- Email: [email protected]
Developmental disruption of the serotonin system alters circadian rhythms
Physiology & Behavior 105 (2012) 257–263
Contents lists available at SciVerse ScienceDirect
Physiology & Behavior journal homepage: www.elsevier.com/locate/phb
Developmental disruption of the serotonin system alters circadian rhythms Erin V. Paulus, Eric M. Mintz ⁎ Department of Biological Sciences, Kent State University, Kent, OH 44242, United States School of Biomedical Sciences, Kent State University, Kent, OH 44242, United States
a r t i c l e
i n f o
Article history: Received 29 September 2010 Received in revised form 3 August 2011 Accepted 24 August 2011 Keywords: Pet-1 Serotonin Suprachiasmatic nucleus Circadian Wheel-running Locomotor
a b s t r a c t Serotonin (5-HT) plays an important role in circadian rhythms, acting to modulate photic input to the mammalian clock, the suprachiasmatic nucleus (SCN), as well as playing a role in non-photic input. The transcription factor Pet-1 is an early developmental indicator of neurons that are destined for a 5-HTergic fate. Mice lacking the Pet-1 gene show a 70% loss of 5-HT immunopositive cell bodies in adult animals. 5-HT neurotoxic lesion studies using 5,7-dihydroxytryptamine (5,7-DHT) have highlighted species-specific differences in response to 5-HT depletion and studies using knockout mice lacking various 5-HT receptors have helped to elucidate the role of individual 5-HT receptors in mediating 5-HT's effects on circadian rhythms. Here we investigate the effects of a developmental disruption of the 5-HT system on the SCN and circadian wheel-running behavior. Immunohistochemical analysis confirmed depletion of 5-HT fiber innervation to the SCN as well as greatly reduced numbers of cell bodies in the raphe nuclei in Pet-1 knockout mice. These mice also display significantly longer free-running periods than wildtype or heterozygote counterparts. In light–dark cycles, knockouts showed a shift in peak wheel running behavior towards the late night as compared to wildtype and heterozygote animals. When kept in constant darkness for 70 days, wildtype animals showed decreases in free-running period over time while the period of knockout animals remained constant. Immunohistochemical analysis for neuropeptides within the SCN indicates that the behavioral changes observed in Pet-1 knockout mice were not due to gross changes in SCN structure. These results suggest that developmental loss of serotonergic input to the clock has long-term consequences for both circadian clock parameters and the temporal organization of activity. © 2011 Elsevier Inc. All rights reserved.
1. Introduction In mammals, behavioral circadian rhythms are generated by an endogenous clock located in the suprachiasmatic nucleus (SCN) within the ventral hypothalamus. Rhythms are entrained to the 24 h light– dark cycle by external cues, with light functioning as the primary cue [1–3]. However, other non-photic stimuli such as temperature, food availability, or activity can also serve as cues for entrainment [4–6]. Serotonin (5-HT) is a neurotransmitter that plays an important role in circadian rhythms, modulating photic input to the SCN as well as playing a role in non-photic phase shifts [7]. One of the 5-HTergic pathways thought to convey non-photic information to the SCN, such as the activity level of the animal, as well as modulate photic effects is input from the median raphe (MR) [8]. There is also evidence suggesting direct 5-HTergic input to the SCN from the dorsal raphe (DR) [9]; the DR can also influence the SCN through an indirect connection via the intergeniculate leaflet (IGL) [8].
⁎ Corresponding author at: Department of Biological Sciences, Kent State University, Kent, OH 44242, United States. Tel.: + 1 330 672 3847; fax: + 1 330 672 3713. E-mail address: [email protected] (E.M. Mintz). 0031-9384/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.physbeh.2011.08.032
5-HT release in the brain is highest during periods of activity, and in nocturnal rodents peak 5-HT release occurs around the time of lights off. When induced to run on running wheels, animals show increases in 5-HT levels during mid-subjective day; suggesting that 5HT release is correlated with arousal and locomotor activity [10]. In turn, locomotor activity attenuates light-induced phase shifts [11]. Interestingly, light-induced phase shifts can also be attenuated by pharmacological treatment with (±)-8-hydroxy-2-(dipropylamino) tetralin hydrobromide (8-OH-DPAT), a 5-HT1A/7 receptor agonist [12], and 3-trifluoromethylphenylpiperazine (TFMPP), a 5-HT1B agonist [13]. This differs from the circadian system's response to systemic injections of a 5-HT agonist during the mid-subjective day, which cause phase advances [14–20]. Conversely, inhibition of 5-HT release and/or administration of a 5-HT receptor antagonist potentiates light-induced phase shifts [21–25]. In hamsters, destruction of 5-HT afferents by 5,7-dihydroxytryptamine (5,7-DHT), a neurotoxin that selectively kills serotonergic neurons, causes lengthening of the active phase and longer periods in constant light (LL), but no change in period length in constant darkness (DD) [26,27]. Mice treated with 5,7-DHT and blinded by enucleation also show lengthening of the active phase and exhibit longer circadian periods than controls. Additionally, these mice show greater total activity and altered distribution of activity within the active
258
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263
phase with delays in onset of peak running activity and increased occurrences of bimodal patterns of activity [28]. These data indicate that disruption of serotonergic neurotransmission can have profound consequences for the expression of circadian rhythms. Pet-1, a transcription factor belonging to the ETS family, is an early developmental indicator of neurons destined for a 5-HTergic fate. Pet-1 expression is restricted to 5-HT neurons in the rat and mouse, appearing at embryonic day 12.5, about half a day before 5-HT is detectable. Pet-1 mRNA colocalizes with tryptophan hydroxylase (TPH) in neurons but not non-neuronal cells [29,30], and is necessary for the maintenance of a serotonergic phenotype in adults [31]. Mice lacking the Pet-1 gene appear normal and have normal feeding behavior, motor learning, balance, coordination but increased aggression and anxiety-like behaviors [32]. These animals also show a significant disruption of the 5-HT system, without effects on other monoamine systems, with a 70% loss of 5-HT immunopositive cell bodies in adult Pet-1 KOs. While the remaining 30% of 5-HT cell bodies retain their neuronal phenotype and appear to be normally positioned, they show little to no protein or mRNA expression for some of the substances that are common to the 5-HT neuron phenotype: TPH, 5-HT transporter, and vesicular monoamine transporter 2. Further, Pet-1 KO mice have a reduced density of 5-HT immunopositive fibers in target fields and the brain tissue content of 5-HT and 5hydroxyindoleacetic acid in the Pet-1 KOs are only 10–15% of WT levels [32]. The effects of the loss of Pet-1 on peripheral serotonin levels are unknown. It is clear from these studies that alteration of 5-HT input to the SCN changes the expression of behavioral circadian rhythms. What is unknown is how developmental disruption of the entire system can affect the SCN and behavioral clock output. The Pet-1 KO mouse provides an opportunity to examine this question, so in this study we examine the expression of behavioral rhythmicity in these mice, with the goal of further clarifying the role of 5-HT in circadian rhythm regulation.
WT animals were identified by DNA fragments of 209 bp while Pet-1 KO animals were defined by DNA fragments of 361 bp. The number of animals of each sex and genotype were counted to assess whether breeding results corresponded to the expected Mendelian ratios. Differences from expected ratios were tested with a G-test of goodnessof-fit while differences between sexes were assessed by a G-test for independence. 2.3. Behavioral monitoring WT (n = 8), heterozygote (HET) (n= 8) and KO (n = 8) animals were placed into individual cages equipped with running wheels so that activity patterns could be monitored. Food and water was available ad libitum. Wheel-running activity was monitored for at least 21 days under each of three light schedules: LD 14:10, DD, and LL using Clocklab (Actimetrics) software. An additional group of animals was profiled in LD 12:12. Actograms for each animal were examined for differences in activity patterns between genotypes. Hourly counts of the number of wheel revolutions per hour were averaged across animals and normalized to the baseline 24-hour activity total for each animal to eliminate the effects of running wheels with different frictional resistances on the results. Periods in LD, DD, and LL were calculated by chi-squared periodogram for each animal using Clocklab software then grouped and averaged according to genotype. For the extended DD experiment, WT (n= 8) and KO animals (n= 6) were housed in LD 12:12 then transferred to and maintained in DD for 70 days. Periods were calculated for 10 day increments, starting at day 11 and ending at day 70, and averaged across genotypes. Statistical analysis for all behavioral experiments was performed using repeated measures ANOVAs, with posthoc comparisons to control groups performed using Dunnett's test. 2.4. Histology
2. Materials and methods 2.1. Animals Pet-1 KO animals on a mixed 129Sv and C57BL/6 background were obtained from Dr. Evan Deneris (Case Western Reserve University, Cleveland, OH) [32]. Pet-1 heterozygote breeding pairs were used to produce all experimental animals, which were 2–5 months of age at the time of experimentation. Heterozygotes were used as Pet-1 homozygotes are deficient in maternal behaviors [33]. Experiments were conducted by matching siblings of different genotypes as much as possible to avoid bias from the mixed genetic background. At the time of weaning, animals were ear-tagged and a tail-snip was collected for DNA extraction and genotyping. Experimental animals were entrained to a 14:10 or 12:12 light–dark (LD) cycle prior to the start of the experiment as noted below. Food and water were available ad libitum and all experimental procedures were approved by the Kent State University Institutional Animal Care and Use Committee. 2.2. Genotyping Tail snips from all animals were placed into an extraction buffer containing 20 mM Tris pH 8.0, 1 mM EDTA, 400 mM NaCl, and 0.5% SDS with 20 mg/mL Proteinase K added in order to digest the snip. Samples were maintained at 55 °C with slow rotation overnight. Supernatant from each sample was spun down in order to pellet DNA, rinsed with ethanol and resuspended in TE buffer. DNA of interest was amplified using PCR with primers from IDT (3′ PET: 5′ GCC TGA TGT TCA AGG AAG ACC TCG G 3′ 5′ PET: 3′ CGC ACT TGG GGG GTC ATT ATC AC 3′ 5′ LOX: 5′ CGG TGG ATG TGG AAT GTG TGC G 3′) and thermocycler conditions beginning with a 3 min hold at 94 °C followed by 42 cycles of 94 °C for 50 s, 62 °C for 30 s and 72 °C for 40 s. PCR product was analyzed using the Agilent 2100 Bioanalyzer system.
