Melatonin Administered During the Fetal Stage Affects Circadian Clock in the Suprachiasmatic Nucleus but Not in the Liver e ,1 Marta Nova kova ,1 Kristy ju, pa n Kubık,2 na Mate ˚ 1 St Pavel Houdek,1 Lenka Polidarova 1 Alena Sumova 1
Department of Neurohumoral Regulations, Institute of Physiology, v.v.i., Academy of Science of the Czech Republic, Videnska 1083, 14220 Prague, Czech Republic
2
Department of Neurophysiology of Memory, Institute of Physiology, v.v.i., Academy of Science of the Czech Republic, Videnska 1083, 14220 Prague, Czech Republic
Received 14 May 2014; accepted 16 July 2014
ABSTRACT: The mammalian circadian system develops gradually during ontogenesis, and after birth, the system is already set to a phase of the mothers. The role of maternal melatonin in the entrainment of fetal circadian clocks has been suggested, but direct evidence is lacking. In our study, intact or pinealectomized pregnant rats were exposed to constant light (LL) throughout pregnancy to suppress the endogenous melatonin and behavioral rhythms. During the last 5 days of gestation, the rats were injected with melatonin or vehicle or were left untreated. After delivery, daily expression profiles of c-fos and Avp in the suprachiasmatic nuclei (SCN), and Per1, Per2, Rev-erba, and Bmal1 in the liver were measured in 1-day-old pups. Due to the LL exposure, no gene expression rhythms were detected in the SCN of untreated pregnant rats or in the SCN and liver of the
INTRODUCTION The mammalian circadian system is composed of hierarchically organized clocks forming a complex regulatory system that helps organisms adapt to cyclic changes in external environment. The circaCorrespondence to: A. Sumova ([email protected]). Conflict of interest: The authors declare no conflict of interest. Ó 2014 Wiley Periodicals, Inc. Published online 17 July 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/dneu.22213
pups. The administration of melatonin to pregnant rats entrained the pups’ gene expression profiles in the SCN, but not in the liver. Melatonin did not affect the maternal behavior during pregnancy. Vehicle injections also synchronized the gene expression in the SCN but not in the liver. Melatonin and vehicle entrained the gene expression profiles to different phases, demonstrating that the effect of melatonin was apparently not due to the treatment procedure per se. The data demonstrate that in pregnant rats with suppressed endogenous melatonin levels, pharmacological doses of melatonin affect the fetal clock in the SCN but not in the liver. VC 2014 Wiley Periodicals, Inc. Develop Neurobiol 75: 131–144, 2015
Keywords: ontogenesis; circadian system; suprachiasmatic nuclei; clock gene; melatonin
dian clocks reside in nearly every, if not all, mammalian cells (for a review, see Schibler et al., 2003). The cellular clock mechanism is dependent on a set of socalled clock genes, namely Per1, Per2, Cry1, Cry2, Bmal1, Clock, Rev-erba, and Rora (for a review, see Takahashi et al., 2008). Cellular clocks are regularly entrained by the master pacemaker located in the suprachiasmatic nuclei (SCN) of the ventral hypothalamus (Ralph et al., 1990). The circadian system develops gradually during the fetal and early postnatal period (for review, see Sumova et al., 2012; Weinert, 2005). 131
132
Houdek et al.
The phase of the developing SCN clock is already set prenatally. Earlier studies provided indirect evidence for prenatal setting of the SCN clock via extrapolation of the phase of the behavioral activity rhythm that pups developed later during postnatal ontogenesis to assess the phase of the fetal clock (Davis and Gorski, 1988; Reppert and Schwartz, 1986a). More recent studies used an approach involving the detection of phase of the gene expression rhythms in the SCN of rat pups immediately after birth, which more closely reflected the phase of the fetal clock (El-Hennamy et al., 2008). All these studies provided evidence that the phase of the SCN clock is set prenatally and that after birth, it runs in synchrony with the maternal circadian system. The prenatal setting of peripheral clocks is much less understood because in contrast to the SCN, the phases of the peripheral clocks do not manifest at the behavioral level, and the development of the rhythms is tissue dependent. Our previous data demonstrated that the phases of the daily profiles of Rev-erba expression in the fetal rat liver (Sladek et al., 2007a) and colon (Polidarova et al., 2014) roughly corresponded to the phases in adult animals; however, after delivery, the phases significantly changed and attained the adult-like state again only after weaning. During the prenatal period, the maternal cues are likely important for phasing of the peripheral clocks after birth because the dynamics of the postnatal changes in the phasing of the colonic clocks in newborn pups that had been reared since birth by foster mothers (maintained in opposite LD cycle compared with that of their original mothers) differed relative to the dynamics in the pups that had been reared by their own mothers, suggesting that the prenatal history matters (Polidarova et al., 2014). The nature of maternal cue(s) that entrain the fetal clocks has been studied for decades; however, the mechanisms are still not completely clear. Melatonin has been the most favorite candidate (see below), although it surely is not the only temporal cue perceived by the fetal clocks. Other likely candidates are maternal behavior and the related temperature and feeding cycles. Shifting the feeding regime of pregnant rats with a surgically removed SCN shifted the phase of the drinking activity rhythms that the pups developed postnatally (Weaver and Reppert, 1989). The manipulation had an immediate effect on the fetal SCN because the phase of gene expression in the SCN of pups was shifted immediately after birth (Novakova et al., 2010). In analogy with the experiments in the SCN-lesioned mothers, the feeding regime was able to entrain the phase of the fetal SCN clock when the maternal SCN was disturbed due to Developmental Neurobiology
