and maintenance. Learning Mem 2013;20:61–74. time in its life history. 18. Kolodkin MH, Auger AP. Sex difference in the expression of DNA Epistasis Traits that depend upon more than one gene, each interacting methyltransferase 3a in the rat amygdala during development. J Neu- with one another. roendocrin 2011;23:577–83. 388 SECTION | XI Controversies and Debate of Generational Epigenetic Inheritance

19. Kurian JR, Olesen KM, Auger AP. Sex differences in epigenetic 44. Anway MD, Cupp AS, Uzumcu M, Skinner MK. Epigenetic trans- regulation of the estrogen receptor-alpha promoter within the devel- generational actions of endocrine disruptors and male fertility. oping preoptic area. Endocrinology 2010;151:2297–305. Science 2005;308:1466–9. 20. Auger AP, Auger CJ. Epigenetic turn ons and turn offs: Chro- 45. Nelson VR, Heaney JD, Tesar PJ, Davidson NO, Nadeau JH. Trans- matin reorganization and brain differentiation. Endocrinology generational epigenetic effects of the Apobec1 cytidine deaminase 2011;152:349–53. deficiency on testicular germ cell tumor susceptibility and embryonic 21. Chung WC, Auger AP. Gender differences in neurodevelopment and viability. Proc Natl Acad Sci 2012;109:E2766–73. epigenetics. Eur J Physiol 2013;465:573–84. 46. Cubas P, Vincent C, Coen E. An epigenetic mutation responsible for 22. Carson R. Silent Spring. Boston: Houghton Mifflin; 1962. natural variation in floral symmetry. Nature 1999;401:157–61. 23. Crews D, Moore MC. Evolution of mechanisms controlling mating 47. Faulk C, Dolinoy DC. Timing is everything: The when and how of behavior. Science 1986;231:121–5. environmentally induced changes in the epigenome of animals. 24. Seok J, Warren HS, Cuenca AG, Mindrinos MN, Baker HV, Xu W, Epigenetics 2011;6:791–7. et al. Genomic responses in mouse models poorly mimic human 48. Dolinoy DC, Jirtle RL. Environmental epigenomics in human health inflammatory diseases. Proc Natl Acad Sci 2013;110:3507–12. and disease. Environ Mol Mutag 2008;49:4–8. 25. Martin B, Ji S, Maudsley S, Mattson MP. Control" laboratory 49. Crews D, Bull JJ. Mode and tempo in environmental sex determina- rodents are metabolically morbid: why it matters. Proc Natl Acad Sci tion in vertebrates. Sem Cell Develop Biol 2009;20:251–5. 2010;107:6127–33. 50. Sakata JT, Crews D. Developmental sculpting of social phenotype 26. Waddington CH. The epigenotype. Endeavour 1942;1:18–20. and plasticity. Neurosci Biobehav Rev 2004;28:95–112. 27. Holliday R. DNA methylation and epigenetic inheritance. Philos 51. Reik W, Dean W, Walter J. Epigenetic reprogramming in mamma- Trans R Soc Lond B Biol Sci 1990;326:329–38. lian development. Science 2001;293:1089–93. 28. Holliday R. Epigenetics: a historical overview. Epigenetics 2006; 52. Ozburn GW, Morrison FO. Development of a DDT-tolerant strain of 1:76–80. laboratory mice. Nature 1962;196:1009–10. 29. Waddington CH. Canalization of development and the inheritance of 53. Poonacha KB, Wentworth BC, Chapman AB. Genetic resistance to acquired characters. Nature 1942;150:563–5. DDT in the Japanese quail (Coturnix coturnix japonica). Poultry 30. Wiley EO. Phylogenetics: The Theory and Practice of Phylogenetic Science 1973;52:841–6. Systematics. Wiley and Sons; 1981. 54. Whitehead A. Comparative genomics in ecological physiology: 31. Lewontin RC. The Triple Helix: Gene, Organism and Environment. toward a more nuanced understanding of acclimation and adaptation. Harvard University Press; 2000. J Exp Biol 2012;215:884–91. 32. Kuo Z-Y. The Dynamics of Behavior Development. New York: 55. Whitehead A, Crawford DL. Neutral and adaptive variation in gene Random House; 1967. expression. Proc Natl Acad Sci 2006;103:5425–30. 