Abstract
A role for 5-hydroxytryptamine (5-HT) or serotonin in sleep has been known for decades but was challenged by recent papers that concluded that the apparent sleep phenotype was secondary to defective thermoregulation. Those studies used mice lacking serotonergic neurons resulting from the loss of function mutations in the gene encoding the LIM homeobox transcription factor 1 (Lmx1b). Here we show that, while Lmx1b mutants failed to keep the physiologic body temperature, they exhibited more activities at the room and elevated temperatures. More importantly, we used mice deficient in the gene encoding tryptophan hydroxylase 2 (Tph2), which could not synthesize 5-HT in the brain. Tph2 mutants were capable of thermoregulation and keeping physiologic body temperature when the environmental temperature was reduced and exhibited significantly more activities at both the room and elevated temperatures. Electroencephalographic (EEG) recording also showed decreased sleep in Tph2-deficient mice. Our results indicate that 5-HT is important for sleep regulation but not thermoregulation.
Introduction
The monoamine neurotransmitter 5-hydroxytryptamine (5-HT), also known as serotonin, is important in multiple physiologic processes. It is surprising that one of the earliest suggested roles for 5-HT, i.e., that in sleep regulation, remains controversial after decades of physiologic and pharmacological studies.
Early studies used p-chlorophenylalanine (pCPA), an inhibitor of tryptophan hydroxylase to deplete 5-HT from cats, rats, monkeys, and humans (reviewed in Jouvet, 1969, 1972). Reduction of central 5-HT levels after pCPA administration in cats (Koella et al., 1968) and rats (Mouret et al., 1968) were found to correlate with reduction of sleep. Damage of serotonergic neurons in the raphe nuclei of cats resulted in a decrease of sleep in a manner proportional to the size of the lesion (Jouvet et al., 1967; Jouvet, 1968). 5-Hydroxytryptophan (5-HTP) is a precursor of 5-HT capable of crossing the blood-brain barrier and can bypass the effect of pCPA to restore the cerebral 5-HT level. Injection of 5-HTP proportionally increased sleep and 5-HT levels in cats whose sleep was disrupted by pCPA (Mouret et al., 1967; Jouvet, 1968; Koella et al., 1968; Pujol et al., 1971). These studies led to the hypothesis that central 5-HT promotes sleep.
However, the discovery that administration of either 5-HT or 5-HTP in animals with normal 5-HT level sometimes did not induce physiologic sleep put the hypothesis in question (Jouvet, 1972). Large doses of 5-HTP increased electroencephalography (EEG) synchronization, a hallmark of neural activity during non-rapid eye movement (NREM) sleep, while suppressing rapid eye movement (REM) sleep in cats (Pujol et al., 1971). In monkeys, 5-HTP caused hyperactivity followed by an inactive period alternating between dozing and staring blankly (Macchitelli et al., 1966). Gain-of-function experiments suggested that 5-HT was not simply promoting sleep or triggering sleep onset.
To further complicate the scenario, unit activities in raphe nuclei increased during waking, progressively decreased when animals transitioned from drowsiness to NREM sleep, and was almost silent during REM sleep (McGinty and Harper, 1976; Trulson and Jacobs, 1979; Lydic et al., 1987). Microdialysis measurement of extracellular 5-HT concentration from the dorsal raphe nucleus in cats showed that the 5-HT level was the highest during waking and progressively decreased from NREM to REM sleep (Portas and McCarley, 1994). However, some atypical raphe neurons showed higher activity during REM or NREM sleep or both in freely moving cats (Sakai and Crochet, 2001).