prolonged exposure to constant light (LL; Novakova et al., 2010) but not under a standard 12 h light/12 h dark cycle when the maternal SCN was functional (Novakova et al., 2010). Importantly for context of this study, the results provided evidence that maternal melatonin was not necessary for the shift of the fetal SCN clock because the shift was achieved in spite of melatonin suppression in LL (Novakova et al., 2010). However, these results also implied that rhythmic signaling via the maternal feeding/behavior-related cues was only redundant to the SCN-dependent signaling. The melatonin produced by the pineal gland represents a direct output of the SCN (for review, see Arendt, 2006). As already mentioned, a role of melatonin in setting the phase of the fetal clocks has been repeatedly suggested because it can easily cross the placenta (Reppert et al., 1979; Velazquez et al., 1992; Yellon and Longo, 1987) and the fetal SCN contain melatonin receptors (Reppert et al., 1988). Experiments that aimed to reveal an impact of the absence of the maternal melatonin signal during the prenatal stage on the synchrony of the pups’ clocks provided inconsistent results, demonstrating either no (Reppert and Schwartz, 1986b) or significant (Bellavia et al., 2006) effects. This disparity was likely due to the fact that in both experiments, the strength of the maternal rhythmic cues that were derived from outside the pineal differed. In contrast, timed injections of pharmacological doses of melatonin to SCNlesioned pregnant hamsters restored synchrony among pups (Davis and Mannion, 1988). Therefore, it seems that the effect of melatonin on the phase of the fetal clocks is apparent only in the absence of other SCN-driven rhythmic cues. The role of melatonin in entrainment of fetal peripheral clocks has been previously studied only in the adrenals (Mendez et al., 2012; Seron-Ferre et al. 2012; Torres-Farfan et al., 2011). Although the entraining effect of melatonin on the fetal SCN clock seems well documented (Davis and Mannion, 1988; Grosse and Davis, 1998), a question remains of how and when melatonin entrains the fetal clock mechanism. In vitro, melatonin was able to shift the rhythms in electrical activity of the adult SCN slices (McArthur et al., 1991); however, thus far, no effect on the phase of the clock gene expression rhythms has been demonstrated for either the fetal or adult SCN. In capuchin monkeys, melatonin suppression due to exposure to LL was associated with changes in the expression levels of clock genes in the fetal SCN (Torres-Farfan et al., 2006). However, due to a limitation in the time points at which the clock gene expression was measured, it was not clear whether the observed change was due to a shift
Melatonin Entrains Fetal Circadian Clocks
or suppression of the clock gene expression. Also, it was not shown whether the melatonin that was administered to pregnant capuchin monkeys affected their behavior, which could theoretically affect the fetal SCN clock independently of melatonin. Hence, a possibility for melatonin to entrain the mothers (as was previously demonstrated for adult animals, Redman et al., 1983; Slotten et al., 1999) and that the imposed rhythm in behavior was responsible for the entrainment of the fetal clock could not be ruled out. All other currently available data demonstrating melatonin entrainment of the fetal clock are only based on the detection of the circadian phases of output rhythms, which develop later postnatally. However, theoretically, it is possible that during the fetal stage, maternal melatonin affects non-SCN structures that may modulate the developing rhythms later postnatally. Therefore, direct evidence of the ability of melatonin to affect the fetal SCN clock is still lacking. Thus, the aim of our present study was to elucidate whether melatonin administered during the prenatal stage may synchronize the gene expression profiles in the SCN and liver of fetuses which were mutually desynchronized by exposure of their mothers to LL.
METHODS Experimental Animals Adult male and female Wistar rats (Velaz s.r.o., Czech Republic) were maintained at a temperature of 21 6 2 C in a light/dark regime with 12 h of light and 12 h of darkness per day (LD12:12); the lights were turned on at 06:00 h and off at 18:00 h. Time was expressed as Zeitgeber time (ZT), with ZT0 corresponding to the time when the lights were turned on. Light was provided by overhead 40-W fluorescent tubes, and the illumination was between 50 and 300 lux, depending on the cage position in the animal room. The animals had free access to food and water throughout the entire experiment. All experiments were approved by Animal Care and Use Committee of the Institute of Physiology in agreement with the Animal Protection Law of the Czech Republic as well as European Community Council directives 86/609/EEC. All efforts were made to alleviate the suffering of the animals.
