Essay: DNA Methylation

DNA methylation only occurs on DNA base cytosine where a methyl group is added to the C base to produce 5-methylcytosine as shown in Figure 2. This causes DNA to be more compact which makes the target gene less accessible and therefore reduces gene expression, which is shown in Figure 1. If a large amount of methylation is present the gene may be shut down completely.
Methyl groups are most commonly added at CpG sites: places where cytosine is followed by a guanine in the gene base sequence (p refers to the phosphate group that links the CG nucleotides together). A CpG island is an area containing many CpG sites, therefore causing genes in this area to be heavily supressed due to the frequency of methylation (Schermelleh, 2018).
Histone Acetylation
Histone acetylation occurs on histone tails: specifically the amino acid lysine, where an acetyl group is added to produce acetyl-lysine, shown in Figure 3. Acetylation is one of the 50 different histone modifications that have been identified; some increase gene expression, while others decrease it. Histone acetylation makes chromatin more accessible for gene transcription, causing increased gene expression. This modification is impermanent and can be removed easily unlike DNA methylation, which can remain on human DNA for a whole lifetime (Carey, 2011).
Research into histone acetylation is more difficult to conduct due to the impermanent nature of this epigenetic change. Additionally, histones tent to have many different types of epigenetic modifications on them at once, which have a variety of forms and effects on gene expression. Despite the complexity of these modifications, some researchers have investigated these epigenetic changes and their correlation with schizophrenia.
A study by Tang, B. et al, looked at histone acetylation at the promoter regions of eight genes that had previously been implicated as contributors to the pathogenesis of schizophrenia. The study involved the use of post mortem brain samples. The samples were split into two groups, where one consisted of specified areas of the prefrontal cortex, with some control subjects and those with schizophrenia and bipolar disorder. The second group was made up of young subjects (18-36 years old), also including controls and schizophrenic and bipolar samples of the prefrontal cortex.
The results found that histone acetylation at the selected residues had patterns in certain loci that were related to disease and age-related effects in normal subjects, as well as those with schizophrenia. Additionally, histone acetylation was evidently linked to the expression levels for eight of the genes that were linked to schizophrenia (Tang, B. et al. 2011).
The samples were matched for age and analysed for pH stability to provide valid samples for the experiment. Additionally, there was no record that these subjects were treated with valproic acid, which would cause disturbance to the study as it can act as a HDAC inhibitor. These control measures, as well as the use of controls in both groups of prefrontal cortex samples greatly increases the validity of the findings however an overall sample size of 82 may be considered to small considering the number of variables and groups being tested. For example, only 6 controls and 6 schizophrenic subjects were used in the 18-36 group. This means further testing is needed to confirm these results.
Studies such as these are necessary to further understand the precise effects of histone modifications on schizophrenia. Such understanding could help the development and efficient use of epigenetic treatments such as histone deacetylase inghibitors, which help bring up gene expression on target genes; a technology which is especially necessary due to the reversibility of histone acetylation (Redfield, 2018).
Studying Epigenetic Modifications
Investigating Blood Samples
Although using blood as a source of peripheral samples has its limitations, there are many ways in which information derived from these investigations has been beneficial. There are two models that explain how studying DNA methylation in blood could be used as a tool in researching schizophrenia (Aberg, K. et al. 2014). The first is the “signature” model, which uses biomarkers in the blood, formed due to the association between schizophrenia and DNA methylation, as a signature that acts as an indication of factors that increase susceptibility to the disease. The markings can be used as a clue for the cause of the disease. This cause could be unrelated to methylation in the brain.
The second model is the “functional mirror site” model, which assumes that methylation in the brain causes the biomarkers found in the blood, and therefore these epigenetic modifications in the brain are proportionally reflected in the appropriate loci in the blood. Such mirror sites are common in the blood, and occur due to events in epigenetic reprogramming that affect the germ line, formation of embryos, environmental factors and genetic variation: all of which can alter the extent of epigenetic modifications in various tissues (Efstratiadis, 1994).
Aberg, K. et al. studied these two models by using haloperidol decanoate: an antipsychotic drug, which blocks dopamine receptors causing anti-hallucinogenic effects. They administered the drug to inbred mice, then performing DNA methylome sequencing in the blood, cortex, and hippocampus. This involved identifying the percentage of the CpG sites that were methylated in the genome of each tissue. The results of this study showed that 65% of CpG sites that were compared showed correlation in changes due to the administered drug, “where concordance rates were similar between blood and brain vs. the two brain tissues.” This had indications that a factor such as the administration of antipsychotic medicine could leave biomarkers in the blood in the form of methylation changes, mirroring those in the brain.