33. Greenberg GP, TY. Prolegomena to praxiology redux: The psychol- 56. Whitehead A, Roach JL, Zhang S, Galvez F. Genomic mechanisms of ogy of Zing-Yang Kuo. In: Valsiner J, editor. Fropm Past to Future: evolved physiological plasticity in killifish distributed along an envi- From instinct to epigenesis: Lessons from Zing-Yang Kuo. Frances L. ronmental salinity gradient. Proc Natl Acad Sci 2011;108:6193–8. Hiatt School of Psychology: Clark University; 2000. pp. 13–37. 57. Skinner MK, Haque CG, Nilsson E, Bhandari R, McCarrey JR. Envi- 34. Honeycutt H. The “enduring mission” of Zing-Yang Kuo to eliminate ronmentally induced transgenerational epigenetic reprogramming the nature-nurture dichotomy in psychology. Develop Psychobiol of primordial germ cells and the subsequent germ line. PLoS One 2011;53:331–242. 2013;8:e66318. 35. Gottleib G. Individual Development and Evolution: The Genesis of Novel 58. Stouder C, Paoloni-Giacobino A. Transgenerational effects of Behavior. Mahwah, N.J.: Lawrence Erlbaum Associates, Inc.; 2002. the endocrine disruptor vinclozolin on the methylation pattern 36. Michel G, Moore C. Developmental Psychobiology: An Interdisci- of imprinted genes in the mouse sperm. Reproduction 2010;139: plinary Science. Cambridge, MA: MIT Press; 1995. 373–9. 37. Nijhout HF. The importance of context in genetics. Amer Sci 2003; 59. Crews D, Gore AC, Hsu TS, Dangleben NL, Spinetta M, Schallert T, 91:416–8. et al. Transgenerational epigenetic imprints on mate preference. Proc 38. West-Eberhard MJ. Developmental Plasticity and Evolution. Oxford Natl Acad Sci 2007;104:5942–6. University Press; 2003. 60. Skinner MK, Anway MD, Savenkova MI, Gore AC, Crews D. Trans- 39. Jaenisch R, Bird B. Epigenetic regulation of gene expression, how generational epigenetic programming of the brain transcriptome and the genome integrates intrinsic and environmental signals. Nature anxiety behavior. PLoS ONE 2008;3:e3745. Genetics Suppl 2003;33:245–54. 61. Walker DM, Gore AC. Transgenerational neuroendocrine disruption 40. Crews D. Epigenetic modifications of brain and behavior: theory and of reproduction. Nat Rev Endocrin 2011;7:197–207. practice. Horm Behav 2011;59:393–8. 62. Eskenazi B, Mocarelli P, Warner M, Samuels S, Vercellini P, Olive 41. Crews D, Gore AC. Epigenetic synthesis: A need for a new paradigm D, et al. Serum dioxin concentrations and endometriosis: a cohort for evolution in a contaminated world. F1000 Reports 2012;4:18. study in Seveso, Italy. Environ Health Perspect 2002;110:629–34. 42. Gore AC. Developmental programming and endocrine disruptor 63. Aoki Y. Polychlorinated biphenyls, polychlorinated dibenzo-p-dioxins, effects on reproductive neuroendocrine systems. Front Neuroendo- and polychlorinated dibenzofurans as endocrine disrupters–what we crin 2008;29:358–74. have learned from Yusho disease. Environ Res 2001;86:2–11. 43. Xiong F, Zhang L. Role of the hypothalamic-pituitary-adrenal axis in 64. Schecter A, Stanley J, Boggess K, Masuda Y, Mes J, Wolff M, et al. developmental programming of health and disease. Front Neuroen- Polychlorinated biphenyl levels in the tissues of exposed and nonex- docrinol 2013;34:27–46. posed humans. Environ Health Perspect 1994;102(Suppl. 1):149–58. Chapter | 26 Current Controversies and Debates 389

65. Yang CY, Yu ML, Guo HR, Lai TJ, Hsu CC, Lambert G, et al. The 85. Ramamoorthy K, Vyhlidal C, Wang F, Chen I, Safe S, McDon- endocrine and reproductive function of the female Yucheng ado- nell DP, et al. Additive estrogenic activities of a binary mixture of lescents prenatally exposed to PCBs/PCDFs. Chemosphere 2005; 2′,4′,6′-trichloro- and 2′,3′,4′,5′-tetrachloro-4-biphenylol. Toxicol 61:355–60. Appl Pharm 1997;147:93–100. 