B2848 - Maternal depression antidepressant use during pregnancy and autism/autistic traits - 2017/23/02
Depression is common in women of childbearing age and pregnancy. Antidepressants are some of the most common medications prescribed by the NHS. Recent studies have estimated that up to 10% of women in some western countries take antidepressants during pregnancy. Recent studies have raised the possibility of an association between antidepressant use during pregnancy with longer term neurodevelopmental problems such as autism. However, maternal depression may independently be associated with adverse neurodevelopment- potentially confounding the association between antidepressant use and autism.
This study aims to investigate the association of maternal depression or taking antidepressants during pregnancy with offspring risk of autism or autistic traits in ALSPAC.
B2843 - Investigating putative risk factors for neurodevelopmental disorders using Mendelian Randomization 14-02-2017 - 125156 - 2017/15/02
Neurodevelopmental disorders (NDDs) are conditions involving perturbed brain development with manifestations ranging from specific to global impairment of developmental domains such as communication, social interaction, motor skills, cognition, activity and emotion. This includes autism spectrum disorder (ASD), attention-deficit hyperactivity disorder (ADHD), disorders of language, learning and motor functions, intellectual disabilities, schizophrenia and bipolar disorder. NDDs are highly prevalent in the population (~5% of children), typically emerge early in life, and are associated with substantial and long term disability.
The aim of this study to investigate the role of potentially modifiable risk factors for NDDs, using Mendelian Randomization (MR). Multiple pre-/perinatal risk factors are associated with NDDs, including nutritional deficiency (vitamin D deficiency, fatty acids), thyroid dysfunction, inflammation and autoimmune conditions, metabolic conditions, stress, smoking and low birth weight. However, results from epidemiological studies have been inconclusive and the causal role of these risk factors has not been established for most NDDs. Studies using methods that strengthen causal inference are important so that the role of these potentially modifiable risk factors can be determined.
In this project we plan to use MR analysis, whereby genetic variants robustly associated with the risk factors above are used as instrumental variables (IV) to examine whether they have a causal effect on NDDs (categorical disorders and dimensional traits). We will use two-sample MR in which the IV-risk factor and IV-outcome associations are obtained from non-overlapping samples (Burgess et al., 2015). First, published GWAS studies will be used to identify genetic variants robustly associated with the risk factors. Second, these genetic variants will be used to calculate a genetic score for each risk factor in an independent target sample (ALSPAC). Standard linear and logistic regression models will be used to estimate the effect of each risk factor on the NDD trait of interest in the ALSPAC target sample, using the genetic scores as an instrument.
We will also employ additional methods such as positive and negative control designs to strengthen causal inference (Lawlor, Tilling & Davey Smith, 2017).
Burgess S, Scott RA, Timpson NJ, Davey Smith G, Thompson SG, Consortium E-I (2015). Using published data in Mendelian randomization: a blueprint for efficient identification of causal risk factors. European Journal of Epidemiology 30, 543–552.
Lawlor DA, Tilling K, Davey Smith G (2017). Triangulation in aetiological epidemiology. International journal of epidemiology. [Epub ahead of print]
B2838 - The developmental role of metabolism appetite and growth in eating disorders exploring novel longitudinal risk pathways - 2017/15/02
The Eating Disorders (full and partial syndrome anorexia nervosa and bulimia nervosa and binge eating disorder) are life-threatening illnesses that start in adolescence and affect between one and two in ten adolescents and young adults. There is a lack in our understanding of why eating disorders develop, which affects our ability to develop treatment and adequately prevent eating disorders. This project builds on our preliminary data showing that metabolism and growth might play a role in the development of eating disorders; and aims to test the hypothesis that metabolic function, appetite and growth factors might precede onset of eating disorders. We also aim to explore whether these factors might be related to individuals’ genetic make up, using novel methodologies and based on novel genetic findings. This study will be fundamental in starting to understand new risk mechanisms for adolescent and young adult eating disorders that can be further investigated in larger and more detailed studies following this project.
