SEARCH

SEARCH BY CITATION

Keywords:

  • development;
  • genes;
  • environment;
  • epigenetics

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Modes of Epigenetic Transmission
  5. Summary and Future Implications
  6. References

Objective

To depict the processes through which animals and human beings engage their environment in continuously evolving states of conflict and cooperation.

Method

Descriptive literature review.

Results

Life history outcomes are more relative than they are absolute. Genetic variations play a crucial role, but heavily influencing behavioral outcomes, psychopathology included, are external cues that epigenetically remodel DNA along experience-dependent signaling pathways. The result is phenotypes that either optimize adjustment, or constrain it.

Discussion

Knowledge of epigenetic mechanisms may help shed new light on the origin of maturational phenotypes underlying eating disorders and why adjusting treatments to these realities warrants our attention. © 2014 Wiley Periodicals, Inc. (Int J Eat Disord 2014; 47:696–704)


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Modes of Epigenetic Transmission
  5. Summary and Future Implications
  6. References

Conrad Waddington invoked the terms “epiphenotype” and “epigenetics” more than 70 years ago in a paper discussing the relationship between genes and their phenotypes.[1] For Waddington, the phenotype—how, exactly, does it reach a mature identity?—was the greater intrigue; yet the question was largely ignored by classical genetics. Curious how, and why, a single embryo was capable of evolving different morphologies depending on its physical environment,[2-4] he rejected the conventional wisdom that gene and phenotype were implicitly correlated. An organism, he argued, could not achieve this level of versatility based solely on genes conserved by natural selection; fast acting physio-chemical adaptations by which cells became committed to specific functional states had to be operating. His point was that a deeper knowledge of the phenotype would ultimately perfect knowledge of the genotype. Waddington was not the first to posit that a complex set of processes guided development. Aristotle famously argued much the same in rejecting the doctrine of preformism—the idea that a life's ultimate identity was predelineated in advance.[5] Returning to Waddington, he referred to the mechanistic processes intervening between gene and its expression the epiphenotype, the formal scientific investigation of these mechanisms, epigenetics.

In today's lexicon, epigenetics is the study of heritable alterations in gene processes without change to the sequence of nucleotides that defines the DNA molecule (excellent summaries of the processes involved can be found elsewhere[6, 7]). In brief, the mechanisms involve editing the protein complexes formed when DNA coils around histones to create a structure (referred to as chromatin) compact enough to fit securely into the cell's nucleus. Chromatin serves a number of crucially important biological functions, among which are DNA replication and repair, and modulation of a gene's expression—activating, or increasing, the gene's biochemical material, or repressing (silencing) the gene's action. Activation occurs when the histone bonds loosen, allowing enzymes greater access to DNA for gene transcription. Silencing of a gene's expression occurs mainly when a methyl group attaches to one of the DNA nucleotides. Importantly, these chemical events are evolutionarily conserved mechanisms by which the organism responds to environmental signals that convey developmentally relevant information; the collective “memory” of these induced modifications to DNA's chromatin structure is referred to as the epigenome. Because epigenetic memories are mitotically stable and can last through cell divisions over a lifetime, they are potentially heritable across generations; thus some of the signaled phenotypes prove long-lasting; on the other hand, others shift rapidly in precise coordination with environmental change to optimize adaptation as new demands surface. Hence there are trade-offs as to how epigenetic marks unfold over time: a current phenotype might adapt well to current demands, but if the gene's product is inflexible, and depending on the adjustments required for fitness in the new environment, future development will be constrained. Discussing life history trade-offs, Godfrey et al.[8] write:

In this latter scenario, when a change in course is needed because the future environment proves vastly different from the predicted one, adaptive change is less feasible. Described by Gluckman et al.[9] as the consequence of environmental mismatch—disparities between the prenatal and rearing environment and those encountered postnatally—gene expressions that induce a phenotype optimally suited for survival in the early offending environment (and which is “predicted” to continue into the future) are maladaptive when the future is not the predicted one (pg. 6).