Genetic depletion of serotonergic neurons was achieved by ePet1-Cre driven knockout of Lmx1b gene (Lmx1bf/f/p) (Zhao et al., 2006). Lmx1bf/f/p mice exhibited increased wakefulness and decreased sleep (Buchanan and Richerson, 2010), but these results were attributed to the failure in maintaining body temperature, which caused the animals to exert hyperactivity to stay warm (Hodges et al., 2008; Buchanan and Richerson, 2010). This conclusion was not consistent with previous findings that suggested serotonergic neurons involved in thermoregulation were confined in the raphe magnus and the medullar nuclei (Berner et al., 1999; Tanaka et al., 2002; Nakamura et al., 2004), whereas the level of insomnia caused by lesion was proportional to the size of destruction, which suggested that 5-HT concentration, rather than specific nuclei, was important for sleep-wake regulation (Jouvet, 1969). A large proportion of raphe serotonergic neurons also release glutamate (Hioki et al., 2010; Liu et al., 2014; Sos et al., 2016), raising the possibility that distinct neurotransmitters are involved in divergent functions of raphe neurons.
We used Tph2 mutant mice, in which the raphe neurons are present but 5-HT could not be synthesized in the brain. We found that the Tph2−/− mice were not defective in thermoregulation at cold or normal ambient temperature. Sleep was decreased in these mutant mice, as shown by locomotive activities and by electroencephalography/electromyography (EEG/EMG) recordings.
Materials and Methods
Experimental Animals.
All procedures and protocols were carried out under the instruction of the Institutional Animal Care and Use Committee at Peking University. Briefly, mice were kept under a 12:12-hour dark/light cycle at the ambient temperature of 25 ± 1°C. Experiments were carried out between 12 and 18 weeks of age. ePet1-Cre, Lmx1bf/f, and Tph2 mice were gifts from Dr. Zhoufeng Chen. Generation, breeding, and genotyping of Lmx1bf/f/p (Zhao et al., 2006) and Tph2−/−(Kim et al., 2014) mice have been described previously (Liu et al., 2011). Primers used for genotyping were AGGCTCCATCCATTCTTCTC and CCACAATAAGCAAGAGGCAC for Lmx1bf/f; ATTTGCCTGCATTACCGGTCG and CAGCATTGCTGTCACTTGGTC for ePet1-Cre; and GGGCATCTCAGGACGTAGTAG, GGGCCTGCCGATAGTAACAC, and GCAGCCAGTAGACGTCTCTTAC for Tph2. All strains were backcrossed with C57B/L for over eight generations in our laboratory before these experiments. For Tph2 mice, only males were used, whereas for Lmx1bf/f/p mice, both males and females were used. Littermates with genotype Lmx1bf/f or Tph2+/+ were used as controls for Lmx1bf/f/p and Tph2−/−, respectively. All mice were housed individually during experiments. Numbers of male and female mice: Fig. 1 (n = 7 for Lmx1bf/f, 7 females, 0 males; n = 4 for Lmx1bf/f/p, 2 females, 2 males; n = 6 for Tph2+/+, 5 females, 1 male; n = 7 for Tph2−/−, 7 females, 0 males), Fig. 2 (n = 7 for Lmx1bf/f, 7 females, 0 males; n = 5 for Lmx1bf/f/p, 3 females, 2 males; n = 7 for Tph2+/+, 6 females, 1 male; n = 9 for Tph2−/−, 8 females, 1 male), Fig. 3 (room temperature: n = 16 for Lmx1bf/f, 12 females, 4 males; n = 12 for Lmx1bf/f/p, 6 females, 6 males; n = 36 for Tph2+/+, 29 females, 7 males; n = 33 for Tph2−/−, 27 females, 6 males. 33°C: n = 15 for Lmx1bf/f, 9 females, 6 males; n = 13 for Lmx1bf/f/p, 7 females, 6 males; n = 30 for Tph2+/+, 21 females, 9 males; n = 29 for Tph2−/−, 20 females, 9 males), Fig. 4 (all mice for EEG were male).