For the examination of SCN neuropeptide expression, animals were given a lethal overdose of sodium pentobarbital intraperitoneally (i.p.) and perfused transcardially with 4% ice-cold paraformaldehyde. Brains were removed and post-fixed overnight before being transferred into PBS. Anterior and posterior portions of brains were sectioned at 40 μm using either a vibratome (for arginine vasopressin (AVP) and 5-HT) or cryostat (for vasoactive intestinal polypeptide (VIP)); if a cryostat was used, brains were first cryoprotected in a 30% sucrose solution. Sections from each animal were then processed for immunohistochemistry for 5-HT (Immunostar rabbit anti-5-HT 1:2000 and Jackson Research Cy2 secondary 1:500) and images of the SCN, median raphe, and dorsal raphe were taken for analysis. 5-HT fluorescent staining was examined to confirm the level of 5-HT depletion in each mouse brain. Levels of AVP and VIP were also examined using immunohistochemistry (Chemicon rabbit anti-AVP 1:50,000 and Immunostar rabbit anti-VIP 1:10,000, Jackson Research biotinylated donkey anti-rabbit secondary 1:500 and processed for peroxidase reaction using a Vector Laboratories DAB kit) and images taken for comparison. ImageJ software was used to perform counts of AVP immunopositive cells in the SCN of WT (n= 5), HET (n= 5) and KO (n= 5) animals. Counts of AVP immunoreactive cells were also performed in the paraventricular nucleus (PVN) for each genotype. Statistical analysis of AVP cell counts was done using a one-way ANOVA. VIP immunopositive cells in the SCN could not be counted due to the density of the labeling which obscured individual cells, so only a qualitative inspection was performed. 3. Results 3.1. 5-HT immunohistochemistry in the SCN, MR, and DR Immunohistochemical staining was performed for 5-HT in order to visualize the extent of 5-HT fiber innervation to the SCN (Fig. 1A–C).
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263
259
Fig. 1. Serotonin (5-HT) immunoreactivity in the SCN (A–C), median raphe (MR, D–F), and dorsal raphe (DR, G–I) in wildtype (A, D, G), heterozygote (B, E, H) and Pet-1 knockout (C, F, I) animals. Images have been adjusted for contrast.
Also, since the MR and DR provide ascending 5-HT innervation to the brain, the density of immunopositive cell bodies in the MR (Fig. 1D–F) and DR (Fig. 1G–I) in WT, HET, and KO animals was examined. 5-HT fiber innervation to the SCN of WT and HET animals remained intact (Fig. 1A and B) while the SCN of KO animals was almost devoid of 5HT-positive fibers (Fig. 1C). As expected, the MR of WT and HET animals showed cell bodies immunopositive for 5-HT (Fig. 1D and E) while staining was almost absent in KO animals (Fig. 1F). Similarly, there were numerous 5-HT-labeled cell bodies in the DR of WT and HET animals (Fig. 1G and H) while the number of 5-HT cells in the DR of KO animals was greatly reduced (Fig. 1I). 3.2. Immunohistochemistry for VIP and AVP Since previous experiments have shown evidence that 5-HT depletion can alter the expression pattern of neuropeptides within the brain [34], the effects of a genetic depletion of 5-HT on VIP and AVP cell distribution were examined. VIP immunoreactivity in the SCN of WT, HET, and KO animals showed no obvious differences between genotypes (Fig. 2A–C), but the density of fiber staining prohibited quantification. AVP immunoreactivity was assayed in the SCN by counting the number of immunoreactive cell bodies (Fig. 2D–F). There were no significant differences in total number of AVP positive cells in the SCN between genotypes: 417.75 ± 32.7 cells in WT (Fig. 2D), 400.6 ±
13.89 cells in HET (Fig. 2E), and 382.25 ± 10.48 cells in KO (Fig. 2F) (F(2,10) = 0.70, p = 0.52). Counts of immunopositive AVP cells in the PVN were also not significantly different between genotypes (data not shown). 3.3. Circadian parameters Representative actograms show the differences in wheel-running patterns between genotypes in LD, DD, and LL (Fig. 3A and B). WT (Fig. 3A) animals displayed an average period of 23.59 ± 0.06 h in DD, and 24.90 ± 0.26 h in LL. HET animals displayed an average period of 23.63 ± 0.07 h in DD, and 25.68 ± 0.18 h in LL. KO (Fig. 3B) animals displayed an average period of 24.01 ± 0.04 h in DD, and 24.73 ± 0.23 h in LL. KO animals had significantly longer periods in DD when compared to both WT and HET animals (F(2,20) = 19.57, p = 0.00002). No significant differences were seen between WT and HET animals in DD or LL, however HET animals showed significantly longer circadian periods than KO in LL (F(2,15) = 3.70, p = 0.0495). No differences were seen between WT and KO animals in LL. Results are summarized in Fig. 4. We attempted to measure mean activity onset, activity offset, phase angle of entrainment (in LD), and alpha (duration of daily activity period) and compare across genotypes based on an analysis of each animal's actogram. However, these measures were highly
Fig. 2. Vasoactive intestinal polypeptide (VIP, A–C) and arginine vasopressin (AVP, D–F) immunoreactivity in the SCN of wildtype (A and D), heterozygote (B and E) and Pet-1 knockout (C and F) animals. Images have been adjusted for brightness.
260
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263
Fig. 3. (A and B) Representative double-plotted actograms of circadian wheel-running activity for wildtype (A) and Pet-1 knockout (B) animals under three photo schedules: 14:10 light:dark (LD), constant darkness (DD) and constant light (LL). Actograms for heterozygote animals resemble those for wildtype and are not shown.
variable, both between and within animals, and did not yield any interpretable data. Similarly, attempts to measure alpha in DD and LL resulted in very large standard deviations within groups.
bouts distinctly associated with lights on and lights off is seen in virtually all KO animals but not in WT animals (Fig. 3).
3.5. Extended DD 3.4. Activity profile In order to determine whether there were differences in partitioning of activity between genotypes, total wheel-running per hour in LD as a percentage of total daily activity was examined (Fig. 5A and B). In 14:10 LD (Fig. 5A), a significant interaction between time and genotype was observed (F(46,391) = 3.86, p b 0.000001), in which WT (n= 6) and HET (n= 6) animals showed a significantly greater percentage of total wheel-running at ZT 15 and ZT 16 than KO (n= 8) animals while KO animals showed a significantly greater percentage of wheel-running behavior in the late portion of the dark phase at ZT 24 (pb 0.05 in each case, Dunnett's test). Although our initial experiments were run with the animals in 14:10 LD, we wanted to assess their profile in 12:12 LD as well to determine whether the late night increases in activity were more pronounced with a longer dark phase. 12:12 LD also represents a more normal housing light cycle for laboratory mice; the original 14:10 LD cycle had been used solely due to compatibility with other experiments going on in the same room. In 12:12 LD (Fig. 5B) a significant interaction between time and genotype was observed (F(23,483) = 7.65, p b 0.000001) in which WT (n= 11) animals showed a significantly greater percentage of total wheel-running at ZT 13, ZT 14, ZT 15, and ZT 16 while KO animals (n= 12) showed a significantly greater percentage of wheel-running behavior in the late portion of the dark phase, at ZT 24 (pb 0.05 in each case, Dunnett's test). Unimodal patterns of activity were seen in WT and HET animals while activity in KO animals appeared bimodal; the pattern of a separation of activity into
Fig. 4. Circadian period (mean ± SEM) in hours of wildtype (WT), heterozygote (HET) and Pet-1 knockout (KO) animals in constant darkness (DD) and constant light (LL). Asterisks denote significance (p b 0.05).
Representative actograms show differences between genotypes in wheel-running behavior over 10 weeks of DD (Fig. 6A and B). WT (Fig. 6A) animals exhibit a shortening of period over time in extended DD while KO (Fig. 6B) animals retain significantly longer periods over time. There were significant effects of both genotype (F(1,12) = 12.12, p = 0.005), length of time in DD (F(5,60) = 6.48, p = 0.00007), and the interaction term (F(5,60) = 5.00, p = 0.0007). In WT animals, circadian periods for days 41–50 (23.14 ± 0.20 h), 51–60 (23.22 ± 0.20 h), and 61–70 (23.04 ± 0.19 h) are significantly shorter than at days 11–20
Fig. 5. Activity profile in 14:10 light:dark (LD) for wildtype, heterozygote, and Pet-1 knockout animals (A) and activity profile in 12:12 LD for wildtype and Pet-1 knockout animals (B). Data is represented as mean percentage of total activity in each 1 h bin throughout the 24 h light–dark cycle, and error bars show the standard error of the mean. Vertical bars indicate lights off. Asterisks denote significance (p b 0.05).
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263
261
Fig. 6. (A and B) Representative double-plotted actograms of circadian rhythm activity in 10 weeks of constant darkness (DD) for wildtype (A) and Pet-1 knockout (B) animals.