Within epigenetics, and in the broader perspective of developmental science, the concept of “mismatch” is a level of explanation critical to understanding the evolution of wellness versus disease: as reflected in the quote, whether a “selected” trait is advantageous or not will depend, at least to some degree, on the future environment's context. Barker[10] made the point explicitly in speculating on the adaptations necessary for fetal survival in a gestational environment limited in nutrients. The solution, he posited, was a “thrifty biology”—a set of induced metabolic adaptations that would better tolerate a necessary rationing of energy provisions. From the standpoint of survivability, frugal management of resources made good sense; just the same, the legacy carried risk of later metabolic pathology (obesity and diabetes) if the fetally stressed adult later encountered environments rich in nutrients and promoting a more sedentary lifestyle. This mismatch, how the epigenetic marks of experience arising in one generation set pathological conditions for the next, is now a central theme in both medicine and psychopathology.

Epigenetic inheritance does not replace classical mechanisms of gene transmission as a causal explanation of disease predisposition; rather, it adds to it by speaking to an obvious biological reality: genome variation certainly influences the broad contours of ontogeny, but environmental conditions influence how genome and brain operate integratively and adaptationally from the moment of conception, shaping different life history outcomes in the process. With a steady stream of evidence showing genetic variation alone appears far from realistic as a causal explanation of psychiatric illness, the role of epigenetic transmission has steadily gained favor in models of developmental risk and resilience.[11] Accordingly, our general aim in contributing to this Special Issue is to illustrate how these mechanisms are potentially important to theorizing about phenotype maldevelopment in eating disorders (ED), and the implications they hold for intervention. Aspects of the discussion will be speculative—epigenetic research in EDs is in its infancy,[12-20] although findings to date are generally consistent with the notion of stress-induced phenotypes, in which certain aspects of psychopathology correlate with more pronounced epigenetic marks. Said differently, phenotypes that foreshadow EDs and which are intrindic to its psychopathology bear characteristics similar those that ensue from stress-induced epigenetic programming.[21] We first offer a concise summary of basic concepts, then summarize findings we believe introduce novel premises for our field to consider.

Modes of Epigenetic Transmission

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Modes of Epigenetic Transmission
  5. Summary and Future Implications
  6. References

Interested readers are directed elsewhere for a more detailed discussion.[11] In brief, epigenetic marks and their induced phenotypes can sustain well into adult life via three basic modes of cross-generation transfer. In fetal programming, the sequence of events runs from environmental conditions during pregnancy, their impact on gene expression, and alteration of fetal somatic cells from which unique behavior outcomes arise. In this transfer, the mother is the host environment, but the imposing events are transmitted to the fetus mostly via the placenta. A second mode of epigenetic information transfer across generations is parent-offspring interaction (thus far studied mainly in rodent mothers and their female offspring): in this transmission, parents interact with offspring, epigenetic marks are imprinted on the offspring's' DNA, a phenotype similar to the parent's is induced, and in time, offspring, now expressing the same phenotype, transmit it to their offspring in much the same way their parents transmitted it to them. The third mode of transmission is the more classic one, occurring when environmental effects register in the germ line and then transfer across generations via sexual reproduction.

Epigenetics has been a watershed in the study of maldevelopment because the mechanisms are amenable to experimental study at unprecedented levels of sophistication. Not only have the associations proven statistically strong, the risk profile resulting from life exposures, adverse life events in particular, is surprisingly broad—beyond psychopathology, including cancer, coronary heart disease, obesity, and diabetes.[22] As Waddington argued, to “know” what the early environment imposed on development is to know something about the pathways that impair it. The generalization may seem extreme in the light of all that remains unknown about development, but the weight of evidence points squarely to the transmissibility of the environment's presence.

Gestational Events

One of us (MS) has reported that the lifetime occurrence of anxiety disorders is elevated in first-degree relatives of adolescents with AN.[23] Apart from the immediate implication—anxiety promoting genes may be a substrate for AN risk—the implications may be broader. We are referring to a considerable literature linking adverse influences on maternal emotional state during pregnancy (by inference, impacting the maternal epigenome) to a host of pre/peri/postnatal offspring outcomes.[6, 24-31] They include: over-reactivity of the hypothalamic-pituitary-adrenal axis (HPA); reduced gestational age, lower birth weight, and birth complications; a fearful, low novelty seeking temperament; hyperactivity; reduced catecholamine and serotonin levels; greater right versus left frontal EEG activation (associated with anxious-phobic-inhibited behavioral tendencies[32]); structural change in brain regions implicated in multiple neurocognitive functions; a preference in adulthood for a less flexible, striatal-based habitized learning strategy versus the more versatile, holistic learning style that depends on integrity of hippocampal systems;[33] and in females, poorer quality of maternal care administered to their own offspring. Clearly, maternal stress leaves its mark on offspring biology and behavior. While genetic inheritance can not be ruled out as an explanation, as we note below some of the phenotypes are reversible through environmental manipulations.