Both Lmx1bf/f/p and Tph2−/− mice exhibited elevated locomotion but maintained body temperature. (A) Mean body temperature (Tcore; top) and activity [arbitrary units (A.U.); bottom] of Lmx1bf/f/p (red) vs. Lmx1bf/f (blue) mice over 24 hours at room temperature (RT). (B) Mean Tcore and activity by phase (light and dark) of Lmx1b mice at RT (P = 0.0061 for both dark and light, Mann-Whitney tests; n = 7 for Lmx1bf/f, n = 4 for Lmx1bf/f/p; data are mean ± S.E.M.). **P = 0.0061. (C) Tcore (top) and activity (arbitrary units bottom) of Tph2−/− (red) vs. Tph2+/+ (blue) mice over 24 hours at RT. (D) Mean Tcore and activity by phase (light and dark) of Tph2 mice at RT (P = 0.0082, Mann-Whitney test for the light phase; n = 6 for Tph2+/+; n = 7 for Tph2−/−; Data are mean ± S.E.M.). **P = 0.0082.
Body temperature was maintained by Tph2−/− mice but not by Lmx1bf/f/p mice when the ambient temperature was 4°C. (A) Tcore of Lmx1bf/f/p (red) vs. Lmx1bf/f (blue) mice recorded for 5 hours at 4°C (mean ± S.E.M., S.E.M. indicated in gray). (B) Activity of Lmx1bf/f/p (red) vs. Lmx1bf/f (blue) mice recorded for 5 hours at 4°C (mean ± S.E.M., S.E.M. indicated in gray). (C) Mean Tcore and activity of Lmx1b mice at 4°C (for mean Tcore, P < 0.01, Mann-Whitney test; n = 7 for Lmx1bf/f, n = 5 for Lmx1bf/f/p; data are mean ± S.E.M.). **P < 0.01. (D) Tcore of Tph2−/− (red) vs. Tph2+/+ (blue) mice recorded for 5 hours at 4°C (mean ± S.E.M., S.E.M. indicated in gray). (E) Activity of Tph2−/− (red) vs. Tph2+/+ (blue) mice recorded for 5 hours at 4°C (mean ± S.E.M., S.E.M. indicated in gray). (F) Mean Tcore and activity of Lmx1b mice at 4°C (n = 7 for Tph2+/+, n = 9 for Tph2−/−; data are mean ± S.E.M.).
Both Lmx1bf/f/p and Tph2 mice exhibited elevated activity at room temperature (RT) and 33°C. (A and B) Mean locomotion distance of Lmx1bf/f/p (red) vs. Lmx1bf/f (blue) mice at RT (P < 0.0001 for both dark and light phase in (B), Mann-Whitney tests; data are mean ± S.E.M.). (C and D) Mean active time of Lmx1bf/f/p (red) vs. Lmx1bf/f (blue) mice at RT (P < 0.0001 for dark phase and P = 0.0112 for light phase in (D), Mann-Whitney tests; data are mean ± S.E.M.; n = 16 for Lmx1bf/f, n = 12 for Lmx1bf/f/p). (E and F) Mean locomotion distance of Tph2−/− (red) vs. Tph2+/+ (blue) mice at RT (P = 0.0002 for dark and P = 0.0006 for light phase in (F), Mann-Whitney tests; data are mean ± S.E.M.). (G and H) Mean active time of Tph2−/− (red) vs. Tph2+/+ (blue) mice at RT (P = 0.0005 for dark phase and P < 0.0001 for light phase in (H), Mann-Whitney tests; data are mean ± S.E.M.; n = 36 for Tph2+/+, n = 33 for Tph2−/−). (I and J) Mean locomotion distance of Lmx1bf/f/p (red) vs. Lmx1bf/f (blue) mice at 33°C (P < 0.01 for both dark and light phase in (J), Mann-Whitney tests; data are mean ± S.E.M.). (K and L) Mean active time of Lmx1bf/f/p (red) vs. Lmx1bf/f (blue) mice at 33°C [P < 0.01 for both dark and light phase in (L), Mann-Whitney tests; data are mean ± S.E.M.; n = 15 for Lmx1bf/f, n = 13 for Lmx1bf/f/p]. (M and N) Mean locomotion distance of Tph2−/− (red) vs. Tph2+/+ (blue) mice at 33°C [P < 0.01 for the light phase in (N), Mann-Whitney tests; data are mean ± S.E.M.]. (O and P) Mean active time of Tph2−/− (red) vs. Tph2+/+ (blue) mice at 33°C [P < 0.01 for the light phase in (P), Mann-Whitney tests; data are mean ± S.E.M.; n = 30 for Tph2+/+, n = 29 for Tph2−/−]. For (A, C, E, G, I, K, M, and O), *P < 0.05; **P < 0.01; ***in F, P = 0.0002 for first column, P = 0.0006 for second column, in H, P = 0.0005; ****in D, P < 0.0001; ‡P < 0.01; †P < 0.001; #P < 0.0001; Mann-Whitney tests.
