PerSPecTiveS. Sleep and circadian rhythm disruption in psychiatric and neurodegenerative disease

Transcription

1 PerSPecTiveS sleep opinion Sleep and circadian rhythm disruption in psychiatric and neurodegenerative disease Katharina Wulff, Silvia Gatti, Joseph G. Wettstein and Russell G. Foster Abstract Sleep and circadian rhythm disruption are frequently observed in patients with psychiatric disorders and neurodegenerative disease. The abnormal sleep that is experienced by these patients is largely assumed to be the product of medication or some other influence that is not well defined. However, normal brain function and the generation of sleep are linked by common neurotransmitter systems and regulatory pathways. Disruption of sleep alters sleep wake timing, destabilizes physiology and promotes a range of pathologies (from cognitive to metabolic defects) that are rarely considered to be associated with abnormal sleep. We propose that brain disorders and abnormal sleep have a common mechanistic origin and that many co-morbid pathologies that are found in brain disease arise from a destabilization of sleep mechanisms. The stabilization of sleep may be a means by which to reduce the symptoms of and permit early intervention of psychiatric and neurodegenerative disease. The association between sleep disruption and abnormal brain function has a long history. Emil Kraepelin, one of the founders of modern psychiatry, noted in his first textbook in 1883 (Ref. 1) that abnormal sleep patterns and mental health are linked. Throughout the 1970s and 1980s, pioneering studies by researchers, such as Papousek 2, Wehr 3 and Wirz-Justice 4,5, highlighted the connections between circadian rhythm disruption, sleep timing and mental health. Despite this long-standing association, cause and effect between sleep disruption and brain function have been much debated without any firm conclusions being reached. The recent advances in our understanding of the neural and genetic basis of sleep and circadian (sleep/circadian) rhythm generation have allowed a re-evaluation of these connections and of our understanding of the importance of sleep to the healthy brain. Sleep is a highly complex state that arises from an interaction between the circadian system and a wake-dependent homeostatic build-up of sleep pressure 6 (Supplementary information S1 (figure)). The sleep/circadian timing systems are the product of complex interactions among multiple brain regions, neurotransmitter systems and modulatory hormones (BOX 1; Supplementary information S1 (figure)). As a consequence, abnormalities in any key neurotransmitter system will impinge on the sleep/circadian timing systems at multiple levels. However, this mechanistic connection between primary pathology and sleep abnormalities is seldom made and sleep abnormalities are mainly credited to a combination of secondary factors that include abnormal light exposure 7 (BOX 2), the side-effects of medication 8, an abnormal social environment, and general discomfort and pain 9. This view is changing, however, as common mechanistic pathways that link brain disorders and sleep are being identified. For example, polymorphisms in clock genes (BOX 3) have been associated with bipolar disorder and alcohol addiction 10, and hypofunction of the orexin (hypocretin) system has been associated with narcolepsy as well as with abnormalities in arousal, feeding and reward behaviours 11. Some of the genetic associations between sleep/circadian rhythm generation and psychiatric and neurological disease are listed in Supplementary information S2 (table). It is worth emphasizing that although there is an increasing number of gene polymorphisms that are associated with psychiatric disorders, recent studies on individuals resilience to stress suggest that long-term environmental factors may have an epigenetic influence on mental health 12. A part of the problem with building specific mechanistic links between sleep/circadian rhythm disruption and brain disorders is that there is considerable variation among individual phenotypes. An additional problem is the lack of long-term and quantitative measures of sleep/circadian disturbances in brain disorders 13. Patients with the same diagnosis, age and gender, and even the same medication, can vary greatly in the severity of their sleep/circadian timing disruption from severe disruption to none at all 14. This Opinion aims to provide a conceptual framework for the consideration of a link between sleep/circadian rhythm disruption and psychiatric disorder and neuro logic disease, and to draw attention to the potential therapeutic benefits of normalizing sleep/circadian rhythms in individuals with brain pathologies. A conceptual framework At the core of any psychiatric disorder or neurodegenerative disease is an abnormality in neurotransmitter signalling. Most therapeutic approaches for the treatment of psychiatric and neurodegenerative disease aim to correct such abnormalities (fig. 1). As sleep/circadian timing is dependent on several neurotransmitter systems, it is not surprising that sleep complaints are reported in more than 80% of patients with depression or schizophrenia and that sleep abnormalities are common in patients with Alzheimer s disease (AD) or Parkinson s disease (PD). Conversely, disruption of sleep/circadian control has wide-spread effects on all aspects of neural and neuroendocrine function. These effects include impaired cognition, impaired emotions, metabolic abnormalities, reduced immunity nature reviews NeuroscieNce volume 11 AuguST Macmillan Publishers Limited. All rights reserved

2 PersPectives and elevated risks of cancer and coronary heart disease (fig. 2). A spectrum of clinical symptoms that are reported for sleep disruption are also routinely reported for pathologies that are co-morbid with brain disorders. However, these pathologies are rarely considered to be associated with sleep disruption. Sleep/circadian rhythm disruption leads to abnormal light exposure and atypical patterns of social behaviour. These disturbances may further destabilize sleep/circadian physiology, exacerbating an already abnormal pattern of neurotransmitter release in the brain. This could lead to a state of internal de-synchronization of numerous hormonal and behavioural rhythms 15. In addition, medication, addiction to drugs and substance abuse may also affect the pattern of neurotransmitter release. The net result might be the substantial disruption of multiple neural and neuroendocrine pathways (fig. 1). Sets of interacting factors, such as those mentioned above, would explain the variable nature of sleep disruptions in psychiatric disorders and neurodegenerative disease. relatively small changes in the environment could be amplified and increase an individual s vulnerability to a pathological state. The relationships that are outlined in fig. 1 highlight the problem of assigning simple cause and effect interpretations to any abnormal sleep phenotype. sleep in psychiatric disease Abnormalities in sleep timing and sleep architecture are recognized as common co-morbidities in numerous psychiatric disorders. Indeed, changes in sleep behaviour are listed as key criteria for the diagnosis of affective (mood) disorders in current classification schemes, which include the Diagnostic and Statistical Manual of Mental Disorders (DSM-Iv) of the American Psychiatric Association and the International Classification of Disorders (ICD-10) of the World Health Organization. The common patterns of abnormal sleep that are found in psychiatric disorders have been summarized elsewhere 13. In fig. 3 we exemplify one of these patterns, namely the delayed sleep phase in schizophrenia. It is important to stress that sleep disruption is tightly associated with an increased susceptibility to a broad range of psychopathological and somatic conditions (fig. 2) and is not merely the result of an individual s frustration at failing to initiate or sustain sleep, or at feeling sleepy at an inappropriate time. Despite the recognition of an association between sleep/circadian rhythm disruption and mental health, mechanistic links remain poorly understood. Interestingly, several of the clock genes (BOX 3) have been linked with abnormal sleep timing and affective disorders (Supplementary information S2 (table)). Furthermore, mood, alertness and Box 1 neurotransmitter activation and inactivation during sleep and wakefulness Sleep and wakefulness are two mutually exclusive vigilance states. Sleep is characterized by poor responsiveness to external stimuli and wakefulness is characterized by consciousness. These vigilance states are the product of the alternated release and inhibition of neurotransmitters. This is regulated by the action of sleep-promoting neurons in the anterior hypothalamus and sleep-inhibiting neurons in the lateral and posterior hypothalamus on the arousal-promoting systems in the brainstem During wake the excitatory neurotransmitters noradrenaline, serotonin, histamine, acetylcholine and orexin are released from their respective neurons in the brainstem, midbrain and basal forebrain structures. The release of the inhibitory neurotransmitters GABA (γ-aminobutyric acid) and galanin from the ventrolateral preoptic nucleus (VLPO) is also suppressed. During sleep the non-rapid eye movement (NREM) and REM states of sleep arise from differential patterns of neurotransmitter release. First, during NREM sleep all aminergic and cholinergic neurotransmitters, and orexin are inhibited through VLPO-mediated GABA and galanin release. This decreases arousal. Once NREM sleep has been initiated, thalamo-cortical rhythms that include θ and δ oscillations and sleep spindles can be detected by electroencephalography (EEG). Second, during REM sleep the release of aminergic neurotransmitters of the brainstem is inhibited. However, acetylcholine that originates from neurons in the brainstem, midbrain and basal forebrain is released. Orexin is also released. The GABA and galanin that are released from the brainstem and VLPO act to inhibit aminergic brainstem neurons. The transition from wake to sleep is dependent on the build-up of the ATP breakdown product adenosine during wakefulness. This occurs in specific brain regions, especially