Soo Hyun Rhee and Angelica Ronald (eds.)Advances in Behavior GeneticsBehavior Genetics of Psychopathology201410.1007/978-1-4614-9509-3_5
© Springer Science+Business Media New York 2014
5. Sleep and Psychopathology: Quantitative and Molecular Genetic Research on Comorbidity
(1)
Department of Psychology, Northumbria Centre for Sleep Research, Faculty of Health and Life Sciences, Northumbria University, Newcastle upon Tyne, NE1 8ST, UK
(2)
Department of Psychology, Goldsmiths, University of London, London, UK
Abstract
Sleep problems are common in children, adolescents and adults and are often associated with psychiatric difficulties. Whilst a great deal of behavioural genetic research has focussed on understanding the genetic and environmental factors contributing exclusively to sleep disturbances and to psychopathology, an emerging body of literature centres on understanding the factors that contribute to the overlap between sleep disturbances and symptoms of psychopathology. The purpose of this chapter is to review some of the most important findings from behavioural genetic research which have assessed the aetiology of sleep-wake characteristics, sleep disorders and associations between sleep disturbances and psychopathology. We present findings from the following: (1) family studies of insomnia which provide an insight into possible familial factors involved in sleep disturbances; (2) quantitative genetic studies assessing the heritability of normal sleep patterns and specific sleep disorders; (3) quantitative genetic studies assessing the overlap in genetic and environmental influences between multiple sleep phenotypes and symptoms of psychopathology; (4) molecular genetic research, with a focus on associations between genes common to both sleep-related and psychopathology-related symptoms; (5) areas of research other than behavioural genetics which have provided an insight into possible shared environmental risk factors for sleep disturbances and psychopathology; and (6) studies positing sleep characteristics as plausible endophenotypes of psychopathology. Finally, we present a research agenda for behavioural geneticists within the sleep arena.
Springer Behaviour Genetic Series
Introduction
Sleep disturbances are an ever increasing problem in today’s society and are common in children, adolescents and adults (Carlson & Cordova, 1999; Ohayon, 2002). Sleep disturbances take many forms (e.g. parasomnias, characterised by abnormalities that occur during the night, such as sleepwalking; and dyssomnias, characterised by difficulties in the duration, timing or quality of sleep, including problems such as insomnia). This chapter largely focusses on sleep disturbances characterised by insomnia-type symptoms which are the difficulties most commonly associated with psychiatric difficulties. Note that throughout this chapter, the term ‘sleep disturbances’ is used to encompass various sleep problems/difficulties, as much of the literature in this area relies on sleep symptom reporting rather than reporting of specific sleep disorders.
Epidemiological data suggest that at least some form of sleep disturbances affects approximately a third of the general population at any one time (Ohayon, 2002). According to some estimations, approximately 6–10 % of the adult population meet the diagnostic criteria for insomnia according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (Morin, LeBlanc, Daley, Gregoire, & Merette, 2006; Ohayon, 1997; Ohayon & Reynolds, 2009). The International Classification of Sleep Disorders (ICSD-2) describes over 30 sleep disorders where insomnia symptoms are also present (American Academy of Sleep Medicine, 2005). Furthermore, the DSM-IV describes no fewer than 25 psychological disorders where difficulties with sleep are listed as a symptom or where problems with sleep are a consequence of another underlying condition (American Psychiatric Association, 2000).
Large-scale community samples have indicated that sleep disturbances are most commonly comorbid with psychiatric disorders such as anxiety and depression (Buysse et al., 1994; Ford & Kamerow, 1989; Weyerer & Dilling, 1991). In one study, around 24 % of individuals with insomnia also exhibited a comorbid anxiety disorder, and around 14 % comorbid major depression as defined by the diagnostic interview schedule (DIS) (Ford & Kamerow, 1989). However, in another study, around 61 % of individuals with insomnia also exhibited an anxiety disorder, and around 69 % a depressive disorder as defined by the International Classification of Diseases (ICD) (Weyerer & Dilling, 1991). In comparison, Buysse et al. (1994) compared comorbidity rates between anxiety and depression as diagnosed by the ICSD, DSM-IV and ICD-10 diagnostic systems. Comorbidity rates ranged from 32 % for sleep disorders associated with mood disorder, 44 % for insomnia related to another mental disorder, to 62 % exhibiting non-organic insomnia (according to the ICSD, DSM-IV and ICD-10, respectively). Sleep disturbances are also often comorbid with a range of other psychiatric and psychological disorders such as alcohol and substance abuse (Breslau, Roth, Rosenthal, & Andreski, 1996; Ford & Kamerow, 1989), personality and psychotic disorders (Nowell et al., 1997) including schizophrenia (Weissman, Greenwald, Nino-Murcia, & Dement, 1997), eating disorders (Nobili et al., 2004), obsessive-compulsive disorder (Insel et al., 1982) and somatisation disorder (Weissman et al., 1997). Indeed, approximately a third of treatment-seeking individuals with a primary complaint of insomnia experience some degree of concurrent psychopathology1 (Morin & Ware, 1996). Of course, sleep disturbances other than insomnia are often comorbid with psychopathology, particularly depression (see Franzen & Buysse, 2008, for a review). Studies have shown elevated rates of depression in numerous sleep disorders including obstructive sleep apnoea (Reynolds et al., 1984)—a disorder characterised by frequent pauses in breathing during sleep (American Academy of Sleep Medicine, 2005); narcolepsy (Reynolds, Christiansen, Taska, Coble, & Kupfer, 1983)—a disorder characterised by frequent unintentional lapses into sleep which are often coupled with cataplexy (American Academy of Sleep Medicine, 2005); the sleep-related ‘periodic limb movement disorder’ restless leg syndrome (Picchietti & Winkelman, 2005); and circadian rhythm disorders such as delayed sleep phase syndrome (Regestein & Monk, 1995). In addition to depression, sleep disruption is a typical feature of other forms of psychopathology, such as schizophrenia (Wulff, Gatti, Wettstein, & Foster, 2010). Schizophrenia is often accompanied by disruption of the circadian system, resulting in a cascade of negative effects on the sleep-wake cycle including phase shifts, fragmented nocturnal sleep, frequent daytime naps, hyposomnia and hypersomnia (Wulff et al., 2010).