The relevance of the described phenotypes—restraint and avoidance in the presence of novel, potentially rewarding environments, ease of arousal under conditions of perceived threat—to the vulnerability substrates in EDs is apparent; but a broader life history implication is suggested. Discussed in Worthman and Kuzawa,[34] fetal stress exposure “selects” for anxiety phenotypes by sensitizing the developing state of limbic circuitry, reducing its threshold for activation.[26, 27] And, as noted, it can also inhibit fetal growth, but as a means of conserving energy allocations. Interestingly, the sensitivity of this metabolic adaptation is thought to be greater in the female embryo.[25] In effect, the “survival phenotype,” the “thrifty” biology, is frequently an anxious phenotype as well. The premise is that in broad terms, females have a finely tuned system for calibrating (and down-regulating) energy needs in the face of stress that males lack. Genotypic variations likely play a role in just how sensitive, how operational, this sexually dimorphic adaptation functions, but the implication is that some females harbor a risk of down-regulating too far, others a risk of regulating insufficiently.

Here, the developmental ecology of anxious children comes into play as another level of explanation; specifically, the effects on anxious-prone children of stress in the postnatal environment. We will address rearing effects in a following section, but for the moment it suffices to say that insofar as anxious children and anxious parents share not only a common inherited risk but also the same living space, genomes, biology, and environment inevitably coevolve. So if we grant that affected parents and children express equally hypersensitive limbic circuitry, a parsimonious assumption is that they are also bound to common worries and similarly exaggerated expectancies; that within their living space, vigilance, and avoidance are, for each, the neurodynamic prerequisites for “safety.” The point is, that in addition to genetic and epigenetic patho-biology, secondary mental structures and attitudes, rather than epiphenomena, are equally important evolving elements of the psychopathology: What the child “sees”—what they believe, think, and “feel”—their attributions, the ingrained cognitive, attitudinal, and perceptual schemas—each become elements in a larger causal chain of events needed to fully understand the whole; why each must be taken into account in explaining enduring illness in some, recovery in others. These are the question asked repeatedly of AN: Why do some patients strike up an unusually persuasive allegiance to their skeleton presence? Why do they attribute “value and necessity” to what others rightly see as abhorrent and nonsensical? Why do they adopt a self-referential psychological identity (“I do not want to recover”) around behaviors that, in some cases, arise from early life adversity, but in general, from biological processes they have little grasp of? Why do they experience the discipline of low weight and inhibition of need as “grounding,” as protective? And beyond these questions is it not conceivable that epigenetic influences on neural biology from prolonged dietary restriction is yet another factor propagating behavioral and mental symptoms?[35-37] Clearly, and as the following sections will further suggest, epigenomic concepts point to a panoramic formulation for what we universally regard as complex illnesses.

Maternal Birth Weight and Prenatal Diet

As far as we know, maternal birth weight has not received consideration as a trait legacy that might influence risk toward EDs. Assuming it to be a proxy for fetal adaptation (granted, a theoretical supposition), can a newborn's birth weight arise from a remote ancestor whose legacy is propagated across multiple generations? Kuzawa[38] described this as a “phenotypic inertia.” Some support is found in research by Ramakrishna et al.[39] that tracked birth weight correlations across generations. It showed that the best predictor of a child's birth weight was the mother's, even when her gestational age and adult height were taken into account; in effect, an early gestational environment appeared to program the intrauterine nutritional environment for offspring birthed over successive generations. Whether this observation has relevance to either AN or BN is unknown and criticisms can be anticipated—what evidence is there of nutritional deprivation or excesses in the remote, intergenerational history of persons with EDs? Little, but as Kuzawa notes, the intergenerational correlations were maintained across a range of fetal birth weights. In other words, they are not solely the consequence of “extreme” exposures—either excess maternal birth weight, or its opposite; rather, the programming machinery appears remarkably sensitive to variation in fetal ecological conditions. Clearly, persons with AN have a metabolic capacity for weight loss that is lacking among those with BN, in whom higher premorbid weight and familial obesity is more likely.[40] Yet the common wish upon clinical presentation is to lose weight. An argument can thus be made that these contrasting developmental patterns reflect, to some degree, the operation of different epigenetically transmitted weight legacies, from which different clinical trajectories ensue.