B2840 - Depot specific effect of pre-pubertal IGF-II on post pubertal fat distribution in ALSPAC cohort - 2017/15/02
Insulin like growth factor II (IGF-II) is a hormone. IGF-II levels in the human circulation are maintained at high levels throughout life unlike those of rodents. IGF-II has been strongly linked to obesity in genetic studies but its metabolic role is still far from understood. We were able to confirm using fat biopsies from children that IGF-II is a key regulator of fat cell biology. Our results suggest it acts as a buffer to excess visceral fat accumulation in children.
We are aiming to identify the role of X chromosome on cardiovascular diseases in ALSPAC (including presence of disease, blood pressure, anthropometric traits including BMI, lipid levels and glycemic traits). X chromosome association analysis will be conducted using ALSPAC genotype data. The Variance explained by the X chromosome will be estimated.
Urinary incontinence (UI) in women is a form of involuntary leakage of urine that affects between 25-45% globally. Several possible UI determinants that have been suggested in the literature include number of children, how they were delivered (vaginally or by C-section), body mass index (BMI), age, hysterectomy, smoking, recurrent urinary tract infection and family history but the reasons for UI remain understudied and poorly understood.
It has been estimated that stress urinary incontinence (SUI) which is an involuntary leakage of urine due to physical efforts ranges from 3% to 25% of elderly women. Quality of life and sexual function are often affected by SUI. The ability to perform activity of daily living may be severely affected by SUI, resulting in embarrassment, social isolation and decrease in health-related quality of life. Women suffering from SUI may avoid participating in physical activity, which in turn impacts on the overall health. Finally, some evidence has shown that up to 50% of women with UI will avoid sexual intimacy with their spouse.
B2833 - Assessing the impact of parental smoking on DNA methylation in offspring sex chromosomes - 2017/08/02
Tobacco smoke has been shown to cause highly reproducible changes to the methylome. This has been observed in response to own smoking but also in offspring exposed to maternal smoking in utero. Sustained maternal smoking during pregnancy is associated with aberrant DNA methylation reported in large-scale epigenome-wide association studies (EWAS) of cord blood (1,2). Associations are also seen with paternal smoking, although these largely attenuate when adjusted for maternal smoking (explained by the correlation of maternal and paternal smoking habits). However, all the EWAS for prenatal smoking exposure performed in cord blood to date have focused on autosomes and sex chromosomes have been previously excluded from the analysis. We propose that methylation changes induced in sex chromosomes may have biological importance and we seek to specifically analyse these in the Accessible Resource for Integrated Epigenomic Studies (ARIES).
We hypothesise that paternal and maternal exposure to cigarette smoke may lead to DNA methylation variation in the offspring sex chromosomes and these aberrations may increase the susceptibility of exposed offspring to adverse perinatal outcomes.
1. Joubert BR, Felix JF, Yousefi P, Bakulski KM, Just AC, Breton C, et al. DNA Methylation in Newborns and Maternal Smoking in Pregnancy: Genome-wide Consortium Meta-analysis. Am J Hum Genet 2016;98(4):680-96.
2. Richmond RC, Simpkin AJ, Woodward G, Gaunt TR, Lyttleton O, McArdle WL, et al. Prenatal exposure to maternal smoking and offspring DNA methylation across the lifecourse: findings from the Avon Longitudinal Study of Parents and Children (ALSPAC). Hum Mol Genet 2015;24(8):2201-17.
B2834 - Understanding the Relationship between Adolescent Biopsychological Development and their Behaviour in the School Context - 2017/08/02
The inclusion and achievement of boys in the school setting is of significant concern, especially for those whose challenging behaviour is so severe that it disrupts the learning and wellbeing of others. With a reduction in specialist places for boys with behavioural, emotional and social difficulties (BESD), schools are expected to use devolved funding to establish support mechanisms ‘in-house’. If this includes exclusion to a behavioural unit, the socio-economic outcomes for individuals is often dire involving educational underachievement, youth unemployment, crime and substance dependence. My proposed study aims to explore factors associated with the challenging behaviour of boys with BESD in the school setting with the aim of informing possible interventions and strategies to positively support their behaviour and inclusion. These include biological factors associated with puberty and psychological factors such as social skills and cognition.
To better understand the causal role of lipoproteins and lipid components in cardiovascular and metabolic disease.
