Sleep homeostatic pressure and PER3 VNTR gene polymorphism influence antidepressant response to sleep deprivation in bipolar depression

Author’s Accepted Manuscript Sleep homeostatic pressure and PER3 VNTR gene polymorphism influence antidepressant response to sleep deprivation in bipolar depression Sara Dallaspezia, Clara Locatelli, Cristina Lorenzi, Adele Pirovano, Cristina Colombo, Francesco Benedetti www.elsevier.com/locate/jad

PII: DOI: Reference:

S0165-0327(15)30833-8 http://dx.doi.org/10.1016/j.jad.2015.11.039 JAD7876

To appear in: Journal of Affective Disorders Received date: 25 August 2015 Revised date: 20 October 2015 Accepted date: 22 November 2015 Cite this article as: Sara Dallaspezia, Clara Locatelli, Cristina Lorenzi, Adele Pirovano, Cristina Colombo and Francesco Benedetti, Sleep homeostatic pressure and PER3 VNTR gene polymorphism influence antidepressant response to sleep deprivation in bipolar depression, Journal of Affective Disorders, http://dx.doi.org/10.1016/j.jad.2015.11.039 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Sleep homeostatic pressure and PER3 VNTR gene polymorphism influence antidepressant response to sleep deprivation in bipolar depression.

Sara Dallaspezia, MD; Clara Locatelli, MD; Cristina Lorenzi, PHD; Adele Pirovano, PHD; Cristina Colombo, MD; Francesco Benedetti, MD

Department of Clinical Neurosciences, Scientific Institute and University Vita-Salute San Raffaele, Milan.

Address correspondence to: Dr. Sara Dallaspezia Istituto Scientifico Ospedale San Raffaele Department of Clinical Neurosciences San Raffaele Turro Via Stamira d’Ancona 20 Milano Italy Tel +39/02/26433156 Fax +39/02/26433265 E-mail [email protected]

Abstract

Background: combined Total sleep deprivation(TSD) and light therapy (LT) cause a rapid improvement in bipolar depression which has been hypothesized to be paralleled by changes in sleep homeostasis. Recent studies showed that bipolar patients had lower changes of EEG theta power after sleep and responders to antidepressant TSD+LT slept less and showed a lower increase of EEG theta power then non-responders. A polymorphism in Per3 gene has been associated with diurnal preference, sleep structure and homeostatic response to sleep deprivation in healthy subjects. We hypothesized that the individual variability in the homeostatic response to TSD could be a correlate of antidepressant response and be influenced by genetic factors. Methods: we administered three TSD+LT cycles to bipolar depressed patients. Severity of depression was rated on Hamilton Depression Rating Scale. Actigraphic recordings were performed in a group of patients. Results: PER3 polymorphism influenced changes in total sleep time (F= 2.24; p=0,024): while PER34/4 and PER34/5 patients showed a reduction in it after treatment, PER3 5/5 subjects showed an increase of about 40 minutes, suggesting a higher homeostatic pressure. The same polymorphism influenced the change of depressive symptomatology during treatment (F=3,72; p=0,028). Limitations: sleep information was recorded till the day after the end of treatment: a longer period of observation could give more information about the possible maintenance of allostatic adaptation. Conclusions: a higher sleep homeostatic pressure reduced the antidepressant response to TSD+LT, while an allostatic adaptation to sleep loss was associated with better response. This process seems to be under genetic control.

Key Words : Bipolar Depression, Antidepressant Sleep Deprivation, Per3 Gene, Sleep Homeostasis