the basal forebrain. High levels of adenosine are correlated with increased sleep pressure. A key function of adenosine is its inhibition of GABAergic basal forebrain neurons, which act to inhibit the sleep-promoting neurons of the VLPO. Disinhibition of the VLPO promotes the release of GABA and galanin, which inhibit the arousal-promoting systems in the brainstem, midbrain and basal forebrain, thereby initiating NREM sleep. cognitive performance in healthy individuals have all been shown to vary with the time of day. Mood is generally low in the morning, best towards the evening and declines with extended wakefulness. This has led to the proposal that mood and cognition, like sleep and wakefulness, are partly regulated by circadian and homeostatic processes (BOX 1; Supplementary information S1 (figure)), and that mood instability arises from an abnormal phase relationship between circadian and homeostatic processes. This abnormal phase relationship is similar to the condition of jet lag. After travel across multiple time zones, the peripheral clocks (BOX 3) receive conflicting timing cues from the suprachiasmatic nucleus (SCn; which is itself in the process of resetting) and from the environment (owing to irregular sleep and or activity schedules), abnormal light exposure and food intake (BOX 2). resetting the normal temporal alignment can take days and during this time period cognitive systems and mood are severely assaulted 16. The rapid antidepressant effects of sleep deprivation, light therapy and an imposed sleep schedule at night 17 have been interpreted as the beneficial effects of re-aligning circadian and homeostatic processes 18. This supports the idea that there are tangible links between sleep/circadian processes and affective disorders 19. unfortunately, this idea has not been examined in any detail. Furthermore, results relating to the role of sleep disturbance as a risk factor in mental health, as a contributing factor in disease progression or even as a target for treatment remain sparse. Some of the more robust associations between sleep/circadian rhythm disruption and psychiatric disease are outlined below. Major depressive disorder. up to 90% of patients who suffer from an acute depressive episode report changes in their sleep profile. These changes are usually described as difficulties in initiating and maintaining sleep during the night. Importantly, persistent insomnia increases the risk of relapse into a new major depressive disorder (MDD) episode 20. Increased sleep disruption for example, that experienced by mothers after childbirth raises the risk of post-partum depression, with poorer sleep quality correlating with more severe depression 21. Interestingly, some of the tricyclic antidepressants (for example, amitriptyline, imipramine, clomipramine and non-tricyclic antidepressants (for example, trazodone or mirtazapine) have very pronounced sedative effects, and are commonly used as hypnotic agents in individuals without depression. 2 AuguST 2010 volume Macmillan Publishers Limited. All rights reserved

3 PersPectives Box 2 light, circadian rhythms, sleep and arousal Under normal circumstances, exposure to the 24-hour pattern of light and dark aligns (entrains) biological rhythms to environmental time. The suprachiasmatic nuclei (SCN) receives light and dark information from the eyes via a direct projection termed the retinohypothalamic tract. In addition, light also modulates arousal brain systems through direct and indirect projections from the eyes 37 (Supplementary information S1 (figure)). In humans and other diurnal species, increased levels of light elevate alertness and low light decreases sleep latency 109. Thus, inappropriate light exposure not only disrupts the circadian timing of sleep, but also alters levels of alertness, vigilance and performance. Exposure to a normal light dark schedule ensures a stable phase relationship of internal rhythms with the external environment 110. The recent discovery of a new photoreceptor system that is maximally sensitive to blue light in the mammalian eye has profoundly altered our understanding of ocular light detection and has provided a new mechanistic understanding of how circadian and sleep processes are regulated by light 37. Blue-light photosensitive retinal ganglion cells (prgcs), which use the photopigment melanopsin, trigger a cascade of neuronal activities in the SCN that are critical for the entrainment of the molecular clock to the local light cycle 111. The SCN then regulates peripheral clocks via both humoral and neuronal pathways 112. The prgcs also project to sleep-promoting neurons in the superior colliculus and the ventrolateral preoptic nucleus (VLPO). These structures are associated with the regulation of sleep, emotions, arousal and higher cognitive function 113. As a result, part of the efficacy of some antidepressants can be attributed to their direct action on sleep. Indeed, hypnotic agents (for example, zolpidem) or agonists of melatonin (for example, agomelatine (valdoxan/melitor/tymanax; Servier); a synthetic melatonin receptor agonist with additional serotonin 2C-receptor blocking potential that is marketed as an antidepressant) have been linked to an improvement in the quality of sleep, the latency to sleep, sleep consolidation and the stability of non-rapid eye movement sleep (nrem sleep) 25 (Supplementary information S1 (figure)). This realization has allowed insomnia and daytime sleepiness in MDD to be managed and, crucially, has led to an improved treatment strategy for affective and neurovegetative (physical, emotional and cognitive) symptoms 26,27. In another study the wake-promoting agent modafinil (Provigil/ Modavigil; Cephalon) was administered to patients with MDD who were only partially responsive to selective serotonin reuptake inhibitor (SSrI) antidepressants. Patients showed significantly improved daytime wakefulness (measured using the Epworth Sleepiness Scale), reduced depressive symptoms (measured using the Hamilton rating Scale for Depression) and reduced fatigue (measured using the Fatigue Severity Scale) 28. These treatment strategies seem to improve the quality of life for these patients and hint at possible mechanistic links between the biology of sleep and depression. Interestingly, shortened latency to rem sleep, increased rem density and increased rem sleep are phenotypes that are predictive for unipolar but not bipolar depression These sleep phenotypes, which also occur in first-degree relatives of patients in these studies, have been suggested as a biological endophenotype to reduce the heterogeneity in the diagnosis of depression 30,32. It is worth noting, however, that only a small number of individuals (10 80) has been included in these studies. Seasonal affective disorder. Seasonal affective disorder (SAD), also known as winter depression or winter blues, is a mood disorder in which individuals have normal mental health throughout most of the year but experience depressive symptoms during a specific season usually in the winter year after year. In the DSM-Iv, SAD is not a unique mood disorder but is a specifier of major depression. Individuals with SAD tend to show excessive daytime sleepiness, have little energy and crave sweets and carbohydrates. They may also feel profoundly depressed. Although there seems to be an overall increase in the prevalence of SAD with increasing latitude and with the shortened daily light periods of the winter season, there is considerable country to country variation 33,34. These inconsistencies in the prevalence of SAD may be the result of the condition being a more severe expression of the naturally occurring seasonal variation in mood that is experienced by a large percentage of the general population as well as the result of variation in the classification of SAD versus depression from country to country. The very high success rate of phototherapy is the strongest evidence for a recurrent, seasonally influenced depressive condition that is a function of an interaction between light and internal circadian and/or circannual rhythms 35. In addition, eye defects have been associated with SAD. For example, patients with SAD show a higher frequency of reduced and abnormal electroretinograms 36, and polymorphisms in the melanopsin photopigment gene (BOX 2) have been linked to SAD (Supplementary information S2 (table)). The melanopsin-containing retinal ganglion cells encode the level of environmental light and project to a range of retino-recipient nuclei that include the SCn and the ventrolateral preoptic nucleus (vlpo) 37. The SCn controls the timing and the vlpo acts as a switch between wakefulness and sleep. These anatomical and mechanistic facts provide evidence that the pathology of SAD might be the result of a circadian/sleep rhythm abnormality. Polymorphisms in some of the key clock genes have also been linked to SAD (Supplementary information S2 (table)). Bipolar disorder. Bipolar disorder is associated with a spectrum of symptoms that range from major or minor depression to major or minor mood elevation (mania and hypomania) and from low-grade mood cycling to full psychosis. Major differences in the frequency of mood cycling and the simultaneous occurrence of depressive mood and manic episodes (mixed states) add to the complexity of a psychiatric diagnosis. Irregular sleep timing and a reduction in total sleep time are triggers for manic episodes 38. Indeed, disruption of the 24-hour sleep wake cycle, shortened sleep and travel across multiple time zones seem to act as important triggers for a relapse into mania in individuals who are predisposed to manic episodes (in 77% of patients) 39. Consequently, instability in the 24-hour sleep wake cycle is postulated as an important component of bipolar disorder, and therapies for the disorder involve regimes of stable and adequate sleep 39. One of the consequences of sleep/circadian disruption might be abnormalities in the stress axis, with particular emphasis on atypical neurotransmitter release (fig. 1). For example, hypercortisolaemia can arise from a breakdown in glucocorticoid receptormediated negative feedback mechanisms in the hypothalamic pituitary adrenal axis (HPA axis). Such pathogenesis has been implicated as a major factor in the development of bipolar disorder and of the neurocognitive deficits that are associated with this condition 40. Substantial therapeutic benefits for the treatment of bipolar disorder have been obtained by lowering the circulating cortisol levels or by blocking the effects of elevated cortisol with antagonists (for example, mifepristone or ru-486 (Mifegyne/Mifepres; Danco laboratories and Exelgyn) 40. nature reviews NeuroscieNce volume 11 AuguST Macmillan Publishers Limited. All rights reserved