Analysis of EEG/EMG patterns reveals that Tph2−/− mice exhibited less sleep at both RT and 33°C. (A and B) Examples of time spent on wake, NREM, and REM by Tph2+/+ (A) and Tph2−/− (B) over 24 hours at RT. (C–E) Mean percentage of wake (C), NREM (D), and REM (E) by phase (light and dark) at RT [P < 0.01 for light phase in (E), Mann-Whitney test; data are mean ± S.E.M.; n = 20 for Tph2+/+, n = 22 for Tph2−/−]. (F and G) Examples of time spent on wake, NREM, and REM by Tph2+/+ (F) and Tph2−/− (G) over 24 hours at 33°C. (H–J) Mean percentage of wake (H), NREM (I), and REM (J) by phase (light and dark) at 33°C [P < 0.01 for dark phase in (H), P < 0.01 for dark phase in (I), P < 0.01 for both dark and light phase in (J); Mann-Whitney test; data are mean ± S.E.M.; n = 20 for Tph2+/+, n = 21 for Tph2−/−]. **P < 0.01.
Core Body Temperature and Gross Motor Activity Measurement.
Core body temperature and gross motor activity were measured by G2 E-Mitter implantable telemetry, acquired by ER4000 energizer/receiver, and analyzed by VitalView software (Starr Life Science, Oakmont, PA). Mice were anesthetized with tribromoethanol (250 mg/kg, i.p.), and electrodes were implanted intraperitoneally at the abdomen. Mice were individually housed and recovered for 1 week before recording. Data were recorded every 6 minutes during experiments. Mice were provided with free access to food and water during recording. For experiments at 4°C, the Lmx1b mice were recorded in the cold room for at most 5 hours (and Tph2 mice for 9 hours) if they had not been removed because of a drop in body temperature.
Video Recording and Processing.
Mice were recorded in their home cages for 3 days, and data acquired in the last 24 hours were used for analysis. For experiments at 33°C, mice were kept in climate chamber (MGC-450HP-2; Yiheng Instruments, Kunshan, China) with 12:12-hour dark/light cycles and adequate food and water supply. Videos were captured at 1 frame per second (fps) and processed with lab-developed software. Briefly, the scale of each video was normalized according to standard grid, the region of interest was specified by hand, the threshold to distinguish mouse from background was selected from a binary representation of the original image, and the centroid of each mouse was automatically calculated and traced for activity. Only mice not moving for a consecutive 40 seconds were considered inactive according to previous studies (Pack et al., 2007), otherwise they were active or between activity bouts. Both the latter two states were counted as active.
EEG/EMG Recording and Processing.
Mice were anesthetized with pentobarbital (50 mg/kg, i.p.) and chronically implanted with EEG and EMG electrodes for polysomnographic recordings according to previous studies (Qu et al., 2010). Two stainless steel screws (1 mm in diameter) were bilaterally inserted through the skull into the cortex (1.0 mm anteriorly from bregma and 1.5 mm laterally to both sides from midline) according to the atlas (Paxinos and Franklin, 2008) to serve as EEG electrodes. Two Teflon-coated insulated stainless steel wires were placed bilaterally into trapezius muscles to serve as EMG electrodes. All electrodes were attached to a microconnector and fixed to the skull with dental cement.