Introduction Recent findings (Duffy et al., 2001) support the classical view that sleep is regulated by two main processes in mammals, a circadian process C and a homeostatic process S (Borbely, 1982; Daan et al., 1984). The circadian distribution of sleep is mainly determined by the primary circadian pacemaker in the suprachiasmatic nucleus (SCN) of the anterior hypothalamus (Klein et al., 1991). The master pacemaker is synchronized to external 24-h light/dark cycles by photic inputs coming from the retina through the retinohypothalamic tract (Moore and Lenn, 1972), and its oscillating output gives time context to peripheral oscillators present in most tissues (Stratmann and Schibler, 2006) thus regulating many physiological processes and behaviours, and ensuring proper entrainment of internal rhythms to the daily light–dark cycle. The homeostatic process regulates the propensity for sleep based on the amount of prior wakefulness. It increases exponentially with the increasing duration of wakefulness, and decreases during non-rapid-eyemovement (NREM) sleep. After wake, the organism thus attempts to regain or compensate for the resource (i.e., sleep) that was previously depleted. In both animals and humans the most detailed description of sleep homeostasis is based on the changes in NREM theta and delta power: the dynamics of homeostatic sleep pressure is reflected by the decreasing activity in low EEG frequencies (<10 Hz), which progressively increase during wake and parallel sleepiness (Knoblauch et al., 2002). Clock genes might play a role in integrating the two processes at a molecular level because they both, set up the negative feed-back loops driving circadian rhythms (Takahashi et al., 2008), and influence the homeostatic regulation of sleep(Franken and Dijk, 2009). Among clock genes, PERIOD3(PER3) shows a 54nucleotide coding region motif repeating in 4 or 5 units (Nadkarni et al., 2005), which influences sleep homeostasis and the homeostatic response to sleep deprivation in healthy subjects (Goel et al., 2009; Maire et al., 2014). Compared with the 4-repeat allele (PER34/4), the longer, 5-repeat allele (PER3 5/5) was associated with worse cognitive performance and with higher sleep propensity including slow-wave activity in the sleep EEG—a putative marker of sleep homeostasis—before and after acute total sleep deprivation. When considering chronic partial sleep deprivation, the PER3 variable number tandem repeat (VNTR) polymorphism was not associated with individual differences in neurobehavioral responses but it was related to one marker of sleep homoeostatic response with PER3 5/5 subjects showing higher homeostatic pressure (Goel et al., 2009; Maire et al., 2014). Bipolar Disorder (BD) is characterized by marked alterations of sleep-wake cycle and sleep structure, and depressive episodes during BD can be treated with clinical interventions targeting sleep and circadian rhythms(Harvey, 2008; Plante and Winkelman, 2008; Wirz-Justice et al., 2013). The chronotherapeutic combination of repeated total sleep deprivation and morning light therapy (TSD+LT) causes a marked and rapid improvement of bipolar depression which leads to sustained remission with response rates comparable to those observed with usual antidepressants (Benedetti et al., 2007a; Bunney and Potkin, 2008). Due of its rapidity and its multi-target mechanism of action, chronotherapeutics has been proposed as a model antidepressant treatment to investigate the neurobiological correlates of rapid antidepressant response (Gillin et al., 2001; Schloesser et al., 2012), and recent advances confirmed its pivotal role in studying biomarkers of antidepressant response (Benedetti and Terman, 2013; Zarate et al., 2013). Changes in sleep homeostasis have been hypothesized, but not proven, to play a major role in the mechanism of action of antidepressants, including TSD (Wirz-Justice and Van den Hoofdakker, 1999). According to Kim and Laposky(Kim et al., 2007), sleep restriction is followed both by homeostatic response (increased sleep pressure and deeper recovery sleep) and by allostatic adaptation which lead to a change in sleep–wake regulation in the context of repeated sleep restriction. According to the S-deficiency model, during depression there is a deficient build-up of the homeostatic process with process C remaining unaffected(Borbely and Wirz-Justice, 1982). TSD transiently increases the level of S to normal, thus counteracting the hyperarousal state in depression, and relapses observed after recovery sleep are supposed to be linked to the return to low levels of S process(Wirz-Justice and Van den Hoofdakker, 1999). Indeed, recent studies showed that (1) patients with BD show lower changes of EEG theta power after sleep; (2) responders to antidepressant TSD+LT showed a lower increase of EEG theta power than non-

responders(Canali et al., 2014); and (3) responders sleep less after TSD than non-responders(Benedetti et al., 2007b). Considering all these issues, we hypothesized that antidepressant response to therapeutic TSD could be correlated to individual variability in the homeostatic sleep response and could be influenced by genetic factors, such as PER3 VNTR polymorphism, which modulate sleep homeostasis. We tested this hypothesis in a homogeneous sample of patients with BD.