4 PersPectives Box 3 The molecular clock and the role of the suprachiasmatic nuclei in sleep A paired cluster of approximately 50,000 neurons within the anterior hypothalamus termed the suprachiasmatic nuclei (SCN) coordinate the circadian rhythm of sleep and act to drive wakefulness throughout the day and sleep during the night. The generation of circadian rhythms at the subcellular level is thought to depend on the activity of a key group of clock genes 114. At the core is a molecular feedback loop in which the proteins brain and muscle ARNT-like 1 (BMAL1; also known as ARNTL) and CLOCK act as transcriptional regulators and drive the expression of the period homologue 1 (PER1), PER2, PER3 and cryptochrome 1 (CRY1) and CRY2 genes. A complex that is formed by PER1, PER2, PER3, CRY1 and CRY2 proteins enters the nucleus and then impedes the BMAL1 and CLOCK-mediated transcriptional drive. The complex therefore acts as a transcriptional repressor of its own genes. Additional feedback loops have been identified that regulate the resetting properties of the core feedback loop and thereby stabilize 24-hour periodicity 114. Such molecular clocks are not confined to the neurons of the SCN but occur in most cells of the body. Thus, the conceptual view of the mammalian circadian system has changed markedly in the last decade from a central clock in the SCN driving 24-hour rhythmicity to a more hierarchical arrangement of multiple peripheral clocks that are synchronized by the SCN 115. Under pathological conditions the SCN and peripheral clocks lose their normal phase relationship and a state of internal desynchronization may develop. If sustained, this state may predispose individuals to disease 116. The SCN is also thought to receive multiple feedbacks from the periphery that include information regarding metabolic status 117 and the levels of activity 118. The net result is a complex arrangement of reciprocal interactions that all contribute to the period, phase and amplitude of an individual s circadian phenotype 15. Several of these clock genes have been linked directly to abnormal sleep/circadian phenotypes 6. For example, in humans a specific point mutation in DEC2 a gene encoding a transcriptional repressor causes a short sleep duration phenotype 119. In mice, a loss of function in FBXL3 (which is involved in phosphorylation-dependent ubiquitination) leads to a stabilization of the CRY proteins, which then leads to a global transcriptional repression of the Per and Cry genes and a long period (26-hour) circadian phenotype 120,121. In humans, the long allele variant of PER3 5/5, which is encoded by PER3 with five repetitions of the variable number tandem repeat (VNTR), has been linked to extreme morning chronotypes, whereas the shorter allele (PER3 4/4 ) is associated with extreme eveningness and delayed sleep phase syndrome 122. These variant forms of PER3 gene are also associated with homeostatic processes of sleep. Homozygosity for either variant allele contributes to differences in sleep homeostasis and electroencephalography (EEG) sleep structure: the longer allele (PER3 5/5 ) is associated with a more rapid sleep onset (which indicates greater sleep pressure) as well as a longer time spent in slow wave sleep (SWS) states and increased θ and α activity during wakefulness (again a strong indication of higher levels of sleep pressure). In addition, subjects who inherited the longer allele performed very poorly in cognitive tests after periods of extended wakefulness 122. This was correlated with increased θ activity and slow eye rolling (both objective markers for sleepiness and inattention). Interestingly, when these allele frequencies of PER3 were examined in patients who were diagnosed with bipolar disorder, the data suggested that early onset bipolar disorder is associated with the longer allele (PER3 5/5 ) and later onset with the shorter allele (PER3 4/4 ) 43. Mice that carry a mutation in the circadian Clock gene (BOX 3) show a decreased need for sleep, increased motor activity, lower anxiety and an increased preference for cocain intake behaviours that are comparable to the mania-state of bipolar disorder. Mood-stabilizing drugs, such as lithium, restore normal behaviour in these mice 41. However, in patients with severe mania (for example, in bipolar disorder type I) sleep/circadian abnormalities persist after mood stabilization and apparent clinical recovery 42. This suggests that multiple interactions contribute to this spectrum disorder. genetic association studies of patients with bipolar disorder have suggested that polymorphisms in some of the clock genes are linked to the frequency of depressive relapses, to positive responses to sleep deprivation and to improved responses to long-term lithium treatment. recently, the long allele variant of the period homologue 3 (PER3) clock gene (PER3 5/5 ) has been linked to the early onset of bipolar disorder type I (Ref. 43) (BOX 3; Supplementary information S2 (table)). Early onset of bipolar disorder type I (diagnosis before the age of 18 years) is considered to be a predictor for a more severe course of the disorder with more psychotic features, a higher frequency of mixed episodes and a poorer overall prognosis. This prediction is based on the poorer prophylactic lithium response of patients with early onset bipolar disorder type I compared with patients with late onset of the disorder (diagnosis above the age of 40 years) 44. The importance of early detection and intervention in the treatment of bipolar affective disorder has led to attempts to identify prodromal features of the disorder in high-risk subjects 45. For example, an analysis of sleep microstructure has identified increased rem density as a vulnerability marker for affective disorders 46. Biological markers allow early detection of disorders in high-risk subjects and provide cues for useful endophenotypes in genetic association studies. Alcoholism. Sleep is disturbed in patients with alcohol problems, on drinking nights and on nights with no alcohol. During periods of heavy drinking and up to two years after the discontinuation of alcohol intake, patients sleep phases show reduced slow wave sleep (SWS) states, suppressed rem sleep, fragmented sleep during the second half of a normal 8-hour bedtime and shortened overall sleep durations 47. rates of insomnia in these patients are reported to be between 40 70%. Interestingly, patients with alcohol addictions who have a good prognosis for recovery are those individuals who tend to return to normal ratios of nrem to rem sleep and who sleep well in general 47. Altered sleep patterns, such as short rem latencies and high rem densities, are thought to predispose individuals to a relapse as they may be more prone to use ethanol to induce sleep 48. Excessive alcohol consumption in addicted and nonaddicted individuals can cause depressivelike states. Indeed, ethanol is classified as a depressant drug as it has sedative effects owing to its facilitation of the response at gaba (γ-aminobutyric acid)-ergic synapses and inhibition of glutamatergic nmdar (N-methyl-d-aspartate receptor) signalling 47. Co-morbid addiction to alcohol in patients with bipolar disorder is common and seems to influence circadian preference: patients who are addicted to alcohol show a predominance of morning chronotypes, whereas patients who are not addicted to alcohol show a predominance of evening chronotypes 49. Studies in mice have indicated a strong indirect association between alcohol consumption and Per2 variant(supplementary information S2 (table)). In Per2-mutant mice, the expression of the membrane glutamate transporter excitatory amino acid receptor 1 (EAA1; also known as grik4) is reduced. EAA1 is primarily expressed in astroglia and acts to clear excess glutamate from the synaptic cleft and to transport glutamate back into the neuron. The resulting hyperglutamatergic state in Per2-mutant mice has been implicated in the aetiology of alcohol dependence 50. Thus, increased alcohol consumption seems to be linked to a dysfunction of the Per2 gene. Additional associations between alcohol preference and both clock genes and sleep genes have been 4 AuguST 2010 volume Macmillan Publishers Limited. All rights reserved