The cable was connected through a slip ring to enable recording from free moving animals. Mice were individually housed for 10 days to recover and allowed to habituate to the recording cable for 3 to 4 days before polygraphic recording started. Each animal was recorded for 24 hours from the onset of dark phase at 7:00 PM.
EEG and EMG signals were amplified and filtered (EEG, 0.5–30 Hz; EMG, 20–200 Hz), digitized at a sampling rate of 128 Hz, and recorded by using SLEEPSIGN (Kissei Comtec, Nagano, Japan) as described earlier (Huang et al., 2005). The processed polygraphic signals were then automatically scored offline by 4-s epochs as wakefulness, REM, and NREM sleep by SLEEPSIGN according to standard criteria (Huang et al., 2005; Kohtoh et al., 2008). The defined sleep-wake states were finally examined manually and corrected, if necessary.
The Mann-Whitney test was carried out with GraphPad Prism [La Jolla, CA]).
Results
Neither Lmx1bf/f/p nor Tph2−/− Mice Significantly Differ in Body Temperature from Wild-Type Mice.
When serotonergic neurons were depleted with Lmx1b-floxp crossed to ePet1-cre (Lmx1bf/f/p), mice were found to have higher body temperature in the dark phase than Lmx1b-floxp (Lmx1bf/f) controls during 12:12-hour dark/light cycles (Hodges et al., 2008). It was not possible to rule out whether this was attributable to heat generated from elevated activity.
To examine whether Tph2-deficient mice were abnormal in thermoregulation, we measured the body temperature and activities of both Lmx1bf/f/p- and Tph2-deficient mice by implantable emitter telemetry during a 24-hour light/dark cycle. All four groups of mice show fluctuations in body temperature, indicated as core temperature, within the 24-hour period: higher during the dark phase and lower during the light phase (Fig. 1). Different from the previous report (Hodges et al., 2008), we did not observe significant difference in body temperature between Lmx1bf/f versus Lmx1bf/f/p mice (Fig. 1, A and B), nor between Tph2+/+and Tph2−/− mice (Fig. 1, C and D).
Movement of each mouse was measured as an indicator for activity simultaneously with the body temperature. All four groups of mice exhibited higher activity levels during the dark phase than those in the light phase (Fig. 1). Lmx1bf/f/p mice exhibited higher level of cumulative activity than Lmx1bf/f controls during both the dark and the light phases (Fig. 1, A and B), whereas the cumulative activity of Tph2−/−mice was only higher than the Tph+/+ during the light phase (Fig. 1, C and D), when nocturnal animals spend most of the time sleeping. Elevated activities were also observed in previous studies (Hodges et al., 2008; Buchanan and Richerson, 2010) and were suggested to increase the body temperature. This explanation is not consistent with our results, because neither Lmx1bf/f/p nor Tph2−/−mice had elevated body temperature compared with their controls.
Neither Lmx1bf/f/p nor Tph2−/− Mice Exhibited Excessive Activity during Cold Challenge Despite the Failure of Lmx1bf/f/p to Maintain Body Temperature.
To test the hypothesis that cold stress is the cause of elevated activity in the Lmx1bf/f/p-deficient mice, as suggested in the previous paper (Buchanan and Richerson, 2010), we measured the body temperature and locomotion of all four groups of mice at 4°C for 5 hours. Lmx1bf/f/p mice indeed failed to maintain the body temperature, which dropped from 36.7 ± 0.3°C to around 25°C (Fig. 2, A and C). By contrast, Tph2−/− mice were able to maintain their body temperature within the same range as the Tph2+/+ mice (Fig. 2, D and F), suggesting that molecules other than 5-HT is involved in thermoregulation. Furthermore, we did not observe significantly elevated activity in either Lmx1bf/f/p or Tph2−/− mice compared with their controls (Fig. 2, B, C, E, and F), which is against the hypothesis that excessive activities are responsible for counteracting body temperature loss.