(23.61 ± 0.07 h) (p b 0.05, Dunnett's test). Periods in KO animals did not differ significantly between time points. Results are summarized in Fig. 7. 3.6. Assessment of breeding ratios To date, we have genotyped 1805 mice. The totals by genotype were 477 WT, 965 HET, and 363 KO animals. This distribution of values differs significantly from the expected 1:2:1 ratio (χ2(2) = 23.06, p b 0.000001). Only about 76% of the expected number of KO animals survived to be weaned and genotyped. There was no significant difference in the ratio of genotypes between males and females (χ2(2) = 2.39, p = 0.30). 4. Discussion In this study, we examined the effects of a partial, developmental disruption of the 5-HT system on 5-HT input to the SCN, circadian output behavior, and neuropeptide distribution within the SCN. Immunohistochemical analysis of 5-HT immunoreactivity shows numerous 5-HT cell bodies in the MR and DR of WT and HET animals as well as extensive 5-HT fiber innervation to the SCN. Pet-1 KO animals show virtually no 5HT labeling in the SCN and greatly reduced numbers of 5-HT cell bodies in the MR and DR. Our results are similar to previously published reports indicating reduced 5-HT cell bodies in the MR and DR and an absence of 5-HT fibers in target fields such as the cortex and hippocampus [32]. Since previous studies have demonstrated an important role for 5-HT in the regulation of behavioral circadian rhythms, we expected to find changes in the circadian wheel-running activity rhythm.
Fig. 7. Circadian period (mean ± SEM) in hours for wildtype and Pet-1 knockout animals in 10 weeks of DD. Mean periods are grouped in ten day bins starting at day 11 and ending at day 70. KO animals have significantly longer periods than WT at all time points. Asterisks denote significant differences between circadian periods (compared to period from days 11–20 bin) within the WT group only (pb 0.05).
Our study found that Pet-1 KO and HET mice could entrain to 14:10 and 12:12 LD schedules. In DD, KO animals had significantly longer periods than WT or HET animals. These findings differ with published reports in Syrian hamsters which indicate that 5-HT depletion, due to 5,7DHT administration, does not change the circadian period in DD but rather, lengthens the period in LL [26,27]. These differences could be due to the loss of 5-HT during development in Pet-1 KO mice compared to destruction of the 5-HT system in adults, or could reflect species differences. Our results are more similar to previously published reports showing that mice blinded by enucleation and treated with 5,7-DHT have longer circadian periods, a longer active phase, delayed onset of peak running activity within the active phase, and increased occurrences of bimodal patterns of activity compared to controls [28]. A lengthening of period in DD as a result of serotonin deficiency could be explained if the serotonergic input to the circadian clock in the SCN has an average effect of phase advancing the clock. Evidence for the phase advancing properties of 5-HT exists in numerous studies showing that the phase response curve to 5-HTergic agonists shows a large phase advance region during the middle of the subjective day [14–20]. At this time of day extracellular serotonin concentrations are normally low [10], but it could be that continuous release of serotonin at a low level during the day results in a daily phase advance, and absence of that stimulation would slow the clock, leading to longer free-running periods. This effect is similar to that seen in 5,7-DHT lesioned mice [28]. However, 5-HT1A [35], 5-HT1B [36], and 5-HT7 [37] receptor knockout mice show no differences in free-running period compared to WT counterparts suggesting that the lengthening of period in Pet-1 KOs is not mediated by any one of these receptor subtypes. The changes in period are therefore likely the result of the loss of 5-HT action on multiple receptors simultaneously. It is also possible that an inducible receptor knockout would show a phenotype if the receptor were eliminated in an adult, but that the loss of that receptor during development leads to compensation by other receptor subtypes. The absence of the shortening of free-running period during extended DD in Pet-1 KO mice may indicate that 5-HT plays a role in long-term changes associated with light deprivation, but our results do not suggest any specific mechanism for this phenomenon. Unfortunately, we do not know the status of 5-HT receptor levels in Pet-1 KO mice. The 5,7-DHT lesion studies suggest, though, that the change in free-running period is not a result of the loss of serotonin during early development, but is a more acute effect derived either from the loss of serotonin itself, or the loss of serotonergic neurons that may exert effects through other signaling molecules in addition to serotonin. Furthermore, the failure of any 5-HT receptor knockouts to mimic the behavioral phenotype seen in the Pet-1 KO mice may indicate that developmental compensation may occur in the case of the knockouts but not in the case of the loss of the transmitter. It is also possible that
262
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263
the phenotype may be a function of the fact that while 5-HT levels are severely reduced in knockouts, some residual 5-HT is still present; therefore, the phenotype may be more representative of a severe reduction in 5-HT rather than its total elimination. An additional hypothesis might be that the changes in period may be mediated by changes in other neurotransmitter systems. Serotonin antagonists increase glutamate release in the SCN [38]; other neurotransmitters may be affected as well. Increased glutamate in the early night in Pet-1 KO mice, a time when serotonin release is high in the WT mice, might result in daily light-like phase delays, thus increasing free-running period. The loss of 5-HT could also alter release of intrinsic neuropeptides such as VIP, GRP, or AVP as well as neuropeptides innervating the SCN such as NPY [39–41]. A second effect of a slower clock might be a shift in the profile of daily activity. In both 14:10 and 12:12 LD cycles, Pet-1 KO animals shift a substantial proportion of their daily wheel-running activity from the early part of the night to the late night. Some KO animals showed wheel-running that extended into the period of lights-on, which may suggest a reduced suppressive effect of light on activity. However, we did not see evidence of a change in lights' effect on the circadian clock in the form of an effect on free-running period in LL in KO animals. HET animals had periods in LL that were not significantly different from WT, though they were significantly longer than KO. A difference between HET and KO is difficult to explain as the HET animals otherwise show a phenotype identical to WT. In the course of observing the circadian behaviors of the mice, we noticed that the free-running period of WT mice appeared to decrease over time while in DD. In order to examine this further, we performed an experiment to examine changes in period over time in extended DD. KO animals maintained significantly longer periods than WT animals and did not show the shortening of circadian period over time seen in WT animals, indicating that the longer periods observed in Pet-1 KOs are a stable effect of 5-HT depletion. It also suggests that the decrease of free-running period in DD in WT animals may be through a mechanism that involves serotonergic feedback on the circadian clock. It was possible that the changes in circadian clock function seen in KO mice were the result of structural changes in the SCN. To examine this possibility, we examined the patterns of VIP and AVP immunoreactivity in the SCN. No obvious differences between genotypes were observed based on hemotoxylin staining of the SCN. In mice, VIP coordinates daily rhythms in the SCN by synchronizing a subset of neurons [42] and AVP expression patterns correspond with the subset of SCN cells that are intrinsically rhythmic [43]. No differences were seen in the distribution of VIP immunoreactivity between genotypes. Likewise, AVP immunoreactivity, which in the mouse is concentrated in the dorsomedial region of the SCN as well as in clusters around the edges of the SCN, did not qualitatively differ between genotypes. In addition, we counted the number of AVP-positive cell bodies and no differences were detected. Previous studies suggest that depletion of 5-HT by administration of p-chlorophenylalanine (pCPA) to pregnant rats causes lasting changes in neuropeptide content within the SCN in progeny, specifically by increasing numbers VIP and AVP immunopositive neurons and the concentration of these peptides within cell bodies [34]. We did not see a comparable effect in Pet-1 KO animals, supporting the idea that lengthening of period in the KO animals is not due to disruption of SCN structure or neurochemical organization. In the course of breeding the mice for our studies, we noted that there were approximately 25% fewer KO animals at weaning age than would be expected from Mendelian genetic ratios, suggesting an increased mortality rate associated with Pet-1 KOs. Hendricks et al. [32] mentioned a non-Mendelian ratio of genotypes due to a loss of Pet-1 KO mice within the first week after birth. Since the stomachs of non-surviving neonates were full of milk, this loss is not thought to be due to a decreased ability to feed. It likely has to do with an impairment of the respiratory system in neonates [44,45]. Recently, it's been reported that Lmx1b flox/flox;ePet-Cre/+ mice, which lack 5-HT neurons,
display apnea which is most severe during the first two weeks of life. Apnea was shown to improve by 2–4 weeks after birth but these animals showed a high rate of perinatal mortality and decreased growth rate until respiratory deficits resolved in surviving mice [46]. This report suggests that Pet-1 KO mice, which also lack 5-HT neurons, may display similar respiratory difficulties leading to the increased mortality seen in KO animals. However, we did not see any deficiencies in overall locomotor activity in Pet-1 KO adults, so we do not believe that this developmental affect is having a material effect on our results regarding the characteristics of circadian clock function. Overall, these data indicate that a dysfunctional 5-HT system causes changes not just in circadian clock period, but also in the distribution of locomotor activity across the circadian cycle. In serotonin-deficient mice, peak levels of running-wheel activity were redistributed from a primarily unimodal pattern into a bimodal matter, with dramatically increased activity levels in the hour immediately prior to lights on. This change may be secondary to the lengthening of circadian period but may also represent an organizational effect of 5-HT on activity patterns. It is also possible that the change in the distribution of activity itself plays a role in the change in circadian period. Therefore, these data suggest that 5-HT plays a significant role in regulating the timing of locomotor activity, and that dysfunction of the 5-HT system may result in changes in behavioral activity patterns.
Acknowledgments We would like to thank Evan Deneris for providing the Pet-1 knockout mice, RoxAnne Murphy for her assistance with the genotyping protocol, and Heather Caldwell for her helpful comments on the manuscript.