However, understanding of the direction of effects between sleep and psychopathology is complicated by the fact that different patterns emerge from studies assessing symptoms in childhood and adulthood (Ford & Kamerow, 1989; Gregory, Caspi, et al., 2005) and because associations are likely to be bidirectional (Franzen & Buysse, 2008). Several longitudinal studies have suggested that sleep disturbances may be a risk factor for the development of chronic psychiatric disorders, such as depression and anxiety (Breslau et al., 1996; Gregory, Rijsdijk, Lau, Dahl, & Eley, 2009; Johnson, Roth, & Breslau, 2006; Livingston, Blizard, & Mann, 1993; Neckelmann, Mykletun, & Dahl, 2007; Perlis, Giles, Buysse, Tu, & Kupfer, 1997), panic disorder (Weissman et al., 1997) in adults, as well as alcohol and drug use from childhood to adolescence (Wong, Brower, Fitzgerald, & Zucker, 2004; Wong, Brower, Nigg, & Zucker, 2010) and aggression and attention problems in children (Gregory & O’Connor, 2002). For example, in a prospective longitudinal study following 490 adoptive and non-adoptive children from ages 4 to 15 years, parent-reported sleep difficulties were concurrently associated with anxiety/depression, attention and aggressive problems at successive time points over the 11-year period (Gregory & O’Connor, 2002). Furthermore, prior sleep difficulties at age 4 years predicted behavioural/emotional problems in mid-adolescence. Additionally, the associations between sleep difficulties and depression and anxiety significantly increased across this developmental time period. Similarly, using prospective epidemiological data from the Dunedin Multidisciplinary Health and Development Study, Gregory and colleagues reported that persistent childhood sleep disturbances predicted anxiety disorder in adulthood, but not depression (Gregory, Caspi, et al., 2005). This study suggests that childhood sleep disturbances could share underlying risk factors with adult anxiety, whilst factors influencing depression may be somewhat distinct. It is possible that increased hyperarousal in childhood could explain the association with adult anxiety and not depression.
Studies using relatively short follow-up periods in adults, however, have suggested that sleep disturbances forecast later depression. In a US study, which assessed the presence of sleep and psychiatric complaints at two time points, individuals with insomnia at baseline were 40 times more likely to develop a new major depression and six times more likely to develop an anxiety disorder at follow-up 1 year later compared to those without insomnia (Ford & Kamerow, 1989). Furthermore, individuals whose insomnia symptoms had reduced at follow-up were also less likely to experience new-onset depression. In a study in the UK, presence of insomnia at baseline was associated with approximately three times greater risk of developing depression 1 year later and twice the risk of developing anxiety disorder (Morphy, Dunn, Lewis, Boardman, & Croft, 2007). Studies using longer follow-up periods (ranging from 3 to 30 years) demonstrate a similar pattern (e.g. Breslau et al., 1996; Chang, Ford, Mead, Cooper-Patrick, Klag, 1997). A recent meta-analysis identified 21 longitudinal studies assessing the prediction of depression from prior insomnia, finding a twofold risk of developing first-onset depression in individuals experiencing prior insomnia compared to individuals with no sleep disturbances (Baglioni et al., 2011). Whilst these studies point to the possibility that sleep disturbances in adults play a role in the pathogenesis of depression and anxiety, evidence also suggests that depressive symptoms are an important risk factor for the development and persistence of sleep disturbances (Patten, Choi, Gillin, & Pierce, 2000; Quan et al., 2005).
Whilst a great deal of behavioural genetic research has focussed on understanding the genetic and environmental factors contributing exclusively to sleep disturbances (e.g. see Barclay & Gregory, 2012, for a review) and to psychopathology (see Plomin, DeFries, McClearn, & McGuffin, 2008), an emerging body of literature centres on understanding the factors that contribute to the overlap between sleep disturbances and symptoms of psychopathology. Using multivariate quantitative genetic techniques, researchers are able to estimate the following: (1) the relative contribution of genetic and environmental influences on each phenotype univariately, (2) the extent to which genetic and environmental influences contribute to the associations between phenotypes, and (3) the overlap in the genetic and environmental influences between each phenotype measured. This latter point is a key piece of information derived from quantitative genetic studies as knowledge of the overlap in the genetic influences between phenotypes has the potential to guide molecular genetic research aimed at identifying candidate genes associated with such phenotypes. Substantial genetic overlap between sleep disturbances and symptoms of psychopathology would suggest that similar candidate genes should be investigated in relation to both phenotypes. Conversely, little overlap would suggest that candidate genes unique to each phenotype should be sought. Likewise, knowledge of the overlap in the environmental influences between phenotypes could provide clues as to where to aim the search for environmental risk factors.