133
of the same intensity as the light of LD12:12. In experiment 1, pregnant rats received regular daily injections of melatonin (i.p., 1 mg/kg b.w.) at ZT11 during the last 5 days of the pregnancy, that is, on E17–21. One control group was treated with vehicle (2% ethanolic saline) at ZT11 on E17–21, and the other control group was left untreated during the entire pregnancy. In experiment 2, the dams underwent surgery to remove their pineal glands before they were mated with males and exposed to LL (for the surgical procedure, see below). During the pregnancy, melatonin or vehicle was administered in the same way as described for Experiment 1. In both experiments, the pups were sacrificed by decapitation at 3-h intervals during the 24 h period of the 1st–2nd postnatal day and samples of tissues (brains and livers in Experiment 1 and only brains in Experiment 2) were collected. The brains were immediately frozen on dry ice and stored at 280 C. The samples of the livers were immersed into RNAlater stabilization reagent (Qiagen, Valencia) and stored at 280 C. A series of control experiments was performed as follows:
i. An effect of exposure to LL on the endogenous melatonin serum levels in dams was determined. Dams were maintained either in the same LL conditions as in Experiments 1 and 2 (n 5 6) or in LD12:12 (n 5 8). After 17 days in LL, the dams were sacrificed at ZT8 (corresponding to the previous daytime), or at ZT16 and ZT20 (corresponding to the previous first and second halves of the nighttime, respectively), and blood samples were collected and processed to measure the melatonin levels by radioimmunoassay as described below. ii. An effect of the injection procedure per se (vehicle injection and handling) to pregnant rats on the serum corticosterone levels was assessed. The pregnant dams (n 5 9) were maintained in LL from the beginning of pregnancy until the gestational day 17. Then, one group of pregnant rats (n 5 5) received a single vehicle injection (i.p.) between ZT11:30 and ZT12:30, that is, at approximately the same time as melatonin or vehicle was administered in Experiments 1 and 2) and was sacrificed 30 min afterward. The other group of dams (n 5 4) was left untreated and was sacrificed at the same time as the rats of the first group. The collected blood samples were processed to determine the levels of corticosterone by radioimmunoassay as described below.
Experimental Protocol
Locomotor Activity Monitoring
Adult female rats were inspected by vaginal smears every day to detect the phases of their estrous cycles. On the night of proestrus, the dams were mated with males, and the day when the vaginal smears were found to be sperm-positive was considered embryonal day (E) 0. From E0 until the end of the experiment, the pregnant rats were maintained in LL
Pregnant rats were maintained individually in cages equipped with infrared movement detectors attached above the center of the top of the cage, thus enabling detection of their locomotor activity across the entire cage. The activity was detected every minute using a circadian activity monitoring system (Dr. H.M. Cooper, INSERM, France). Developmental Neurobiology
134
Houdek et al.
Double-plotted actograms were generated for visualization of data.
Pinealectomy Pinealectomy (PINX) of adult female rats for Experiment 2 was performed using deep anesthesia (ketamine 85 mg/kg b.w., xylazine 10 mg/kg b.w., and atropine 0.1 mg/kg b.w.). The hair on the head was shaved, the skin was incised, and a hole in the skull was bored at the lambda position by a micro drill. The pineal gland was removed by fine forceps. The surgical wound was closed with the bone fragment being returned back, and the skin was sutured. The threads were removed 2 weeks after the surgery. The recovery period took approximately 1 month until the sutures were no longer visible. Thereafter, mating with male rats was realized as described above. The completeness of the pineal removal was assessed after the Experiment 2 was terminated via (i) visual inspection post mortem and (ii) detection of the plasma melatonin levels using radioimmunoassay. Before the blood sampling, the PINX dams were exposed to LD12:12 for 10 days and then they were sacrificed, and the blood samples were collected around midnight, that is, between ZT17 and ZT19, when the highest endogenous melatonin levels were expected.