66. Kashyap MP, Singh AK, Kumar V, Tripathi VK, Srivastava RK, 86. Martinez JM, Stephens LC, Jones LA. Long-term effects of neonatal Agrawal M, et al. Monocrotophos induced apoptosis in PC12 cells: exposure to hydroxylated polychlorinated biphenyls in the BALB/ role of xenobiotic metabolizing cytochrome P450s. PLoS One cCrgl mouse. Environ Health Perspect 2005;113:1022–6. 2011;6:e17757. 87. Willingham E, Crews D. The red-eared slider turtle: An animal model 67. Guillette Jr LJ, Edwards TM. Environmental influences on fertility: can for the study of low-doses and mixtures. Amer Zool 2000;40:421–8. we learn lessons from studies of wildlife? Fertil Steril 2008;89:e21–4. 88. Bergeron JM, Willingham E, Osborn CT, Rhen T, Crews D. Devel- 68. McEwen BS. Brain on stress: how the social environment gets under opmental synergism of steroidal estrogens in sex determination. the skin. Proc Natl Acad Sci 2012;109(Suppl. 2):17180–5. Environ Health Perspect 1999;107:93–7. 69. Romeo RD. Adolescence: a central event in shaping stress reactivity. 89. Jeng YJ, Kochukov M, Watson CS. Combinations of physiologic Dev Psychobiol 2010;52:244–53. estrogens with xenoestrogens alter calcium and kinase responses, 70. McCormick CM, Mathews IZ, Thomas C, Waters P. Investigations of prolactin release, and membrane estrogen receptor trafficking in rat HPA function and the enduring consequences of stressors in adoles- pituitary cells. Environmental Health 2010;9:61. cence in animal models. Brain Cogn 2010;72:73–85. 90. Larsson K, Sodersten P, Beyer C, Morali G, Perez-Palacios G. 71. Mani SK, Lara H, Romero C, Handa RJ, Wu TJ. Advances in neuro- Effects of estrone, estradiol and estriol combined with dihydrotes- endocrine mechanisms. Horm Metab Res 2012;44:561–2. tosterone on mounting and lordosis behavior in castrated male rats. 72. Crews D, Gillette R, Scarpino SV, Manikkam M, Savenkova MI, Horm Behav 1976;7:379–90. Skinner MK. Epigenetic transgenerational inheritance of altered 91. Watson CS, Jeng YJ, Kochukov MY. Nongenomic actions of estra- stress responses. Proc Natl Acad Sci 2012;109:9143–8. diol compared with estrone and estriol in pituitary tumor cell signal- 73. Schecter A, Pavuk M, Amirova DA, Grosheva EI, Papke Ryan JJ, ing and proliferation. FASEB J 2008;22:3328–36. Adibi J, et al. Characterization of dioxin exposure in firefighters, resi- 92. Wynne-Edwards KE, Edwards HE, Hancock TM. The human fetus dents, and chemical workers in the Irkutsk Region of Russian Sibe- preferentially secretes corticosterone, rather than cortisol, in response ria. Chemosphere 2002;47:147–56. to intra-partum stressors. PLoS One 2013;8:e63684. 74. Schecter A, Tiernan T. Occupational exposure to polychlorinated 93. Gaylord DW, Sheenan DM, Young JF, Mattison DR. The threshold dioxins, polychlorinated furans, polychlorinated biphenyls, and dose question in teratogenesis. Teratology 1988;38:389–91. biphenylenes after an electrical panel and transformer accident in an 94. Sheehan DM, Willingham EJ, Bergeron JM, Osborn CT, Crews D. No office building in Binghamton. NY. Environ Health Perspect 1985;60: threshold dose for estradiol-induced sex reversal of turtle embryos: 305–13. How little is too much? Environ Health Perspect 1999;107:155–9. 75. Schecter A, Pavuk M, Papke O, Ryan JJ, Birnbaum L, Rosen R. Poly- 95. Vandenberg LN, Maffini MV, Sonnenschein C, Rubin BS, Soto AM. brominated diphenyl ethers (PBDEs) in U.S. mothers’ milk. Environ Bisphenol-A and the great divide: a review of controversies in the Health Perspect 2003;111:1723–9. field of endocrine disruption. Endocr Rev 2009;30:75–95. 76. Schecter A, Colacino J, Sjodin A, Needham L, Birnbaum L. Parti- 96. Dietrich DR, Aulock SV, Marquardt H, Blaauboer B, Dekant W, tioning of polybrominated diphenyl ethers (PBDEs) in serum and Kehrer J, et al. Scientifically unfounded precaution drives European milk from the same mothers. Chemosphere 2010;78:1279–84. Commission’s recommendations on EDC regulation, while defying 77. McLachlan JA. Environmental signaling: what embryos and evolu- common sense, well-established science and risk assessment prin- tion teach us about endocrine disrupting chemicals. Endocrine Rev ciples. Chemico-Biological Inter 2013;205:A1–5. 