The purpose of this chapter is to review some of the most important findings from behavioural genetic research which have assessed the aetiology of sleep-wake characteristics, sleep disorders as well as associations between sleep disturbances and psychopathology. To this end, we present findings from the following: (1) family studies of insomnia which provide an insight into possible familial factors involved in sleep disturbances; (2) quantitative genetic studies assessing the heritability of normal sleep patterns as well as specific sleep disorders; (3) quantitative genetic studies assessing the overlap in genetic and environmental influences between multiple sleep phenotypes and symptoms of psychopathology; (4) molecular genetic research, with a focus on associations between genes common to both sleep-related and psychopathology-related symptoms; (5) areas of research other than behavioural genetics which have provided an insight into the possible shared environmental risk factors for sleep disturbances and psychopathology; and (6) studies positing sleep characteristics as plausible endophenotypes of psychopathology. Finally, we present a research agenda for behavioural geneticists within the sleep arena. We emphasise the need to investigate the aetiology of the following: associations between a broader scope of sleep disorders and psychopathological disturbances, longitudinal associations between sleep and psychopathology as well as exploiting recent advances in the genetics field by exploring epigenetic processes and utilising newer generation sequencing methods to further our understanding of the genetics of sleep and associated phenotypes.
Family Studies of Insomnia
Population-based and clinical samples assessing insomnia have repeatedly demonstrated increased risk of self-reported insomnia symptoms in individuals with a family history of insomnia, reflecting possible genetic and/or shared environmental effects (Bastien & Morin, 2000; Beaulieu-Bonneau, LeBlanc, Merette, Dauvilliers, & Morin, 2007; Dauvilliers et al., 2005; Drake, Scofield, & Roth, 2008; Hauri & Olmstead, 1980). One study reported that 72.7 % of individuals with insomnia (compared to only 24.1 % of control participants with no insomnia) had a positive family history of insomnia (Dauvilliers et al., 2005). In several studies, this trend has been shown to be stronger in individuals with an early age of onset in childhood or adolescence and that mothers are often the most frequently afflicted first-degree relative (Bastien & Morin, 2000; Beaulieu-Bonneau et al., 2007; Dauvilliers et al., 2005; Hauri & Olmstead, 1980). Given that accumulating evidence suggests a female predisposition to insomnia (with a meta-analysis suggesting that females present 1.41 times greater risk of insomnia compared to males, see Zhang & Wing, 2006), this latter finding is perhaps not surprising, but is of particular interest as it sheds light on the possible mode of inheritance (suggesting a possible role for X-linked genes or a parent of origin effect on imprinting). Such a finding could also reflect the possibility that mothers may be particularly important environmental role models.
Quantitative Genetic Research on Normal Sleep-Wake Characteristics and Sleep Disorders
Twin Studies
Normal Sleep-Wake Characteristics
A wealth of behavioural genetic research has focussed on understanding factors contributing to variability in normal sleep-wake characteristics (such as sleep stage organisation and characteristics of the sleep electroencephalogram) and sleep disorders (such as insomnia and circadian rhythm sleep disorders). Twin studies using objective methods to understand the aetiology of sleep-wake behaviour have typically used polysomnography (PSG) in order to characterise sleep. PSG typically involves the measurement of electrical activity using electroencephalography (EEG), electrooculography (EOG) and electromyography (EMG) to examine indices of sleep staging and microstructure during sleep. Using PSG, several lines of evidence converge on the finding that genetic influences (as indicated by greater MZ twin correlations in comparison to DZ twins) are important for determining particular characteristics of, and the amount of time spent in, stage 2 sleep, slow-wave sleep and rapid eye movement (REM) sleep (Ambrosius et al., 2008; Chouvet, Blois, Debilly, & Jouvet, 1983; Gould, Austin, & Cook, 1978; Hori, 1986; Linkowski, Kerkhofs, Hauspie, Susanne, & Mendlewicz, 1989; Webb & Campbell, 1983). Indeed, one study suggested that the EEG spectral composition of sleep appears to be one of the most highly heritable traits with heritability estimated at around 96 % (De Gennaro et al., 2008).
The majority of work focussing on sleep-wake characteristics in the normal range, however, has focussed on subjective reports of sleep duration, timing and quality. In a sample of 18-month-old twin pairs, shared environmental effects were found to contribute largely to sleep duration (Brescianini et al., 2011). A similar pattern of results has been observed in a sample of school-aged children, where shared and non-shared environmental influences were important for child-reported sleep duration, rather than genetics (Gregory, Rijsdijk, & Eley, 2006). Interestingly, in the same study, it was found that sleep duration reported by parents was attributable largely to genetic factors. In adults, the heritability of phenotypes such as sleep duration, sleep onset latency, sleep timing and sleep efficiency has been estimated between around 20 and 46 % in numerous studies (Barclay, Eley, Buysse, Rijsdijk, & Gregory 2010; Boomsma, van Someren, Beem, de Geus, & Willemsen, 2008; de Castro, 2002; Heath, Kendler, Eaves, & Martin, 1990; Partinen, Kaprio, Koskenvuo, Putkonen, & Langinvainio, 1983; Watson, Buchwald, Vitiello, Noonan, & Goldberg, 2010). For example, in a sample of 2,238 monozygotic (MZ) and 4,545 dizygotic (DZ) young adult twins from the Finnish Twin Cohort, Partinen and colleagues were the first to report on the heritability of subjective sleep duration and sleep quality estimating overall heritability of both phenotypes at 44 % (Partinen et al., 1983). However, when comparing participants aged 18–24 years with those aged 25+, the authors found that genetic influences appeared to be smaller in the younger aged cohort. Our own research has reflected this finding in a sample of young adult twins (with a mean age of 22 years) from the G1219Twins study, where shared and non-shared environmental influences were found to influence sleep duration with no evidence of a role for genetics (Barclay, Eley, Buysse, Rijsdijk, et al., 2010). It is possible that this lack of genetic effect on sleep duration reflects the fact that the amount of time one sleeps is largely under voluntary control. Furthermore, in young adults, there may be social pressures to stay out late and sleep late in the day, thus attenuating the influence of genes.