In situ Hybridization The pup heads, which had been maintained at 280 C, were sectioned in a cryostat into five series of 12-mm-thick slices throughout the entire rostrocaudal extent of the SCN, and the sections were organized in an alternating order on slides. The sections were further processed for in situ hybridization to determine the profiles of c-fos mRNA and Avp hnRNA in the SCN. The cDNA fragments of rat c-fos (1160 bp; corresponding to nucleotides 141–1300 of the sequence in GenBank accession number X06769) and Avp (506 bp; identical to nucleotides 796–1302 of the intronic sequence in GenBank accession no. X01637) were used as templates for in vitro transcription of complementary RNA probes. The c-fos fragment–containing vector was generously donated by Dr. Tom Curran (Children’s Hospital of Philadelphia, Pennsylvania). The Avp cDNA was cloned in our laboratory. Probes were labeled using 35S-UTP, and the in situ hybridizations were performed as previously described (El-Hennamy et al., 2008). The sections were hybridized for 20 h at 60 C. Following a posthybridization wash, the sections were dehydrated in ethanol and dried. Finally, the slides were exposed to BIOMAX MR film (Kodak) for 10–14 days and developed using the ADEFOMIX-S developer and ADEFOFIX fixer (ADEFO-CHEMIE Gmbh, Germany). Brain sections from all experimental groups were processed simultaneously under identical conditions. Autoradiographs of the sections were analyzed using an image analysis system (Image Pro, Olympus, New Hyde Park, NY) and the relative optical density (OD) of the speDevelopmental Neurobiology
cific hybridization signal was detected. For each animal, the mRNA or hnRNA was quantified bilaterally, at the midcaudal SCN section containing the strongest hybridization signal. Each measurement was corrected for nonspecific background by subtracting the OD values from the same adjacent area in the hypothalamus. The background signal of that adjacent area served as an internal standard; it was consistently low and did not exhibit marked changes with the time of day. Finally, the slides were counterstained with cresyl violet to check for the presence and midcaudal position of the SCN in each section. The OD for each animal was calculated as a mean of values for the left and right SCN.
RNA Isolation and Real-Time RT-PCR Total RNA was isolated using an RNeasy Mini kit (Qiagen, Valencia) according to the manufacturer’s instructions. RNA concentrations were determined using spectrophotometry at 260 nm, and the RNA quality was assessed by electrophoresis on a 1.5% agarose gel. Moreover, the integrity of randomly selected samples of total RNA was tested using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara). The real-time RT-PCR method, which was used to detect mRNA levels of Per1, Per2, Rev-erba, Bmal1, Dbp, and Wee1, was described previously (Polidarova et al., 2011; Sladek et al., 2007a). Briefly, 1 mg of total RNA was reverse transcribed using a SuperScript VILO cDNA Synthesis kit (Life Technologies, USA). Diluted cDNA was then amplified on a LightCycler (Roche, Basel, Switzerland) using a Express SYBR Green qPCR Supermix Universal kit (Life Technologies, USA) and the corresponding primers. Relative quantification was achieved using a standard curve and subsequently normalizing the measured gene expression to b2-microglobulin. The housekeeping gene b2-microglobulin has been used for normalization previously (Sladek et al., 2007a; Sladek et al., 2007b), and its expression was stable throughout the 24 h and did not vary markedly between the analyzed tissues.
Radioimmunoassay The collected blood samples (5–10 mL per animal) were left to clot for 45 min at room temperature while protected from light. After centrifugation (1800 g, 30 min), the serum was separated and stored at 220 C until the RIA was performed. For assessment of the serum melatonin and corticosterone levels, the Melatonin Research RIA kit (LDN, Germany) and Rat Corticosterone H3 kit (MP Biomedicals, CA), respectively were used according to the guidelines of the manufacturers.
Statistical Analysis For analysis of the gene expression profiles, data were tested for rhythmic and nonrhythmic expression by fitting to two alternative regression models, that is, either a horizontal straight line (null hypothesis) or a single cosine curve (alternative hypothesis), which was defined by the
Melatonin Entrains Fetal Circadian Clocks equation Y 5 mesor 1 [amplitude 3 cos(2 3 p 3 (X-acrophase)/wavelength)] with a constant wavelength of 24 h. The extra sum-of-squares F test was used for comparison, and when the p value exceeded 0.05, the cosine curve parameters were calculated. The amplitude (i.e., the difference between the peak or trough and the mean value of a cosine curve), acrophase (i.e., the phase angle of the peak of a cosine curve), mesor (i.e., the average value around which the variable oscillates), and coefficient of determination R2 (i.e., goodness of fit) were detected. The leastsquares regression method implemented in the Prism 5 software (GraphPad, La Jolla) was applied. The mean daily levels of melatonin were compared between two experimental groups of rats by Student’s t test, and the results of the Student’s t-test are expressed as tx values (with x in the lower index representing degrees of freedom) and as p values (level of significance), with p < 0.05 required for significance.