The sample for this study used 759 schizophrenia cases and 738 controls. The samples went through multiple stages of genome sequencing including and the massive data set produced was “processed using a specifically designed data analysis pipeline”. Some of the most crucial findings were subsequently replicated using other sequencing methods in independent case-control participants.
Due to this large sample size, control measures and repeatability, this study provides a highly reliable and thorough demonstration of how testing blood samples of schizophrenic patients could be used to improve disease management through the investigation of biomarkers, in this case using the functional mirror site model.
Limitations To Epigenetic Studies
An important limitation of studies into epigenetics in schizophrenia is the inability to obtain brain tissue for examination, therefore requiring the use of peripheral samples mainly obtained from blood (Peedicayil, J. 2012). A problem with this is that whole blood samples with different cell types are often analysed. This may be an explanation for the varied findings in investigations that use these kinds of samples, as a cell type unrelated to schizophrenia may be selected for analysis.
Additionally, the changes in gene expression resulting from epigenetic changes are usually a cumulative effect of a multitude of epigenetic factors. This means that tools that are often used to control the epigenome as a whole through “global inhibitors of DNA methyltransferases and histone deacetylases” provide a limited way of understanding how the individual epigenetic properties affect gene expression and subsequently the behaviour of a patient. Therefore, the lack of tools to regulate the epigenome restricts the ability to fully understand the processes that lead to epigenetic changes, ultimately stifling the development of epigenetically targeted treatments. (Day, 2014)
Nutrition
Healthy and balanced nutrition is known to be an important aspect of healthy human development. When studying nutrition and its effects on developing of schizophrenia many obstacles can be identified, especially when conducting medical research. In order to study nutrition and schizophrenia, research would primarily need to be conducted during early human growth, as prenatal nutrition would have the greatest and most apparent effect on neurological development. Obstacles in conducting this research include the ethical problems of putting certain individuals at greater risk of schizophrenia, or the inability to thoroughly examine prenatal nutrition due to the lack of available test subjects.
Due to this difficulty in conducting research, historical events such as famines, can occasionally be found to provide the strongest evidence linking prenatal nutrition and schizophrenia. Natural experiments that provided invaluable knowledge into the effects of prenatal starvation include the Dutch Hunger Winter, and the Chinese Famine.
The Dutch Famine
The Dutch Famine occurred during a 6-month period in 1944-1945 as a result of a Nazi blockade in western Holland. An unusually cold winter had already caused the canals used to transport food to freeze over. The harsh weather conditions, combined with rigorous prevention of food and fuel shipments by German forces, resulted in the average daily calorie intake dropping from approximately 1500 to 700. The 4.5 million affected survived largely due to soup kitchens, while the estimated 18000 casualties were reported to mostly have been elderly men.
In spite of the famine, detailed records of birth outcomes and food rations were maintained. Health outcomes continued to be maintained decades after the famine, allowing the data of those who had been through gestation during the famine to be examined. Furthermore, as “the height of the famine was brief” and “clearly circumscribed in time”, the birth outcomes during that period would have been reliably known to have resulted from nutritional deficiency, since the outcomes before and after that period of time would have acted as controls. (Brown, A., Susser, E., 2008). Additionally, the nutritional intakes in the southern and northern regions of the country were unaffected during the 6-month period of the famine, providing further controls for the investigation (Stein et al. 1975). Overall this provides great validity to the study since all neurodevelopmental findings can be known to have nutritional causes, however the exact types of nutritional deficiencies or possible toxins in the diet could not be identified, which limits the investigation.
There may also be some unknown factors such as prenatal stress, which the mothers would likely experience due to a combination of war and famine. It is worth noting however that other areas of the country also experienced the hardships of the war and malnutrition, without the result of increased incidence of neural defects in birth cohorts.