Twin studies assessing the subjective timing of sleep have focussed on indices of circadian rhythms such as diurnal preference. Diurnal preference refers to one’s preference towards morningness or eveningness, and several studies, including our own, converge on the finding that genetic influences account for around half of the observed variability in this phenotype (Barclay, Eley, Buysse, Archer, & Gregory, 2010; Hur, 2007; Hur, Bouchard, & Lykken, 1998; Koskenvuo, Hublin, Partinen, Heikkila, & Kaprio, 2007). Using independent samples of adolescents (mean age 17 years) and adults (mean age 48 years), Vink and colleagues reported on the genetic overlap on diurnal preference across time (Vink, Groot, Kerkhof, & Boomsma, 2001). Whilst the overall proportion of variance accounted for by genetic influences was comparable across age groups, the genetic correlation between them was only around 0.50, suggesting that substantially different genetic factors are important for diurnal preference in adolescence as compared to middle-aged adulthood.
Sleep Disorders
Numerous twin studies have aimed to determine the aetiology of a range of clinical sleep disorders in childhood and adulthood including parasomnias such as sleepwalking, bruxism (teeth grinding), sleep talking, nightmares, night terrors and enuresis (bed-wetting) and dyssomnias such as primary insomnia, narcolepsy, sleep-disordered breathing including obstructive sleep apnoea and the sleep-related ‘periodic limb movement disorder’ restless legs syndrome.
Parasomnias
In children, Gregory and colleagues reported that around half of the variability in a composite measure of parasomnias in 8-year-old twins (incorporating behaviours such as teeth grinding and sleep talking) was attributable to genetic factors (Gregory, 2008). Other studies report similar conclusions relating to the role of genetic factors when focussing on specific parasomnias in children. Perhaps the most extensive work within this area comes from the Finnish Twin Cohort, where heritability has been estimated between 40 and 44 % for disorders including sleepwalking (Hublin, Kaprio, Partinen, Heikkila, & Koskenvuo, 1997), bruxism (Hublin, Kaprio, Partinen, & Koskenvuo, 1998b), sleep talking (Hublin, Kaprio, Partinen, & Koskenvuo, 1998c), nightmares (Hublin, Kaprio, Partinen, & Koskenvuo, 1999) and enuresis (Hublin, Kaprio, Partinen, & Koskenvuo, 1998a). As well as including data on retrospective reports of childhood parasomnias (as described above), the Finnish Twin Cohort also contains a wealth of data on adulthood parasomnias, making it possible to examine longitudinal associations. Generally, studies of adulthood parasomnias are somewhat in line with those in childhood, finding that genetic influences account for between 36 and 53 % of adult symptoms (Hublin et al., 1998b, 1998c, 1999). What is interesting from this series of studies, however, is that the authors noted high genetic correlations between these parasomnias over time (ranging from rA = 0.75 − 0.95), suggesting that similar genetic factors contribute to the stability of disturbances. Conversely, overlap in the non-shared environmental influences over time was somewhat lower (ranging from rE = 0.57 − 0.75) suggesting that, whilst there appears to be some overlap in the environmental influences on these parasomnias from childhood to adulthood, there also appears to be evidence for time-specific influences.
Dyssomnias
In children, quantitative genetic research on dyssomnias has tended to conceptualise sleep disturbances more generally rather than differentiating between symptoms. Van den Oord and colleagues first reported on the heritability of sleep disturbances assessed by the Child Behavior Checklist (CBCL) in a sample of 3-year-old twins (Van den Oord, Boomsma, & Verhulst, 2000). The authors noted that genetic influences accounted for 61 % of variability in sleep disturbances in this age group. Using data on a composite measure of dyssomnias in a sample of 8-year-old twins, Gregory reported a similar heritability estimate in the region of 70 % (Gregory, 2008). Within this study, the same sample of twins was followed up at age 10 years, making it possible to assess longitudinal associations. A significant genetic correlation between dyssomnias at age 8 and dyssomnias at age 10 (rA = 0.46) suggested that although to some extent the same genetic factors contribute to the stability of sleep disturbances over time, new genetic influences also come into play.
The heritability of dyssomnias has also been the focus of several adult twin studies. Contrary to the work in children, studies of adult dyssomnias have focussed on specific problems. One study investigated specific insomnia symptoms such as ‘trouble falling asleep’, ‘trouble maintaining sleep’, ‘waking several times per night’, ‘early morning awakening’ and ‘waking up feeling tired and worn out’ in a large sample of adult male twins (McCarren, Goldberg, Ramakrishnan, & Fabsitz, 1994). Genetic influence contributed between 21 and 42 % to these symptoms. In another study, heritability of insomnia (assessed by one question assessing trouble falling asleep or maintaining sleep) was estimated at 57 % in a sample of adults (Watson, Goldberg, Arguelles, & Buchwald, 2006). In both studies, the remaining source of variance was attributable to the non-shared environment.