B2812 - Epigenetic gestational age clock predictors and associations with child growth trajectories - 2017/25/01
DNA methylation is a chemical change to our DNA that influences the extent to which our genes are expressed. Environmental factors can affect DNA methylation. DNA methylation has previously been used to accurately predict age, and the difference between a person's age and the age predicted by their DNA methylation has been shown to be associated with mortality and various health outcomes. Recently, this approach has also been used to identify a set of DNA methylation markers that predict gestational age at birth. This measure of DNA methylation gestational age has been shown to be associated with birth weight, independently of gestational age, sex and ancestry. In this project, we will look at whether maternal factors such as maternal smoking, alcohol use, age, and educational level are associated with DNA methylation gestational age, and we will assess the association of DNA methylation gestational age with childhood growth trajectories.
Potentially harmful elements (PHE) in the environment, either derived from natural or man-made sources, are known to have impacts on human health. The British Geological Survey’s Geochemical Baseline (G-BASE) survey quantifies the concentrations of some 50 elements in soils across the UK. BGS’s medical geologists have studied how much of these contaminants are available for absorption into the systemic circulation of the human body through different exposure routes. There is, however, a lack of information on the relationship between the concentrations of the PHEs in soil in a specified area and the concentration of the PHE in humans living in the same location. In order to obtain this information detailed studies of the body burdens of PHEs in human subjects (as measured by human biomarkers, e.g. blood, urine, finger/toenails and hair) is required alongside of high resolution geochemical sampling of the local soils.
This information will provide medical practitioners and risk assessors with the necessary information required to assess what concentration of PHE in soil constitutes a safe level
Metabolomics provide the opportunity to characterise the state of an organism at the molecular level. Genomics focus on identifying genes, or areas of the genome, responsible for the characteristics of an organism. Combining metabolomics and genomics for human health has the ability to help us identify the functional characteristics of individual genes and to also understand the genetic basis of metabolic variables.
- Association of dietary factors with cardiometabolic health Possible mediating role of DNA methylation and metabolome - 2017/25/01
Cardiometabolic disease (CMD) is a major public health problem. It is well known that CMD is caused by interactions between environmental and genetic factors. Additionally, recent studies on mechanism of CMDs have demonstrated that changes in molecular intermediates (e.g. levels of metabolites and DNA methylation) affect prevalence of CMDs. However, there has been limited evidence to suggest causal effect of dietary factors on CMDs, mediated by molecular intermediates. The aim of current study is to estimate relationship between dietary factors and cardiometabolic health, considering effect of molecular mediation including levels of metabolite and DNA methylation. This study can contribute to enhance an understanding of causal effect of dietary factors on CMDs and clarify the mechanism of diet-metabolome-genome-epigenome-health outcomes and contribute to prevent CMDs.
Asthma and other wheezing disorders is one of the main causes for health care utilization in childhood with an increasing prevalence in westernized countries in recent decades. Concomitantly, the use of vegetable oils and grain feeding of livestock has resulted in an increase in the intake of n-6 polyunsaturated fatty acids (PUFA) and a decrease in n-3 PUFA, especially the long chain PUFA (LCPUFA) eicosapentaenoic acid (20:5n-3, EPA) and docosahexaenoic acid (22:6n-3, DHA) particularly found in fish and this has been hypothesized to be a potential cause of childhood asthma. LCPUFA influence immune regulation and some observational studies have reported an association between n-3 LCPUFA deprived diet during pregnancy and increased risk of asthma and related disorders in the offspring. Randomized controlled trials (RCTs) of n-3 LCPUFA supplementation to pregnant women have generally been under-powered and shown ambiguous results. We recently conducted a double-blind RCT of n-3 LCPUFA supplementation during third trimester of pregnancy in an unselected group of 743 pregnant Danish women from the COPSAC2010 cohort showing that n-3 LCPUFA supplementation in pregnancy was associated with reduced risk of asthma in the offspring (Bisgaard et al. NEJM 2016). Specifically, the preventive effect of supplementation was highest in children of mothers with intake and blood-levels of EPA+DHA prior to the intervention and with FADS gene risk variants (minor allele at rs1535). A similar tendency was seen for allergic rhinitis.