RESULTS LL Abolishes the Rhythms in Behavior, Plasma Melatonin Levels and Gene Expression in the SCN of Pregnant Rats, As Well As the Rhythms in Gene Expression in the SCN of Newborn Pups As expected, exposure to LL abolished the circadian rhythm in behavior in all pregnant rats. At the beginning, the locomotor activity became free running with a long period for approximately the first 2 weeks, and later, during the last week of pregnancy, all rats became completely arrhythmic [Fig. 1(A)]. To detect an effect of the 3-week exposure to LL on endogenous melatonin levels, melatonin was assayed in the serum of rats at three different time points, namely at the times corresponding to the previous daytime (ZT8) and the previous nighttime (ZT16 and 20). The daily melatonin levels were compared with those in control animals that remained in LD12:12 and sampled at the corresponding time points [Fig. 1(B)]. The daily melatonin levels expressed as a mean of the three samples during the 24 h were significantly lower in the animals maintained in LL compared with those maintained in LD12:12 (t-test, F12 5 9.894, p < 0.001). The data confirmed that in the pregnant rats, the endogenous pineal melatonin production was significantly reduced throughout the 24-h period. Additionally, in the SCN of the pregnant rats maintained in LL, no significant rhythms in c-fos and Avp gene expression were detected [Fig. 1(C)]. To determine whether the lack of the maternal rhythmicity also affects the fetal SCN, the daily profiles of c-fos and
135
Avp gene expression were determined in the SCN of newborn pups at P1 [Fig. 1(C)]. Cosinor analysis did not reveal significant circadian rhythms in the expression of any of these genes (c-fos: p 5 0.4473; Avp: p 5 0.3343). These results demonstrate that due to exposure to the LL for 3 weeks, the maternal SCN clock did not drive the circadian rhythm in locomotor activity, pineal melatonin production and the SCN c-fos and Avp gene expression. The SCN clocks of the newborn pups did not exhibit circadian rhythms in c-fos and Avp gene expression.
The Injection Procedure Per Se May Affect the Behavior and Corticosterone Levels in Some But Not All Pregnant Rats Maintained in LL Inspection of the individual locomotor activity records revealed that repeated injections (vehicle and/or melatonin) to pregnant rats at E17–21 affected the behavior of some of the animals. To analyze whether the injections entrained locomotor activity, the actograms for the interval of the entire pregnancy and v2 periodograms for the interval of the last 5 days when injections were administered to each pregnant rat were evaluated. In most of the animals, the behavior was not affected by administration of vehicle [Fig. 2(A,C)] or by melatonin injections [Fig. 2(B,D)], and the rats exhibited arrhythmic behavior that was undisturbed by the injections. In a minority of the rats, the effect of the injections was manifested by slightly increased activity immediately after each injection (2 out of 8 vehicle-treated and in 3 out of 9 melatonin-treated animals). Only one out of eight vehicle-treated and one out of nine melatonin-treated animals exhibited a significant 24-h period in locomotor activity during the 5 days of injections. Obviously, the proportion of animals affected by the injections was approximately the same in the vehicle- and melatonin-treated groups. Therefore, the repeated melatonin injections did not significantly entrain the behavioral rhythms of the pregnant rats. The vehicle contained ethanolic saline because of melatonin solubility, and thus, we could not exclude a possibility that these injections were slightly painful to the pregnant rats. Therefore, in a separate experiment, we tested whether the i.p. injections represented a mild stressor to the animals. In untreated controls (n 5 4), the serum corticosterone levels measured at ZT11:30 were 490.0 6 62.8 ng/mL. In the vehicle-injected animals (n 5 5), the serum corticosterone levels measured 30 min after the injection Developmental Neurobiology
136
Houdek et al.
Figure 1 Effect of constant light on locomotor activity, melatonin levels in blood and gene expression in the SCN. (A) A representative double-plotted actogram of a pregnant rat maintained in constant light throughout the entire gestation, that is, from embryonic day (E)1 until E22. (B) The melatonin levels in the serum (pg/mL) of pregnant rats maintained in light/dark regime (LD12:12) or constant light (LL) and sampled at 3 time points a day. ZT8 represents the day, ZT16 represents the first half of night, and ZT20 represents the second half of night. (C) Daily profiles of c-fos and Avp expression in the SCN of mothers after delivery that were maintained in LL throughout entire pregnancy (upper panel) and in their 1-day-old pups (lower panel). The data are expressed as the relative OD measured in the individual SCN. The data were tested by cosinor analysis for rhythmic expression, and the horizontal straight lines demonstrate the absence of significant circadian rhythms. Time is expressed in Zeitgeber time (ZT). For more details, see Materials and Methods.
at ZT11 varied among the pregnant rats; in three animals, the levels were approximately the same as in controls (528.7 6 17.9 ng/mL), but in two animals, the levels were remarkably increased (2407.7 and Developmental Neurobiology
1278.7 ng/mL). The data suggested that at least in some (but not in all) pregnant rats, the i.p. injections of vehicle/melatonin affected the behavior and basal serum corticosterone levels.