Several twin studies have focussed on the heritability of other dyssomnias including narcolepsy (Kaprio, Hublin, Partinen, Heikkila, & Koskenvuo, 1996), sleep-disordered breathing (Carmelli, Colrain, Swan, & Bliwise, 2004; Carmelli, Swan, & Bliwise, 2001), obstructive sleep apnoea and restless legs syndrome (Desai, Cherkas, Spector, & Williams, 2004). Although figures vary somewhat between disorders, the evidence generally points to the finding that genetic and non-shared environmental factors are important determinants of most disorders of sleep (for a more comprehensive review of quantitative genetic studies of sleep and sleep disorders, see Barclay & Gregory, 2012).
Quantitative Genetic Research on the Associations Between Sleep Disturbances and Psychopathology in Children
Despite the evident interest from large-scale epidemiological studies in investigating concurrent and longitudinal associations between sleep disturbance and symptoms of psychopathology in children (see above), relatively little work has focussed on understanding the aetiology of these associations. The few twin studies that have been conducted to assess such associations in children have focussed on the co-occurrence and overlap in the aetiological influences between broadly defined sleep disturbances and behavioural and emotional difficulties. In one study, Van den Oord and colleagues assessed concurrent associations between parent-reported oppositional, withdrawn/depressed, aggressive, anxious, overactive and sleep disturbances in a sample of 446 MZ and 912 DZ 3-year-old twin pairs assessed by the CBCL (Van den Oord et al., 2000). The authors reported moderate phenotypic correlations between sleep and other behavioural syndromes (r = 0.25 − 0.39) with the highest correlations between sleep disturbances and oppositional problems, followed by anxious problems. When assessing the aetiology of the phenotypic associations between syndromes, genetic influences were non-existent for the association between sleep disturbances and withdrawn/depressed symptoms. However, for the remaining phenotypic associations, shared environmental influences were the major contributor whilst the contribution of non-shared environmental influences was relatively small. When comparing the overlap in the aetiological influences between syndromes, the authors observed no shared environmental effects unique to the individual syndromes, but rather suggested that the same shared environmental factor was contributing to all. On the other hand, non-shared environmental influences demonstrated greater specificity between syndromes. Likewise, genetic effects were largely syndrome specific, particularly with regard to sleep disturbances and withdrawn/depressed problems, exhibiting 72 % and 62 % unique genetic effects, respectively.
Such findings highlight that the family-wide environment has a significant influence on the development of sleep as well as behavioural/emotional problems in vulnerable young children. This emphasises the importance of specifying shared environmental influences that increase risk for such problems. To this end, Gregory and colleagues addressed the contribution of a variety of putative shared environmental influences on both sleep disturbances and anxiety, finding particularly strong associations with family disorganisation and maternal depression (Gregory, Eley, O’Connor, Rijsdijk, & Plomin, 2005). Furthermore, both family disorganisation and maternal depression accounted for a substantial proportion of the phenotypic association between sleep disturbances and anxiety, with the remaining variance in the association largely due to unspecified family influences and genetics.
In contrast to the studies reviewed above which highlighted the importance on the shared environment on the associations between sleep and anxiety/depression in early childhood, in a sample of 8-year-old twins, Gregory and colleagues reported substantial genetic overlap between sleep disturbances (parent report) and depression symptoms (child self-report). This association was almost entirely explained by genetic influences rather than the shared environment (Gregory, Rijsdijk, Dahl, McGuffin, & Eley, 2006). These contrasting results could reflect developmental changes in the aetiology of the associations between sleep and depression from pre-school to school-aged children. For example, genes that play only a small part in sleep and depression in early childhood may become ‘switched on’ (hence more important) later on. However, using data from the same sample of twins at age 10 years, Gregory reported that whilst genetic influences appeared most important in explaining the phenotypic correlation at age 8 years, at age 10 years the non-shared environment appeared to carry most of the effect (Gregory et al., 2009). It is possible that this is explained by the increase in non-shared environmental influences that children are exposed to as they get older—a finding which is typical within the field of behavioural genetic research more generally (see Plomin et al., 2008).
Gregory and colleagues also investigated the direction and aetiology of the associations between sleep disturbances and depression longitudinally (Gregory et al., 2009). Prior sleep disturbances at age 8 years predicted depression at age 10 years, whilst the converse was not true. Whilst this longitudinal association was of small effect (r = 0.20), it was to some extent explained by shared genetic factors. That is, genetic effects that contributed to sleep disturbances at age 8 years were carried over to influence depression symptoms at age 10. Findings such as this have the potential to inform the development of prevention and/or intervention programmes for depression and are consistent with the possibility that early treatment of sleep disturbances may be protective against later problems. Furthermore, this finding suggests that similar genes may be responsible for both sleep disturbances and depression symptoms in childhood. Thus, investigation of genetic variants that increase susceptibility to both sleep disturbances and depression would appear to be worthy candidates for exploration (a discussion of plausible candidate genes involved in sleep and psychopathology is provided later in this chapter). A recent study assessed the heritability and overlap of insomnia symptoms, overanxious disorder and depression in a sample of twins aged between 8 and 16 years (Gehrman et al., 2011). Symptomology was assessed by clinician ratings from a clinical interview administered to both parents and youth. In line with much of the adult literature on the heritability of insomnia (e.g. see Barclay & Gregory, 2012, for a review), genetic factors accounted for around a third of variance in sleep disturbances (on a combined measure of parental and youth reports), with the remaining source of variance due to the non-shared environment. In multivariate models, the authors reported that non- shared environmental influences were to some extent unique to insomnia. This is perhaps not surprising, since the types of environmental factors that may contribute to sleep disturbance (e.g. noise due to living near a busy road) are unlikely to contribute to anxiety and/or depression-type symptoms. Genetic influences, on the other hand, were found to be entirely shared with those influencing anxiety and depression. Whilst this is consistent with the findings from Gregory and colleagues, these findings show support for genetic overlap between clinician-rated insomnia and psychopathology in children and adolescence using a combination of parent-reported and child-reported data.