2001;22:319–41. 97. Bergman A, Andersson AM, Becher G, van den Berg M, Blumberg 78. Guillette Jr LJ. Endocrine disrupting contaminants–beyond the B, Bjerregaard P, et al. Science and policy on endocrine disrupters dogma. Environ Health Perspect 2006;114(Suppl. 1):9–12. must not be mixed:a reply to a “common sense” intervention by toxi- 79. Crews D, Willingham E, Skipper JK. Endocrine disruptors: Present cology journal editors. Environ Health 2013;12:69. issues, future directions. Quart Rev Biol 2000;75:243–60. 98. Grandjean P, Ozonoff D. Transparency and translation of science in 80. Soto AM, Rubin BS, Sonnenschein C. Endocrine disruption and the a modern world. Environ Health 2013;12:70. female. In: Gore AC, editor. Endocrine-Disrupting Chemicals: From 99. Bourguignon JP, Soder O. Endocrine Disruptors: A Most Relevant Basic Research to Clinical Practice. Totowa, NJ: Humana Press; 2007. Issue for the Pediatric Endocrinologist. Horm Res Paediatr 2013. 81. Korach KS, Sarver P, Chae K, McLachlan JA, McKinney JD. Estrogen http://dx.doi.org/10.1159/000355610. receptor-binding activity of polychlorinated hydroxybiphenyls: Con- 100. Gore AC, Balthazart J, Bikle D, Carpenter DO, Crews D, Czernichow formationally restricted structural probes. Mol Pharm1988;33:120–6. P, et al. Policy decisions on endocrine disruptors should be based on 82. Berthoud H-R. Synergy: a concept in search of a definition. Endocri- science across disciplines:a response to Dietrich et al. Endocrinology nology 2013;154:3974–7. 2013;154:2987–8. 83. Sonnenschein C, Soto AM. An updated review of environmental 101. Gore AC. An international riposte to naysayers of endocrine-disrupting estrogen and androgen mimics and antagonists. J Steroid Biochem chemicals. Endocrinology 2013;154:3955–6. Mol Biol 1998;65:143–50. 102. Wolstenholme JT, Edwards M, Shetty SR, Gatewood JD, Taylor JA, 84. Bergeron JM, Crews D, McLachlan JA. PCBs as environmental Rissman EF, et al. Gestational exposure to bisphenol A produces estrogens: Turtle sex determination as a biomarker of environmental transgenerational changes in behaviors and gene expression. Endo- contamination. Environ Health Perspect 1994;102:780–1. crinology 2012;153:3828–38. 390 SECTION | XI Controversies and Debate of Generational Epigenetic Inheritance

103. Wolstenholme JT, Taylor JA, Shetty SR, Edwards M, Connelly JJ, 118. Uphouse L. Effects of chlordecone on neuroendocrine function of Rissman EF. Gestational exposure to low dose bisphenol A alters female rats. Neurotoxicol 1985;6:191–210. social behavior in juvenile mice. PloS One 2011;6:e25448. 119. Warren DW, Ahmad N, Rudeen PK. The effects of fetal exposure to 104. Watson CS, Alyea RA, Cunningham KA, Jeng YJ. Estrogens of mul- 1,2-dibromo-3-chloropropane on adult male reproductive function. tiple classes and their role in mental health disease mechanisms. Inter Biol Reprod 1988;39:707–16. J Women’s Health 2010;2:153–66. 120. Goldman JM, Stoker TE, Perreault SD, Cooper RL, Crider MA. 105. Patrick J, Challis J, Natale R, Richardson B. Circadian rhythms in Influence of the formamidine pesticide chlordimeform on ovulation maternal plasma cortisol, estrone, estradiol, and estriol at 34 to 35 in the female hamster:Dissociable shifts in the luteinizing hormone weeks’ gestation. Amer J Obstet Gyn 1979;135:791–8. surge and oocyte release. Toxicol Appl Pharm 1993;121:279–90. 