On this point, while no firm basis exists for assuming prenatal diet insufficiencies are a normative fetal ecology in EDs, their epigenetic imprints have been studied. Research on the Dutch Hunger Winter,[41] the Nazi imposed famine in the Netherlands, has linked periconceptual exposure to reduced methylation of the insulin-like growth factor 2 gene (IGF2),[42] which is involved in cellular differentiation and, interestingly, is also expressed during fear extinction.[43] How epigenetic remodeling of these molecular processes might impact the development of systems that regulate emotional memory or fear persistence is unknown; but worth noting are postmortem studies that show a correspondingly positive relationship between brain weight (in males) and DNA methylation at IGF2, and a negative association between brain weight and schizophenia.[44, 45] Again, the point is the important role of social factors that mediate in utero diet, which in turn, and depending on the dietary components involved, can influence epigenetic mechanisms that differentially program disease risk or resilience for a variety of physical and mental health profiles.

Early Life Exposure to Toxicants

The same subtlety is seen. We live in a world of toxicants, a number of which are known to disrupt endocrine function, and to confer risk of anxious and depressive behavior in children, particulary females, even when the in utero dose of exposure is low.[46] Noting this, Kundakovic et al.[47] showed in mice that these negative effects were largely attenuated if the effected mice received high levels of maternal care postnatally. The role of toxicant exposure in psychopathology has been little studied.

Natural Disasters

Again, rarely do they impose on the life history of persons with EDs, but they, too, are another window into the imprinting of stress on developmental outcomes. “Project Ice Storm” assessed women who were pregnant, or became pregnant, shortly after a series of devastating ice storms that hit southern Quebec in January 1998. Following the storm, King et al.[48] assembled a cohort of mothers whose pregnancies were impacted. Thus far, findings from Project Ice Storm offer substantial evidence of multiple deleterious effects: shorter gestation length, lower birth weight, dermatoglyphic asymmetry, slower cognitive and language development, and obesity. Intriguingly, a very recent assessment of “Ice Storm children,” at 13 years of age, revealed increased risk of maladaptive eating behaviors, especially those in the bulimia spectrum.[49] The finding is intriguing in the light of evidence linking indices of gestational stress to increased risk of later EDs.[50, 51]

Postnatal Rearing

Interest in the impact of postnatal emotional care on developmental outcomes dates back to seminal investigations of “critical periods” and the role of favorable early experiences in attenuating stress reactivity.[52, 53] Since then, a large literature has accumulated documenting how early life parental care induces offspring phenotypes, some of which are sustained throughout development via modulation of brain systems that regulate anxiety, fear, and reward motivation across a web of anatomical regions sensitive to epigenetic programming.[6, 7, 54-59] Not only do the perturbations register in somatic cells, they can be inherited through the germ line, and they need not be extreme to leave a mark on gene expression.

In brief summary, the chain of events runs from: (a), quality of the postpartum care mother delivers to offspring, (b) to change in offspring hormone activity, (c) to DNA remodeling, (d) the induced phenotype. In rodent studies, care is measured as the amount of licking and grooming (LG) the newborn pup receives. When high, glucocorticoid receptor (GR) expression in the offspring hippocampus is increased, which in turn, assures: effective negative feedback suppression of stress hormone release from the hypothalamus and pituitary; increased expression of oxytocin (OT) and estrogen (ER) receptors, sensitivity to which primes the future mother's postpartum investment in offspring care; and increased hippocampal neuronal survival, dendrite formation, and synaptogenesis, which in turn, enhance cognitive performance and mediate stress resilience. When LG is low, GR, OT, and ER are reduced, with correspondingly greater stress arousal, neophobia, and avoidance motivation (and in females, reduced future care giving). As to the epigenetic marks involved, high maternal LG signals reduced methylation (less repression) of the promoter region of the GR and ER genes, thereby increasing their receptor densities. When LG is low, there is greater methylation (greater silencing) of these promoter regions. In short, low maternal care during the early life environment leaves an imprint that induces potentially long-lasting anxious, low reward seeking, low maternal care offspring behavior. Importantly, the fact that anxiety aggregates in AN families gains new significance in this light: the possibility that anxiety proneness ramifies across development via multiple biological and experiential pathways.