2. Methods The sample included 105 consecutively admitted inpatients (F=66, M=39) affected by Bipolar Disorder, current depressive episode without psychotic features (DSM-IV criteria; SCID interview). Inclusion criteria were a baseline Hamilton Depression Rating Scale (HDRS) score of 18 or higher; absence of other diagnoses on Axis I; absence of mental retardation on Axis II; absence of pregnancy, history of epilepsy, major medical and neurological disorders; no treatment with long-acting neuroleptic drugs in the last three months before admission; no treatment with neuroleptics in the last month before admission; absence of a history of drug or alcohol dependency or abuse within the last six months. Physical examinations, laboratory tests and electrocardiograms were performed at admission. After complete description of the study to the subjects a written informed consent was obtained. The study was approved by the local ethical committee (ASL Citta' di Milano). At study outset 32 patients were on ongoing lithium medication, 3 were taking selective serotonin reuptake inhibitors, and 3 venlafaxine. Chronotherapeutic interventions were added to ongoing treatment, which was continued without changes in drug dose. Patients were assuming a lithium dosage causing a blood therapeutic dosage ranging from 0.5 to 0.7 mEq/l. All patients were administered three consecutive TSD cycles (day 1-6); each cycle was composed of a period of 36 hours awake. On days 1, 3, and 5 patients were totally sleep deprived from 07:00 h until 19.00 h of the following day. They were then allowed to sleep during the night of days 2, 4, and 6. Patients were administered LT (exposure for 30 minutes to a 10000 lux light) at 03:00 a.m. during the TSD night and in the morning after recovery sleep, half an hour after the awake time. The TSD nights were carried out in a room with 80 lux ambient light where a nurse monitored patients. During the TSD days, patients were not allowed to go to bed and they stayed in the ward common rooms (i.e. television room, smoker room, canteen) or garden. During the recovery night patients did not sleep in a research laboratory but in a normal hospital room with two beds and lights off. There was a dim light in the corridor of the ward and a small glass in the doors of the rooms. Each room had one window with roll-up shutters which were closed during the recovery nights, and open during the day. Severity of depression was rated on Hamilton Depression Rating Scale (HDRS) by trained raters during treatment (day 1,2,3 and 7). A group of 75 patients was instructed to wear activity monitors (Mini Motionlogger Actigraphs by Ambulatory Monitoring, Inc., Ardsley, NY) on their non-dominant wrist. They never removed actigraphs during the study period. Data were collected in Zero-Crossing Mode using 1-min epochs. Patients started actimetry soon after hospitalization, before psychiatric treatment of their condition was started, and they wore actigraphic devices till the end of the three TSD cycles. Following usual polysomnographic standards, adaptation was allowed, and the first day and night were not analyzed. Automated sleep scoring (UCSD ZeroCrossing algorithm) were performed with Action 4 software (v1.7, Ambulatory Monitoring, Inc., Ardsley, NY). We performed Total Sleep Time (the period between the moment when the patients first fell asleep and the moment when they definitively woke up), Sleep Time (the actual duration of sleep) during the night and Wake time during the night. The sleep scoring were measured in minutes.

Genomic DNA was extracted from leucocytes using Illustra blood genomicPrep (GE Healthcare). PCR was performed with the 5’-TGTCTTTTCATGTGCCCTTACTT-3’ and 5’-TGTCTGGCATTGGAGTTTGA-3’ primers. The PCR reaction was carried out in a 10 microl volume containing 150 ng genomic DNA, 5 pM of each primer, 200 mM each dNTP, 1x PCR Hot Start Buffer (Eppendorf, Italy), and 0.025 U/ml of Hotmaster Taq DNA polymerase (Eppendorf, Italy). After an initial step of 6 min at 94°C, 35 cycles of amplification (40 sec. at 94°C, 30 sec. at 60°C, 40 sec. at 70°C) and a final extension step of 12 min at 70°C were performed. An aliquot of PCR product was analysed on agarose gel electrophoresis. This analysis allowed us to distinguish the 5 repeats allele (401 bp) from the 4 repeats allele (347 bp) generated by the ins/del polymorphism. Two separate analyses of variance (ANOVA) were performed in the context of the General Linear Model (McCulloch et al., 2008; Timm and Kim, 2006). Since total sleep time (TST) is linked to sleep homeostasis (Franken and Dijk, 2009), to study changes in sleep homeostasis a repeated measure analysis was performed with TST (at baseline, after the first, second and third recovery nights, and during the night of free sleep after the end of treatment) as dependent variable, and time and PER3 VNTR as factors. We included age and duration of illness in years as nuisance factors. To study clinical response, a repeated measure analysis was performed with Delta HDRS scores (change from the baseline after the first night of TSD, after the first recovery night, and after the end of treatment) as dependent variable, and time, and PER3 VNTR as factors, also considering presence of lithium treatment (dichotomous variable current lithium treatment, no current lithium treatment) and basal HDRS score as covariates. The significance of the effect of the single independent factors on the dependent variable was estimated (least squares method) by parametric estimates of predictor variables and following standard computational procedures(Hill and Lewicki, 2006). Finally, the correlation between changes in HDRS scores and changes in sleep duration was investigated.