5 PersPectives identified. For example, enforced wakefulness leads to changes in the expression of Per2 and the Clock-controlled D site albumin promoter-binding protein gene (Dbp; a Pro- and acidic-rich region (PAr) leu zipper transcription factor and a component of the circadian output pathway) in the forebrain of mice, suggesting that these circadian genes may have a role in the homeostatic regulation of sleep 51. Importantly, Dbp expression is increased in the frontal cortex of alcohol-preferring rats 52, and in humans this gene has been mapped closely to a genetic linkage locus for bipolar disorder on chromosome 19q13 (Ref. 53). Addiction or substance abuse Co-morbid pathologies Abnormal light dark exposure Disrupted social behaviour Neuropathology Psychiatric disorder and neurodegenerative disease Abnormal neurotransmitter release Sleep/circadian rhythm disruption Stress axis activation Medication Schizophrenia. The neuropathology of schizophrenia is characterized by the dysfunction of inhibitory neuronal circuits, reduction of cortical neuropil, a lower number of neurons in the hippocampus and childhood-onset cortical grey matter loss 54. Elevated cortisol levels are also reported in patients with schizophrenia. However, many treatment paradigms focus exclusively on the modulation of dopamine- and serotonin-mediated synaptic signalling, often with only limited success 55. In view of such wide-spread alterations in brain architecture and physiology, it is hardly surprising that abnormal sleep has been described in patients with schizophrenia since the 1920s and, more recently, has been documented as a common feature in previously unmedicated patients who are now undergoing treatment 56 and in patients who are undergoing long-term antipsychotic treatment 23. Abnormal sleep patterns in schizophrenia include reductions in rem latency, rem density, sleep efficiency, total sleep time and the duration of nrem stage 4, and an increase in sleep latency 57 (fig. 3b,3d). Multiple circadian abnormalities have also been recorded and these include delayed phase, advanced phase, free-running and/ or irregular sleep timing patterns (K.W. and r.g.f, unpublished observations). Sleep disorders are reported in 30 80% of patients with schizophrenia and are one of the most common symptoms of the disorder. Patients with poor sleep also score badly on many quality of life clinical subscales 57. Adults and older patients with schizophrenia comment that an improvement in sleep is one of their highest priorities during treatment 58. Improvement of sleep quality is frequently correlated with an improvement in negative symptoms 59. Social isolation and/or a lack of social constraint are routinely suggested as the cause of the sleep disruption in schizophrenia. This seems unlikely, however, as Figure 1 The complex relationship between neuropathology (psychiatric disorder or neurodegenerative disease), an abnormal pattern of neurotransmitter release and circadian and sleep timing. At the core of all neurological disease is an abnormality (of some level) in one or more of the neurotransmitter systems. Such an abnormality probably impinges on the circadian and sleep (circadian/sleep) timing as this system is itself regulated across different brain regions and by a range of neurotransmitters 123. Disruption of sleep/circadian control all aspects of neural and neuroendocrine function. This will result in co-morbid pathologies (FiG. 2). Sleep/circadian rhythm disruption commonly leads to abnormal social behaviour and light exposure and will feedback to further destabilize sleep/ circadian physiology. Many of these abnormalities will activate the stress axis, which will distort physiological responses even further. This distortion, the impact of medication, and perhaps addiction and substance abuse contribute to the disruption of the major neural and neuroendocrine networks. sleep problems have been shown in patients who follow a fixed routine, whereas individuals without schizophrenia and who are released from any fixed routine do not show such severe sleep wake timing disruption (K.W. and r.g.f, unpublished observations). genome-wide association studies have suggested that numerous genes may be involved in the development of schizophrenia. Several of these genes are linked to sleep and/or circadian regulation. For example, variants of disrupted in schizophrenia 1 (DISC1) have been linked to an inability to experience pleasure (anhedonia) 60 and the homeostatic regulation of sleep in animal models 61. A polymorphism in cyclic AMP (camp)-specific 3,5 -cyclic phosphodiesterase 4D (PDE4D) has been associated with sleepiness in one cohort study 62 and with schizophrenia in another study 63. Furthermore, the T3111C single nucleotide polymorphism of the human CLOCK gene has been linked to abnormalities in dopamine regulation 38 and to schizophrenia 64. Synaptosomal-associated protein 25 (SnAP25) is a neuron-specific SnAP receptor (SnArE) protein that is essential for normal synaptic vesicle release from presynaptic nerve terminals (Supplementary information S2 (table)) and has been shown to play an important part in light-signalling to the circadian system 65. Interestingly, in humans SnAP25 has been linked to schizophrenia by genetic association and linkage analysis and levels of SnAP25 have been shown to be reduced in the hippocampus of patients with schizophrenia 66. In mice, abnormalities in SnAP25 induced a range of psychiatric-like phenotypes, which included hyperactivity, alterations in social interaction and deficits in spatial memory 67. In recent in vitro studies, inhibition of synaptic vesicle recycling by botulinum toxin A or dynasore (which block exocytosis and endocytosis, respectively) in organotypic cultures of the SCn resulted in abnormal patterns of circadian gene expression in SCn neurons 68. Anxiety, panic, obsessive-compulsive disorder and post-traumatic stress disorder. The relationship between disrupted sleep and various anxiety-related disorders is well recognized and is part of the current, textrevised DSM-Iv (DSM-Iv-Tr) diagnosis for post-traumatic stress disorder (PTSD). Cause and effect, however, can be difficult to untangle for example, insomnia predisposes individuals to anxiety, which precipitates sleep disruption and this disruption then increases the likelihood of panic. Conversely, clinical research has shown that sleep problems may actually precede conditions such as anxiety and depression 21,69. In contrast to MDD, patients with PTSD and nature reviews NeuroscieNce volume 11 AuguST Macmillan Publishers Limited. All rights reserved

6 PersPectives Attention Reduced ability to concentrate and to continue performing Difficulties sustaining attention and alertness Drowsiness Microsleep Emotional responses Feeling states Stress Overt behaviour Exhaustion Increased irritability Mood fluctuations Depressed mood Bodily sensations of pain and being chilled Cardiovascular disease obsessive-compulsive disorder (OCD) frequently show hyper-vigilance and problems in falling asleep 70. As a result, the diagnostic discrimination between anxiety disorder and delayed sleep phase syndrome (DSPS) is difficult. neuropeptide S (nps) has been linked in recent studies to anxiety and sleep. For example, rodent studies have shown that nps suppresses anxiety and appetite 71, induces wakefulness and hyperactivity 72,73 and plays a substantial part in the amelioration of conditioned fear 74. In humans, nps suppresses sleep and promotes wakefulness, and has been implicated in the modulation of emotional energy and arousal states 75. nps receptors (npsrs) are widely distributed in the brain for example npsrs Disorders of the hypothalamo-pituitary adrenal axis Cognitive responses Memory Decreased working memory capacity Reduced memory of facts Reduced recall of events or episodes Somatic responses Risk of cancer Frustration, anger Increased impulsivity Mania and increased risk taking Decreased motor performance Increased stimulant and sedative use Alcohol use and misuse Metabolic problems Executive functions Reduced ability to multi task Reduced decision making Reduced creativity and productivity Risk of diabetes II Reduced immunity to disease and viral infection Figure 2 The health consequences of shortened or reduced sleep and desynchronized circadian rhythms, classified by emotional, cognitive and somatic responses. For a full list and references see Supplementary information S3 (table). are located in the amygdala (the centre for emotional processing) and the nearby locus coeruleus (the arousal-promoting area). nps has also been shown to co-localise with glutamate, which is the main excitatory neurotransmitter of the brain. A specific polymorphism in the npsr (Asn107Ile) of men but not of women results in a gain of function of the receptor by increasing its sensitivity approximately tenfold. The less active isoform of the npsr (Asn107) occurs less frequently in male patients with panic disorder and this suggests that nps may have a potential protective function in panic disorders 76. Collectively, these results support the idea that nps may be part of a neuro transmitter system that modulates sleep and wake timings, and anxiety levels. sleep in neurodegenerative disease Abnormal patterns of sleep timing and sleep architecture are consistently associated with neurodegenerative disease. The phenotype is variable and the specific mechanistic connections remain poorly defined, in the same way as mental health 77. unsurprisingly, when neurodegenerative disease affects brain structures and the neurotransmitters that are involved in sleep/circadian rhythms, sleep disruption is the result (BOX 1). neurodegenerative disease is usually progressive and irreversible, but the treatment of sleep abnormalities is emerging as a possible approach for the improvement of the overall condition of patients. In some neurodegenerative diseases the treatment of sleep abnormalities may even slow the progression of physical and mental decline. Below, we consider several neurodegenerative diseases that are commonly associated with sleep/circadian rhythm disruption as well as those in which sleep stabilization has been shown to be beneficial. Alzheimer s disease. The impact of fragmented night-time sleep in patients with AD is very debilitating to both patients and care-givers and is a primary reason for patient institutionalization. The general neurodegenerative process in AD undoubtedly alters many aspects of sleep and circadian control, and certain changes have been suggested as markers of disease progression 78. Brain nuclei in the anterior hypothalamus (for example, vlpo and SCn), lateral and posterior hypothalamus and the basal forebrain contain key regulatory circuits for the control of sleep and vigilance, and degeneration of these regions in patients with AD almost certainly contributes to their sleep problems. Specifically, the SCn shows a higher level of neuro pathology in patients with AD compared with individuals without AD who were of the same age 79. In addition, degeneration of cholinergic neurons in the basal forebrain is one of the first biochemical changes that is noted in patients with AD. Activation of the cerebral cortex during rem sleep is dependent on the basal forebrain s cholinergic innervations to the cerebral cortex. A disruption in this circuit is the likely cause of rem sleep disruption in AD patients 80. Key interactions that ensure nrem sleep promotion include the disinhibition of vlpo neurons by adenosinergic inhibition of gaba-containing basal forebrain neurons and the subsequent vlpo-driven gaba-mediated inhibition of diencephalic 6 AuguST 2010 volume Macmillan Publishers Limited. All rights reserved