Increased Activity in Lmx1bf/f/p Mice was Not Diminished at Elevated Ambient Temperature.
Since telemetry only offered a gross indicator of activity by recording the number of times that animals changed their locations, we also measured locomotion by video recording. Each group of mice was recorded in home cages at either room temperature or 33°C for a 12:12-hour dark/light cycle. All mice retained normal 24-hour circadian cycles with more activities during the dark phase than the light phase. At room temperature, both Lmx1bf/f/p and Tph2−/− mice continuously displayed more activities than their controls. Lmx1bf/f/p mice moved a distance of 134.1 ± 13.1 m, whereas Lmx1bf/f mice moved 30.9 ± 5.0 m during the dark phase and 27.9 ± 3.5 m versus 7.8 ± 1.3 m during the light phase (Fig. 3, A and B), and also spent more time moving than Lmx1bf/f (Fig. 3, C and D). Similarly, the distances traveled by Tph2−/− versus Tph2+/+ mice (Fig. 3, E and F) and duration of locomotion differed significantly (Fig. 3, G and H). Similar to a recent publication using adult-specific conditional knockout of Tph2 gene (Whitney et al., 2016), the active state of Tph2−/− lasted for more than 1 hour after the light was on, whereas it was not observed in Lmx1bf/f/p mice at 25°C (Fig. 3, A, C, E, and G).
In contrast to the previous report of thermosensitive sleep recovery of Lmx1bf/f/p mice at 33°C versus the room temperature (Buchanan and Richerson, 2010), we observed that Lmx1bf/f/p mice displayed higher activity levels during both the dark and light phases at 33°C (Fig. 3, I and J). Similar to our finding with telemetry (Fig. 1), Tph2−/− mice exhibited elevated activity only during the light phase at 33°C (Fig. 3, M–P). There was no significant differences in distance or duration during the dark phase at 33°C (Fig. 3, M–P). The diminished difference in locomotion between Tph2−/− and Tph2+/+ mice during the dark phase was not due to decreased activity with Tph2−/−, because the locomotion distances traveled at higher temperature increased with both Tph2+/+ and Tph2−/− groups, but the increase with Tph2+/+ mice was significantly larger (Supplemental Fig. S1). While Lmx1bf/f/p mice exhibited prolonged activity at 33°C, Tph2−/− mice displayed even higher level of activity after light on (Fig. 3, I, K, M, and O).
REM Sleep Was Decreased in Tph2−/− Mice.
To directly verify whether 5-HT was involved in regulating sleep, we recorded the EEG/EMG to distinguish wakefulness, NREM, and REM sleep in Tph2+/+ and Tph2−/− mice for a 24-hour cycle at either room temperature or 33°C. The EEG signal is high in frequency and low in voltage during wakefulness and REM sleep, and these features are reversed during NREM sleep with low frequency and high voltage. The EMG signal is high during wakefulness and low during both REM and NREM sleep, which, when combined with the EEG signal, unambiguously distinguishes the three states of sleep-wake cycle.
To our surprise, at room temperature, Tph2−/− mice did not exhibit significantly more wakefulness than Tph2+/+ mice (Fig. 4, A–C). Neither did they differ in NREM sleep (Fig. 4, A–D). At room temperature, Tph2−/− mice showed significantly less REM than Tph2+/+ during the light phase but not the dark phase (Fig. 4, A, B, and E).
At 33°C, interestingly, the differences of wakefulness, NREM, and REM become larger between Tph2+/+ and Tph2−/− mice. During the dark phase, Tph2−/− mice stayed awake for longer periods (Fig. 4, F–H) and spent less time in both NREM (Fig. 4, F, G, and I) and REM sleep (Fig. 4, F, G, and J). During the light phase, Tph2−/− mice also had less REM sleep than the Tph2+/+ (Fig. 4, F, G, and J). We did not observe a significant difference in the amount of wakefulness (Fig. 4, F–H) or NREM sleep (Fig. 4, F, G, and I) during the light phase at 33°C. These results indicate that the sleep phenotype of Tph2−/− mice is present at 33°C.