Melatonin Entrains Fetal Circadian Clocks
137
Figure 2 Analysis of locomotor activity in pregnant rats maintained in constant light and injected with vehicle (A,C) and melatonin (B,D) during the last 5 days of gestation. Representative doubleplotted actograms demonstrate the locomotor activity of a rat treated with vehicle (A) and of a rat treated with melatonin (B). The vertical bars depict the days and time of the treatment. Individual v2 periodograms for the same animals as in A and B were calculated from the 5 days of injections with vehicle (C) and melatonin (D). The diagonal line depicts the threshold for significant period of the rhythm (p < 0.001).
Repeated Injections of Melatonin to Pregnant Rats Maintained in LL Affect the Gene Expression in the Fetal SCN Daily profiles of c-fos and Avp expression were determined in the SCN of 1-day-old pups that were born to mothers maintained in LL and injected with melatonin/vehicle as described in Experiment 1. The data were fitted with cosine curves (Fig. 3). The results of the cosinor analysis, namely the acrophases, amplitudes, mesors, R2, and p values, are summarized in Table 1. The c-fos and Avp expression profiles exhibited significant circadian rhythms in the SCN of the 1day-old pups born to mothers maintained in LL and
injected with vehicle or melatonin. However, the phases of the rhythms in the c-fos expression (Fig. 3; upper part) differed significantly between the melatonin- and vehicle-treated groups (t-test, F124 5 19.544, p < 0.001). Similarly, the phases of Avp expression profiles (Fig. 3; lower part) also significantly differed in between the pups born to the mothers injected with vehicle and those born to mothers injected with melatonin, (t-test, F124 5 22.559, p < 0.01). Therefore, the repeated vehicle injections to pregnant rats were able to synchronize the c-fos and Avp expression in the fetal SCN and the rhythms peaked approximately 6 and 10 h after the injections, respectively. The repeated melatonin injections synchronized the c-fos and Avp expressions and the rhythms Developmental Neurobiology
138
Houdek et al.
Figure 3 Daily profiles of c-fos (upper panel) and Avp (lower panel) expression in the SCN of 1-day-old pups born to mothers that were maintained in constant light throughout the entire gestation and were treated with vehicle (left) and melatonin (right). The data are expressed as values of the relative OD measured in the individual SCN. The data were fit with cosine curves to determine the presence or absence of circadian rhythms. The sinusoid curves demonstrated significant circadian rhythms. The parameters of the rhythms are summarized in Table 1. Time is expressed in ZT, and the apostrophe in ZT120 means that it is 24 h after the ZT12 when the sampling of the pups started.
peaked approximately 12 and 13 h after the injection, respectively.
Repeated Injections of Melatonin to Pregnant PINX Rats Entrain Gene Expression in the Fetal SCN Only dams whose pineal glands were completely removed and their nighttime serum melatonin levels were reduced to a minimum (23.55 6 3.13 pg/mL) were used in the study. Similarly to intact rats, the PINX pregnant rats maintained in LL also did not exhibit a circadian rhythm in locomotor activity, and the injection procedure had only a minor effect on the behavior. During the 5 days of injections, periodograms detected a significant 24-h period in 0 out of 5 melatonin-treated and 1 out of 6 vehicle-treated PINX pregnant rats. Daily profiles of c-fos and Avp expression were detected in the SCN of 1-day-old pups born to the PINX mothers maintained in LL as described in Experiment 2. The data were fitted with cosine curves Developmental Neurobiology
(Fig. 4), and results of the cosinor analysis are summarized in Table 1. In the SCN of 1-day-old pups born to PINX mothers repeatedly injected with vehicle, the c-fos expression (Fig. 4; upper part) was completely desynchronized. However, in the pups born to mothers that were repeatedly injected with melatonin, the c-fos expression exhibited a significant circadian rhythm. For the Avp expression (Fig. 4; lower part), only a weak rhythm was present in the SCN of the pups born to mothers repeatedly injected with vehicle. In contrast, after repeated melatonin injections, a circadian rhythm of Avp expression with a higher amplitude and R2 value was detected. Moreover, the acrophases of the Avp expression rhythms in the SCN of the pups born to vehicle- and melatonin-treated mothers were significantly different (t-test, F124 5 29.334, p < 0.001). Therefore, in the SCN of the pups born to PINX mothers, the vehicle injections were not able to synchronize the c-fos expression, and only weakly synchronized the Avp expression in the fetal SCN; the Avp expression rhythm peaked approximately
0.063 6 0.003 – – 0.008 0.7611 0.022 6 0.001 23.7 6 0.6 0.014 6 0.002 0.627
Department of Neurohumoral Regulations, Institute of Physiology, v.v.i., Academy of Science of the Czech Republic, Videnska 1083, 14220 Prague, Czech Republic
2
Department of Neurophysiology of Memory, Institute of Physiology, v.v.i., Academy of Science of the Czech Republic, Videnska 1083, 14220 Prague, Czech Republic
Received 14 May 2014; accepted 16 July 2014