106. Young LJ, Crews D. Comparative neuroendocrinology of steroid 121. Kelce WR, Monosson E, Gamcsik MP, Laws SC, Gray Jr LE. Envi- receptor gene expression and regulation:relationship to physiology ronmental hormone disruptors:evidence that vinclozolin develop- and behavior. Trends Endocrin Metab 1995;6:317–23. mental toxicity is mediated by antiandrogenic metabolites. Toxicol 107. Zala SM, Penn DJ. Abnormal behaviours induced by chemical Appl Pharm 1994;126:276–85. pollution:A review of the evidence and new challenges. Anim Behav 122. Gray Jr LE, Ostby JS, Kelce WR. Developmental effects of an envi- 2004;68:649–64. ronmental antiandrogen:the fungicide vinclozolin alters sex differen- 108. Willingham E, Crews D. Sex reversal effects of environmentally rele- tiation of the male rat. Toxicol Appl Pharm 1994;129:46–52. vant xenobiotic concentrations in the red-eared slider turtle, a species 123. Murakami M, Hosokawa S, Yamada T, Harakawa M, Ito M, Koyama with temperature-dependent sex determination. Gen Comp Endocrin Y, et al. Species-specific mechanism in rat Leydig cell tumorigenesis 1999;113:429–35. by procymidone. Toxicol Appl Pharm 1995;131:244–52. 109. Podreka S, Georges A, Maher B, Limpus C. The environmental con- 124. Sodersten P. Sexual differentiation:do males differ from females in taminant DDE fails to influence the outcome of sexual differentia- behavioral sensitivity to gonadal hormones? Progr Brain Res 1984; tion in the marine turtle Chelonia mydas. Environ Health Perspect 61:257–70. 1998;106:185–8. 125. Bigsby R, Chapin RE, Daston GP, Davis BJ, Gorski J, Gray LE, 110. Clark EJ, Norris DO, Jones RE. Interactions of gonadal steroids and et al. Evaluating the effects of endocrine disruptors on endo- pesticides (DDT, DDE) on gonaduct growth in larval tiger salaman- crine function during development. Environ Hlth Persp 1999; ders, Ambystoma tigrinum. Gen Comp Endocr 1998;109:94–105. 107(Suppl. 4):613–8. 111. Gray Jr LE, Ostby J, Ferrell J, Rehnberg G, Linder R, Cooper R, 126. Guerrero-Bosagna C, Sabat P, Valladares L. Environmental signal- et al. A dose-response analysis of methoxychlor-induced alterations ing and evolutionary change: can exposure of pregnant mammals to of reproductive development and function in the rat. Fund Appl environmental estrogens lead to epigenetically induced evolutionary Toxicol 1989;12:92–108. changes in embryos? Evol Devel 2005;7:341–50. 112. Gray Jr LE, Ferrell JM, Ostby JS. Alteration of behavioral sex dif- 127. Guerrero-Bosagna C, Valladares L. Endocrine disruptors, epigeneti- ferentiation by exposure to estrogenic compounds during a critical cally induced changes, and transgenerational transmission of charac- neonatal period:effects of zearalenone, methoxychlor, and estradiol ters and epigenetic states. In: Gore AC, editor. Endocrine-Disrupting in hamsters. Toxicol Appl Pharm 1985;80:127–36. Chemicals: From Basic Research to Clinical Practice. Totowa, NJ: 113. Gray Jr LE. Neonatal chlordecone exposure alters behavioral sex Humana Press; 2007. differentiation in female hamsters. Neurotoxicol 1982;3:67–79. 128. Howdeshell K, Hotchkiss A, Thayer K, Vandenbergh J, vom 114. Brown HE, Salamanca S, Stewart G, Uphouse L. Chlordecone Saal F. Exposure to bisphenol A advances puberty. Nature (Kepone) on the night of proestrus inhibits female sexual behavior in 1999;401:763–4. CDF-344 rats. Toxicol Appl Pharm 1991;110:97–106. 129. Clark MM, Crews D, Galef Jr BG. Concentrations of sex steroid 115. Uphouse L, Eckols K, Sierra V, Kolodziej M, Brown H. Failure of hormones in pregnant and fetal Mongolian gerbils. Physiol Behav chlordecone (Kepone) to induce behavioral estrus in adult ovariecto- 1991;49:239–43. mized rats. Neurotoxicol 1986;7:127–41. 