Discussion
5-HT and central serotonergic neurons in the raphe nuclei have been implicated in multiple physiologic processes, which complicated the endeavor to distinguish one role from another and also made it confusing whether a role was played by 5-HT or by serotonergic neurons but with other transmitters. The function of 5-HT in sleep-wake regulation has remained under debate for many years, with recent papers concluding that the apparent sleep phenotype in mice lacking serotonergic neurons was explained by the primary role of serotonergic neurons in thermoregulation (Buchanan and Richerson, 2010). Here we provide multiple pieces of evidence indicating that the role of 5-HT in sleep is not attributable to its role in thermoregulation.
Mice depleted of serotonergic neurons were found to be defective in body temperature control as they were not able to keep normal temperature during coldness challenge (Hodges et al., 2008). However, a paper with a Tph2−/− strain, which deleted the coding part of Tph2 gene exon 1 and 2, did not report body temperature changes in Tph2-deficient mice (Alenina et al., 2009). We used two types of mouse mutants to study the roles of serotonergic neurons and 5-HT in sleep regulation. One of them, Lmx1bf/f/p crossed with ePet1-cre, is the same as has been studied previously (Buchanan and Richerson, 2010). Mice lacking serotonergic neurons showed higher body temperature during the dark phase and were unable to maintain normal body temperature when challenged by coldness at 4°C. We also found that the temperature of Lmx1bf/f/p mice dropped dramatically at 4°C within 5 hours (Fig. 2). Since hindbrain ePet1-positive neurons also contain other neurotransmitters such as glutamate and neuropeptides such as substance P, we used Tph2−/− mice to study the role of 5-HT (Zhao et al., 2006). We found that the body temperature of Tph2−/− mice did not differ from that of wild-type (WT) mice and remained at the normal range at 4°C. Another study found that, although the body temperature of the Tph2−/− mutants dropped lower than the WT at the beginning of cold challenge, it recovered slowly to the same level as the WT at the end of the 4th hour (Alenina et al., 2009), supporting that mice lacking 5-HT were not defective in thermoregulation.
Previous work with Lmx1bf/f/p mice demonstrated that increased wakefulness and decreased sleep with the Lmx1bf/f/p disappeared as the ambient temperature increased from 24°C to 33°C, leading to the conclusion that elevated activity and reduced sleep was due to deregulation of body temperature and perception of 24°C as cold stress. However, our results do not support this conclusion. First of all, another study from the same group demonstrated that Lmx1bf/f/p mice preferred ambient temperature of 30.6°C, which was similar to the Lmx1bf/f (Hodges et al., 2008), suggesting that temperature perception in Lmx1bf/f/p mice was not different from the WT. Secondly, despite decreased body temperature at 4°C, no elevated activity of Lmx1bf/f/p mice was observed by us during cold challenge (Fig. 2), raising the possibility that activity was not increased in these mice to generate heat. Moreover, Lmx1bf/f/p mice exhibited higher level of locomotion activity at both room temperature and 33°C. Similar to a recent publication with the Tph2 gene conditionally deleted in the raphe nuclei of adult mice (Whitney et al., 2016), the Tph2−/− mice exhibited higher activity level in locomotion and activity time during both the dark and light phases at the room temperature, whereas elevation was only observed during the light phase at 33°C (Fig. 3). Activity was increased in Tph2−/− mice as well as Lmx1bf/f/p at both the normal and warm ambient temperatures, suggesting a direct role of 5-HT in sleep-wake regulation. These results demonstrated that absence of 5-HT did not dramatically disrupt body temperature control. Therefore, the elevated activity in Tph2−/−mice could not be attributed to hypothermia. Thirdly, the EEG/EMG measurements further support our point. Compared with Tph2+/+ mice, REM duration was reduced in Tph2−/− mice at both room temperature and 33°C (Fig. 4). The changes in wakefulness and NREM sleep duration are less clear. We only observed marginal difference between Tph2−/− and Tph2+/+ mice at room temperature. It was not significant until the ambient temperature was 33°C and only during the dark phase. The different tendency of changes from room temperature to 33°C between activity and sleep/wake duration may indicate an increase in passive wakefulness with Tph2−/− and increased activity with Tph2+/+at warm temperatures.