References [1] Rusak B, Zucker I. Neural regulation of circadian rhythms. Physiol Rev 1979;59: 449–526. [2] Moore RY. Organization and function of a central nervous system circadian oscillator: the suprachiasmatic hypothalamic nucleus. Fed Proc 1983;42:2783–9. [3] Klein DC, Moore RY, Reppert SM. Suprachiasmatic nucleus: the mind's clock. USA: Oxford University Press; 1991. [4] Ruby NF, Burns DE, Heller HC. Circadian rhythms in the suprachiasmatic nucleus are temperature-compensated and phase-shifted by heat pulses in vitro. J Neurosci 1999;19:8630–6. [5] Abe H, Rusak B. Anticipatory activity and entrainment of circadian rhythms in Syrian hamsters exposed to restricted palatable diets. Am J Physiol Regul Integr Comp Physiol 1992;263:116–24. [6] Reebs SG, Mrosovsky N. Effects of induced wheel running on the circadian activity rhythms of Syrian hamsters: entrainment and phase response curve. J Biol Rhythms 1989;4:39–48. [7] Morin LP. Serotonin and the regulation of mammalian circadian rhythmicity. Ann Med 1999;31:12–33. [8] Meyer-Bernstein EL, Morin LP. Differential serotonergic innervation of the suprachiasmatic nucleus and the intergeniculate leaflet and its role in circadian rhythm modulation. J Neurosci 1996;16:2097–111. [9] Yamakawa GR, Antle MC. Phenotype and function of raphe projections to the suprachiasmatic nucleus. Eur J Neurosci 2010;31:1974–83. [10] Dudley TE, Dinardo LA, Glass JD. Endogenous regulation of serotonin release in the hamster suprachiasmatic nucleus. J Neurosci 1998;18:5045–52. [11] Ralph MR, Mrosovsky N. Behavioral inhibition of circadian responses to light. J Biol Rhythms 1992;7:353–9. [12] Rea MA, Glass JD, Colwell CS. Serotonin modulates photic responses in the hamster suprachiasmatic nuclei. J Neurosci 1994;14:3635–42. [13] Pickard GE, Rea MA. TFMPP, a 5HT1B receptor agonist, inhibits light-induced phase shifts of the circadian activity rhythm and c-Fos expression in the mouse suprachiasmatic nucleus. Neurosci Lett 1997;231:95–8. [14] Tominaga K, Shibata S, Ueki S, Watanabe S. Effects of 5-HT1A receptor agonists on the circadian rhythm of wheel-running activity in hamsters. Eur J Pharmacol 1992;214:79–84. [15] Edgar DM, Miller JD, Prosser RA, Dean RR, Dement WC. Serotonin and the mammalian circadian system: II. Phase-shifting rat behavioral rhythms with serotonergic agonists. J Biol Rhythms 1993;8:17–31. [16] Cutrera RA, Ouarour A, Pevet P. Effects of the 5-HT1a receptor agonist 8-OH-DPAT and other non-photic stimuli on the circadian rhythm of wheel-running activity in hamsters under different constant conditions. Neurosci Lett 1994;172:27–30.
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263 [17] Cutrera RA, Saboureau M, Pevet P. Phase-shifting effect of 8-OH-DPAT, a 5HT1A/5-HT7 receptor agonist, on locomotor activity in golden hamster in constant darkness. Neurosci Lett 1996;210:1–4. [18] Mintz EM, Gillespie CF, Marvel CL, Huhman KL, Albers HE. Serotonergic regulation of circadian rhythms in Syrian hamsters. Neuroscience 1997;79:563–9. [19] Horikawa K, Yokota S, Fuji K, Akiyama M, Moriya T, Okamura H, et al. Nonphotic entrainment by 5-HT1A/7 receptor agonists accompanied by reduced Per1 and Per2 mRNA levels in the suprachiasmatic nuclei. J Neurosci 2000;20:5867–73. [20] Horikawa K, Shibata S. Phase-resetting response to (±) 8-OH-DPAT, a serotonin 1A/7 receptor agonist, in the mouse in vivo. Neurosci Lett 2004;368:130–4. [21] Rea MA, Barrera J, Glass JD, Gannon RL. Serotonergic potentiation of photic phase shifts of the circadian activity rhythm. Neuroreport 1995;6:1417–20. [22] Smart CM, Biello SM. WAY-100635, a specific 5-HT1A antagonist, can increase the responsiveness of the mammalian circadian pacemaker to photic stimuli. Neurosci Lett 2001;305:33–6. [23] Gannon RL. Serotonergic serotonin1A mixed agonists/antagonists elicit largemagnitude phase shifts in hamster circadian wheel-running rhythms. Neuroscience 2003;119:567–76. [24] Lall GS, Harrington ME. Potentiation of the resetting effects of light on circadian rhythms of hamsters using serotonin and neuropeptide Y receptor antagonists. Neuroscience 2006;141:1545–52. [25] Kessler EJ, Sprouse J, Harrington ME. NAN-190 potentiates the circadian response to light and speeds re-entrainment to advanced light cycles. Neuroscience 2008;154:1187–94. [26] Smale L, Michels KM, Moore RY, Morin LP. Destruction of the hamster serotonergic system by 5, 7-DHT: effects on circadian rhythm phase, entrainment and response to triazolam. Brain Res 1990;515:9–19. [27] Morin LP, Blanchard J. Depletion of brain serotonin by 5, 7-DHT modifies hamster circadian rhythm response to light. Brain Res 1991;566:173–85. [28] Mistlberger RE, Bossert JM, Holmes MM, Marchant EG. Serotonin and feedback effects of behavioral activity on circadian rhythms in mice. Behav Brain Res 1998;96:93–9. [29] Hendricks T, Francis N, Fyodorov D, Deneris ES. The ETS domain factor Pet-1 is an early and precise marker of central serotonin neurons and interacts with a conserved element in serotonergic genes. J Neurosci 1999;19:10348–56. [30] Pfaar H, von Holst A, Vogt Weisenhorn DM, Brodski C, Guimera J, Wurst W. mPet1, a mouse ETS-domain transcription factor, is expressed in central serotonergic neurons. Dev Genes Evol 2002;212:43–6. [31] Liu C, Maejima T, Wyler SC, Casadesus G, Herlitze S, Deneris ES. Pet-1 is required across different stages of life to regulate serotonergic function. Nat Neurosci 2010;13:1190–8.
263
[32] Hendricks TJ, Fyodorov DV, Wegman LJ, Lelutiu NB, Pehek EA, Yamamoto B, et al. Pet-1 ETS gene plays a critical role in 5-HT neuron development and is required for normal anxiety-like and aggressive behavior. Neuron 2003;37:233–47. [33] Lerch-Haner JK, Frierson D, Crawford LK, Beck SG, Deneris ES. Serotonergic transcriptional programming determines maternal behavior and offspring survival. Nat Neurosci 2008;11:1001–3. [34] Mirochnik V, Bosler O, Tillet Y, Calas A, Ugrumov M. Long-lasting effects of serotonin deficiency on differentiating peptidergic neurons in the rat suprachiasmatic nucleus. Int J Dev Neurosci 2005;23:85–91. [35] Smith VM, Sterniczuk R, Phillips CI, Antle MC. Altered photic and non-photic phase shifts in serotonin1A receptor knockout mice. Neuroscience 2008;157: 513–23. [36] Sollars PJ, Ogilvie MD, Rea MA, Pickard GE. 5-HT1B receptor knockout mice exhibit an enhanced response to constant light. J Biol Rhythms 2002;17:428–37. [37] Gardani M, Biello SM. The effects of photic and nonphotic stimuli in the 5-HT7 receptor knockout mouse. Neuroscience 2008;152:245–53. [38] Srkalovic G, Selim M, Rea MA, Glass JD. Serotonergic inhibition of extracellular glutamate in the suprachiasmatic nuclear region assessed using in vivo brain microdialysis. Brain Res 1994;656:302–8. [39] Glass JD, Guinn J, Kaur G, Francl JM. On the intrinsic regulation of neuropeptide Y release in the mammalian suprachiasmatic nucleus circadian clock. Eur J Neurosci 2010;31:1117–26. [40] Francl JM, Kaur G, Glass JD. Regulation of vasoactive intestinal polypeptide release in the suprachiasmatic nucleus circadian clock. Neuroreport 2010;21:1055–9. [41] Francl JM, Kaur G, Glass JD. Roles of light and serotonin in the regulation of gastrin-releasing peptide and arginine vasopressin output in the hamster SCN circadian clock. Eur J Neurosci 2010;32:1170–9. [42] Aton SJ, Colwell CS, Harmar AJ, Waschek J, Herzog ED. Vasoactive intestinal polypeptide mediates circadian rhythmicity and synchrony in mammalian clock neurons. Nat Neurosci 2005;8:476–83. [43] Antle MC, Silver R. Orchestrating time: arrangements of the brain circadian clock. Trends Neurosci 2005;28:145–51. [44] Cummings KJ, Li A, Deneris ES, Nattie EE. Bradycardia in serotonin-deficient Pet1−/− mice: influence of respiratory dysfunction and hyperthermia over the first 2 postnatal weeks. Am J Physiol Regul Integr Comp Physiol 2010;298:R1333–42. [45] Erickson JT, Shafer G, Rossetti MD, Wilson CG, Deneris ES. Arrest of 5HT neuron differentiation delays respiratory maturation and impairs neonatal homeostatic responses to environmental challenges. Respir Physiol Neurobiol 2007;159:85–101. [46] Hodges MR, Wehner M, Aungst J, Smith JC, Richerson GB. Transgenic mice lacking serotonin neurons have severe apnea and high mortality during development. J Neurosci 2009;29:10341–9.
