Epigenetics: responding to life by switching genes on and off

Session 3 of the Progress Educational Trust's annual conference (PET), held on Wednesday 18 November 2009 at Clifford Chance, was chaired by Dian Donnai, Professor of Medical Genetics at the University of Manchester, and started with a talk by Karen Temple, Professor of Medical Genetics and Honorary Consultant in Clinical Genetics at the University of Southampton and Wessex Clinical Genetics Service. Professor Temple gave an intriguing talk on the influence of parents - not directly through the genes we inherit from them but as a result of epigenetics and genetic imprinting. In clinical genetics, this has the potential to introduce a new layer of complexity.

Epigenetics involves interactions of chemical groups (most commonly the methyl group, CH₃) with DNA, and that can alter how our genes work. So although the genes themselves remain intact and capable of functioning, genetic imprinting, a type of epigenetic regulation, means that around 60 of our genes are expressed selectively because one of our two copies is silenced. Which copy is silenced is sex-specific, so that only the gene copy, or allele, that was inherited from the mother or from the father is expressed, but never both.

Professor Temple illustrated how this influence takes effect by describing a case study of a child with a monoepigenic disease. The child in question was born with neonatal and transient diabetes. This is a condition that a neonate will recover from by one-year old, but which is known to recur later in life. Normally, classical genetic tests - which involve looking at the chromosomes under a microscope - would reveal the chromosomal anomalies related to the condition. But in this case they did not. It was later discovered that for chromosome 6, rather than having inherited one copy from its mother and the other from its father as would usually be the case, the baby had inherited both copies of chromosome 6 from its father.

Professor Temple said that she thinks of the epigenetic control of genes as an electric light. For a light bulb to work, it needs to be manufactured properly, but, regardless of this, if the switch has not been turned on then the light bulb will not work. Epigenetics is similar, in the sense that it is the process by which genes are switched on and off. Like any switch, an epigenetic switch is not fixed, but reversible. Adding methyl groups to a gene makes it difficult for that gene to be read by the cell's machinery, and effectively stops it from being expressed. Removing the methyl groups means the gene is expressed normally again. Many of the 60 genes known to be genetically imprinted tend to promote fetal growth when they are expressed from the father's allele, and to restrict this when expressed from the mother's allele. The gene implicated in the condition seen in Professor Temple's case study was PLAGL-1. This is normally an imprinted gene. But it is only methylated, or switched off, on the mother's copy of chromosome 6. In contrast, the PLAGL-1 gene on the chromosome 6 inherited from the father is always expressed. By inheriting both chromosomes from its father, the baby with neonatal diabetes ended up with two functional PLAGL-1 genes instead of just one, which meant the gene was being overexpressed, and the baby effectively overdosed. Other children who were subsequently tested were found to have genes that had lost their parent-of-origin-specific methylation mark, so that the mother's allele acted just like the father's. One of the goals of epigenetic research is now to try to reverse this pattern of methylation, so that babies with transient diabetes are no longer at risk of the late-onset diabetes that they would otherwise experience later in life.

Professor Temple indicated that the process of in vitro fertilisation (IVF) is somewhat associated with methylation disturbance or imprinting defects. She said that if we looked into the reasons why couples have IVF, it is because this intervention gets around natural mechanisms of preventing pregnancies where there may be a genetic problem. But given that IVF is used widely, understanding the genetics of epigenetics will help further knowledge of the biology and the causes behind these disorders. It may also be that epigenetics plays a part in responding to environmental factors implicated in some increasingly prevalent conditions, like diabetes and obesity.

Dr Jonathan Mill, Lecturer in Psychiatric Epigenetics at the Institute of Psychiatry, King's College London, followed Professor Temple with even more fascinating data. In his talk, 'Beyond A, C, G and T: Epigenetics and psychiatric disorders', Dr Mill described his investigations into exactly how much of a role epigenetics plays in mental health. Studies using monozygotic, or 'identical' twins have pointed to the possibility that disorders like alcoholism, schizophrenia, Alzheimer's, autism, major affective disorder and dyslexia are heritable and not solely a consequence of environmental conditions. But although there have been many genome wide association studies published on psychiatric disorders, very little variation has actually been found in the genes that accounts for these patterns of inheritance. Dr Mill said that these conditions are not like diabetes, because with psychiatric disorders he would be looking for very, very small effect sizes. It is very also difficult to explain psychiatric disorders, and how they can change over the life course of a person, either by genes or by environment.