Variants in the maternal fatty acid desaturase (FADS) gene region (rs1535 and closely correlated SNPs) have been reported to be associated with EPA and DHA levels during pregnancy. Child genotype has been shown to have smaller but independent effect on cord blood levels of EPA and DHA. This gene variation therefore provide an opportunity to study genetically the hypothesis that prenatal n-3 LCPUFA levels affect the risk of asthma in the offspring. This would support that supplementation with n-3 LCPUFA during pregnancy could prevent asthma in the offspring and therefore could have large importance for asthma prevention.
Unlike other conditions, there is no single test to establish a diagnosis of asthma. Instead a diagnosis is made by gathering information from an individual, and using that information to weigh up how likely they are to have asthma. Information can be gathered from the symptoms described, past or family history, findings from a physical examination, lung function and other clinical tests. In some cases, a diagnosis of asthma is clear from the information given and treatment can be started. However, for a large number of people, the diagnosis isn’t clear, and more information is needed before asthma can be ruled in or out.
The fact that asthma isn’t always straightforward to diagnose can lead to problems. Individuals can be wrongly diagnosed: being told they have asthma when in fact they don’t. Also known as over-diagnosis, the main consequences are unnecessary treatment, exposing patients to medicines with unhelpful long term effects and at a financially cost.
Improving the way in which asthma is diagnosed is very important. Identifying the most important aspects of information that can be gathered and using them to make a prediction score would help Doctors to weigh up the likelihood of asthma more easily and accurately.
I will perform a genome wide association study in which I hope to identify variants within the genome that may confer 'resilience', as measured by a strengths and difficulties questionnaire administered at age 11. This may identify genes which also confer an ability to cope with adverse life events, illness and birth defects such as cleft lip and palate. This will ultimately be useful for identifying those individuals who may need more interventions/close monitoring during adverse situations.
Psychiatric disorders are a major health concern with numerous social and financial burdens (WHO, 2012). Disorders such as anxiety and depression are common all over the world and estimated lifetime prevalence rates range between 21%-29% (Kessler et al., 2005). Likewise, psychiatric disorders such as depression are associated with poorer physical health such as cardiovascular, metabolic and lung diseases, as well as being comorbid with other diseases such as anxiety (Bishop et al., 2004; Hessler, 2006). Understandably, there has been a great emphasis to try to understand the aetiology of these disorders in a bid to develop interventions and enhance existing treatments. Additionally, many psychiatric diseases that occur later in life (e.g., adolescence/early adulthood) originate from early life, and research has focussed on identifying potential risk factors with an end for developing treatments and interventions at an early stage (Bould et al., 2014). To date, much of the research on psychiatric disorders have used measures on depression, anxiety, temperament and wellbeing. Given the wealth of this type of data within ALSPAC, it seems rationale to concentrate on these phenotypes and build upon existing research.
There are many exposures that have been found to be associated with both immediate depression and depression at a later stage of life. Such exposures range from stress, relationships to job loss and financial difficulties (to name but a few). However, one important exposure for depression is the impact of neighbourhoods (Diez Roux and Mair, 2010; Mair et al., 2008; Matheson et al., 2006). To date, a wealth of longitudinal and cross-sectional data has investigated the impact of neighbourhoods on depression and mostly found that neighbourhoods are associated with increased depression (Graif et al., 2016; Kim, 2008). However, there is evidence that suggests that the neighbourhood may also protect against depression as well in certain contexts (Blair et al., 2014). Much of the research to date on neighbourhoods and depression has centred on certain aspects of the neighbourhood such as proximity to crime (Cutrona et al., 2005) or the urban environment (Gariepy et al., 2015). These are important stratifications to be made as it is still unclear how much these varying aspects like crime or urban build within an environment contribute to psychiatric disorders like depression. Nevertheless, it appears that the impact of the neighbourhood is important in the aetiology of depression and warrants further research. What remains unclear are the mechanisms that may underpin this association and what causal pathways lie on this association between neighbourhoods and depression.