Contents lists available at SciVerse ScienceDirect
Physiology & Behavior journal homepage: www.elsevier.com/locate/phb
Developmental disruption of the serotonin system alters circadian rhythms Erin V. Paulus, Eric M. Mintz ⁎ Department of Biological Sciences, Kent State University, Kent, OH 44242, United States School of Biomedical Sciences, Kent State University, Kent, OH 44242, United States
a r t i c l e
i n f o
Article history: Received 29 September 2010 Received in revised form 3 August 2011 Accepted 24 August 2011 Keywords: Pet-1 Serotonin Suprachiasmatic nucleus Circadian Wheel-running Locomotor
a b s t r a c t Serotonin (5-HT) plays an important role in circadian rhythms, acting to modulate photic input to the mammalian clock, the suprachiasmatic nucleus (SCN), as well as playing a role in non-photic input. The transcription factor Pet-1 is an early developmental indicator of neurons that are destined for a 5-HTergic fate. Mice lacking the Pet-1 gene show a 70% loss of 5-HT immunopositive cell bodies in adult animals. 5-HT neurotoxic lesion studies using 5,7-dihydroxytryptamine (5,7-DHT) have highlighted species-specific differences in response to 5-HT depletion and studies using knockout mice lacking various 5-HT receptors have helped to elucidate the role of individual 5-HT receptors in mediating 5-HT's effects on circadian rhythms. Here we investigate the effects of a developmental disruption of the 5-HT system on the SCN and circadian wheel-running behavior. Immunohistochemical analysis confirmed depletion of 5-HT fiber innervation to the SCN as well as greatly reduced numbers of cell bodies in the raphe nuclei in Pet-1 knockout mice. These mice also display significantly longer free-running periods than wildtype or heterozygote counterparts. In light–dark cycles, knockouts showed a shift in peak wheel running behavior towards the late night as compared to wildtype and heterozygote animals. When kept in constant darkness for 70 days, wildtype animals showed decreases in free-running period over time while the period of knockout animals remained constant. Immunohistochemical analysis for neuropeptides within the SCN indicates that the behavioral changes observed in Pet-1 knockout mice were not due to gross changes in SCN structure. These results suggest that developmental loss of serotonergic input to the clock has long-term consequences for both circadian clock parameters and the temporal organization of activity. © 2011 Elsevier Inc. All rights reserved.
1. Introduction In mammals, behavioral circadian rhythms are generated by an endogenous clock located in the suprachiasmatic nucleus (SCN) within the ventral hypothalamus. Rhythms are entrained to the 24 h light– dark cycle by external cues, with light functioning as the primary cue [1–3]. However, other non-photic stimuli such as temperature, food availability, or activity can also serve as cues for entrainment [4–6]. Serotonin (5-HT) is a neurotransmitter that plays an important role in circadian rhythms, modulating photic input to the SCN as well as playing a role in non-photic phase shifts [7]. One of the 5-HTergic pathways thought to convey non-photic information to the SCN, such as the activity level of the animal, as well as modulate photic effects is input from the median raphe (MR) [8]. There is also evidence suggesting direct 5-HTergic input to the SCN from the dorsal raphe (DR) [9]; the DR can also influence the SCN through an indirect connection via the intergeniculate leaflet (IGL) [8].
⁎ Corresponding author at: Department of Biological Sciences, Kent State University, Kent, OH 44242, United States. Tel.: + 1 330 672 3847; fax: + 1 330 672 3713. E-mail address: [email protected] (E.M. Mintz). 0031-9384/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.physbeh.2011.08.032
5-HT release in the brain is highest during periods of activity, and in nocturnal rodents peak 5-HT release occurs around the time of lights off. When induced to run on running wheels, animals show increases in 5-HT levels during mid-subjective day; suggesting that 5HT release is correlated with arousal and locomotor activity [10]. In turn, locomotor activity attenuates light-induced phase shifts [11]. Interestingly, light-induced phase shifts can also be attenuated by pharmacological treatment with (±)-8-hydroxy-2-(dipropylamino) tetralin hydrobromide (8-OH-DPAT), a 5-HT1A/7 receptor agonist [12], and 3-trifluoromethylphenylpiperazine (TFMPP), a 5-HT1B agonist [13]. This differs from the circadian system's response to systemic injections of a 5-HT agonist during the mid-subjective day, which cause phase advances [14–20]. Conversely, inhibition of 5-HT release and/or administration of a 5-HT receptor antagonist potentiates light-induced phase shifts [21–25]. In hamsters, destruction of 5-HT afferents by 5,7-dihydroxytryptamine (5,7-DHT), a neurotoxin that selectively kills serotonergic neurons, causes lengthening of the active phase and longer periods in constant light (LL), but no change in period length in constant darkness (DD) [26,27]. Mice treated with 5,7-DHT and blinded by enucleation also show lengthening of the active phase and exhibit longer circadian periods than controls. Additionally, these mice show greater total activity and altered distribution of activity within the active
258
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263
phase with delays in onset of peak running activity and increased occurrences of bimodal patterns of activity [28]. These data indicate that disruption of serotonergic neurotransmission can have profound consequences for the expression of circadian rhythms. Pet-1, a transcription factor belonging to the ETS family, is an early developmental indicator of neurons destined for a 5-HTergic fate. Pet-1 expression is restricted to 5-HT neurons in the rat and mouse, appearing at embryonic day 12.5, about half a day before 5-HT is detectable. Pet-1 mRNA colocalizes with tryptophan hydroxylase (TPH) in neurons but not non-neuronal cells [29,30], and is necessary for the maintenance of a serotonergic phenotype in adults [31]. Mice lacking the Pet-1 gene appear normal and have normal feeding behavior, motor learning, balance, coordination but increased aggression and anxiety-like behaviors [32]. These animals also show a significant disruption of the 5-HT system, without effects on other monoamine systems, with a 70% loss of 5-HT immunopositive cell bodies in adult Pet-1 KOs. While the remaining 30% of 5-HT cell bodies retain their neuronal phenotype and appear to be normally positioned, they show little to no protein or mRNA expression for some of the substances that are common to the 5-HT neuron phenotype: TPH, 5-HT transporter, and vesicular monoamine transporter 2. Further, Pet-1 KO mice have a reduced density of 5-HT immunopositive fibers in target fields and the brain tissue content of 5-HT and 5hydroxyindoleacetic acid in the Pet-1 KOs are only 10–15% of WT levels [32]. The effects of the loss of Pet-1 on peripheral serotonin levels are unknown. It is clear from these studies that alteration of 5-HT input to the SCN changes the expression of behavioral circadian rhythms. What is unknown is how developmental disruption of the entire system can affect the SCN and behavioral clock output. The Pet-1 KO mouse provides an opportunity to examine this question, so in this study we examine the expression of behavioral rhythmicity in these mice, with the goal of further clarifying the role of 5-HT in circadian rhythm regulation.
WT animals were identified by DNA fragments of 209 bp while Pet-1 KO animals were defined by DNA fragments of 361 bp. The number of animals of each sex and genotype were counted to assess whether breeding results corresponded to the expected Mendelian ratios. Differences from expected ratios were tested with a G-test of goodnessof-fit while differences between sexes were assessed by a G-test for independence. 2.3. Behavioral monitoring WT (n = 8), heterozygote (HET) (n= 8) and KO (n = 8) animals were placed into individual cages equipped with running wheels so that activity patterns could be monitored. Food and water was available ad libitum. Wheel-running activity was monitored for at least 21 days under each of three light schedules: LD 14:10, DD, and LL using Clocklab (Actimetrics) software. An additional group of animals was profiled in LD 12:12. Actograms for each animal were examined for differences in activity patterns between genotypes. Hourly counts of the number of wheel revolutions per hour were averaged across animals and normalized to the baseline 24-hour activity total for each animal to eliminate the effects of running wheels with different frictional resistances on the results. Periods in LD, DD, and LL were calculated by chi-squared periodogram for each animal using Clocklab software then grouped and averaged according to genotype. For the extended DD experiment, WT (n= 8) and KO animals (n= 6) were housed in LD 12:12 then transferred to and maintained in DD for 70 days. Periods were calculated for 10 day increments, starting at day 11 and ending at day 70, and averaged across genotypes. Statistical analysis for all behavioral experiments was performed using repeated measures ANOVAs, with posthoc comparisons to control groups performed using Dunnett's test. 2.4. Histology
2. Materials and methods 2.1. Animals Pet-1 KO animals on a mixed 129Sv and C57BL/6 background were obtained from Dr. Evan Deneris (Case Western Reserve University, Cleveland, OH) [32]. Pet-1 heterozygote breeding pairs were used to produce all experimental animals, which were 2–5 months of age at the time of experimentation. Heterozygotes were used as Pet-1 homozygotes are deficient in maternal behaviors [33]. Experiments were conducted by matching siblings of different genotypes as much as possible to avoid bias from the mixed genetic background. At the time of weaning, animals were ear-tagged and a tail-snip was collected for DNA extraction and genotyping. Experimental animals were entrained to a 14:10 or 12:12 light–dark (LD) cycle prior to the start of the experiment as noted below. Food and water were available ad libitum and all experimental procedures were approved by the Kent State University Institutional Animal Care and Use Committee. 2.2. Genotyping Tail snips from all animals were placed into an extraction buffer containing 20 mM Tris pH 8.0, 1 mM EDTA, 400 mM NaCl, and 0.5% SDS with 20 mg/mL Proteinase K added in order to digest the snip. Samples were maintained at 55 °C with slow rotation overnight. Supernatant from each sample was spun down in order to pellet DNA, rinsed with ethanol and resuspended in TE buffer. DNA of interest was amplified using PCR with primers from IDT (3′ PET: 5′ GCC TGA TGT TCA AGG AAG ACC TCG G 3′ 5′ PET: 3′ CGC ACT TGG GGG GTC ATT ATC AC 3′ 5′ LOX: 5′ CGG TGG ATG TGG AAT GTG TGC G 3′) and thermocycler conditions beginning with a 3 min hold at 94 °C followed by 42 cycles of 94 °C for 50 s, 62 °C for 30 s and 72 °C for 40 s. PCR product was analyzed using the Agilent 2100 Bioanalyzer system.