3. Results Clinical and demographic characteristics of the sample are resumed in Table 1. Observed genotype frequencies were: PER34/4 36/105 (34.28%), PER34/5 53/105 (50.48%), PER35/5 16/105 (15.24%). The sample was in Hardy-Weinberg equilibrium (χ2=0.237, p=0.63). Sleep characteristic are resumed in Table2. After the TSD+LT tretment, 68 patients (65% of subjects) responded to treatment (HDRS score 50% reduction). Severity of depression showed a rapid pattern of change in responders soon after the first TSD, with lower effects in non-responders (Figure 1). Albeit not reaching the response criterium, changes in non responders were however significant (post-hoc Newman Keuls test, day 1 vs day 7 p<0.001). The antidepressant effect of treatment was influenced by the PER3 VNTR polymorphism (Figure 2). The GLM ANOVA showed a significant effect of the whole model on changes in severity of depression at every time point (respectively R2=0.118, F6,98=3.10, p=0.0051 after first TSD night ; R2=0.094; F6,98=5.42; p=0.015 after the first recovery night and R2=0.196, F6,98=3.24, p=0.00011 at the end of treatment), indicating that the included factors significantly explained the observed improvement of depression. PER3 VNTR polymorphism had a significant main effect on delta HDRS changes (F=3.72; d.f. 2,98 p=0.028), with PER34/4 and PER34/5 subjects showing a better final reduction of HDRS scores compared to PER35/5 (Figure 3), as confirmed by post-hoc comparisons (Newman-Keuls critical ranges test: PER5/5 vs PER34/4 p=0.025, vs PER34/5 p=0.058). Basal severity of illness (basal HDRS scores) had a significant main effect on delta HDRS changes (F=19.17; d.f. 1,98 p=0.00003) with higher basal severity of illness meaning higher decrease of depressive symptomatology at every timepoint (after the first night of TSD: p=0.001, Beta=0.308; after the first recovery night: p=0.001, Beta=0.309; after the end of treatment: p=0.00001, Beta=0.403). The PER3 VNTR polymorphism also influenced changes in total sleep time before/after treatment (Figure 3). The GLM ANOVA showed a significant effect of the whole model (R2=0.083, F4,70=2.696, p=0.037), indicating that the included factors significantly explained the observed changes in TST. Compared to the baseline night, patients increased their TST during the three recovery nights after each TSD, as confirmed by

a significant effect of time (F= 3.45; d.f. 4,70 p=0.009). PER3 VNTR polymorphism showed a significant interaction with time in affecting changes in TST (F= 2.24; d.f. 8,70 p= 0.024). While at baseline TST did not differ among genotype groups, after treatment, that means after the end of three TSD cycles and so during the night after the last recovery night, PER34/4 and PER34/5 patients showed a reduction in TST (mean 45 and 20 minutes, respectively) compared to baseline, while PER35/5 subjects showed an increase in TST of about 38 minutes (Figure 3). Post-hoc comparison (Newman-Keuls critical ranges test) confirmed that PER3 VNTR polymorphism influenced TST after the end of treatment (PER3 5/5 vs PER34/4 p=0.033 and PER35/5 vs PER34/5 p=0.026), but not at other timepoints. No difference was found in TST before and after treatment between PER34/4 and PER34/5 subjects. Changes of depression severity and of TST were significantly associated. The overall benefit from treatment (delta HDRS scores before/after treatment) showed an inverse relationship with changes of total sleep time (Pearson’s r=0.302; p= 0.008), with patients showing a reduction of sleep duration during the night after treatment also showing the best clinical improvement.

4.Discussion We observed that the PER3 VNTR polymorphism influences both response to antidepressant chronotherapeutics (worse effects in PER35/5) and TST (more TST in PER35/5) after the repeated SD treatment. We also observed a negative correlation between clinical improvement and TST after treatment. Given that TST is linked to sleep homeostasis (Franken and Dijk, 2009), and the PER3 VNTR influences human sleep homeostatic rebound after SD (Viola et al., 2007), these findings suggest that factors affecting sleep homeostasis also influence the antidepressant response to chronotherapeutics, and that patients showing a higher homeostatic pressure after SD also had a worse response to treatment. This finding is in agreement with previous research by our group showing increased sleep duration during treatment in non-responders (Benedetti et al., 2007b). Consistent with our observation, in healthy humans subjects sleep restriction protocols yielded evidence for a faster sleep pressure build-up in PER35/5 compared to PER34/4, as shown by more deep sleep and slow wave activity during night sleep (Goel et al., 2009; Viola et al., 2007), and by a higher nap sleep efficiency and subjective quality (Maire et al., 2014). Moreover, PER35/5 subjects were found to have an higher cognitive impairment after SD, with widespread reduced cortical activations and a higher need of recruitment of supplemental cortical regions in respect to PER34/4 during a working memory task (Vandewalle et al., 2009). According to the model of Dijk and Archer (Dijk and Archer, 2010), the PER3 VNTR genotypes could then differ in their time constants of the build-up and the dissipation of sleep pressure, which in turn affects the interaction with the circadian process. Moreover, PER35/5 individuals exhibited a more pronounced alerting response to light, also attenuating electroencephalographic activity in the theta range (Chellappa et al., 2012), thus suggesting a higher sensibility to a range of environmental factors that influence circadian rhythms and sleep homeostasis. In agreement with the higher dependence of behaviour on clock molecular characteristics in patients with mood disorders (Benedetti and Terman, 2013; McClung, 2013), the sensitive PER35/5 genotype has also been associated with BD itself (Karthikeyan et al., 2014). A linear mechanistic explanation of the effects of PER3 gene variants on antidepressant response cannot yet be defined, but several mechanisms may contribute to the observed effects. PER3 is one of the paralogs of the Period family of genes, which is upregulated by CLOCK/BMAL1 heterodimers, but then represses this upregulation in an autoregulatory transcription-translation feedback loop using PER/CRY heterodimers to interact with CLOCK/ BMAL1. Its post-transcriptional mRNA stability modulation provided by several proteins, influences the flexibility of oscillation amplitude, the robustness of the period, and the phase for circadian mPer3 expression (Kim et al., 2015). PER3 VNTR occurs exclusively in simiiforme, and not in prosimian, primates, with a range of 2 to 11 copies of the primitive unit across species, suggesting an evolutionary history which parallels the control of sleep and circadian phenotype patterns in diurnal primates (Sabino et al., 2014). Tandem repeats can influence binding sites, chromatin structure, gene transcription, stability and activity of the gene product, and, in the case of proteins, protein-protein interactions and fine-

