Introduction

Early life experience has enduring effects. But exactly how does experience come to program certain behaviors? Epigenetics, a level of control "on top of" genetics, offers a mechanism that sits at the interface between nature and nurture. By physically and chemically altering the chromosomal material in response to the environment, epigenetics provides a possible pathway from an event in the outside world to a persistent change in the organism—a way that such diverse stimuli as a mother's love, a father's diet, or drug abuse could stably alter gene expression. Research into the relationship between behavior and epigenetic changes in the brain is just beginning: What causes them? And can they be undone?

Epigenetics also provides an avenue for investigation into neuropsychiatric diseases, where purely genetic approaches may come up short. Epigenetic processes have been implicated in depression, schizophrenia, and suicide; neurodevelopmental disorders including autism and Rubinstein-Taybi syndrome; and neurodegenerative diseases such as Alzheimer's disease. Epigenetic regulation has also recently been found to play a role in learning and memory, where the mechanisms operate in a surprisingly dynamic and rapid manner. Aberrations of epigenetic processes may underlie the cognitive dysfunction common to many disorders.

The 2-day conference on "Behavioral Epigenetics," one of the first in this pioneering field, brought together molecular biologists and behavioral scientists who don't normally share a stage. The event at the University of Massachusetts, Boston, on October 29–30, 2010, was jointly sponsored by the New York Academy of Sciences and The Warren Alpert Medical School of Brown University, and will be the focus of an upcoming Online Meeting Report to be published in the Annals of the New York Academy of Sciences.

Celia Moore, a psychobiologist, professor, and director of the Developmental Sciences Research Center at the University of Massachusetts, Boston, opened the conference by pointing to the demise of the nature-nurture dichotomy and the emergence of a more integrated model of development in which epigenetic mechanisms likely play a central role. Understanding these mechanisms could also open the door to new treatments for mental illness. "There are many reports of [epigenetic] histone deacetylase (HDAC) inhibitors basically curing every neurologic and psychiatric syndrome in animal models that's been studied to date," says Eric Nestler, molecular psychiatrist at the Mount Sinai School of Medicine and co-organizer of the conference. Turning these enzymes into human pharmaceuticals is harder; the related drugs that are in the pipeline for cancer are broad-acting and very toxic, cautions Nestler. "But maybe there are subtypes of these enzymes that can be targeted." Studies have already shown that the epigenetic effects seen in rat pups that have been neglected by their mothers can be reversed with drugs.

Background and perspectives

The word "epigenetics" was coined by Conrad Waddington in the 1940s to refer to genetic aspects of the unfolding of phenotype during development: Cells with identical DNA differentiate into liver and brain cells, muscles and skin cells, and then they stay that way. A liver cell's daughter cell inherits not just a sequence of genes, but also the epigenetic state of being a liver cell.

Epigenetics has since broadened to encompass a number of (sometimes incompatible) ideas. Some scientists define epigenetics as heritable changes in gene function that cannot be explained by changes in gene sequence. The heritability can be between generations of an organism as well as within cell lineages. Mechanisms of chromatin biology that can account for such stable (and sometimes environmentally sensitive) changes in gene function continue to be elucidated, and a common use of "epigenetics" now refers to those mechanisms, whether they are heritable or not.

At this conference, talks focused on DNA methylation and histone modification, a subset of mechanisms that mediate gene expression by changing the structure of chromatin. Every mammalian cell packs about two meters of DNA, tightly wound around histone proteins, into a single, microscopic nucleus. An inaccessible gene won't be expressed. In DNA methylation, methyl groups covalently bind to DNA and repress transcription. In histone modification, the various molecules that attach to the tails of the histone proteins form a complex regulatory code.

An open chromatin structure is accessible for gene transcription; a condensed one is not. Here DNA (in orange) winds around histone proteins to form nucleosomes. Each histone's N-terminal tail can be post-transcriptionally modified. Addition of acetyl (A) groups by histone acetyl transferases (HATs) relaxes the chromatin structure; histone deacetylases (HDACs) condense it. Methylation of the DNA itself by DNA methyltransferases (DNMTs) usually represses transcription.

The extent to which behavior affects and is affected by these epigenetic mechanisms was the subject of the present conference. Alongside extraordinary examples in the animal literature, participants discussed the application of behavioral epigenetics to humans, on whom research is still extremely limited by insufficient access to brain tissue. Whether there is transgenerational inheritance of behavioral traits remains controversial in both humans and animals, but there are very well-developed models of how environmental stimuli can create stable, long-term change within an animal's lifetime.