For the examination of SCN neuropeptide expression, animals were given a lethal overdose of sodium pentobarbital intraperitoneally (i.p.) and perfused transcardially with 4% ice-cold paraformaldehyde. Brains were removed and post-fixed overnight before being transferred into PBS. Anterior and posterior portions of brains were sectioned at 40 μm using either a vibratome (for arginine vasopressin (AVP) and 5-HT) or cryostat (for vasoactive intestinal polypeptide (VIP)); if a cryostat was used, brains were first cryoprotected in a 30% sucrose solution. Sections from each animal were then processed for immunohistochemistry for 5-HT (Immunostar rabbit anti-5-HT 1:2000 and Jackson Research Cy2 secondary 1:500) and images of the SCN, median raphe, and dorsal raphe were taken for analysis. 5-HT fluorescent staining was examined to confirm the level of 5-HT depletion in each mouse brain. Levels of AVP and VIP were also examined using immunohistochemistry (Chemicon rabbit anti-AVP 1:50,000 and Immunostar rabbit anti-VIP 1:10,000, Jackson Research biotinylated donkey anti-rabbit secondary 1:500 and processed for peroxidase reaction using a Vector Laboratories DAB kit) and images taken for comparison. ImageJ software was used to perform counts of AVP immunopositive cells in the SCN of WT (n= 5), HET (n= 5) and KO (n= 5) animals. Counts of AVP immunoreactive cells were also performed in the paraventricular nucleus (PVN) for each genotype. Statistical analysis of AVP cell counts was done using a one-way ANOVA. VIP immunopositive cells in the SCN could not be counted due to the density of the labeling which obscured individual cells, so only a qualitative inspection was performed. 3. Results 3.1. 5-HT immunohistochemistry in the SCN, MR, and DR Immunohistochemical staining was performed for 5-HT in order to visualize the extent of 5-HT fiber innervation to the SCN (Fig. 1A–C).
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263
259
Fig. 1. Serotonin (5-HT) immunoreactivity in the SCN (A–C), median raphe (MR, D–F), and dorsal raphe (DR, G–I) in wildtype (A, D, G), heterozygote (B, E, H) and Pet-1 knockout (C, F, I) animals. Images have been adjusted for contrast.
Also, since the MR and DR provide ascending 5-HT innervation to the brain, the density of immunopositive cell bodies in the MR (Fig. 1D–F) and DR (Fig. 1G–I) in WT, HET, and KO animals was examined. 5-HT fiber innervation to the SCN of WT and HET animals remained intact (Fig. 1A and B) while the SCN of KO animals was almost devoid of 5HT-positive fibers (Fig. 1C). As expected, the MR of WT and HET animals showed cell bodies immunopositive for 5-HT (Fig. 1D and E) while staining was almost absent in KO animals (Fig. 1F). Similarly, there were numerous 5-HT-labeled cell bodies in the DR of WT and HET animals (Fig. 1G and H) while the number of 5-HT cells in the DR of KO animals was greatly reduced (Fig. 1I). 3.2. Immunohistochemistry for VIP and AVP Since previous experiments have shown evidence that 5-HT depletion can alter the expression pattern of neuropeptides within the brain [34], the effects of a genetic depletion of 5-HT on VIP and AVP cell distribution were examined. VIP immunoreactivity in the SCN of WT, HET, and KO animals showed no obvious differences between genotypes (Fig. 2A–C), but the density of fiber staining prohibited quantification. AVP immunoreactivity was assayed in the SCN by counting the number of immunoreactive cell bodies (Fig. 2D–F). There were no significant differences in total number of AVP positive cells in the SCN between genotypes: 417.75 ± 32.7 cells in WT (Fig. 2D), 400.6 ±
13.89 cells in HET (Fig. 2E), and 382.25 ± 10.48 cells in KO (Fig. 2F) (F(2,10) = 0.70, p = 0.52). Counts of immunopositive AVP cells in the PVN were also not significantly different between genotypes (data not shown). 3.3. Circadian parameters Representative actograms show the differences in wheel-running patterns between genotypes in LD, DD, and LL (Fig. 3A and B). WT (Fig. 3A) animals displayed an average period of 23.59 ± 0.06 h in DD, and 24.90 ± 0.26 h in LL. HET animals displayed an average period of 23.63 ± 0.07 h in DD, and 25.68 ± 0.18 h in LL. KO (Fig. 3B) animals displayed an average period of 24.01 ± 0.04 h in DD, and 24.73 ± 0.23 h in LL. KO animals had significantly longer periods in DD when compared to both WT and HET animals (F(2,20) = 19.57, p = 0.00002). No significant differences were seen between WT and HET animals in DD or LL, however HET animals showed significantly longer circadian periods than KO in LL (F(2,15) = 3.70, p = 0.0495). No differences were seen between WT and KO animals in LL. Results are summarized in Fig. 4. We attempted to measure mean activity onset, activity offset, phase angle of entrainment (in LD), and alpha (duration of daily activity period) and compare across genotypes based on an analysis of each animal's actogram. However, these measures were highly
Fig. 2. Vasoactive intestinal polypeptide (VIP, A–C) and arginine vasopressin (AVP, D–F) immunoreactivity in the SCN of wildtype (A and D), heterozygote (B and E) and Pet-1 knockout (C and F) animals. Images have been adjusted for brightness.
260
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263
Fig. 3. (A and B) Representative double-plotted actograms of circadian wheel-running activity for wildtype (A) and Pet-1 knockout (B) animals under three photo schedules: 14:10 light:dark (LD), constant darkness (DD) and constant light (LL). Actograms for heterozygote animals resemble those for wildtype and are not shown.
variable, both between and within animals, and did not yield any interpretable data. Similarly, attempts to measure alpha in DD and LL resulted in very large standard deviations within groups.
bouts distinctly associated with lights on and lights off is seen in virtually all KO animals but not in WT animals (Fig. 3).
3.5. Extended DD 3.4. Activity profile In order to determine whether there were differences in partitioning of activity between genotypes, total wheel-running per hour in LD as a percentage of total daily activity was examined (Fig. 5A and B). In 14:10 LD (Fig. 5A), a significant interaction between time and genotype was observed (F(46,391) = 3.86, p b 0.000001), in which WT (n= 6) and HET (n= 6) animals showed a significantly greater percentage of total wheel-running at ZT 15 and ZT 16 than KO (n= 8) animals while KO animals showed a significantly greater percentage of wheel-running behavior in the late portion of the dark phase at ZT 24 (pb 0.05 in each case, Dunnett's test). Although our initial experiments were run with the animals in 14:10 LD, we wanted to assess their profile in 12:12 LD as well to determine whether the late night increases in activity were more pronounced with a longer dark phase. 12:12 LD also represents a more normal housing light cycle for laboratory mice; the original 14:10 LD cycle had been used solely due to compatibility with other experiments going on in the same room. In 12:12 LD (Fig. 5B) a significant interaction between time and genotype was observed (F(23,483) = 7.65, p b 0.000001) in which WT (n= 11) animals showed a significantly greater percentage of total wheel-running at ZT 13, ZT 14, ZT 15, and ZT 16 while KO animals (n= 12) showed a significantly greater percentage of wheel-running behavior in the late portion of the dark phase, at ZT 24 (pb 0.05 in each case, Dunnett's test). Unimodal patterns of activity were seen in WT and HET animals while activity in KO animals appeared bimodal; the pattern of a separation of activity into
Fig. 4. Circadian period (mean ± SEM) in hours of wildtype (WT), heterozygote (HET) and Pet-1 knockout (KO) animals in constant darkness (DD) and constant light (LL). Asterisks denote significance (p b 0.05).
Representative actograms show differences between genotypes in wheel-running behavior over 10 weeks of DD (Fig. 6A and B). WT (Fig. 6A) animals exhibit a shortening of period over time in extended DD while KO (Fig. 6B) animals retain significantly longer periods over time. There were significant effects of both genotype (F(1,12) = 12.12, p = 0.005), length of time in DD (F(5,60) = 6.48, p = 0.00007), and the interaction term (F(5,60) = 5.00, p = 0.0007). In WT animals, circadian periods for days 41–50 (23.14 ± 0.20 h), 51–60 (23.22 ± 0.20 h), and 61–70 (23.04 ± 0.19 h) are significantly shorter than at days 11–20
Fig. 5. Activity profile in 14:10 light:dark (LD) for wildtype, heterozygote, and Pet-1 knockout animals (A) and activity profile in 12:12 LD for wildtype and Pet-1 knockout animals (B). Data is represented as mean percentage of total activity in each 1 h bin throughout the 24 h light–dark cycle, and error bars show the standard error of the mean. Vertical bars indicate lights off. Asterisks denote significance (p b 0.05).
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263
261
Fig. 6. (A and B) Representative double-plotted actograms of circadian rhythm activity in 10 weeks of constant darkness (DD) for wildtype (A) and Pet-1 knockout (B) animals.