ABSTRACT: The mammalian circadian system develops gradually during ontogenesis, and after birth, the system is already set to a phase of the mothers. The role of maternal melatonin in the entrainment of fetal circadian clocks has been suggested, but direct evidence is lacking. In our study, intact or pinealectomized pregnant rats were exposed to constant light (LL) throughout pregnancy to suppress the endogenous melatonin and behavioral rhythms. During the last 5 days of gestation, the rats were injected with melatonin or vehicle or were left untreated. After delivery, daily expression profiles of c-fos and Avp in the suprachiasmatic nuclei (SCN), and Per1, Per2, Rev-erba, and Bmal1 in the liver were measured in 1-day-old pups. Due to the LL exposure, no gene expression rhythms were detected in the SCN of untreated pregnant rats or in the SCN and liver of the
INTRODUCTION The mammalian circadian system is composed of hierarchically organized clocks forming a complex regulatory system that helps organisms adapt to cyclic changes in external environment. The circaCorrespondence to: A. Sumova ([email protected]). Conflict of interest: The authors declare no conflict of interest. Ó 2014 Wiley Periodicals, Inc. Published online 17 July 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/dneu.22213
pups. The administration of melatonin to pregnant rats entrained the pups’ gene expression profiles in the SCN, but not in the liver. Melatonin did not affect the maternal behavior during pregnancy. Vehicle injections also synchronized the gene expression in the SCN but not in the liver. Melatonin and vehicle entrained the gene expression profiles to different phases, demonstrating that the effect of melatonin was apparently not due to the treatment procedure per se. The data demonstrate that in pregnant rats with suppressed endogenous melatonin levels, pharmacological doses of melatonin affect the fetal clock in the SCN but not in the liver. VC 2014 Wiley Periodicals, Inc. Develop Neurobiol 75: 131–144, 2015
Keywords: ontogenesis; circadian system; suprachiasmatic nuclei; clock gene; melatonin
dian clocks reside in nearly every, if not all, mammalian cells (for a review, see Schibler et al., 2003). The cellular clock mechanism is dependent on a set of socalled clock genes, namely Per1, Per2, Cry1, Cry2, Bmal1, Clock, Rev-erba, and Rora (for a review, see Takahashi et al., 2008). Cellular clocks are regularly entrained by the master pacemaker located in the suprachiasmatic nuclei (SCN) of the ventral hypothalamus (Ralph et al., 1990). The circadian system develops gradually during the fetal and early postnatal period (for review, see Sumova et al., 2012; Weinert, 2005). 131
132
Houdek et al.
The phase of the developing SCN clock is already set prenatally. Earlier studies provided indirect evidence for prenatal setting of the SCN clock via extrapolation of the phase of the behavioral activity rhythm that pups developed later during postnatal ontogenesis to assess the phase of the fetal clock (Davis and Gorski, 1988; Reppert and Schwartz, 1986a). More recent studies used an approach involving the detection of phase of the gene expression rhythms in the SCN of rat pups immediately after birth, which more closely reflected the phase of the fetal clock (El-Hennamy et al., 2008). All these studies provided evidence that the phase of the SCN clock is set prenatally and that after birth, it runs in synchrony with the maternal circadian system. The prenatal setting of peripheral clocks is much less understood because in contrast to the SCN, the phases of the peripheral clocks do not manifest at the behavioral level, and the development of the rhythms is tissue dependent. Our previous data demonstrated that the phases of the daily profiles of Rev-erba expression in the fetal rat liver (Sladek et al., 2007a) and colon (Polidarova et al., 2014) roughly corresponded to the phases in adult animals; however, after delivery, the phases significantly changed and attained the adult-like state again only after weaning. During the prenatal period, the maternal cues are likely important for phasing of the peripheral clocks after birth because the dynamics of the postnatal changes in the phasing of the colonic clocks in newborn pups that had been reared since birth by foster mothers (maintained in opposite LD cycle compared with that of their original mothers) differed relative to the dynamics in the pups that had been reared by their own mothers, suggesting that the prenatal history matters (Polidarova et al., 2014). The nature of maternal cue(s) that entrain the fetal clocks has been studied for decades; however, the mechanisms are still not completely clear. Melatonin has been the most favorite candidate (see below), although it surely is not the only temporal cue perceived by the fetal clocks. Other likely candidates are maternal behavior and the related temperature and feeding cycles. Shifting the feeding regime of pregnant rats with a surgically removed SCN shifted the phase of the drinking activity rhythms that the pups developed postnatally (Weaver and Reppert, 1989). The manipulation had an immediate effect on the fetal SCN because the phase of gene expression in the SCN of pups was shifted immediately after birth (Novakova et al., 2010). In analogy with the experiments in the SCN-lesioned mothers, the feeding regime was able to entrain the phase of the fetal SCN clock when the maternal SCN was disturbed due to Developmental Neurobiology