7 PersPectives and ascending brainstem arousal systems (BOX 1). If the basal forebrain and the vlpo are impaired, the basal forebrain-mediated disinhibition and the inhibition that the vlpo exerts on the ascending brainstem arousal systems may also be weakened. This leads to the observed diminution of SWS and the increased nrem stage 1 sleep. In addition, K complexes and sleep spindles, which are specific features of nrem stage 2 sleep, are poorly formed, are of low amplitude, have a shorter duration and are less numerous. A link between sleep spindle number and verbal memory consolidation during sleep has been proposed in healthy subjects 81. Patients with AD also have a high prevalence of day-time sleeping 82 and this can disrupt the homeostatic drivers for sleep (Supplementary information S1 (figure)). A notable feature of AD is sundowning which is a tendency to be confused and agitated in the late afternoon and evening. This could be related to mental and physical exhaustion at the end of the day, however, reduced light at this time of day could also contribute to impaired levels of cognitive alertness. In addition, patients who are institutionalized show a greater lack of a robust daytime light signal. This lack of a robust signal promotes daytime somnolence and fragmented bouts of sleep at night, predisposes individuals to freerunning sleep patterns (BOX 2), and ultimately results in internal temporal desynchrony, not least in the pattern of melatonin release from the pineal gland. Melatonin release is known to be lower than normal in patients with AD 83. In age-matched, healthy individuals, the administration of melatonin has been shown to decrease sleep onset latency and increase daytime alertness, whereas results in patients with AD are mixed. After melatonin treatment some patients with AD show improved sleep quality (less interrupted sleep and reduced daytime sleepiness and agitation), some show no effects on sleep and some become more aggressive. Melatonin effects are, in part, dependent on the phase of the circadian cycle 84 and so the different responses to melatonin treatment may be due to differences in the internal circadian phase of patients when melatonin is administered. This hypothesis is supported by the observation that melatonin administration in combination with light therapy in institutionalized patients with AD results in improved night-time sleep, higher daytime activity and less daytime sleep 85. Thus, the use of melatonin in combination with a 9:00 c 9:00 Local clock time 15:00 19:00 23:00 Local clock time 15:00 19:00 23:00 07:00 09:00 07:00 09:00 b MVT REM Awake NREM NREM d S1 S2 S3 S4 MVT REM Awake S1 S2 S3 S4 bright light exposure could prove to be a valuable therapeutic tool for the improvement of sleep quality in patients with AD. Parkinson s disease. The earliest description of the shaking palsy by James Parkinson included a reference to disturbed sleep 86. night-time sleep disturbances as well as daytime sleepiness occur in PD 87. It has been estimated that 80 90% of patients with PD have a sleep disorder that affects their ability to fall asleep and stay asleep, their dreams, motor activity during sleep, post-sleep behaviour or daytime somnolence. There is also some evidence for a disrupted pattern of melatonin release and that weak light entrainment signals (for example, in the winter or during institutionalization) can exacerbate the condition 88. Pathologically, PD is characterized by neuronal cell loss with lewy Local clock time 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 Local clock time 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 Figure 3 rest activity profiles and hypnograms from two human subjects, illustrating normal and disrupted circadian behaviour. a rest activity cycles were recorded using wrist actigraphs from a healthy unemployed subject with a regular sleep phase between 22:00 and 08:00 local time. each row represents a day of 24 hours and each line starts at 09:00. The green shaded area highlights sleep in the normal range (23:00 07:00) and bedtime (23:00) is indicated by the red line. b electroencephalography (eeg)-derived hypnogram from a home, overnight recording of a 41-year-old healthy, unemployed man showing a normal sleep structure across the entire sleep period (from 01:00 to 08:00). The sleep hypnogram illustrates the cyclical pattern of brain activity, which consists of four basic vigilance states: wakefulness (shown in red); non-rapid eye movement sleep (NreM) stages one (S1) and S2 (shown in yellow); slow wave sleep (SWS; also referred to as deep sleep), which consists of S3 and S4 of NreM (shown in green); and rem sleep (shown in blue). c rest activity from a subject with schizophrenia. This shows circadian rhythm disruption with a delayed sleep phase. Sleep onset times range between 04:00 and 06:00 and wake times occur between 1500 and 1600 hours on most days. d Hypnogram from a home, overnight recording of a 49-year-old patient with schizophrenia. it is important to note the longer sleep duration and reduction in SWS (no stage four sleep) that is often seen in schizophrenia. MvT, movement (physical activity without waking up). body formation in the substantia nigra leading to striatal dopamine deficiency 89. However, degeneration with lewy body formation also occurs in the brainstem nuclei (for example, the pedunculopontine nucleus), which are critical in thalamocortical arousal 77. Degeneration of these nuclei seems to lead to the disruption of basic rem and nrem sleep architecture. Substantial evidence suggests that abnormal rem sleep behaviour precedes Parkinsonism and dementia by several years 87. It might therefore provide a useful marker during the early phase of PD and present an opportunity for early intervention of this disease. Preliminary studies show that the use of melatonin in combination with bright light exposure is useful for improving sleep quality in PD, but more studies are needed to provide statistically robust conclusions 88. nature reviews NeuroscieNce volume 11 AuguST Macmillan Publishers Limited. All rights reserved

8 PersPectives Huntington s disease. In addition to PD, other diseases of the basal ganglia, such as Huntington s disease (HD), also show progressive abnormalities of sleep and circadian function as well as atrophy in key brain structures that are involved in the regulation of sleep for example, the brain stem and lateral hypothalamus Transgenic r6/2 mice that carry the HD mutation have provided important insights into the molecular basis of circadian rhythm disruption in HD (Supplementary information S2 (table)). In this mouse model, pharmacological restoration and maintenance of rest-activity cycles correlate with a normalized pattern of circadian gene expression in the SCn and a marked improvement in the cognitive deficits that are seen in these animals 94, 95. Multiple sclerosis. reduced sleep latency and daytime sleepiness are common features of multiple sclerosis (MS) and correlate with increased levels of pain, fatigue and depression It should also be noted that treatment with β-interferon (IFnβ), which is thought to suppress autoimmune responses in MS, may exacerbate sleep disruption because sleep efficiency is markedly decreased on the night following IFnβ injections 99. Circadian rhythm disruption has not been widely reported in MS 96, suggesting that MS lesions do not occur frequently within the SCn 100. By contrast, brainstem demyelination and midbrain lesions are common in MS and are associated with migraines and headaches but not with sleep disturbance. However, lesions to the lateral hypothalamus result in low orexin (hypocretin) levels in the cerebrospinal fluid and result in hypersomnia, which has been treated successfully with the anti-inflammatory drug prednisolone. lesions in the reticular formation of the medulla oblongata (the centre for automatic breathing control) give rise to sleep-disordered breathing 101. As the interactions of the multiple neuronal systems that are involved in sleep generation and sleep wake control are complex, demyelination in one or more of these neuronal systems could profoundly affect sleep and arousal in MS. In patients with MS, sleep stabilization may improve the quality of life, but to our knowledge no systematic studies have investigated the effectiveness of this treatment. sleep as a therapeutic target If the relationships that are depicted in fig.1 are broadly correct, then stabilization of the sleep/circadian system in patients with the conditions and diseases that are described above should have a positive effect. Evidence is emerging that such stabilization does indeed prove beneficial. For example, bright-light phototherapy has been used as an entrainment signal for the circadian system and has been shown to alleviate some of the symptoms of several mood disorders that include SAD, unipolar depression and bipolar depression 102. In addition to Glossary Bipolar disorder A disorder characterized by abrupt mixed states of mood from an energetic elevated mood (termed mania or, if milder, hypomania) to a deep depressive state. Bipolar disorder type 1 classification is based on the occurrence of at least one manic episode, with or without the occurrence of a major depressive episode. Bipolar disorder type 2 is characterized by at least one hypomania episode and one major depressive state. Chronotype An individual s preference for daytime or night-time activities (also referred to as morningness and eveningness or larks and owls, respectively). Morning types wake up early and are most alert in the first part of the day, whereas evening types are most alert in the late evening hours and prefer to go to bed late. Circadian phase A particular reference point in the circadian cycle. for example, the onset of sleep. Circadian system (Also known as process C). The entire molecular, cellular and physiological basis for the generation of circadian rhythms in an organism. Diagnostic and statistical manual of mental disorders (DSM-IV). A manual published by the American Psychiatric Association that provides diagnostic criteria for mental health disorders. DSM-IV-TR is the most recent, text-revised version published in Diurnal An activity or process that occurs during the daytime (during light). Electroencephalography (eeg). A measure of the electrical activity of the brain that can be used to define different wake, NReM and ReM sleep states. Endophenotype A special type of biomarker. In mental health, it is the division of behavioural symptoms into recognizable phenotypes with a clear genetic association. Entrainment The process by which an organism s circadian rhythm is synchronized to an environmental rhythm such as the light dark cycle. Hypothalamic-pituitary-adrenal axis A complex set of direct influences and feedback interactions among the hypothalamus, pituitary gland and adrenal glands. The hypothalamic-pituitary-adrenal (HPA) axis constitutes a major part of the neuroendocrine system that controls reactions to stress. light, melatonin can also act to entrain the circadian system. Melatonin administration in conjunction with phototherapy has shown some promising albeit preliminary improvement of sleep patterns in dementia 85. However, melatonin administration seems not to be a powerful therapeutic agent for mood regulation 7. The value of melatonin treatment for sleep stabilization in affective disorders and schizophrenia remains to International Classification of Disorders (ICD-10). A disease classification published by the World Health Organization that provides diagnostic criteria for mental health disorders. The ICD-10 classification consists of 10 main groups. K complexes A brief, negative high-voltage peak, usually greater than 100 μv. Like sleep spindles, K-complexes are another characteristic of stage two sleep. Light therapy (Also known as phototherapy). Consists of exposure to daylight or artificial light (provided by a light box). Light exposure is of a defined intensity and is given at a specific time. Light therapy has been used to treat circadian rhythm disorders, such as delayed sleep phase syndrome, and can also be used to treat seasonal affective disorder. Major depressive disorder A disorder characterized by severe, highly persistent depression and a loss of interest or pleasure in everyday activities. It is often associated with lack of appetite, chronic fatigue and sleep disturbances. There is an increased risk of suicide. Narcolepsy A chronic sleep disorder (or dyssomnia). In relation to sleep, the condition is characterized by excessive daytime sleepiness whereby the individual experiences extreme fatigue at inappropriate times and may fall asleep. Non-rapid eye movement sleep (NReM). There are four distinct stages of NReM sleep (NReM 1 4) defined on the basis of eeg or polysomnography and other characteristics that are seen in each stage. Prodromal An early symptom (or set of symptoms) that might indicate the start of a disease before specific symptoms occur. Selective serotonin reuptake inhibitors (Also known as serotonin-specific reuptake inhibitors). A class of compounds typically used as antidepressants in the treatment of depression, anxiety disorders and some personality disorders. These inhibitors increase the extracellular level of the neurotransmitter serotonin by inhibiting its reuptake into the presynaptic cell. This increases the level of serotonin that is available to bind to the postsynaptic receptor. Sleep spindles Stage 2 sleep is characterized by sleep spindles that signify a burst of brain activity which is visible on an electroencephalogram (eeg) ranging from 11 to 16 Hz. 8 AuguST 2010 volume Macmillan Publishers Limited. All rights reserved

9 PersPectives be determined. In combination with light and/or melatonin, social cues can also be useful in regulating the circadian/sleep system. Timed activities can influence daily patterns of light exposure and can modify the timing of behaviour through associative learning and reinforcement. Meal timing, for example, is a strong stimulus for the synchronization of peripheral circadian rhythms in animals and humans 103,104 and could prove valuable if incorporated into cognitive behavioural therapy (CBT) paradigms 105. Conclusions Psychiatric disorder and neurodegenerative disease are often concurrent with some form of sleep/circadian rhythm disruption and there is increasing evidence for a mechanistic overlap between these neuropathologies and the basic control mechanisms of sleep/circadian timing. Because sleep/circadian disturbance is the most commonly reported sign that precedes the onset of many psychiatric disorders and neurodegenerative diseases, an individual s sleep biology may prove to be useful in the identification of risk factors and markers of vulnerability during the early phase of such psychiatric and neurodegenerative conditions. Despite this knowledge, very little use has been made of sleep/circadian disruption as a marker for the early detection and intervention of psychiatric disorders and neurodegenerative diseases in high-risk subjects. relatively few studies have attempted to stabilize sleep wake timing in psychiatric disorder and neurodegenerative disease using light and/or melatonin treatments. However, results from these studies show a reduction of primary symptoms and co-morbid pathologies, such as cognitive decline, attentional failures, immune suppression and metabolic problems, in response to these treatments (fig. 2). There is a need to investigate the neurological dysfunctions that underlie psychiatric disorders and neurodegenerative diseases to better understand whether sleep/circadian system disruption can be classified as a cause or effect of these diseases, and to understand how the stabilization of these systems may affect relapse of recurring disorders, such as bipolar disorder, and disease progression of neurodegenerative diseases. We propose that a more integrated consideration of sleep disruption in brain disorders will result in a clearer understanding of the broader health problems that are associated with these conditions. Furthermore, application of this information to routine clinical practice could substantially improve the quality of life of patients and their care-givers. Katharina Wulff and Russell G. Foster are members of the Circadian and Visual Neuroscience group, Nuffield Laboratory of Ophthalmology, University of Oxford, Headley Way, Oxford, OX3 9DU, UK. Silvia Gatti and Joseph G. Wettstein are at the Neuroscience Research, F. Hoffmann-La Roche Ltd., 4070 Basel, Switzerland. Correspondence to R.G.F. russell.foster@eye.ox.ac.uk doi: /nrn2868 Published online 14 July Kraeplin, E. Compendium der Psychiatrie zum Gebrauch fuer Studierende und Aerzte (Abel Verlag, Leipzig, 1883). 2. Papousek, M. [Chronobiological aspects of cyclothymia (author s transl)]. Fortschr. Neurol. Psychiatr. Grenzgeb. 43, (1975). 3. Wehr, T. A., Sack, D., Rosenthal, N., Duncan, W. & Gillin, J. C. Circadian rhythm disturbances in manicdepressive illness. Fed. Proc. 42, (1983). 4. Wirz-Justice, A., Puhringer, W. & Hole, G. Sleep deprivation and clomipramine in endogenous depression. Lancet 2, 912 (1976). 5. Wirz-Justice, A. et al. Sleep deprivation: effects on circadian rhythms of rat brain neurotransmitter receptors. Psychiatry Res. 5, (1981). 6. Cirelli, C. The genetic and molecular regulation of sleep: from fruit flies to humans. Nature Rev. Neurosci. 10, (2009). 7. Wirz-Justice, A. Biological rhythm disturbances in mood disorders. Int. Clin. Psychopharmacol. 21, (Suppl. 1) (2006). 8. Wirz-Justice, A., Haug, H. J. & Cajochen, C. Disturbed circadian rest activity cycles in schizophrenia patients: an effect of drugs? Schizophr. Bull. 27, (2001). 9. Smith, M. T., Perlis, M. L., Smith, M. S., Giles, D. E. & Carmody, T. P. Sleep quality and presleep arousal in chronic pain. J. Behav. Med. 23, 1 13 (2000). 10. Takahashi, J. S., Hong, H. K., Ko, C. H. & McDearmon, E. L. The genetics of mammalian circadian order and disorder: implications for physiology and disease. Nature Rev. Genet. 9, (2008). 11. Mieda, M. & Sakurai, T. Integrative physiology of orexins and orexin receptors. CNS Neurol. Disord. Drug Targets 8, (2009). 12. Feder, A., Nestler, E. J. & Charney, D. S. Psychobiology and molecular genetics of resilience. Nature Rev. Neurosci. 10, (2009). 13. Wulff, K., Porcheret, K., Cussans, E. & Foster, R. G. Sleep and circadian rhythm disturbances: multiple genes and multiple phenotypes. Curr. Opin. Genet. Dev. 19, (2009). 14. Armitage, R. Sleep and circadian rhythms in mood disorders. Acta Psychiatr. Scand. Suppl. 115, (2007). 15. Schibler, U. The daily timing of gene expression and physiology in mammals. Dialogues Clin. Neurosci. 9, (2007). 16. Katz, G., Durst, R., Zislin, Y., Barel, Y. & Knobler, H. Y. Psychiatric aspects of jet lag: review and hypothesis. Med. Hypotheses 56, (2001). 17. Benedetti, F., Barbini, B., Colombo, C. & Smeraldi, E. Chronotherapeutics in a psychiatric ward. Sleep Med. Rev. 11, (2007). 18. Wirz-Justice, A. & Van den Hoofdakker, R. H. Sleep deprivation in depression: what do we know, where do we go? Biol. Psychiatry 46, (1999). 19. Bunney, J. N. & Potkin, S. G. Circadian abnormalities, molecular clock genes and chronobiological treatments in depression. Br. Med. Bull. 86, (2008). 20. Pigeon, W. R. et al. Is insomnia a perpetuating factor for late-life depression in the IMPACT cohort? Sleep 31, (2008). 21. Posmontier, B. Sleep quality in women with and without postpartum depression. J. Obstet. Gynecol. Neonatal Nurs. 37, ; (2008). 22. Asnis, G. M. et al. Zolpidem for persistent insomnia in SSRI-treated depressed patients. J. Clin. Psychiatry 60, (1999). 23. Krystal, A. D., Thakur, M. & Roth, T. Sleep disturbance in psychiatric disorders: effects on function and quality of life in mood disorders, alcoholism, and schizophrenia. Ann. Clin. Psychiatry 20, (2008). 24. Kasper, S. et al. Efficacy of the novel antidepressant agomelatine on the circadian rest-activity cycle and depressive and anxiety symptoms in patients with major depressive disorder: a randomized, doubleblind comparison with sertraline. J. Clin. Psychiatry 71, (2010). 25. Lopes, M. C., Quera-Salva, M. A. & Guilleminault, C. Non-REM sleep instability in patients with major depressive disorder: subjective improvement and improvement of non-rem sleep instability with treatment (Agomelatine). Sleep Med. 9, (2007). 26. Olie, J. P. & Kasper, S. Efficacy of agomelatine, a MT1/MT2 receptor agonist with 5-HT2C antagonistic properties, in major depressive disorder. Int. J. Neuropsychopharmacol. 