130. Clark MM, Crews D, Galef Jr BG. Androgen mediated effects of 116. Gellert RJ. Kepone, mirex, dieldrin, and aldrin:estrogenic activity male fetuses on the behavior of dams late in pregnancy. Develop and the induction of persistent vaginal estrus and anovulation in rats Psychobiol 1993;26:25–35. following neonatal treatment. Environ Res 1978;16:131–8. 131. Freund J, Brandmaier AM, Lewejohann L, Kirste I, Kritzler M, 117. Ohtake F, Takeyama K, Matsumoto T, Kitagawa H, Yamamoto Y, Kruger A, et al. Emergence of individuality in genetically identical Nohara K, et al. Modulation of oestrogen receptor signalling by mice. Science 2013;340:756–9. association with the activated dioxin receptor. Nature 2003;423: 132. Crews D, McLachlan JA. Epigenetics, evolution, endocrine disrup- 545–50. tion, health, and disease. Endocrinology 2006;147:S4–10. Glossary

Canalization Developmental process proposed by Waddington by Epigenotype A term coined by Waddington to suggest a process by which the cell, tissue, organ, or organism initially is subject to which a cell may differentiate as a consequence of changes in the internal and external forces but, through time, less subject to influ- internal and external environments; a potential for change that is ences from these forces, ultimately leading to reduced variability heritable and the basic genetic program of that cell. Refers to a in naturally occurring species. stable pattern of gene expression that is outside the actual base Context-dependent epigenetic modification Epigenetic modifica- pair sequence of DNA. tion caused by experiences during sensitive periods of life (prena- Epimutations Differential presence of epigenetic marks that lead to tal, early postnatal, adolescence) that modify the biological system altered genome activity. but that are not intrinsically heritable in the absence of continued Epiphenotype A concept developed by E.O. Wiley as a means to exposure. Such experiences can include passive or active exposures evaluate the total or essential characters of an individual at any and have long-lasting effects on the individual’s subsequent life point in time in its life history. history. These modifications can be perpetuated across generations Epistasis Traits that depend upon more than one gene, each interact- due to persistence of the causal environmental factor such that each ing with one another. generation is exposed to the same conditions. F0, F1, F2, F3 Various generations (parent, offspring, grand-offspring, CpG island A genomic region that contains a high frequency of cy- great-grand-offspring). tosine and guanine base pairs together with a phosphodiester bond. Genetic assimilation A concept attributed to Waddington in which De novo methylation Addition of methyl groups to new positions phenotypes that are environmentally malleable become genetical- on a DNA molecule. ly fixed and no longer require the original environmental stimulus Endocrine-disrupting chemical (EDC) Anthropogenic chemical(s) in order to be manifested. that interfere(s) with any aspect of hormone action. Germline-dependent epigenetic modification Epigenetic modi- Epialleles A group of otherwise identical genes that differ in the de- fication arising from exposure to an environmental cue during gree of methylation and produce novel phenotypes that are herit- germline development. By being incorporated into the germline, able across generations. the effect is manifested each generation, even in the absence of Epiallelic variants Epigenetically distinct versions of a single allele. continued exposure to the causative agent. Epigenetic transgenerational inheritance Germline-mediated in- Homeorhesis Stability of the process of development itself. heritance of epigenetic information between generations in the ab- Homeostasis Stability of a final steady state. sence of direct environmental influences that leads to phenotypic variation.

391