To understand what the other possibilities are, we have to take a step back and look at what DNA is, at how it is packaged, and at how genes are accessed by the cell's machinery. In fact, how DNA is packaged is really what mediates phenotypic changes, like the onset of a disease. Methylation is associated with DNA compaction, a process which renders it inaccessible switching off gene expression or causing it to be expressed at very low levels. This is important because even if you had two identical strands of DNA, how it is packaged will make the difference between significant gene expression and no expression; between large amounts of protein being produced or else no protein produced. It is also important to consider when and where genes are expressed, because even genes that carry no unfavourable mutations, may be rendered useless or harmful if they are expressed at the wrong time or in the wrong place. This means that the methylation - the epigenetic regulation - that either 'locks a gene up' or releases it to be freely expressed is likely to be a key mechanism in some pathological conditions.

A loss of methylation that should be imprinted in a parent-of-origin-specific manner is known to directly affect brain weight. Our genome can acquire or lose methyl groups sitting on the DNA more easily than it can make changes to the actual DNA hardware. Therefore, rather than just genome-wide-association studies that scrutinise changes in the hardware, genome-wide epigenetic scans that look for methylation or lack of methylation are being done using the DNA of individuals affected by psychiatric conditions. And because it is very, very rare to find complete concordance in monozygotic twins for psychiatric disorders, epigenetic profiling studies in twins are also currently being performed. Dr Miller said that when studying complex disease, environment can interact with genes; and genetic variation maps upon epigenetic variation. For him, the real value in looking at epigenetics is linking up nature and nurture.

The final talk of the session, 'Epigenetics and early life experience' was given by Professor Marcus Pembrey, Founding Chair of the Progress Educational Trust and Emeritus Professor of Paediatric Genetics at the Institute of Child Health. He began by describing a study which showed that life expectancy can be predicted by looking at socioeconomic status as a child, so that an adverse social economic position in childhood is associated with higher mortality rate.

Professor Pembrey considered the possibilities that might explain this association between fetal or childhood experience and adult health and well being. Explanations could include social patterning, that is, a learnt lifestyle that reinforces habits, or inherited variations in your or your parents' DNA that cause both the early situation and the later outcomes. But it could also be that your early experience alters the pattern of gene actively through epigenetic mechanisms (such as methylation of DNA which then persists throughout life); or else some mechanism or mechanisms yet to be discovered. While all of these factors are likely to play a role, social patterning and epigenetic mechanisms can converge in the prenatal period. For example, nicotine, alcohol, and famine are all known to cause epigenetic changes. The DNA methylation changes associated with periconceptional famine have lasted around 60 years carried in the stem cells. Epigenetic states have also been shown to be transmitted in the inheritance of a cancer-associated MLH1 germ-line epigenetic change in two families. Although the mediating mechanism is unknown, male-line transgenerational responses have been observed in humans and are triggered by exposures in early life.

Professor Pembrey gave examples to illustrate the transmission of epigenetic changes from one generation to the next - how parents' environment might have an effect on the future wellbeing of their children and grandchildren. The first was from the Avon Longitudinal Study of Parents and Children, research which Professor Pembrey played a key part in establishing and which is widely regarded as the most comprehensive general birth cohort in the world. He described data demonstrating that the age at which a man starts smoking affects the birth-weight and early growth of his sons - an effect not correlated to maternal smoking. The earlier father starts smoking, the larger the body mass index (BMI) of his future sons. However, there was no significant effect in daughters. The study also showed that the food supply of paternal grandparents affects the mortality of grandchildren. The relationship is not straightforward, but may make more sense in the light of genetic imprinting, where genes are switched off in a sex-specific manner. For example, if a paternal grandmother had a good supply of food, it was found that her granddaughters died off earlier; but if a paternal grandfather had a good food supply, his grandsons did the same.

Studies also show that variations in maternal behaviour are heritable in rats, but not genetic. Rats differ in their licking and grooming and arched-back nursing of newborn pups. While the offspring of non-maternal mother rats become stressed and fearful, the pups of maternal mothers are less stressed and continue this maternal behaviour when they become mothers. But stressed and fearful pups continue the cycle of neglect with their own pups. However, this is not 'genetic', as adopting the pups of a maternal mother to a non-maternal one causes these pups to become fearful and stressed, and vice versa - and the pups will proceed to continue the cycle of behaviour that they were exposed to.

Interestingly, human DNA methylation studies of the equivalent (NR3C1) gene in the brains of suicide and road traffic victims found methylation signatures that told of a history of child abuse in some victims and not in others. So while the environment we are exposed to in early life may not change our genome, studies like these can ask the question of whether ancestral early-life exposure can change the epigenome, and change it in a heritable way.

Professor Pembrey concluded that although it is early days for epigenetic research, understanding epigenetic responses during development is on the horizon. The differences in DNA methylation that has already been observed in adults in relation to prenatal starvation, childhood abuse and childhood social economic circumstances has now established the feasibility of epigenetic epidemiology using birth cohorts - including the epigenetic study of effects that may be transmitted through families over several generations.