10, (2007). 27. Quera-Salva, M. A., Lemoine, P. & Guilleminault, C. Impact of the novel antidepressant agomelatine on disturbed sleep wake cycles in depressed patients. Hum. Psychopharmacol. 25, (2010). 28. Fava, M. et al. Modafinil augmentation of selective serotonin reuptake inhibitor therapy in MDD partial responders with persistent fatigue and sleepiness. Ann. Clin. Psychiatry 19, (2007). 29. Kupfer, D. J. et al. REM sleep, naps, and depression. Psychiatry Res. 5, (1981). 30. Modell, S., Huber, J., Holsboer, F. & Lauer, C. J. The munich vulnerability study on affective disorders: risk factors for unipolarity versus bipolarity. J. Affect. Disord. 74, (2003). 31. Rao, U. et al. Heterogeneity in EEG sleep findings in adolescent depression: unipolar versus bipolar clinical course. J. Affect. Disord. 70, (2002). 32. Gregory, A. M., Cox, J., Crawford, M. R., Holland, J. & Haravey, A. G. Dysfunctional beliefs and attitudes about sleep in children. J. Sleep Res. 18, (2009). 33. Levitt, A. J. & Boyle, M. H. The impact of latitude on the prevalence of seasonal depression. Can. J. Psychiatry 47, (2002). 34. Axelsson, J., Ragnarsdottir, S., Pind, J. & Sigbjornsson, R. Daylight availability: a poor predictor of depression in Iceland. Int. J. Circumpolar Health 63, (2004). 35. Wirz-Justice, A. Chronobiology and psychiatry. Sleep Med. Rev. 11, (2007). 36. Lavoie, M. P. et al. Evidence of a biological effect of light therapy on the retina of patients with seasonal affective disorder. Biol. Psychiatry 66, (2009). 37. Hankins, M. W., Peirson, S. N. & Foster, R. G. Melanopsin: an exciting photopigment. Trends Neurosci. 31, (2008). 38. McClung, C. A. Circadian genes, rhythms and the biology of mood disorders. Pharmacol. Ther. 114, (2007). 39. Plante, D. T. & Winkelman, J. W. Sleep disturbance in bipolar disorder: therapeutic implications. Am. J. Psychiatry 165, (2008). 40. Young, A. H. Antiglucocoticoid treatments for depression. Aust. N. Z. J. Psychiatry 40, (2006). 41. Roybal, K. et al. Mania-like behavior induced by disruption of CLOCK. Proc. Natl Acad. Sci. USA 104, (2007). 42. Salvatore, P. et al. Circadian activity rhythm abnormalities in ill and recovered bipolar I disorder patients. Bipolar Disord. 10, (2008). 43. Benedetti, F. et al. A length polymorphism in the circadian clock gene Per3 influences age at onset of bipolar disorder. Neurosci. Lett. 445, (2008). 44. Bellivier, F. et al. Age at onset in bipolar I affective disorder: further evidence for three subgroups. Am. J. Psychiatry 160, (2003). 45. Conus, P. et al. The proximal prodrome to first episode mania a new target for early intervention. Bipolar Disord. 10, (2008). nature reviews NeuroscieNce volume 11 AuguST Macmillan Publishers Limited. All rights reserved

10 PersPectives 46. Modell, S., Ising, M., Holsboer, F. & Lauer, C. J. The Munich vulnerability study on affective disorders: premorbid polysomnographic profile of affected high-risk probands. Biol. Psychiatry 58, (2005). 47. Roehrs, T. & Roth, T. Sleep, sleepiness, sleep disorders and alcohol use and abuse. Sleep Med. Rev. 5, (2001). 48. Krystal, A. D. Treating the health, quality of life, and functional impairments in insomnia. J. Clin. Sleep Med. 3, (2007). 49. Hatonen, T., Forsblom, S., Kieseppa, T., Lonnqvist, J. & Partonen, T. Circadian phenotype in patients with the co-morbid alcohol use and bipolar disorders. Alcohol Alcohol. 43, (2008). 50. Spanagel, R. et al. The clock gene Per2 influences the glutamatergic system and modulates alcohol consumption. Nature Med. 11, (2005). 51. Franken, P., Thomason, R., Heller, H. C. & O Hara, B. F. A non-circadian role for clock-genes in sleep homeostasis: a strain comparison. BMC Neurosci. 8, 87 (2007). 52. Rodd, Z. A. et al. Candidate genes, pathways and mechanisms for alcoholism: an expanded convergent functional genomics approach. Pharmacogenomics J. 7, (2007). 53. Francks, C. et al. Population-based linkage analysis of schizophrenia and bipolar case-control cohorts identifies a potential susceptibility locus on 19q13. Mol. Psychiatry 15, (2008). 54. Hendler, T., Bleich-Cohen, M. & Sharon, H. Neurofunctional view of psychiatry: clinical brain imaging revisited. Curr. Opin. Psychiatry 22, (2009). 55. Stahl, S. M. & Buckley, P. F. Negative symptoms of schizophrenia: a problem that will not go away. Acta Psychiatr. Scand. 115, 4 11 (2007). 56. Van Cauter, E. et al. Circadian and sleep-related endocrine rhythms in schizophrenia. Arch Gen Psychiatry 48, (1991). 57. Cohrs, S. Sleep disturbances in patients with schizophrenia: impact and effect of antipsychotics. CNS Drugs 22, (2008). 58. Auslander, L. A. & Jeste, D. V. Perceptions of problems and needs for service among middle-aged and elderly outpatients with schizophrenia and related psychotic disorders. Community Ment. Health J. 38, (2002). 59. Hofstetter, J. R., Lysaker, P. H. & Mayeda, A. R. Quality of sleep in patients with schizophrenia is associated with quality of life and coping. BMC Psychiatry 5, 13 (2005). 60. Tomppo, L. et al. Association of variants in DISC1 with psychosis-related traits in a large population cohort. Arch. Gen. Psychiatry 66, (2009). 61. Sawamura, N. et al. Nuclear DISC1 regulates CREmediated gene transcription and sleep homeostasis in the fruit fly. Mol. Psychiatry 13, (2008). 62. Gottlieb, D. J., O Connor, G. T. & Wilk, J. B. Genome-wide association of sleep and circadian phenotypes. BMC Med. Genet. 8, (Suppl. 1) 9 (2007). 63. Tomppo, L. et al. Association between genes of disrupted in schizophrenia 1 (DISC1) interactors and schizophrenia supports the role of the DISC1 pathway in the etiology of major mental illnesses. Biol. Psychiatry 65, (2009). 64. Takao, T., Tachikawa, H., Kawanishi, Y., Mizukami, K. & Asada, T. CLOCK gene T3111C polymorphism is associated with Japanese schizophrenics: a preliminary study. Eur. Neuropsychopharmacol. 17, (2007). 65. Ding, J. M. et al. Resetting the biological clock: mediation of nocturnal circadian shifts by glutamate and NO. Science 266, (1994). 66. Young, C. E. et al. SNAP-25 deficit and hippocampal connectivity in schizophrenia. Cereb. Cortex 8, (1998). 67. Oliver, P. L. & Davies, K. E. Interaction between environmental and genetic factors modulates schizophrenic endophenotypes in the Snap-25 mouse mutant blind drunk. Hum. Mol. Genet. 18, (2009). 68. Deery, M. J. et al. Proteomic analysis reveals the role of synaptic vesicle cycling in sustaining the suprachiasmatic circadian clock. Curr. Biol. 19, (2009). 69. Monti, J. M. & Monti, D. Sleep disturbance in generalized anxiety disorder and its treatment. Sleep Med. Rev. 4, (2000). 70. Papadimitriou, G. N. & Linkowski, P. Sleep disturbance in anxiety disorders. Int. Rev. Psychiatry 17, (2005). 71. Xu, Y. L. et al. Neuropeptide S: a neuropeptide promoting arousal and anxiolytic-like effects. Neuron 43, (2004). 72. Reinscheid, R. K. & Xu, Y. L. Neuropeptide S as a novel arousal promoting peptide transmitter. FEBS J. 272, (2005). 73. Okamura, N. & Reinscheid, R. K. Neuropeptide S: a novel modulator of stress and arousal. Stress 10, (2007). 74. Leonard, S. K. et al. Pharmacology of neuropeptide S. in mice: therapeutic relevance to anxiety disorders. Psychopharmacology (Berlin) 197, (2008). 75. Reinscheid, R. K. Neuropeptide S: anatomy, pharmacology, genetics and physiological functions. Results Probl. Cell Differ. 46, (2008). 76. Okamura, N. et al. Gender-specific association of a functional coding polymorphism in the Neuropeptide S receptor gene with panic disorder but not with schizophrenia or attention-deficit/hyperactivity disorder. Prog. Neuropsychopharmacol. Biol. Psychiatry 31, (2007). 77. Bohnen, N. I. & Albin, R. L. The cholinergic system and Parkinson disease. Behav. Brain Res. 7 Jan 2010 (doi: /j.bbr ). 78. Jeong, J. EEG dynamics in patients with Alzheimer s disease. Clin. Neurophysiol. 115, (2004). 79. Hofman, M. A. & Swaab, D. F. Alterations in circadian rhythmicity of the vasopressin-producing neurons of the human suprachiasmatic nucleus (SCN) with aging. Brain Res. 651, (1994). 80. Vitiello, M. V., Prinz, P. N., Williams, D. E., Frommlet, M. S. & Ries, R. K. Sleep disturbances in patients with mild-stage Alzheimer s disease. J. Gerontol. 45, M131 M138 (1990). 81. Clemens, Z., Fabo, D. & Halasz, P. Overnight verbal memory retention correlates with the number of sleep spindles. Neuroscience 132, (2005). 82. Tractenberg, R. E., Singer, C. M. & Kaye, J. A. Characterizing sleep problems in persons with Alzheimer s disease and normal elderly. J. Sleep Res. 15, (2006). 83. Zhou, J. N., Liu, R. Y., Kamphorst, W., Hofman, M. A. & Swaab, D. F. Early neuropathological Alzheimer s changes in aged individuals are accompanied by decreased cerebrospinal fluid melatonin levels. J. Pineal Res. 35, (2003). 84. Rajaratnam, S. M. & Arendt, J. Health in a 24-h society. Lancet 358, (2001). 85. Dowling, G. A. et al. Melatonin and bright-light treatment for rest-activity disruption in institutionalized patients with Alzheimer s disease. J. Am. Geriatr. Soc. 56, (2008). 86. Parkinson, J. An essay on the shaking palsy J. Neuropsychiatry Clin. Neurosci. 14, (2002); discussion in 14, 222 (2002). 87. Arnulf, I., Leu, S. & Oudiette, D. Abnormal sleep and sleepiness in Parkinson s disease. Curr. Opin. Neurol. 21, (2008). 88. Willis, G. L., Kelly, A. M. & Kennedy, G. A. Compromised circadian function in Parkinson s disease: enucleation augments disease severity in the unilateral model. Behav. Brain Res. 193, (2008). 89. Metzler-Baddeley, C. A review of cognitive impairments in dementia with Lewy bodies relative to Alzheimer s disease and Parkinson s disease with dementia. Cortex 43, (2007). 90. Kassubek, J. et al. Topography of cerebral atrophy in early Huntington s disease: a voxel based morphometric MRI study. J. Neurol. Neurosurg. Psychiatry 75, (2004). 91. Rosas, H. D. et al. Evidence for more widespread cerebral pathology in early HD: an MRI-based morphometric analysis. Neurology 60, (2003). 92. Morton, A. J. et al. Disintegration of the sleep wake cycle and circadian timing in Huntington s disease. J. Neurosci. 25, (2005). 93. Goodman, A. O. & Barker, R. A. How vital is sleep in Huntington s disease? J. Neurol. 257, (2010). 