Serotonergic receptors have been studied for their roles in sleep regulation (reviewed in Monti, 2011). For example, 5-HT1A receptor is thought to act as a presynaptic self- inhibitory receptor, and REM sleep was increased in mouse mutants lacking 5-HT1AR, which was opposite to the effect of 5-HT depletion (Boutrel et al., 2002). Consistently, Inhibition of 5-HT1A receptors by antagonist WAY100635 increased REM sleep, whereas the agonist 8-hydroxy-2-(di-n-propylamino)tetralin (8-OH-DPAT) decreased REM sleep in both mice and cats (Portas et al., 1996; Boutrel et al., 2002). REM was also increased in mice lacking 5-HT1BR (Boutrel et al., 1999), which is thought to act as an inhibitory receptor, too. NREM sleep was decreased, wakefulness was increased, but REM sleep was not affected in 5-HT2AR mutant mice (Popa et al., 2005). Wakefulness was increased, NREM decreased, but REM was unaffected in mutants lacking the 5-HT2C receptor (Frank et al., 2002), which is similar to the sleep phenotype of mice lacking 5-HT2AR. REM was decreased in mice lacking the 5-HT7 receptor (Hedlund et al., 2005). In the future, it will be ideal to study mutants of 5-HT receptors of identical genetic background and with identical methods under the same conditions.
In summary, we found that dysfunction in thermoregulation could not explain the elevated wakefulness and decreased sleep level in animals lacking 5-HT, in that: 1) animals deficient in 5-HT synthesis were able to maintain body temperature in cold environment, 2) animals depleted of serotonergic neurons preferred the same ambient temperature as the WT mice, 3) animals lacking serotonergic neurons exhibited more activity at an elevated ambient temperature, and 4) animals deficient in 5-HT synthesis woke more and slept less even at an elevated ambient temperature. Thus, our study supports a role of 5-HT in sleep-wake regulation independent of thermoregulation.
Acknowledgments
The authors are grateful to Dr. Zhoufeng Chen for sharing the Tph2 mice and to Yunxia Si and Xiangyun Yue for animal husbandry.
Authorship Contributions
Participated in research design: Zhang, Yan, Huang, Rao.
Conducted experiments: Zhang, Yan, Luo.
Performed data analysis: Zhang, Yan, Luo.
Wrote or contributed to the writing of the manuscript: Zhang, Yan, Rao.
Footnotes
- Received November 22, 2017.
- Accepted April 3, 2018.
We are grateful to the National Natural Science Foundation of China (Project 31421003 to Y.R.), the Beijing Innovation Center for Genomics, the Beijing Municipal Commission of Science and Technology (No. Z151100003915121) for grant support.
↵1 X.Z. and H.Y. contributed equally to this work as first authors.
↵
This article has supplemental material available at molpharm.aspetjournals.org.
Abbreviations
- 5-HT
- 5-hydroxytryptamine
- 5-HTP
- 5-hydroxytryptophan
- EEG
- electroencephalographic
- EMG
- electromyography
- Lmx1b
- LIM homeobox transcription factor 1
- NREM
- non-rapid eye movement
- pCPA
- p-chlorophenylalanine
- REM
- rapid eye movement
- Tph2
- tryptophan hydroxylase 2
- WT
- wild type
- Copyright © 2018 by The American Society for Pharmacology and Experimental Therapeutics