(23.61 ± 0.07 h) (p b 0.05, Dunnett's test). Periods in KO animals did not differ significantly between time points. Results are summarized in Fig. 7. 3.6. Assessment of breeding ratios To date, we have genotyped 1805 mice. The totals by genotype were 477 WT, 965 HET, and 363 KO animals. This distribution of values differs significantly from the expected 1:2:1 ratio (χ2(2) = 23.06, p b 0.000001). Only about 76% of the expected number of KO animals survived to be weaned and genotyped. There was no significant difference in the ratio of genotypes between males and females (χ2(2) = 2.39, p = 0.30). 4. Discussion In this study, we examined the effects of a partial, developmental disruption of the 5-HT system on 5-HT input to the SCN, circadian output behavior, and neuropeptide distribution within the SCN. Immunohistochemical analysis of 5-HT immunoreactivity shows numerous 5-HT cell bodies in the MR and DR of WT and HET animals as well as extensive 5-HT fiber innervation to the SCN. Pet-1 KO animals show virtually no 5HT labeling in the SCN and greatly reduced numbers of 5-HT cell bodies in the MR and DR. Our results are similar to previously published reports indicating reduced 5-HT cell bodies in the MR and DR and an absence of 5-HT fibers in target fields such as the cortex and hippocampus [32]. Since previous studies have demonstrated an important role for 5-HT in the regulation of behavioral circadian rhythms, we expected to find changes in the circadian wheel-running activity rhythm.
Fig. 7. Circadian period (mean ± SEM) in hours for wildtype and Pet-1 knockout animals in 10 weeks of DD. Mean periods are grouped in ten day bins starting at day 11 and ending at day 70. KO animals have significantly longer periods than WT at all time points. Asterisks denote significant differences between circadian periods (compared to period from days 11–20 bin) within the WT group only (pb 0.05).
Our study found that Pet-1 KO and HET mice could entrain to 14:10 and 12:12 LD schedules. In DD, KO animals had significantly longer periods than WT or HET animals. These findings differ with published reports in Syrian hamsters which indicate that 5-HT depletion, due to 5,7DHT administration, does not change the circadian period in DD but rather, lengthens the period in LL [26,27]. These differences could be due to the loss of 5-HT during development in Pet-1 KO mice compared to destruction of the 5-HT system in adults, or could reflect species differences. Our results are more similar to previously published reports showing that mice blinded by enucleation and treated with 5,7-DHT have longer circadian periods, a longer active phase, delayed onset of peak running activity within the active phase, and increased occurrences of bimodal patterns of activity compared to controls [28]. A lengthening of period in DD as a result of serotonin deficiency could be explained if the serotonergic input to the circadian clock in the SCN has an average effect of phase advancing the clock. Evidence for the phase advancing properties of 5-HT exists in numerous studies showing that the phase response curve to 5-HTergic agonists shows a large phase advance region during the middle of the subjective day [14–20]. At this time of day extracellular serotonin concentrations are normally low [10], but it could be that continuous release of serotonin at a low level during the day results in a daily phase advance, and absence of that stimulation would slow the clock, leading to longer free-running periods. This effect is similar to that seen in 5,7-DHT lesioned mice [28]. However, 5-HT1A [35], 5-HT1B [36], and 5-HT7 [37] receptor knockout mice show no differences in free-running period compared to WT counterparts suggesting that the lengthening of period in Pet-1 KOs is not mediated by any one of these receptor subtypes. The changes in period are therefore likely the result of the loss of 5-HT action on multiple receptors simultaneously. It is also possible that an inducible receptor knockout would show a phenotype if the receptor were eliminated in an adult, but that the loss of that receptor during development leads to compensation by other receptor subtypes. The absence of the shortening of free-running period during extended DD in Pet-1 KO mice may indicate that 5-HT plays a role in long-term changes associated with light deprivation, but our results do not suggest any specific mechanism for this phenomenon. Unfortunately, we do not know the status of 5-HT receptor levels in Pet-1 KO mice. The 5,7-DHT lesion studies suggest, though, that the change in free-running period is not a result of the loss of serotonin during early development, but is a more acute effect derived either from the loss of serotonin itself, or the loss of serotonergic neurons that may exert effects through other signaling molecules in addition to serotonin. Furthermore, the failure of any 5-HT receptor knockouts to mimic the behavioral phenotype seen in the Pet-1 KO mice may indicate that developmental compensation may occur in the case of the knockouts but not in the case of the loss of the transmitter. It is also possible that
262
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263
the phenotype may be a function of the fact that while 5-HT levels are severely reduced in knockouts, some residual 5-HT is still present; therefore, the phenotype may be more representative of a severe reduction in 5-HT rather than its total elimination. An additional hypothesis might be that the changes in period may be mediated by changes in other neurotransmitter systems. Serotonin antagonists increase glutamate release in the SCN [38]; other neurotransmitters may be affected as well. Increased glutamate in the early night in Pet-1 KO mice, a time when serotonin release is high in the WT mice, might result in daily light-like phase delays, thus increasing free-running period. The loss of 5-HT could also alter release of intrinsic neuropeptides such as VIP, GRP, or AVP as well as neuropeptides innervating the SCN such as NPY [39–41]. A second effect of a slower clock might be a shift in the profile of daily activity. In both 14:10 and 12:12 LD cycles, Pet-1 KO animals shift a substantial proportion of their daily wheel-running activity from the early part of the night to the late night. Some KO animals showed wheel-running that extended into the period of lights-on, which may suggest a reduced suppressive effect of light on activity. However, we did not see evidence of a change in lights' effect on the circadian clock in the form of an effect on free-running period in LL in KO animals. HET animals had periods in LL that were not significantly different from WT, though they were significantly longer than KO. A difference between HET and KO is difficult to explain as the HET animals otherwise show a phenotype identical to WT. In the course of observing the circadian behaviors of the mice, we noticed that the free-running period of WT mice appeared to decrease over time while in DD. In order to examine this further, we performed an experiment to examine changes in period over time in extended DD. KO animals maintained significantly longer periods than WT animals and did not show the shortening of circadian period over time seen in WT animals, indicating that the longer periods observed in Pet-1 KOs are a stable effect of 5-HT depletion. It also suggests that the decrease of free-running period in DD in WT animals may be through a mechanism that involves serotonergic feedback on the circadian clock. It was possible that the changes in circadian clock function seen in KO mice were the result of structural changes in the SCN. To examine this possibility, we examined the patterns of VIP and AVP immunoreactivity in the SCN. No obvious differences between genotypes were observed based on hemotoxylin staining of the SCN. In mice, VIP coordinates daily rhythms in the SCN by synchronizing a subset of neurons [42] and AVP expression patterns correspond with the subset of SCN cells that are intrinsically rhythmic [43]. No differences were seen in the distribution of VIP immunoreactivity between genotypes. Likewise, AVP immunoreactivity, which in the mouse is concentrated in the dorsomedial region of the SCN as well as in clusters around the edges of the SCN, did not qualitatively differ between genotypes. In addition, we counted the number of AVP-positive cell bodies and no differences were detected. Previous studies suggest that depletion of 5-HT by administration of p-chlorophenylalanine (pCPA) to pregnant rats causes lasting changes in neuropeptide content within the SCN in progeny, specifically by increasing numbers VIP and AVP immunopositive neurons and the concentration of these peptides within cell bodies [34]. We did not see a comparable effect in Pet-1 KO animals, supporting the idea that lengthening of period in the KO animals is not due to disruption of SCN structure or neurochemical organization. In the course of breeding the mice for our studies, we noted that there were approximately 25% fewer KO animals at weaning age than would be expected from Mendelian genetic ratios, suggesting an increased mortality rate associated with Pet-1 KOs. Hendricks et al. [32] mentioned a non-Mendelian ratio of genotypes due to a loss of Pet-1 KO mice within the first week after birth. Since the stomachs of non-surviving neonates were full of milk, this loss is not thought to be due to a decreased ability to feed. It likely has to do with an impairment of the respiratory system in neonates [44,45]. Recently, it's been reported that Lmx1b flox/flox;ePet-Cre/+ mice, which lack 5-HT neurons,
display apnea which is most severe during the first two weeks of life. Apnea was shown to improve by 2–4 weeks after birth but these animals showed a high rate of perinatal mortality and decreased growth rate until respiratory deficits resolved in surviving mice [46]. This report suggests that Pet-1 KO mice, which also lack 5-HT neurons, may display similar respiratory difficulties leading to the increased mortality seen in KO animals. However, we did not see any deficiencies in overall locomotor activity in Pet-1 KO adults, so we do not believe that this developmental affect is having a material effect on our results regarding the characteristics of circadian clock function. Overall, these data indicate that a dysfunctional 5-HT system causes changes not just in circadian clock period, but also in the distribution of locomotor activity across the circadian cycle. In serotonin-deficient mice, peak levels of running-wheel activity were redistributed from a primarily unimodal pattern into a bimodal matter, with dramatically increased activity levels in the hour immediately prior to lights on. This change may be secondary to the lengthening of circadian period but may also represent an organizational effect of 5-HT on activity patterns. It is also possible that the change in the distribution of activity itself plays a role in the change in circadian period. Therefore, these data suggest that 5-HT plays a significant role in regulating the timing of locomotor activity, and that dysfunction of the 5-HT system may result in changes in behavioral activity patterns.
Acknowledgments We would like to thank Evan Deneris for providing the Pet-1 knockout mice, RoxAnne Murphy for her assistance with the genotyping protocol, and Heather Caldwell for her helpful comments on the manuscript.