tuning protein conformation, thus possibly leading to the the fine-tuning of circadian phenotypes and homeostatic sleep regulation which is observed in diurnal primates (Sabino et al., 2014). Previous research showed that (1) in animal models wake is associated with an effective increase of synaptic weights, and deep slow wave sleep is related with synaptic downscaling, leading to a loss of synaptic weights to restore the synaptic homeostasis along a circadian pattern (Vyazovskiy et al., 2008); that (2) in humans SD promote changes in cortical excitability which parallel increased build-up of synaptic weights during wake (Huber et al., 2012); and that (3) in patients with BD this process of increased cortical excitability after wake, and decreased after sleep, correlates with antidepressant response to SD+LT, with higher cortical excitability predicting and correlating over time with better response to treatment(Canali et al., 2014) . If cortical neuroplasticity is a necessary correlate of response to treatment, it can then be hypothesized that carriers of the PER35/5 genotype could loose, during their massive homeostatic sleep rebound, the increased cortical neuroplasticity induced by SD+LT. Actually, the genotype groups which showed the best antidepressant response (PER34/4 and PER34/5) even showed a paradoxical reduction of TST after the repeated SD treatment, suggesting adaptive allostatic sleep responses with attenuated or non-significant increases in sleep time which are usually observed after chronic sleep restriction (Carskadon and Dement, 1981; Kim et al., 2007), possibly mediated by norephinephrine system (Kim et al., 2013). Moreover, beyond its role in the molecular machinery of the master clock in the suprachiasmatic nucleus, PER3 has tissue specific functions in regulating endogenous periods and internal alignment of circadian clocks (Pendergast et al., 2012). At the cellular level, Per3 is a checkpoint protein involved in cell proliferation and apoptosis, whose overexpression leads to inhibition of cell proliferation and apoptotic cell death (Im et al., 2010). PER3 VNTR affects breast cancer risk, possibly by interacting with hormones in tumorigenesis and/or affecting peripheral tissue sensitivity to hormones (Zhu et al., 2005), and is possibly involved in other types of cancer (Geng et al., 2015). The clock gene machinery could then provide a mechanism for the control of circadian gene expression and of response to stimuli at the cellular level (Kondratov et al., 2006). In brain tissues, clock gene variants have been shown to influence neural responses to stimuli in areas which control emotions and behaviour (Benedetti et al., 2008). Moreover, lithium salts, the gold standard of mood stabilizers, were shown to ameliorate neurophysiological deficits observed in ClockΔ19 mice, which is an animal model of mania (Dzirasa et al., 2010), and clock gene manipulations in animal models markedly alter neurotransmitter function . The PER3 VNTR could then possibly modify brain activity by influencing the neurotransmitter systems targeted by SD+LT (Benedetti and Smeraldi, 2009) and by direct effects at the cellular level, which associate, at the behavioural level, with sleep rebound and with the depressive phenotype. Overall, these data confirmed that sleep homeostatic pressure can influence antidepressant response to chronotherapeutics, thus renew interest in sleep-related measures as biomarkers in mood disorders (Bassetti et al., in press). The major limitation of this study is the brief duration of sleep information records: since the difference in sleep characteristic between the genotypic groups seemed to be present after the end of treatment, a longer period of observation could give more information about the possible maintenance of allostatic adaptation. Future researches are needed to clarify this issue.