One study took a cohort from births with central nervous system (CNS) anomalies in the famine cities during a period where the height of the famine corresponded to early gestation for the cohort. The Dutch psychiatric registry was used to compare cohorts exposed and unexposed to the famine. Data was taken from the Dutch National Psychiatric Registry for the years 1970-1992, during which the subjects would have been between 24 and 48 years old. They tracked the diagnosis of the birth cohort and compared their outcome decades later, with those unaffected by the famine. The results of the study found that the birth cohort exposed to the famine in early gestation had a two-fold increase in cumulative risk of schizophrenia (Brown, A., Susser, E. 2008).
The study first looked at archived biological specimens from the birth cohorts and investigated how levels of homocysteine in the blood correlated with development of schizophrenia. Homocysteine is a common amino acid generally derived from eating meat and is associated with low levels of vitamins B6, B12 and folate. Homocysteine is also produced as a part of the methylation process in the body (Hart, 2012).
Results found that elevated levels of homocysteine in the first trimester correlated with a two-fold increase of schizophrenia in offspring, although due to a small sample size, the finding failed to be statistically significant. However, for higher homocysteine levels in the third trimester the same two-fold increase of schizophrenia was found, and the larger number of subjects, as well as adjustment for potential cofounders such as race, smoking and age allowed a statistically significant result.
Other neurodevelopmental findings study of the Dutch Famine included increased incidence of congenital neural defects among a birth cohort conceived during the famine (Susser, E. et al. 1996). These included spina bifida: a neural tube defect where the spine and spinal cord do not develop fully, causing a gap in the spine, and anencephaly: a serious birth defect where a baby is born without parts of the brain and skull (NHS 2017)(CDC 2017). These findings further support the correlation between prenatal famine and schizophrenia, as malnutrition has been shown to have adverse effects on early neurological development.
The Chinese Famine
The Chinese famine of 1959-1961 claimed the lives of around 30-40 million people, and affected every province in China. Although primarily caused by drought and severe weather conditions, it was worsened by a political upheaval known as the ‘Great Leap Forward’, which involved the adoption of ineffective agricultural methods, and agricultural labour being diverted onto other purposes, such as steel production, leaving very little food for workers (Kucha, G., Llewellyn, J. 2015).
Unlike in the Dutch Famine, monthly data on calorie intake was not available and the conditions and severity of the food shortage were not imposed consistently throughout the population. This makes the study generally less reliable than that of the Dutch famine, however the use of birth rates as a measure of malnutrition levels proved to be effective in determining the gestational periods of individuals. Furthermore, the large sample size allowed exploration of how different levels of malnutrition may lead to varied health outcomes in birth cohorts.
Research was carried out in one of the provinces most affected by the famine: Wuhu region of Anhui, where rates of schizophrenia were measured in those born before, during and after the famine. Wuhu and the surrounding six provinces were all served by one psychiatric hospital, and reports from that hospital were examined for the years 1971-2001.
Results showed that among those born during the most severe famine years there was once again a two-fold risk of developing schizophrenia later in life. The severity of the famine was measured by the birth rates at the time, which from historical records of 1959-1961 were shown to be less than one-third of the average from 1956-1959. The risk of developing schizophrenia later in life increased from 0.84% in 1959 to 2.15% in 1960 and 1.81% in 1961 (St Clair, et al. 2005). These results therefore support the findings of the Dutch famine study and provide further evidence for the link between prenatal malnutrition and risk of developing schizophrenia.
In spite of limited data recorded during the famine, the use of the data and its large sample size has been effective in supporting existing hypotheses regarding environmental changes and schizophrenia. As with the Dutch famine study, the specific types of deficiencies, or combinations of deficiencies, could not be determined, which limits the amount of information that can be derived from looking at famine in general. Other studies have been conducted to find out the exact deficiencies that lead to certain conditions.
Potential Nutritional Deficiencies
Folate is a term for chemically similar trace compounds, which are crucial for many biological processes. This includes biosynthesis of various chemical compounds through transfer of one-carbon units from donor molecules (Lucock, 2000). The human body cannot synthesise folates, therefore they must be acquired from the diet.
Decreased intake of folic acid has been shown to cause decreased methylation in genes vital to neurological development. For example, those exposed to the Dutch famine during early gestation, had lower DNA methylation levels of insulin-like growth factor 2 (IGF2) gene in comparison to same-sex siblings who were unexposed. Additionally, in a study comparing IGF2 gene methylation levels in children aged 17 months, some of whose mothers had taken folic acid between conception and early pregnancy, it was found that children with mothers who used folic acid had significantly higher methylation of the IGF2 gene (Peedicayil, J. 2012). This suggests that increased expression due to lower methylation of the IGF2 gene, as a result of lower intake of folic acid is one of the possible causes of schizophrenia in the affected members of the Dutch famine birth cohort.