However, what is apparent from studies assessing sleep in children is the importance of considering the method of assessment. Two of the studies reviewed above investigated differences in child- and parent-reported data on sleep (Gehrman et al., 2011; Gregory, Rijsdijk, & Eley, 2006). Both studies found that, as compared to parent report, child report and adolescent report demonstrated more frequent sleep disturbances and yielded higher estimates of the non-shared environment. Explanations for the discrepancy between child-reported and parent-reported data on the prevalence of sleep disturbance include the possibility that (1) parents lack an awareness of their child’s sleeping patterns and consequently underestimate the extent of sleep disturbance and (2) children’s reports of their own sleep are inaccurate. Furthermore, it is worth highlighting that when different reporters rate each twin within a pair (such as in self-report), the non-shared environmental component is increased compared to when the same reporter rates each twin (as in parent report). In contrast, parent reports may result in higher genetic estimates if they only detect and report more severe cases which may be more genetically influenced. This highlights the need to consider the method of assessment when interpreting research on childhood sleep patterns.
Quantitative Genetic Research on the Associations Between Sleep Disturbances and Psychopathology in Adults
Genetically speaking, several lines of evidence point to the possibility that anxiety and depression are different manifestations of the same underlying genetic factors (e.g. Middeldorp, Cath, Van Dyck, & Boomsma, 2005). There is now a growing body of evidence demonstrating genetic overlap between anxiety, depression and insomnia suggesting that sleep disturbances may also, to some extent, be a part of this cluster (see below). Behavioural genetic research investigating associations between sleep and psychopathology in adults has tended to conceptualise sleep disturbances, anxiety and depression subjectively and has largely focussed on individual differences in the normal range rather than in samples with clinically significant disorders. Our own research investigated the overlap in the aetiological influences between subjective reports of sleep quality (using the Pittsburgh Sleep Quality Index (PSQI): Buysse, Reynolds, Monk, Berman, & Kupfer, 1989), anxiety and depression from a population-based sample of young adult twins and siblings (aged between 18 and 27 years; mode age = 22 years) (Gregory, Buysse, et al., 2011). Our findings converge on those observed in children and adolescence. We reported substantial genetic overlap between sleep disturbance and anxiety (rA = 0.58) and sleep disturbance and depression (rA = 0.68). The phenotypic associations between sleep disturbances and anxiety and depression were moderate (r = 0.39 and r = 0.50, for anxiety and depression, respectively) but were largely accounted for by genetic factors (74 % and 58 %, respectively). Non-shared environmental correlations between sleep, anxiety and depression, whilst significant, demonstrated far greater specificity, in line with the findings reported above in children and adolescence (Gehrman et al., 2011).
Contrastingly, Kendler and colleagues assessed the specificity of the genetic influences on subjective sleep difficulties associated with worrying or feeling miserable (assessed by the two items, ‘worrying kept me awake’ and ‘miserable, difficulty with sleep’), anxiety and depression symptoms in a large sample of adult twins (aged between 18 and 88 years) from the Australian Twin Registry (Kendler, Heath, Martin, & Eaves, 1987). Using a factor analytic approach to identify common factors amongst symptoms, the authors noted that whilst anxiety and depression appeared to stem from a common genetic factor, a separate independent genetic factor loaded onto the sleep items. Thus, the authors conclude that the genetic influences that may contribute to sleep disturbances are largely separable from those influencing anxiety and depression. Likewise, environmental influences on sleep symptoms were largely separable from those on anxiety and depression. As part of the same survey, Heath and colleagues reported on associations between sleep disturbances, defined using John’s Sleep Questionnaire—a more comprehensive questionnaire tapping into a range of sleep difficulties, such as the overall quality, variability and duration of sleep, as well as initial insomnia, night waking and daytime napping—and anxiety and depression (Heath, Eaves, Kirk, & Martin, 1998). Using DeFries-Fulker linear regression techniques (a method designed to assess heritability of extreme scores using data from probands) rather than structural equation modelling, the authors demonstrated that genetic effects on sleep disturbance were to some extent mediated through associations with anxiety and depression in females and anxiety in males. However, despite this evidence for shared genetic effects between sleep and psychopathology, much of the variance in sleep disturbance was explained by sleep-specific genetic effects independent of these intervening variables. In combination, these studies provide mixed evidence as to the overlap in the genetic factors on sleep disturbances, anxiety and depression in adulthood. It is likely that there are to some extent shared genetic effects between sleep and psychopathology, whilst some evidence suggests that symptom-specific effects are also important. Investigation of genetic variants associated with both sleep disturbances and symptoms of psychopathology may provide an insight into the mechanisms underlying their co-occurrence.