Supporting this idea is evidence that some between-generation transmissions via maternal rearing are nongenomic. If a pup birthed by a low LG mother is crossfostered by a mother whose LG is high, the offspring's trajectory reflects the rearing mother's phenotype. Likewise, offspring birthed by a high LG mother but raised by one with low LG exhibit characteristics indistinguishable from pups who receive low LG care from birth. What this work eloquently shows is not only is the operative transmitting factor valence (quality) of the rearing environment, the negative effects of poor rearing are remediable.

As for stress that impinges later in life, Champagne[58] notes that the imprints are equally persuasive, and they, too, are reversible, sometimes in ways maladaptive. Specifically, prolonging maternal-female offspring contact by extending the weaning period, or exposing female offspring to postweaning social isolation, lowers offspring OT and ER expression levels and reduces the quality of care they later provide to their own offspring. Analogously, enriching the postweaning social environment of offspring who received low LG at birth reverses the negative effects of this postpartum exposure. Maternal care behavior has also been shown to vary by the social context of weaning: contact with other lactating females can enrich care, whereas it can change in sex-specific fashion in reaction to gender composition of the litter.[60] Finally, it has been shown that phenotypes mediated by early life maternal care can be transformed entirely during adulthood by pharmacologic and dietary modulators of DNA methylation and demythylation that effectively remove initial epigenetic marks.[61] Specifically, after receiving these modulators, stress resilient adult rats who were reared by high LG mothers revert to fearful behavior, whereas high anxious rats who were reared by low LG mothers suddenly show reduced anxiety behaviors. The implications—considering in our clinical work the rearing impact of opposite-sex siblings, early life and current intrafamilial strains, the future potential of novel pharmacotherapies—warrant thought.

Paternal Effects and the Role of Germ Line Inheritance

Fathers matter. While their impacts are not easily separated from the effects just described, recent evidence supports the conceivability that paternal phenotype variation as well as exposure to diet insuffiencies and toxicants can influence offspring outcomes through germ line inheritance. Alter et al.[62] differentiated virgin, genetically identical male mice from an inbred strain based on open field activity and then mated them with females from the same strain. In spite of having no postnatal contact with their offspring, fathers with higher levels of open-field activity (a proxy for less anxiety) had daughters, but not sons, with the same activity phenotype; the high activity paternal phenotype also correlated with higher brain and hippocampal weight in offspring of both sexes. In Franklin et al.,[63, 64] male mice exposed to intermittent, unpredictable maternal separation and then bred over several generations were shown to transmit increased stress responsivity and serotonin release in the frontal cortex from a generation in which neither offspring nor their inherited sperm were stress exposed. Lambrot et al.[65] fed male mice either a folate-sufficient or folate-deficient diet (folate being a methyl donor for DNA methylation) and observed differential methylation in the sperm epigenome and negative birth outcomes in offspring.

Beyond effects passed through the germ line is the further possibility that a paternal phenotype can differentially effect how a mother invests in her offspring. In other words, male behavior impacts the ecology of maternal-child care. Mashoodh et al.[66] exposed genetically identical males to one of two starkly different postweaning inducible environments—social isolation, versus enrichment. Females whose offspring were the product of mating with males exposed to early life social enrichment spent significantly greater time in LG of their offspring; in other words, a paternal early life imprint on maternal care giving. The unknown here is whether the effect is mediated by genetic traits passed from father to offspring, which calibrate aspects of prenatal and postnatal maternal care via growth factors secreted from the placenta, or by variations in postpartum behavior displays (e.g., offspring frequency of vocalizations, activity level, sucking strength).[58-60] Nonetheless, the findings are consistent with predictions from evolutionary biology that females will allocate more investments toward offspring sired by a desired male. Here again is yet another example of how complexly interactive the mechanistic pathways leading to vulnerability can be, and how little they have been explored up to now in EDs.