The link between CNS abnormalities and schizophrenia from the Dutch cohort also gives evidence for the role of folates in schizophrenia. It has consistently been found that prenatal folate supplementation has an effect on neural tube defects, as dietary intervention can decrease their likelihood by as much as 80% (Brown, A., Susser, E. 2008). This strongly suggests that prenatal neurological development during famines is greatly affected by lack of folates in the diet. This idea has been vital in bringing about foods fortified with folates in many countries to ensure improvement of prenatal health (Scott, 2018).
Along with folates, iron deficiency has been identified as a possible cause of poor prenatal neurological development. During pregnancy, the need to grow the placenta and the foetus, supply the foetus with oxygen, as well as the need to increase maternal red blood cell mass, means a larger supply of iron is needed (Beard, J., Connor, J. 2003). If there isn’t a sufficient amount of iron provided from maternal stores, there is a danger of reduced haemoglobin levels, and of developing anaemia. This could lead to foetal hypoxia, which can have adverse effects on brain development. The hippocampus is known to be vulnerable to the effects of hypoxia, which have been implicated as causes for increased susceptibility to schizophrenia (Cannon, et al. 2002).
There may also be pathways independent of anaemia, in which iron deficiency may affect brain development, which could lead to schizophrenia. Iron metabolism is important in dopaminergic neurotransmission. The hormone dopamine plays a key role in the pathophysiology of schizophrenia (Brown, A., Susser, E., 2008).
Twin Studies
Scientists have long been aware that genetics plays a role in determining who develops schizophrenia. One reason for this if one of a pair of monozygotic twins has schizophrenia, the other has a 50% chance of developing it as well, compared to the usual 1% chance in the general population.
Monozygotic twins are two offspring that result from one zygote that split into equal halves during an early stage of development, giving rise to two individuals with the same blood group, genetic constitution, age, maternal environment and gender (Farlex Medical Dictionary 2012). As monozygotic twins share the same DNA sequence, they are ideal candidates for studying the epigenetic factors in disease aetiology. Additionally, identical twins are usually matched in the environments they are raised in. All of these factors make them the best-matched control available (Bell, J., Spector, T. 2011). This is because the identical genetic codes in two individuals allows for the investigation into discordance in phenotypic expressions, excluding those caused by genetic differences.
Figure 5 shows that the more closely related two individuals are to someone with schizophrenia, the more likely they are to develop it themselves. The most important parts of the graph are the two bars at the bottom from which compare the concordance rates of identical and non-identical (fraternal) twins. Non-identical twins share the same developmental environment but are genetically no more similar than any other pair of siblings.
The comparison between the two types of twins is important because twins in a pair generally share quite similar developmental environments. If schizophrenia were mainly caused by environmental factors we would expect concordance rates for the disease to be fairly similar between identical and non-identical twins. Instead we see that if one non-identical twin develops schizophrenia, the other twin has a 17% chance of doing the same, while in identical twins this jumps to nearly 50%. The almost three-fold higher risk for identical versus non-identical twins tells us that there is a major genetic component to schizophrenia. However, since the concordance for identical twins is not 100% due to their identical genetic makeup, epigenetics must play a large role alongside genetics (Carey, 2011).
A study by Castellani et al. used blood samples from a pair of monozygotic twins discordant for schizophrenia, and tested for differentially methylated regions (DMRs) across the genome. Differentially methylated regions are areas with different levels of methylation between two samples. They also compared results with methylation found in the genome of their parents. The blood samples from two families were used, and their pedigrees are shown in Figure 4 below.
Genome-wide results showed the presence of DMRs between the monozygotic twins of both families. Some of these DMRs were shared with parents whereas others appeared to be newly formed during the lifetime of the twin due to environmental exposures or random events, and were not present in the parental genomes. There were 27 genes that appeared to be affected by DMRs. The genes affected and unaffected are shown as a Venn diagram in Figure 4 below. Two of the networks that were identified were “cell death and survival” as well as a “cellular movement and immune cell trafficking”. Evidence prior to this study has found that both of these networks play a role in the pathogenesis of schizophrenia.