Identifying causal pathways to disease are vital for developing effective treatments and interventions. Many cross-sectional studies suffer from confounding which make it hard to establish true effects along these pathways. However, through the use of longitudinal data, we can account for some of this confounding which can aid in understanding and identification of these causal pathways. A recent systematic review of the literature identified only 14 longitudinal studies investigating the association between neighbourhoods and depression (Blair et al., 2014). Many of these studies used small sample sizes are were mainly homogeneous to the USA. In order to fully understand the association between neighbourhoods and depression, more studies need to use longitudinal data that can incorporate confounding variables that can in turn detect true and meaningful effects.
Finally, whilst research into the effects of neighbourhoods on depression have uncovered important findings, there are still a lot of unexplained variance in the data that has yet to be explained. One potential explanation for this is the role of genetics. It is well established that both genetics and the environment contribute to psychiatric disorders such as depression (Munafò, 2015). Recent research has identified a role for the genetics of schizophrenia predicting neighbourhood deprivation (Gage et al., 2016). Promisingly, findings conducted from a recent GWAS have identified genetics associated with depression and neuroticism (Obkay et al., 2016). In order to fully understand the effects of neighbourhoods on depression, it is important to establish if and how genetics can influence this relationship. This project will utilise data from the ALSPAC study to investigate the impact of the neighbourhood on depression whilst taking into account both genetic and environmental data.
Self-harm in young people is a major problem. As many as 1-in-6 teenagers have self-harmed at some point in their lives, and self-harm is particularly common between the ages of 10 and 16 years. Young people who self-harm seem to be more likely to do poorly in a number of ways in early adulthood, including being more likely to have a mental health problem, and to use substances like alcohol and drugs. It is also the strongest known risk factor for suicide. Although self-harm is very common in young people, we know little about what causes it. A better understanding of the things that increase the risk for self-harm, and how they work could help us to develop more effective ways to prevent self-harm and support young people when they need it.
Research shows that facing bad experiences early in life (such as physical and sexual abuse, and emotional neglect) increases the chances that a young person will self-harm. However we do not yet understand how this works. We know that the social environment (including the occurrence of adversity) can have an impact on the body's internal working, such as inflammation, changes to chemicals around genes, and hormones, and that these things in turn are involved in suicide and self-harm. In this study, we aim to investigate whether early bad experiences are associated with three different biological processes (inflammation, alterations to DNA, and puberty (stage and timing)) and explore whether these factors are, in turn, associated with self-harm in adolescence. Our research will also investigate whether associations between inflammation, alterations to DNA, puberty and self-harm are real, or whether they might be explained by other factors. This research could help to identify potential markers of future self-harm risk, as well as possible targets for treatment for young people who self-harm.
We also plan to increase understanding about whether there are clearly distinct types of self-harm. Some researchers have proposed that it may be useful to classify self-harm according to whether or not the person has suicidal ideas, and has self-harmed with the intention of ending their life. However, it can be difficult to judge the extent to which someone wants to end their life and so we are not sure how useful it is to try and judge this in clinical practice. We plan to study whether the genes of young people who self-harm differ according to whether or not they have harmed with suicidal intent. This work will inform future research studies of self-harm, and could help practitioners working with young people who self-harm.
We will disseminate our research findings at key national and international conferences. We will target conferences in the fields of epidemiology and suicide. These include the Congress of the International Federation for Psychiatric Epidemiology, and the European Symposium for Suicide and Suicidal Behaviour – widely regarded as the highest quality international conference for clinicians, service users, the voluntary sector and researchers working on suicide research. In addition, we will present our findings internally, at seminars within the School of Social and Clinical Medicine, and the MRC Integrative Epidemiology Unit. These well-established seminar series are attended by a wide range of academics both within and outside the University of Bristol.
We will seek to publish manuscripts arising from this proposal in high impact, open access journals. To ensure that our findings reach a wide audience, we will work with the University and ALSPAC press offices to facilitate the dissemination of research findings through press releases, sent to local, national and international media outlets. We will communicate our research findings to study participants via the ALSPAC cohort website (http://www.bristol.ac.uk/alspac/participants/), and will prepare an article for the ALSPAC newsletter, which is sent to over 10,000 families.