References Bassetti, C.L., Ferini-Strambi, L., Brown, S., Adamantidis, A., Benedetti, F., Bruni, O., Cajochen, C., DolencGroselj, L., Ferri, R., Gais, S., Huber, R., Khatami, R., Lammers, G.J., Luppi, P.H., Manconi, M., Nissen, C., Nobili, L., Peigneux, P., Pollmacher, T., Randerath, W., Riemann, D., Santamaria, J., Schindler, K., Tafti, M.,

Van Someren, E., Wetter, T.C., in press. Neurology and psychiatry: waking up to opportunities of sleep. State of the art and clinical/research priorities for the next decade. Eur. J. Neurol. Benedetti, F., Barbini, B., Colombo, C., Smeraldi, E., 2007a. Chronotherapeutics in a psychiatric ward. Sleep Med Rev 11, 509-522. Benedetti, F., Dallaspezia, S., Fulgosi, M.C., Barbini, B., Colombo, C., Smeraldi, E., 2007b. Phase advance is an actimetric correlate of antidepressant response to sleep deprivation and light therapy in bipolar depression. Chronobiol Int 24, 921-937. Benedetti, F., Radaelli, D., Bernasconi, A., Dallaspezia, S., Falini, A., Scotti, G., Lorenzi, C., Colombo, C., Smeraldi, E., 2008. Clock genes beyond the clock: CLOCK genotype biases neural correlates of moral valence decision in depressed patients. Genes Brain Behav. 7, 20-25. Benedetti, F., Smeraldi, E., 2009. Neuroimaging and genetics of antidepressant response to sleep deprivation: implications for drug development. Curr. Pharm. Des. 15, 2637-2649. Benedetti, F., Terman, M., 2013. Much ado about...a moody clock. Biol. Psychiatry 74, 236-237. Borbely, A.A., 1982. A two process model of sleep regulation. Hum Neurobiol 1, 195-204. Borbely, A.A., Wirz-Justice, A., 1982. Sleep, sleep deprivation and depression. A hypothesis derived from a model of sleep regulation. Hum Neurobiol 1, 205-210. Bunney, J.N., Potkin, S.G., 2008. Circadian abnormalities, molecular clock genes and chronobiological treatments in depression. Br Med Bull. Canali, P., Sferrazza Papa, G., Casali, A.G., Schiena, G., Fecchio, M., Pigorini, A., Smeraldi, E., Colombo, C., Benedetti, F., 2014. Changes of cortical excitability as markers of antidepressant response in bipolar depression: preliminary data obtained by combining transcranial magnetic stimulation (TMS) and electroencephalography (EEG). Bipolar Disord 16, 809-819. Carskadon, M.A., Dement, W.C., 1981. Cumulative effects of sleep restriction on daytime sleepiness. Psychophysiology 18, 107-113. Chellappa, S.L., Viola, A.U., Schmidt, C., Bachmann, V., Gabel, V., Maire, M., Reichert, C.F., Valomon, A., Gotz, T., Landolt, H.P., Cajochen, C., 2012. Human melatonin and alerting response to blue-enriched light depend on a polymorphism in the clock gene PER3. J. Clin. Endocrinol. Metab. 97, E433-437. Daan, S., Beersma, D.G., Borbely, A.A., 1984. Timing of human sleep: recovery process gated by a circadian pacemaker. Am J Physiol 246, R161-183. Dijk, D.J., Archer, S.N., 2010. PERIOD3, circadian phenotypes, and sleep homeostasis. Sleep Med Rev 14, 151-160. Duffy, J.F., Rimmer, D.W., Czeisler, C.A., 2001. Association of intrinsic circadian period with morningnesseveningness, usual wake time, and circadian phase. Behav Neurosci 115, 895-899. Dzirasa, K., Coque, L., Sidor, M.M., Kumar, S., Dancy, E.A., Takahashi, J.S., McClung, C.A., Nicolelis, M.A., 2010. Lithium ameliorates nucleus accumbens phase-signaling dysfunction in a genetic mouse model of mania. J Neurosci 30, 16314-16323. Franken, P., Dijk, D.J., 2009. Circadian clock genes and sleep homeostasis. Eur J Neurosci 29, 1820-1829. Geng, P., Ou, J., Li, J., Wang, N., Xie, G., Sa, R., Liu, C., Xiang, L., Liang, H., 2015. Genetic association between PER3 genetic polymorphisms and cancer susceptibility: a meta-analysis. Medicine (Baltim.) 94, e568. Gillin, J.C., Buchsbaum, M., Wu, J., Clark, C., Bunney, W., Jr., 2001. Sleep deprivation as a model experimental antidepressant treatment: findings from functional brain imaging. Depress Anxiety 14, 37-49. Goel, N., Banks, S., Mignot, E., Dinges, D.F., 2009. PER3 polymorphism predicts cumulative sleep homeostatic but not neurobehavioral changes to chronic partial sleep deprivation. PLoS One 4, e5874. Harvey, A.G., 2008. Sleep and circadian rhythms in bipolar disorder: seeking synchrony, harmony, and regulation. Am. J. Psychiatry 165, 820-829. Hill, T., Lewicki, P., 2006. Statistics: methods and applications. A comprehensive reference for science, industry, and data mining. General Linear Models, StatSoft, Tulsa (OK) Chapter 18, 245-276. Horne, J.A., Ostberg, O., 1976. A self-assessment questionnaire to determine morningness-eveningness in human circadian rhythms. Int J Chronobiol 4, 97-110. Huber, R., Maki, H., Rosanova, M., Casarotto, S., Canali, P., Casali, A.G., Tononi, G., Massimini, M., 2012. Human Cortical Excitability Increases with Time Awake. Cereb. Cortex. Im, J.S., Jung, B.H., Kim, S.E., Lee, K.H., Lee, J.K., 2010. Per3, a circadian gene, is required for Chk2 activation in human cells. FEBS Lett. 584, 4731-4734.