This stability is what makes epigenetics an attractive candidate to explain the developmental origins of health and disease, one of the inspirations for this conference for co-organizer Barry Lester of The Warren Alpert Medical School of Brown University and Women & Infants Hospital of Rhode Island. Whole lifetimes can be marked by experiences in the womb. For example, low birth-weight babies gestated during the Dutch "Hunger Winter" famine of 1944–1945 had a higher incidence of metabolic and cardiovascular disorders when they grew up, possibly because their development was primed for scarcity. "The concept that the fetus is actually making adaptations through programming to 'prepare' for the postnatal environment in response to signals raises a whole host of fascinating questions," Lester said. "In the case of fetal origins we're primarily talking about undernutrition. But can these effects or similar effects be produced by environmental insults or factors other than undernutrition? What might be the underlying mechanisms? Is there applicability of this kind of model, beyond chronic disease, to behavior?" In fact, low birth weight is associated not only with chronic disease, but also with schizophrenia, depression, and psychological distress; and it is a basic tenet of psychology that early experiences can mark us for life.

The recent rapid growth of epigenetics.

While thousands of studies of epigenetics have been conducted over the last 40 years, the application of epigenetics to the study of behavior is just beginning. Lester reported on a literature search of citations which found only 96 articles to date on behavioral epigenetics. These articles addressed substance use, psychiatric illness, learning/memory, neurodevelopment, parenting, stress, and neurodegenerative disorders. Most were on mice, rats, and humans (though one used snails). The majority of studies analyzed brain samples (about 15 of them human); some studied blood, and a few collected cheek swabs, semen, or other tissues. Lester also broke the literature down by the epigenetic mechanisms that were studied (DNA methylation, histone modification, genomic imprinting, or noncoding RNA) and by the particular genes investigated.

Neuropsychiatry Session and Post-Conference Think Tank

Speakers:
Ted Abel, University of Pennsylvania
Lisa Monteggia, University of Texas Southwestern Medical Center
Ian Maze, The Rockefeller University
David H. Skuse University College London, UK

Highlights

• The attention span of a newborn human is correlated with DNA methylation patterns in the placenta.
• Drugs that affect epigenetic states have affected memory, cognition, anxiety, depression, and addiction in animal models.
• Chronic cocaine use makes mice more vulnerable to stress from social defeat through epigenetic mechanisms.
• The parental origin of the X chromosome may affect expression of some X-linked genes in the brain.

Histone deacetylase inhibitors

Drugs that target epigenetic processes have been effective in treating a remarkable number of animal diseases and are in clinical trials for use against human cancers. A few have already been FDA-approved. The problem with them is their lack of specificity: If you're going to use a drug that changes your genome, it's best to know which genes it's affecting.

Histone deacetylase (HDAC) inhibitors are currently very promising targets for behavioral and psychiatric drug development, said Ted Abel of the University of Pennsylvania. HDACs are diverse—valproic acid, long prescribed for epilepsy and bipolar disorder, turns out to be one of them—and many new inhibitors specific to particular HDACs have been developed just in the last few years.

HDAC inhibitors have prevented neuronal death in flies with the fly version of Huntington disease (Stephan et al., 2001) and have restored learning and memory in a mouse model of neurodegeneration (Guan et al., 2009). They have ameliorated cognitive deficits in mice with the neurodevelopmental disorder Rubinstein-Taybi syndrome (Alarcón et al., 2004). They have reversed aspects of Parkinson's disease in the fly (Kontopoulos et al., 2006) and improved memory in aging mice (Peleg et al., 2010). And in a pilot study on the use of HDAC inhibitors to treat symptoms of autism, valproic acid reduced symptoms of irritability in children (Hollander et al., 2010).

It's hard to know exactly how they're working. Histone deacetylases remove acetyl groups from any protein, not just histones. They affect thousands of proteins and hundreds of genes. So which of them are the ones that are changing behavior? Several of the speakers at this conference are trying to find out.

Future pharmaceuticals

In Ted Abel's lab, research on how histone acetylation modulates memory (some of which was conducted by speaker Marcelo Wood) is leading to potential targets for memory enhancement—the specific genes and gene products that the histone modifications regulate. His goal is to treat the cognitive impairments common to many neurodegenerative and psychiatric disorders, from aging to schizophrenia, that existing therapies do not address.

Lisa Monteggia of University of Texas Southwestern Medical Center is studying a mouse model of Rett syndrome, a syndrome that is one of the leading causes of mental retardation and autistic behavior in girls. Rett syndrome is the result of loss of function mutations in the gene for methyl-CpG-binding-protein-2 (MeCP2), an intimate part of the epigenetic machinery that binds to methylated DNA. Although Rett syndrome is a neurodevelopmental disorder, loss of MeCP2 function in the adult brain leads to functional deficits recapitulating aspects of the Rett syndrome phenotype. Symptoms appear to be triggered by functional deficits at the synaptic level. Monteggia is tracking synaptic function back to DNA methylation by deleting the MeCP2 gene from adult mouse brains. She is studying regulation of synaptic function by MeCP2-linked gene regulatory processes such as DNA methylation and histone deacetylases.