Summary and Future Implications

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Modes of Epigenetic Transmission
  5. Summary and Future Implications
  6. References

If there is one overarching theme in epigenetics, it is that everything about development matters; it is the principle “fact of life”: brain and genome continuously sift through and assimilate information from the environment, from events seemingly mundane to ones intense. How these experiences are delivered and received shape and reshape the functionality of our biology, and thus our mental life. In short, life history evolution is inevitably a mix of both probabilistic and strategic events[8]: gene circuits “feel” these events, then exploit—“take advantage” if you will—the information flow in shaping ontogeny, in much the same way that genotypes—the base pair sequence of DNA—influence to greater or lesser degree how plastic, or constrained, a life history structure will be. Ultimately, crucial to the overall calculus is how these multiple, ever changing gene-environment exchanges modulate development from gestation onward, and whether, over time, the adaptations fit well, or poorly, to the realities of future demands. Indeed, as robust as epigenetic mechanisms in ontogeny are proving to be, some now argue that much of causality in psychopathology is ulitimately a “relational” causality.[67]

Stress is the central theme of epigenetics. As it pertains to EDs, we know it is a common precedent of binge eating; we see it often in the development of persons with BN and it has been implicated in the conversion from AN to BN.[68] But its relevance extends just as surely to the phenomenology of EDs. In the anxious child, averse to change, regimented and hypersensitive to error, not only is the threshold for alarm, neurally and psychologically, set low, the signals are ubiquitous. Stress is ever-present, unavoidable: a need to change schools; a change of residence; new siblings; illness in a loved one; arguments between parents, divorce, and alienation; homework assignments and impending exams; estrangement from classmates; a body whose sudden change heralds puberty—stress pervades every turn of development and for persons prone to EDs the effects are more likely to arouse limbic circuits that register alarm and avoidance. How epigenetic mechanisms are involved, even from the point of conception, in evolving these early life phenotypes, how they later bridge to clinical symptoms—how they impair accommodation to a radically changing environment, how acute caloric restriction might also prime, or modulate, anxiogenic circuits and constrain reward motivation—the idea that an early life metabolic pacemaker may, in some, support unusually extreme weight loss—these are novel questions for our field to consider. At least some evidence is suggestive. For example, AN and anxiety have been shown to share genetic variations in common,[69] which accords with associations between childhood generalized anxiety disorder in persons with AN and clinical features suggestive of a more virulent trajectory—lower lifetime BMI and more extreme fasting and exercise.[70-72]

And as for the prospects of refining our psychological interventions, knowing that stress in the family domain can skew development toward fear-based temperaments and learning styles grounded in environmentally bound stimulus-response associations, we would do well to note of research outside of our field showing how optimized treatment models targeting problematic familial environments have proven effective in reducing symptom morbidity in a variety of severe psychopatholgies.[73-75] Regardless of our chosen focus—understanding the phenomenology of EDs, maximizing treatment effects—the challenge is unique: EDs, AN in particular, not only bear the influence of fear and habit systems that are biologically and adaptationally self-organizing, they self-perpetuate in systems of avowed beliefs and attitudes.

The findings we have reviewed inform the challenge and offer new lines of thought for its further study. As to the prospects they hold for intervention, knowledge that epigenetic phenotypes are sensitive to future environments supports the premise that when EDs emerge against a background rife with uncertainty, with unchecked emotion, unpredictable security, the odds are greater that these fear-based, safety rituals will be more entrenched. Just the same, to the degree that similar phenotypes have been shown to be reversible if the environmental tone can be shifted, models of family intervention targeting inter- and intrapersonal stress remediation at the many levels it occurs in family life may yield a similar promise when applied to high-stress ED families. And the same promise holds for developing novel molecular interventions that focus on re-editing early life epigenetic marks. Emerging evidence that post-transcriptional methylation of RNA molecules may be just as critical to the regulation of gene expression as DNA epigenetics offers yet another vista for future biological applications.[76] Here, creative applications of animals models may prove useful.[77, 78]

The authors are grateful to Drs. Kelly Klump and Glenn Waller for comments on the original manuscript.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Modes of Epigenetic Transmission
  5. Summary and Future Implications
  6. References