Knowledge of genetic overlap could be used to differentiate insomnia with no other presenting symptoms and insomnia with clinically comorbid conditions. The presence or absence of sleep disturbances in depression, such as early morning awakening, difficulty initiating or maintaining sleep, may signal different neurobiological processes underlying sleep disturbances presenting exclusively and sleep disturbances presenting with comorbidities. The possibility that there are also additional genetic effects on insomnia independent of psychopathology, however, may lead to the discovery of genetic effects implicated exclusively in insomnia without comorbidities. Such information may be informative for refining current diagnostic classification systems for insomnia, such as DSM-5 which is scheduled for publication in 2013 (American Psychiatric Association, 2010).
Studies investigating the aetiological overlap between objective measures of sleep with anxiety and depression are, to our knowledge, non-existent. Studies comparing laboratory-based methods and self-reports to assess sleep have consistently shown that subjective sleep complaints do not always correlate with polysomnographically defined problems (see Gregory, Cousins, et al., 2011, for an example of this phenomenon in adolescents). Generally, individuals with insomnia significantly overestimate their sleep onset latency and underestimate the quantity and quality of their sleep compared to objective data (Carskadon et al., 1976; Edinger & Fins, 1995; Manconi et al., 2010)—a pattern which has also been found in healthy adults (Baker, Maloney, & Driver, 1999). Determining the comorbidity between objectively defined sleep disturbances (by techniques such as actigraphy or polysomnography) and anxiety and depression is therefore essential in order to determine whether the pattern of overlap in the aetiological influences on subjectively defined sleep disturbances is also reflected in objective measures. Such an investigation will also enable us to determine whether the associations between psychopathology and subjectively defined problems are simply a result of a possible pessimistic and inaccurate reporting style inherent in subjectively poor sleepers, individuals with insomnia and depressed individuals.
Molecular Genetic Research on the Associations Between Sleep Disturbances and Psychopathology
Molecular genetic research aimed at identifying candidate genes associated with sleep and psychopathology has largely centred on genes implicated in monoamine neurotransmission, particularly the serotonin and monoamine oxidase systems, and a core group of CLOCK genes. A review of some of the most important findings in relation to sleep, psychopathology and these candidate genes is presented below.
Serotonin, Sleep and Psychopathology
The neurotransmitter, serotonin (5HT), has been shown to have numerous functions, modulating sleep-wake activity cycles and circadian rhythm regulation as well as cognition, mood, emotion, motor function and appetite (Adrien, 2002; Cools, Roberts, & Robbins, 2008; Portas, Bjorvatn, & Ursin, 2000; Ursin, 2002). Serotonin is a prime candidate to investigate mechanisms underlying the association between sleep and psychopathology since extensive research has linked genes in the serotonin system with numerous forms of psychopathology including depression, anxiety and their various subtypes (see below).
Evidence for a relationship between sleep and psychopathology comes from numerous studies which demonstrate that these phenotypes share common neurobiological substrates. For example, the sleep-wake cycle is largely governed by the serotonergic system. Serotonergic neurons are thought to be most active during wakefulness, contribute to the build-up of sleep propensity, and deactivate in the transition from wake to sleep onset (Adrien, 1995). Furthermore, 5HT has been associated with sleep stage characteristics and sleep regulation. In one study, reducing 5HT2 receptor activity (receptors that bind serotonin) increased slow-wave sleep (SWS) and slow-wave activity (SWA) (Landolt et al., 1999). Additionally, administration of selective serotonin reuptake inhibitors (SSRIs) has been shown to increase stages 1 and 2 sleep at the expense of SWS and REM sleep and is associated with increased sleep latency, yet greater sleep disturbances and poorer sleep quality assessed objectively by polysomnography (Oberndorfer, Saletu-Zyhlarz, & Saletu, 2000).
The role of the serotonergic system in mood states, particularly depression and anxiety, whilst unequivocal, is complex. Depressed patients show decreased serotonergic activity, and depressive symptoms are often effectively treated with SSRIs to enhance serotonin neurotransmission (see Maes & Meltzer, 1995, for a review). The role of serotonin in anxiety traits is evidenced by the fact that the benzodiazepines, drugs which decrease serotonin function, have anxiety-reducing properties (Charney, Woods, Goodman, & Heninger, 1987). Yet, SSRIs have also been shown to have therapeutic effects for many anxiety disorders (Baldwin et al., 2005). Understanding of the complex interaction between sleep and psychopathology can be achieved by examining research investigating the interplay between these phenotypes. For example, following acute sleep deprivation, particularly the suppression of REM sleep, a short-term enhancement of mood regulation and a reduction of depressive symptoms are often observed (Van den Hoofdakker, 1997; Wirz-Justice & Van den Hoofdakker, 1999). It is thought that sleep deprivation mimics the effects of antidepressant drugs, leading to increases in serotonin, and consequently alleviating depressive symptoms. However, other studies have suggested that the beneficial effects of sleep deprivation on depressive symptoms may be dependent on chronotype, thus providing further evidence for an interaction with circadian processes (Selvi, Gulec, Agargyn, & Besiroglu, 2007). Furthermore, instability of mood in depressed patients may be explained by an instability of phase relationships between the homeostatic sleep drive and processes controlled by the circadian system (Wirz-Justice & Van den Hoofdakker, 1999).