References [1] Rusak B, Zucker I. Neural regulation of circadian rhythms. Physiol Rev 1979;59: 449–526. [2] Moore RY. Organization and function of a central nervous system circadian oscillator: the suprachiasmatic hypothalamic nucleus. Fed Proc 1983;42:2783–9. [3] Klein DC, Moore RY, Reppert SM. Suprachiasmatic nucleus: the mind's clock. USA: Oxford University Press; 1991. [4] Ruby NF, Burns DE, Heller HC. Circadian rhythms in the suprachiasmatic nucleus are temperature-compensated and phase-shifted by heat pulses in vitro. J Neurosci 1999;19:8630–6. [5] Abe H, Rusak B. Anticipatory activity and entrainment of circadian rhythms in Syrian hamsters exposed to restricted palatable diets. Am J Physiol Regul Integr Comp Physiol 1992;263:116–24. [6] Reebs SG, Mrosovsky N. Effects of induced wheel running on the circadian activity rhythms of Syrian hamsters: entrainment and phase response curve. J Biol Rhythms 1989;4:39–48. [7] Morin LP. Serotonin and the regulation of mammalian circadian rhythmicity. Ann Med 1999;31:12–33. [8] Meyer-Bernstein EL, Morin LP. Differential serotonergic innervation of the suprachiasmatic nucleus and the intergeniculate leaflet and its role in circadian rhythm modulation. J Neurosci 1996;16:2097–111. [9] Yamakawa GR, Antle MC. Phenotype and function of raphe projections to the suprachiasmatic nucleus. Eur J Neurosci 2010;31:1974–83. [10] Dudley TE, Dinardo LA, Glass JD. Endogenous regulation of serotonin release in the hamster suprachiasmatic nucleus. J Neurosci 1998;18:5045–52. [11] Ralph MR, Mrosovsky N. Behavioral inhibition of circadian responses to light. J Biol Rhythms 1992;7:353–9. [12] Rea MA, Glass JD, Colwell CS. Serotonin modulates photic responses in the hamster suprachiasmatic nuclei. J Neurosci 1994;14:3635–42. [13] Pickard GE, Rea MA. TFMPP, a 5HT1B receptor agonist, inhibits light-induced phase shifts of the circadian activity rhythm and c-Fos expression in the mouse suprachiasmatic nucleus. Neurosci Lett 1997;231:95–8. [14] Tominaga K, Shibata S, Ueki S, Watanabe S. Effects of 5-HT1A receptor agonists on the circadian rhythm of wheel-running activity in hamsters. Eur J Pharmacol 1992;214:79–84. [15] Edgar DM, Miller JD, Prosser RA, Dean RR, Dement WC. Serotonin and the mammalian circadian system: II. Phase-shifting rat behavioral rhythms with serotonergic agonists. J Biol Rhythms 1993;8:17–31. [16] Cutrera RA, Ouarour A, Pevet P. Effects of the 5-HT1a receptor agonist 8-OH-DPAT and other non-photic stimuli on the circadian rhythm of wheel-running activity in hamsters under different constant conditions. Neurosci Lett 1994;172:27–30.
E.V. Paulus, E.M. Mintz / Physiology & Behavior 105 (2012) 257–263 [17] Cutrera RA, Saboureau M, Pevet P. Phase-shifting effect of 8-OH-DPAT, a 5HT1A/5-HT7 receptor agonist, on locomotor activity in golden hamster in constant darkness. Neurosci Lett 1996;210:1–4. [18] Mintz EM, Gillespie CF, Marvel CL, Huhman KL, Albers HE. Serotonergic regulation of circadian rhythms in Syrian hamsters. Neuroscience 1997;79:563–9. [19] Horikawa K, Yokota S, Fuji K, Akiyama M, Moriya T, Okamura H, et al. Nonphotic entrainment by 5-HT1A/7 receptor agonists accompanied by reduced Per1 and Per2 mRNA levels in the suprachiasmatic nuclei. J Neurosci 2000;20:5867–73. [20] Horikawa K, Shibata S. Phase-resetting response to (±) 8-OH-DPAT, a serotonin 1A/7 receptor agonist, in the mouse in vivo. Neurosci Lett 2004;368:130–4. [21] Rea MA, Barrera J, Glass JD, Gannon RL. Serotonergic potentiation of photic phase shifts of the circadian activity rhythm. Neuroreport 1995;6:1417–20. [22] Smart CM, Biello SM. WAY-100635, a specific 5-HT1A antagonist, can increase the responsiveness of the mammalian circadian pacemaker to photic stimuli. Neurosci Lett 2001;305:33–6. [23] Gannon RL. Serotonergic serotonin1A mixed agonists/antagonists elicit largemagnitude phase shifts in hamster circadian wheel-running rhythms. Neuroscience 2003;119:567–76. [24] Lall GS, Harrington ME. Potentiation of the resetting effects of light on circadian rhythms of hamsters using serotonin and neuropeptide Y receptor antagonists. Neuroscience 2006;141:1545–52. [25] Kessler EJ, Sprouse J, Harrington ME. NAN-190 potentiates the circadian response to light and speeds re-entrainment to advanced light cycles. Neuroscience 2008;154:1187–94. [26] Smale L, Michels KM, Moore RY, Morin LP. Destruction of the hamster serotonergic system by 5, 7-DHT: effects on circadian rhythm phase, entrainment and response to triazolam. Brain Res 1990;515:9–19. [27] Morin LP, Blanchard J. Depletion of brain serotonin by 5, 7-DHT modifies hamster circadian rhythm response to light. Brain Res 1991;566:173–85. [28] Mistlberger RE, Bossert JM, Holmes MM, Marchant EG. Serotonin and feedback effects of behavioral activity on circadian rhythms in mice. Behav Brain Res 1998;96:93–9. [29] Hendricks T, Francis N, Fyodorov D, Deneris ES. The ETS domain factor Pet-1 is an early and precise marker of central serotonin neurons and interacts with a conserved element in serotonergic genes. J Neurosci 1999;19:10348–56. [30] Pfaar H, von Holst A, Vogt Weisenhorn DM, Brodski C, Guimera J, Wurst W. mPet1, a mouse ETS-domain transcription factor, is expressed in central serotonergic neurons. Dev Genes Evol 2002;212:43–6. [31] Liu C, Maejima T, Wyler SC, Casadesus G, Herlitze S, Deneris ES. Pet-1 is required across different stages of life to regulate serotonergic function. Nat Neurosci 2010;13:1190–8.
263
[32] Hendricks TJ, Fyodorov DV, Wegman LJ, Lelutiu NB, Pehek EA, Yamamoto B, et al. Pet-1 ETS gene plays a critical role in 5-HT neuron development and is required for normal anxiety-like and aggressive behavior. Neuron 2003;37:233–47. [33] Lerch-Haner JK, Frierson D, Crawford LK, Beck SG, Deneris ES. Serotonergic transcriptional programming determines maternal behavior and offspring survival. Nat Neurosci 2008;11:1001–3. [34] Mirochnik V, Bosler O, Tillet Y, Calas A, Ugrumov M. Long-lasting effects of serotonin deficiency on differentiating peptidergic neurons in the rat suprachiasmatic nucleus. Int J Dev Neurosci 2005;23:85–91. [35] Smith VM, Sterniczuk R, Phillips CI, Antle MC. Altered photic and non-photic phase shifts in serotonin1A receptor knockout mice. Neuroscience 2008;157: 513–23. [36] Sollars PJ, Ogilvie MD, Rea MA, Pickard GE. 5-HT1B receptor knockout mice exhibit an enhanced response to constant light. J Biol Rhythms 2002;17:428–37. [37] Gardani M, Biello SM. The effects of photic and nonphotic stimuli in the 5-HT7 receptor knockout mouse. Neuroscience 2008;152:245–53. [38] Srkalovic G, Selim M, Rea MA, Glass JD. Serotonergic inhibition of extracellular glutamate in the suprachiasmatic nuclear region assessed using in vivo brain microdialysis. Brain Res 1994;656:302–8. [39] Glass JD, Guinn J, Kaur G, Francl JM. On the intrinsic regulation of neuropeptide Y release in the mammalian suprachiasmatic nucleus circadian clock. Eur J Neurosci 2010;31:1117–26. [40] Francl JM, Kaur G, Glass JD. Regulation of vasoactive intestinal polypeptide release in the suprachiasmatic nucleus circadian clock. Neuroreport 2010;21:1055–9. [41] Francl JM, Kaur G, Glass JD. Roles of light and serotonin in the regulation of gastrin-releasing peptide and arginine vasopressin output in the hamster SCN circadian clock. Eur J Neurosci 2010;32:1170–9. [42] Aton SJ, Colwell CS, Harmar AJ, Waschek J, Herzog ED. Vasoactive intestinal polypeptide mediates circadian rhythmicity and synchrony in mammalian clock neurons. Nat Neurosci 2005;8:476–83. [43] Antle MC, Silver R. Orchestrating time: arrangements of the brain circadian clock. Trends Neurosci 2005;28:145–51. [44] Cummings KJ, Li A, Deneris ES, Nattie EE. Bradycardia in serotonin-deficient Pet1−/− mice: influence of respiratory dysfunction and hyperthermia over the first 2 postnatal weeks. Am J Physiol Regul Integr Comp Physiol 2010;298:R1333–42. [45] Erickson JT, Shafer G, Rossetti MD, Wilson CG, Deneris ES. Arrest of 5HT neuron differentiation delays respiratory maturation and impairs neonatal homeostatic responses to environmental challenges. Respir Physiol Neurobiol 2007;159:85–101. [46] Hodges MR, Wehner M, Aungst J, Smith JC, Richerson GB. Transgenic mice lacking serotonin neurons have severe apnea and high mortality during development. J Neurosci 2009;29:10341–9.