94. Pallier, P. N. et al. Pharmacological imposition of sleep slows cognitive decline and reverses dysregulation of circadian gene expression in a transgenic mouse model of Huntington s disease. J. Neurosci. 27, (2007). 95. Pallier, P. N. & Morton, A. J. Management of sleep/ wake cycles improves cognitive function in a transgenic mouse model of Huntington s disease. Brain Res. 1279, (2009). 96. Taphoorn, M. J. et al. Fatigue, sleep disturbances and circadian rhythm in multiple sclerosis. J. Neurol. 240, (1993). 97. Attarian, H. P., Brown, K. M., Duntley, S. P., Carter, J. D. & Cross, A. H. The relationship of sleep disturbances and fatigue in multiple sclerosis. Arch. Neurol. 61, (2004). 98. Gallup, A. C., Gallup, G. G. Jr. & Feo, C. Yawning, sleep, and symptom relief in patients with multiple sclerosis. Sleep Med. 11, (2010). 99. Mendozzi, L., Tronci, F., Garegnani, M. & Pugnetti, L. Sleep disturbance and fatigue in mild relapsing remitting multiple sclerosis patients on chronic immunomodulant therapy: an actigraphic study. Mult. Scler. 16, (2010) Ceccarelli, A. et al. T2 hypointensity in the deep gray matter of patients with benign multiple sclerosis. Mult. Scler. 15, (2009) Auer, R. N., Rowlands, C. G., Perry, S. F. & Remmers, J. E. Multiple sclerosis with medullary plaques and fatal sleep apnea (Ondine s curse). Clin. Neuropathol. 15, (1996) Wirz-Justice, A., Benedetti, F. & Terman, M. Chronotherapeutics for Affective Disorders (Karger, Basel, 2009) Damiola, F. et al. Restricted feeding uncouples circadian oscillators in peripheral tissues from the central pacemaker in the suprachiasmatic nucleus. Genes Dev. 14, (2000) Goetz, F. et al. Timing of single daily meal influences relations among human circadian rhythms in urinary cyclic AMP and hemic glucagon, insulin and iron. Experientia 32, (1976) Riemann, D. et al. The hyperarousal model of insomnia: a review of the concept and its evidence. Sleep Med. Rev. 14, (2010) Pace-Schott, E. F. & Hobson, J. A. The neurobiology of sleep: genetics, cellular physiology and subcortical networks. Nature Rev. Neurosci. 3, (2002) Diniz Behn, C. G., Kopell, N., Brown, E. N., Mochizuki, T. & Scammell, T. E. Delayed orexin signaling consolidates wakefulness and sleep: physiology and modeling. J. Neurophysiol. 99, (2008) Behn, C. G., Brown, E. N., Scammell, T. E. & Kopell, N. J. Mathematical model of network dynamics governing mouse sleep-wake behavior. J. Neurophysiol. 97, (2007) Lockley, S. W. et al. Short-wavelength sensitivity for the direct effects of light on alertness, vigilance, and the waking electroencephalogram in humans. Sleep 29, (2006) Scheer, F. A., Wright, K. P. Jr, Kronauer, R. E. & Czeisler, C. A. Plasticity of the intrinsic period of the human circadian timing system. PLoS ONE 2, e721 (2007) Meijer, J. H., Michel, S. & Vansteensel, M. J. Processing of daily and seasonal light information in the mammalian circadian clock. Gen. Comp. Endocrinol. 152, (2007) Buijs, R. M., van Eden, C. G., Goncharuk, V. D. & Kalsbeek, A. The biological clock tunes the organs of the body: timing by hormones and the autonomic nervous system. J. Endocrinol. 177, (2003) Lupi, D., Oster, H., Thompson, S. & Foster, R. G. The acute light-induction of sleep is mediated by OPN4- based photoreception. Nature Neurosci. 11, (2008) Ko, C. H. & Takahashi, J. S. Molecular components of the mammalian circadian clock. Hum. Mol. Genet. 15, R271 R277 (2006) Schibler, U. The 2008 Pittendrigh/Aschoff lecture: peripheral phase coordination in the mammalian circadian timing system. J. Biol. Rhythms 24, 3 15 (2009) Corradini, I., Verderio, C., Sala, M., Wilson, M. C. & Matteoli, M. SNAP-25 in neuropsychiatric disorders. Ann. N. Y. Acad. Sci. 1152, (2009). 10 AuguST 2010 volume Macmillan Publishers Limited. All rights reserved

11 PersPectives 117. Yi, C. X. et al. Ventromedial arcuate nucleus communicates peripheral metabolic information to the suprachiasmatic nucleus. Endocrinology 147, (2006) Malek, Z. S., Sage, D., Pevet, P. & Raison, S. Daily rhythm of tryptophan hydroxylase-2 messenger ribonucleic acid within raphe neurons is induced by corticoid daily surge and modulated by enhanced locomotor activity. Endocrinology 148, (2007) He, Y. et al. The transcriptional repressor DEC2 regulates sleep length in mammals. Science 325, (2009) Maywood, E. S., O Neill, J. S., Chesham, J. E. & Hastings, M. H. Minireview: the circadian clockwork of the suprachiasmatic nuclei analysis of a cellular oscillator that drives endocrine rhythms. Endocrinology 148, (2007) Godinho, S. I. et al. The after-hours mutant reveals a role for Fbxl3 in determining mammalian circadian period. Science 316, (2007) Dijk, D. J. & Archer, S. N. PERIOD3, circadian phenotypes, and sleep homeostasis. Sleep Med. Rev. 14, (2010) Reghunandanan, V. & Reghunandanan, R. Neurotransmitters of the suprachiasmatic nuclei. J. Circadian Rhythms 4, 2 (2006). Acknowledgements The work is supported by the National Institute for Health Research (NIHR). Biomedical Research Centre, Oxford, UK, and The Wellcome Trust, London, UK. We would like to thank G. Goodwin (head of the Department of Psychiatry, University of Oxford, UK), C. Kennard (head of the Department of Clinical Neurology, University of Oxford, UK), K. Porcheret (Nuffield Laboratory of Ophthalmology, University of Oxford, UK), K. Davies and P. Oliver (Medical Research Council Functional Genomics Unit, University of Oxford, UK) for their valuable input during the preparation of this manuscript. Competing interests statement The authors declare no competing financial interests. DATABAses entrez Gene: Clock Dbp DISC1 Per2 PER3 UniProtKB: eaa1 BMAL1 ifnβ NPS PDe4D SNAP25 FURTHeR information Author s homepage: supplementary information see online article: S1 (figure) S2 (table) S3 (table) All links Are AcTive in The online pdf nature reviews NeuroscieNce volume 11 AuguST Macmillan Publishers Limited. All rights reserved

12 SUPPLEMENTARY INFORMATION In format provided by Foster et al. (august 2010) mg) during the day it can induce sleepiness and impair cognitive performance 5. Melatonin, and its synthetic agonists, can also shift the circadian timing of the sleep/wake cycle 6, 7. It is noteworthy that neurons in the SCN have a high concentration of melatonin receptors and melatonin will suppress their activity in vitro 8. Receptors for melatonin are also expressed on multiple neuroendocrine cells within the hypothalamus and so melatonin has been implicated in the regulation of numerous elements of the hypothalamo-pituitary (HP) axis 9. References S1 (figure) Scheme of circadian and homeostatic interaction to regulate sleep. Diagram illustrating the key components in the generation and maintenance of the sleep/wake cycle and its relationship to mood and cognition. Mood and cognition are directly modulated by sleep and the circadian system and directly influenced by both light and social interactions 1. Sleep is regulated by two broad mechanisms involving both the 24h body clock (circadian system; known as process C) and a wake-dependent homeostatic build-up of sleep pressure (also called process S) 2. The circadian pacemaker located within the suprachiasmatic nucleus (SCN) coordinates the timing of wakefulness throughout the day and sleep during the night. This 24h rhythm interacts with the homeostatic drive for sleep, whereby the sleep pressure increases during wake and dissipates during sleep. This process has been likened to an hourglass oscillator. The circadian and homeostatic drivers regulate the multiple neurotransmitter and brain systems involved in sleep and arousal. Sleep-wake behaviour in turn feeds back upon the circadian pacemaker and homeostat. These components are modulated by light which acts to entrain the circadian pacemaker to the environmental light/ dark cycle, acutely suppress melatonin production from the pineal and acutely elevate or suppress levels of arousal. Finally, social activities such as meal times or forced awakening by an alarm clock will drive sleep-wake activity. Core body temperature and melatonin are also important in the initiation and consolidation of sleep in humans and may be linked physiologically. The administration of melatonin has been shown to cause vasodilation of peripheral blood vessels coinciding with an increase of peripheral skin temperature and a drop in core body temperature of ~ 1 C 3. Melatonin synthesis and release is regulated by a multi-synaptic pathway originating in the SCN. Melatonin levels rise shortly after dusk, when sleep is initiated and body temperature drops and falls in anticipation of dawn. If individuals are exposed to relatively bright light (~ 2000 lux) at night, melatonin synthesis is acutely and fully inhibited. Thus, melatonin levels broadly reflect the pattern of light/dark exposure. Attempts to sleep during the declining phase of melatonin and the rising phase of core body temperature, as with night shift workers, usually results in a shorter and less well consolidated sleep episode 4. If melatonin is administered (~ Wirz-Justice, A., Benedetti, F. & Terman, M. Chronotherapeutics for Affective Disorders (Karger, Basel, 2009). 2. Krueger, J.M. et al. Sleep as a fundamental property of neuronal assemblies. Nat Rev Neurosci (2008). 3. Gilbert, S.S., van den Heuvel, C.J., Kennaway, D.J. & Dawson, D. Peripheral heat loss: a predictor of the hypothermic response to melatonin administration in young and older women. Physiol Behav 66, (1999). 4. Dijk, D.J. & Duffy, J.F. Circadian regulation of human sleep and age-related changes in its timing, consolidation and EEG characteristics. Ann Med 31, (1999). 5. Dollins, A.B. et al. Effect of pharmacological daytime doses of melatonin on human mood and performance. Psychopharmacology (Berl) 112, (1993). 6. Rajaratnam, S.M. et al. Melatonin agonist tasimelteon (VEC-162) for transient insomnia after sleeptime shift: two randomised controlled multicentre trials. Lancet (2008). 7. Arendt, J. & Skene, D.J. Melatonin as a chronobiotic. Sleep Med Rev 9, (2005). 8. Stehle, J., Vanecek, J. & Vollrath, L. Effects of melatonin on spontaneous electrical activity of neurons in rat suprachiasmatic nuclei: an in vitro iontophoretic study. J Neural Transm 78, (1989). 9. Lincoln, G.A. Decoding the nightly melatonin signal through circadian clockwork. Mol Cell Endocrinol 252, (2006). nature reviews neuroscience