In order to maximize dissemination of our work, we will engage with academic researchers in different disciplines and the public through participation in public engagement events. For example, ALSPAC has hosted a number of successful participant events where we have communicated findings from previous studies with this cohort. We have also discussed our work at local public events such as the Bristol Festival of Ideas (http://www.ideasfestival.co.uk/events/life-project/). We will continue to participate at public events and science fairs, and will work with the University of Bristol’s Centre for Public Engagement to promote engagement of our work beyond academia.
We will use social media (Twitter, Instagram, and Facebook) to engage with participants and the public in partnership with ALSPAC’s social media manager. To maximise the reach of our work, we will disseminate finings via several Twitter accounts including the Bristol suicide and self-harm group (@SASHBristol), the MRC IEU (@mrc_ieu), ALSPAC (@CO90s), as well as personal twitter accounts held by applicants.
DG has close links with WHO’s suicide lead (Dr Alexandra Fleischmann) and through her has advised WHO on updates to suicide-related codes in the International Classification of Diseases (ICD) and Psychiatrists in the USA on DSM revisions. Findings from this work could contribute to revisions of these classification systems. DG is also member of the Department of Health’s National Suicide Prevention Advisory Group and a Samaritans trustee. PM is a member of the Population & Health Services Working Group, advising the Department of Health on “A Ten Year Strategy for Mental Health Research”. 2016-2017. TF is a member of the Child and Adolescent Faculty Executive Committee (CAPFEC) and is on the board of the Association for Child and Adolescent Mental Health (ACAMH). We will present findings relevant to self-harm /suicide prevention to these forums so that they feed into national policy in this area.
A recent study by Liu et al (http://www.nature.com/mp/journal/vaop/ncurrent/full/mp2016192a.html) developed an accurate biomarker of alcohol intake, using the levels of DNA methylation at 144 CpG sites. Such a biomarker has the potential to provide a more reliable estimate of alcohol use in future studies. As self-reported alcohol intake may be bias or otherwise unreliable, such a biomarker may serve as a valuable supplement to self-reported data, especially in epidemiological studies where careful estimation of alcohol use is needed.
We seek to validate this alcohol intake biomarker in an independent cohort, and rigorously assess its performance and modeling assumptions. Further, we'll seek to extend it's application to predictions across the life-course.
B2819 - Markers of prediabetes psychotic experiences and inflammation A cohort study from ALSPAC - 2017/19/01
Patients with schizophrenia have shortened life expectancy, with mortality rates twice that of the general population. Causes for this extend beyond suicide, accidents and risk-taking behaviour. Research indicates that physical illnesses, including heart disease and type 2 diabetes mellitus (T2DM) account for a majority of the increased mortality risk. The prevalence of T2DM in schizophrenia is increased by around one-third compared with the general population.
There is a growing body of research to suggest that links between the two illnesses may go deeper than the previously assumed - related solely due to either a side-effect of medications used to treat schizophrenia, reduced access to healthcare in patients with schizophrenia, and poor diet/sedentary lifestyle. A recent meta-analysis in Lancet Psychiatry by the lead proposal author has shown that participants with first-episode psychosis (the first time a generally younger person goes to hospital with symptoms of schizophrenia) are more likely to have signs of early diabetes than healthy matched controls. All participants in the meta-analysis had very limited exposure to antipsychotic medication, reducing the impact of the potential confounder.
There is a growing body of work showing that both schizophrenia and diabetes may indeed feature an inflammatory component, and this may be shared. The inflammatory component in patients with a potentially earlier form of schizophrenia - termed 'psychotic experiences', has previously been examined in the ALSPAC cohort by one of the proposal authors, with positive findings.
It would therefore be of great interest to examine whether any possible disease link begins earlier than the 'first psychotic episode', by examining patients suffering 'psychotic experiences'. With the noted possible inflammatory disease links between diabetes and schizophrenia, it would be of great interest to examine whether those with psychotic experiences and abnormal inflammatory profiles also have abnormal glycaemic indices.
Finally, it would be interesting to examine whether markers of early diabetes can predict the onset of psychotic experiences.