Karthikeyan, R., Marimuthu, G., Ramasubramanian, C., Arunachal, G., BaHammam, A.S., Spence, D.W., Cardinali, D.P., Brown, G.M., Pandi-Perumal, S.R., 2014. Association of Per3 length polymorphism with bipolar I disorder and schizophrenia. Neuropsychiatr Dis Treat 10, 2325-2330. Kim, S.H., Lee, K.H., Kim, D.Y., Kwak, E., Kim, S., Kim, K.T., 2015. Rhythmic control of mRNA stability modulates circadian amplitude of mouse Period3 mRNA. J. Neurochem. 132, 642-656. Kim, Y., Chen, L., McCarley, R.W., Strecker, R.E., 2013. Sleep allostasis in chronic sleep restriction: the role of the norepinephrine system. Brain Res 1531, 9-16. Kim, Y., Laposky, A.D., Bergmann, B.M., Turek, F.W., 2007. Repeated sleep restriction in rats leads to homeostatic and allostatic responses during recovery sleep. Proc Natl Acad Sci U S A 104, 10697-10702. Klein, D.C., Moore, R.Y., Reppert, S.M., 1991. Suprachiasmatic Nucleus: The Mind's Clock. Knoblauch, V., Krauchi, K., Renz, C., Wirz-Justice, A., Cajochen, C., 2002. Homeostatic control of slow-wave and spindle frequency activity during human sleep: effect of differential sleep pressure and brain topography. Cereb Cortex 12, 1092-1100. Kondratov, R.V., Shamanna, R.K., Kondratova, A.A., Gorbacheva, V.Y., Antoch, M.P., 2006. Dual role of the CLOCK/BMAL1 circadian complex in transcriptional regulation. Faseb J 20, 530-532. Maire, M., Reichert, C.F., Gabel, V., Viola, A.U., Strobel, W., Krebs, J., Landolt, H.P., Bachmann, V., Cajochen, C., Schmidt, C., 2014. Sleep ability mediates individual differences in the vulnerability to sleep loss: evidence from a PER3 polymorphism. Cortex 52, 47-59. McClung, C.A., 2013. How might circadian rhythms control mood? Let me count the ways. Biol. Psychiatry 74, 242-249. McCulloch, C.E., Searle, S.R., Neuhaus, J.M., 2008. Generalized, Linear, and Mixed Models. John Wiley & Sons. Moore, R.Y., Lenn, N.J., 1972. A retinohypothalamic projection in the rat. J Comp Neurol 146, 1-14. Nadkarni, N.A., Weale, M.E., von Schantz, M., Thomas, M.G., 2005. Evolution of a length polymorphism in the human PER3 gene, a component of the circadian system. J Biol Rhythms 20, 490-499. Pendergast, J.S., Niswender, K.D., Yamazaki, S., 2012. Tissue-specific function of Period3 in circadian rhythmicity. PLoS ONE 7, e30254. Plante, D.T., Winkelman, J.W., 2008. Sleep disturbance in bipolar disorder: therapeutic implications. Am J Psychiatry 165, 830-843. Sabino, F.C., Ribeiro, A.O., Tufik, S., Torres, L.B., Oliveira, J.A., Mello, L.E., Cavalcante, J.S., Pedrazzoli, M., 2014. Evolutionary history of the PER3 variable number of tandem repeats (VNTR): idiosyncratic aspect of primate molecular circadian clock. PLoS ONE 9, e107198. Schloesser, R.J., Martinowich, K., Manji, H.K., 2012. Mood-stabilizing drugs: mechanisms of action. Trends Neurosci. 35, 36-46. Stratmann, M., Schibler, U., 2006. Properties, entrainment, and physiological functions of mammalian peripheral oscillators. J Biol Rhythms 21, 494-506. Takahashi, J.S., Hong, H.K., Ko, C.H., McDearmon, E.L., 2008. The genetics of mammalian circadian order and disorder: implications for physiology and disease. Nat Rev Genet 9, 764-775. Timm, N., Kim, K., 2006. Univariate and Multivariate General Linear Models: Theory and Applications with SAS. Springer. Vandewalle, G., Archer, S.N., Wuillaume, C., Balteau, E., Degueldre, C., Luxen, A., Maquet, P., Dijk, D.J., 2009. Functional magnetic resonance imaging-assessed brain responses during an executive task depend on interaction of sleep homeostasis, circadian phase, and PER3 genotype. J Neurosci 29, 7948-7956. Viola, A.U., Archer, S.N., James, L.M., Groeger, J.A., Lo, J.C., Skene, D.J., von Schantz, M., Dijk, D.J., 2007. PER3 polymorphism predicts sleep structure and waking performance. Curr Biol 17, 613-618. Vyazovskiy, V.V., Cirelli, C., Pfister-Genskow, M., Faraguna, U., Tononi, G., 2008. Molecular and electrophysiological evidence for net synaptic potentiation in wake and depression in sleep. Nat. Neurosci. 11, 200-208. Wirz-Justice, A., Benedetti, F., Terman, M., 2013. Chronotherapeutics for Affective Disorders. A Clinician’s Manual for Light and Wake Therapy. II edition. Karger, Basel. Wirz-Justice, A., Van den Hoofdakker, R.H., 1999. Sleep deprivation in depression: what do we know, where do we go? Biol Psychiatry 46, 445-453.