Many functional genetic polymorphisms of serotonin genes have been identified in the literature (see D’Souza & Craig, 2008, for a review). Perhaps the most abundant candidate gene association studies linking genes shared between sleep and psychopathology have focussed on the serotonin transporter gene-linked polymorphic region (5HTTLPR). A common 44 base pair (bp) deletion of 5HTTLPR, which constitutes a ‘short’ (‘S’) allele, reduces transcriptional activity of the 5HTT gene, thus leading to decreased serotonin reuptake. A handful of studies have investigated the role of 5HTTLPR in relation to sleep, and those to date have provided contradictory results. One study found a significantly greater frequency of ‘S’ alleles in a group of insomnia patients as compared to controls (Deuschle et al., 2010). Contrastingly, our own research, which investigated the association between 5HTTLPR genotype and poor sleep quality (assessed by the PSQI) in a sample of young adult twins and siblings, found that ‘long-long’ (‘LL’) homozygotes conferred greater risk for sleep disturbances as compared to carriers of at least one ‘S’ allele (Barclay, Eley, Mill, et al., 2011). Such inconsistency in findings in relation to 5HTTLPR is also reflected in the research on psychopathology. For example, the ‘S’ allele has repeatedly been associated with increased anxiety, neuroticism, depressive symptoms and impaired response to antidepressant medication (Collier et al., 1996; Lesch et al., 1996; Serretti, Kato, De Ronchi, & Kinoshita, 2007). In contrast, one study indicated that the ‘L’ allele confers greater risk for anxiety-related traits in 13–14-year-olds (Jorm et al., 2000). Other studies, including 2 meta-analyses, have failed to provide any support for an association between 5HTTLPR and depression and anxiety-related traits (e.g. Jorm et al., 1998; Munafo, Freimer, et al., 2009; Schinka, Busch, & Robichaux-Keene, 2004).
Another area of research providing mixed evidence focusses on the interrelationship between sleep, anxiety, depression and genotype-dependent effects on treatment response. One study showed that administration of SSRIs can have adverse effects, often inducing insomnia-type symptoms in some participants, and that this differential response to treatment is in part dependent on 5HTTLPR genotype (Perlis et al., 2003). Specifically, the ‘S’ allele was associated with adverse treatment response. Similarly, a study investigating the effects of sleep deprivation on bipolar depression demonstrated that ‘LL’ homozygotes responded significantly better, experiencing a better mood amelioration than ‘S’ allele carriers (Benedetti et al., 1999). Contrastingly, although not focusing on sleep, Eley and colleagues demonstrated beneficial effects of the ‘S’ allele on response to a psychological treatment for anxiety disorder in children (Eley et al., 2012).
In addition to focusing on genetic influences on numerous phenotypes, including sleep and psychopathology, recently studies have focussed on the interaction between genetic and environmental factors. Gene-environment interaction (GxE) can be described as genotypic sensitivity to high-risk environments (Plomin et al., 2008). In other words, GxE exists when genetic vulnerability to a trait is moderated by the environment. GxE research on 5HTTLPR has typically focussed on determining whether allelic differences confer risk for sleep difficulties and psychopathology in the face of environmental adversity. In relation to sleep, Brummett and colleagues demonstrated that individuals carrying one or two copies of the ‘S’ allele experienced significantly poorer sleep quality than ‘LL’ homozygotes—with the effect only significant in individuals experiencing the chronic stress of caregiving for a parent or spouse with dementia (Brummett, Krystal, Ashley-Koch, et al., 2007). However, in our own study, we found no evidence for GxE between 5HTTLPR genotype and negative life events on sleep quality (Barclay, Eley, Mill, et al., 2011). It is possible that the conceptualisation of the environmental risk factor accounted for discrepant results. To our knowledge, no other studies have investigated the interaction between 5HTTLPR genotype and environmental risk on sleep, underscoring the necessity for further research into this area.
Within the psychopathology literature, the ‘S’ allele has been associated with greater symptoms of depression and suicidal tendencies in individuals experiencing stressful life events or environmental risk (Caspi et al., 2003; Eley et al., 2004). Although numerous replication attempts of this finding are evident in the literature, there is significant controversy within this field and recent meta-analyses have provided contradictory results. Two meta-analyses claimed that evidence for the interaction between 5HTTLPR genotype and negative life event exposure in depression was negligible, due in large part to the studies in question being underpowered (Munafo, Durrant, Lewis, & Flint, 2009; Risch et al., 2009). A more recent meta-analysis, however, found support for this effect, and the authors suggested that the discrepancy between their own and the previous meta-analyses resulted from differences in inclusion criteria (Karg, Burmeister, Shedden, & Sen, 2011). It is also likely that contradictory findings can be explained by the various conceptualisations of environmental adversity between studies (Uher & McGuffin, 2010). Regardless of the direction of these effects, these studies provide support for the hypothesis that 5HT plays a role in the underlying mechanism linking sleep and psychopathology, but discrepancies in patterns of results particularly highlight that its role is complex.
Related posts:
Genetic Influences on Depression and Anxiety in Childhood and Adolescence
Progress in Understanding the Causes of Autism Spectrum Disorders and Autistic Traits: Twin Studies from 1977 to the Present Day
Genetics of Substance Use Disorders
Future Directions in Genetics of Psychiatric Disorders
Additional Evidence for Meaningful Etiological Distinctions Within the Broader Construct of Antisocial Behavior
Gene–Environment Interactions (G×E) in Behavioral Genetics
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