Ian Maze, now at the Rockefeller University, has identified a critical role for epigenetic mechanisms in addiction. Cocaine is thought to hijack the brain's reward system; Maze found that it does so in part by downregulating G9a, an enzyme that methylates histones. Histone methylation is thought to be particularly long-lasting, as is a vulnerability to relapse. But "If I overexpress G9a, which blocks transcriptional induction, I can actually decrease an animal's preference for cocaine," Maze said.

In Eric Nestler's social defeat model of depression, repeatedly bullied mice lose all interest in social interaction. Prozac reverses the effect.

Drug addiction makes both people and mice more vulnerable to depression. For a mouse chronically exposed to cocaine, it takes only eight (instead of ten) episodes of being bullied by a bigger mouse to send it into a funk. In work he did in Eric Nestler's lab at the Mount Sinai School of Medicine, Maze showed that cocaine and social stress lead to depression through the same pathway. He was able to restore mouse resilience, creating anti-depressant-like effects, by intervening downstream of G9a.

Parental contributions to X-linked gene expression

David Skuse of University College London is examining sexually dimorphic risk factors in social communication disorders that lie along a continuum running from autism to much less severe behavioral traits in the general population. People with these disorders have trouble recognizing and responding to the feelings of others, they tend to be inflexible in their behaviors, and they have verbal communication difficulties, among other traits. Across this span, males are more likely to be vulnerable to social communication disorders than females, particularly at the higher end of the verbal IQ range.

One possible source of this behavioral difference is the expression of X-linked genes. Males receive their X chromosome from their mothers; hence all their X-linked traits are derived from maternal genes. Females receive one X chromosome from each parent. X-inactivation silences expression of one X chromosome in each cell, resulting in mosaic expression of the paternal and maternal X-linked genes across tissues. Another layer of regulation is genomic imprinting, a phenomenon in which a particular paternal or maternal allele is silenced by methylation of DNA near the gene.

To examine the role of paternal or maternal X-linked gene expression on social communication disorders, Skuse and his colleagues assessed females with a single X chromosome, a condition known as Turner syndrome. Females who inherited the paternal chromosome were found to be similar to XX females in their score on a social and communication disorder (SCDC) test, as well as on tests that assessed empathy and behavioral inhibition. In contrast, females with a maternal X chromosome had scores more similar to males for all these factors.

Skuse’s team created the mouse equivalent of Turner syndrome—animals with only a single paternally-derived or maternally-derived X chromosome. In a test of perseverative and association responses, indicative of the autism spectrum disorder traits of rigidity and failure to learn rapidly, mice with the maternally-derived X chromosome made more perseverative errors and learned new associations less quickly than those with paternally-derived X chromosomes. At the molecular level, the group found an imprinted gene XLr3b that was associated with increased perseverative behavior in males. The gene is preferentially expressed in males (from the maternally-derived X chromosome). Interestingly, deletions of the region containing the gene’s human homolog, FAM9B, were associated with autism in three females, though expression of the gene has thus far only been found in the testes.

Recently, a study by Catherine Dulac’s group at Harvard University found that there is widespread preferential transcription of genes of maternal or paternal origin in mouse brain tissue, including preferential expression of maternally-derived Xlr3b in the pre-optic area of the brain, which associated with behavior. Taken together, these results suggest that epigenetic mechanisms may result in differences in gene expression in daughters versus sons. Some of the affected genes may play a role in brain development, behavior, and disease.

Think tank: Should epigenetics be its own discipline?

Research in epigenetics has exploded over the last decade, but some of the speakers were questioning whether it should be considered a field unto itself. Hans Reul suggested that each research question—cancer, development, memory—can benefit from incorporating epigenetic approaches independently: It's just another mechanism. Michael Meaney would just as soon subsume epigenetics under a wider umbrella. Genomics is already such a moving target, he said, that "trying to subdivide a field that studies its regulation doesn't make a whole lot of sense." So why not just join the club? Barry Lester observed that applying epigenetics to the field of behavior brings up issues that wouldn't otherwise surface. "That's a good point," Meaney acknowledged. Ed Tronick pointed out that given what is now known about environmental effects, "joining the club" may not be sufficient because there needs to be as much effort put into characterizing the environment as has gone into characterizing molecular mechanisms. At the moment environmental "phenotyping" is crude even in experimental studies, and what may seem like stable epigenetic changes may actually be the result of stable environments renewing the epigenetic change.