Zarate, C.A., Jr., Mathews, D.C., Furey, M.L., 2013. Human biomarkers of rapid antidepressant effects. Biol. Psychiatry 73, 1142-1155. Zhu, Y., Brown, H.N., Zhang, Y., Stevens, R.G., Zheng, T., 2005. Period3 structural variation: a circadian biomarker associated with breast cancer in young women. Cancer Epidemiol Biomarkers Prev 14, 268-270.

Figure Legends

Figure 1: HDRS scores during treatment in responder and non responder patients

Figure 2: DELTA (from the baseline) HDRS scores during treatment according to genotype groups

Figure 3: Total sleep time during treatment according to genotype groups

Table 1 : Clinical and demographic characteristics of the sample (mean +/- standard deviation) divided according to PER3 VNTR polymorphism groups.

PER3 VNTR polymorphism genotype 4/4

4/5

PER3

(n=36)

PER3

(n=53)

PER3

Age (years)

47.94

+/- 9.76

47.26

+/- 10.37 44.75

Duration of illness (years)

14.8

+/- 10.1

15.77

+/- 8.76

One-way ANOVA

5/5

(n=16)

F (2,102)

p

+/- 10.76

0.551

0.578

14.12

+/- 8.75

0.117

0.89

3 Previous depressive episodes (n)

3.48

+/- 2.39

6.41

+/- 5.72

6

+/- 4.28

4.367

0.015

Previous manic episodes (n)

2.77

+/- 2.38

4.36

+/- 4.85

3.47

+/- 3.23

1.753

0.178

Duration of current episode (weeks)

31.3

+/-

20.1

+/- 23.33 17.77

+/- 13.04

1.662

0.195

6 HDRS score – Baseline

20.3

44.64 +/- 3.84

19.64

+/- 3.67

18.62

+/- 2.60

1.277

0.283

+/- 5.38

12.09

+/- 5.14

14.81

+/- 4.72

1.951

0.147

+/- 5.33

7.07

+/- 5.80

10

+/- 5.78

1.827

0.166

3 HDRS score – day after the first SD night HDRS score – end of treatment

11.9 7 7.05

Table 2: Sleep characteristics of the sample (mean +/- standard deviation) divided according to PER3 VNTR polymorphism groups.

PER3 VNTR polymorphism genotype 4/4

PER3

(n=33)

PER3

4/5

(n=31)

PER3

5/5

One-way ANOVA (n=11)

F

p

(2,73) Total Sleep Time Before Treatment

489.03

+/- 52.20 499.84

+/- 52.20 547.36

+/- 60.73

3.091

0.052

+/- 92.24 581.1

+/-

5.274

0.007

(minutes) Total Sleep Time After Treatment

444.2

+/-

(minutes)

4

111.68

Sleep Time Before Treatment (minutes)

419.5

+/- 90.57 414.7

7

2

Sleep Time After Treatment (minutes)

368.6 4

481.45

200.36 449

+/- 50.71

0.582

0.561

480.36

+/- 209.1

3.040

0.054

130.13

102.65

+/-

415.32

+/103.71 +/-

Wake Time Before Treatment (minutes)

69.45

+/- 64.92 85.12

+/- 71.75 95.54

+/- 80.81

0.727

0.487

Wake Time After Treatment (minutes)

74.09

+/- 89.40 66.35

+/- 46.32 100.7

+/- 87.97

0.871

0.422

3

Highlights 1. Sleep deprivation and light therapy cause an improvement in bipolar depression 2. Sleep homeostatic response to chronotherapeutic correlates to antidepressant response

3. Sleep homeostatic response to chronotherapeutic seems to be under genetic control