prolonged exposure to constant light (LL; Novakova et al., 2010) but not under a standard 12 h light/12 h dark cycle when the maternal SCN was functional (Novakova et al., 2010). Importantly for context of this study, the results provided evidence that maternal melatonin was not necessary for the shift of the fetal SCN clock because the shift was achieved in spite of melatonin suppression in LL (Novakova et al., 2010). However, these results also implied that rhythmic signaling via the maternal feeding/behavior-related cues was only redundant to the SCN-dependent signaling. The melatonin produced by the pineal gland represents a direct output of the SCN (for review, see Arendt, 2006). As already mentioned, a role of melatonin in setting the phase of the fetal clocks has been repeatedly suggested because it can easily cross the placenta (Reppert et al., 1979; Velazquez et al., 1992; Yellon and Longo, 1987) and the fetal SCN contain melatonin receptors (Reppert et al., 1988). Experiments that aimed to reveal an impact of the absence of the maternal melatonin signal during the prenatal stage on the synchrony of the pups’ clocks provided inconsistent results, demonstrating either no (Reppert and Schwartz, 1986b) or significant (Bellavia et al., 2006) effects. This disparity was likely due to the fact that in both experiments, the strength of the maternal rhythmic cues that were derived from outside the pineal differed. In contrast, timed injections of pharmacological doses of melatonin to SCNlesioned pregnant hamsters restored synchrony among pups (Davis and Mannion, 1988). Therefore, it seems that the effect of melatonin on the phase of the fetal clocks is apparent only in the absence of other SCN-driven rhythmic cues. The role of melatonin in entrainment of fetal peripheral clocks has been previously studied only in the adrenals (Mendez et al., 2012; Seron-Ferre et al. 2012; Torres-Farfan et al., 2011). Although the entraining effect of melatonin on the fetal SCN clock seems well documented (Davis and Mannion, 1988; Grosse and Davis, 1998), a question remains of how and when melatonin entrains the fetal clock mechanism. In vitro, melatonin was able to shift the rhythms in electrical activity of the adult SCN slices (McArthur et al., 1991); however, thus far, no effect on the phase of the clock gene expression rhythms has been demonstrated for either the fetal or adult SCN. In capuchin monkeys, melatonin suppression due to exposure to LL was associated with changes in the expression levels of clock genes in the fetal SCN (Torres-Farfan et al., 2006). However, due to a limitation in the time points at which the clock gene expression was measured, it was not clear whether the observed change was due to a shift
Melatonin Entrains Fetal Circadian Clocks
or suppression of the clock gene expression. Also, it was not shown whether the melatonin that was administered to pregnant capuchin monkeys affected their behavior, which could theoretically affect the fetal SCN clock independently of melatonin. Hence, a possibility for melatonin to entrain the mothers (as was previously demonstrated for adult animals, Redman et al., 1983; Slotten et al., 1999) and that the imposed rhythm in behavior was responsible for the entrainment of the fetal clock could not be ruled out. All other currently available data demonstrating melatonin entrainment of the fetal clock are only based on the detection of the circadian phases of output rhythms, which develop later postnatally. However, theoretically, it is possible that during the fetal stage, maternal melatonin affects non-SCN structures that may modulate the developing rhythms later postnatally. Therefore, direct evidence of the ability of melatonin to affect the fetal SCN clock is still lacking. Thus, the aim of our present study was to elucidate whether melatonin administered during the prenatal stage may synchronize the gene expression profiles in the SCN and liver of fetuses which were mutually desynchronized by exposure of their mothers to LL.
METHODS Experimental Animals Adult male and female Wistar rats (Velaz s.r.o., Czech Republic) were maintained at a temperature of 21 6 2 C in a light/dark regime with 12 h of light and 12 h of darkness per day (LD12:12); the lights were turned on at 06:00 h and off at 18:00 h. Time was expressed as Zeitgeber time (ZT), with ZT0 corresponding to the time when the lights were turned on. Light was provided by overhead 40-W fluorescent tubes, and the illumination was between 50 and 300 lux, depending on the cage position in the animal room. The animals had free access to food and water throughout the entire experiment. All experiments were approved by Animal Care and Use Committee of the Institute of Physiology in agreement with the Animal Protection Law of the Czech Republic as well as European Community Council directives 86/609/EEC. All efforts were made to alleviate the suffering of the animals.
133
of the same intensity as the light of LD12:12. In experiment 1, pregnant rats received regular daily injections of melatonin (i.p., 1 mg/kg b.w.) at ZT11 during the last 5 days of the pregnancy, that is, on E17–21. One control group was treated with vehicle (2% ethanolic saline) at ZT11 on E17–21, and the other control group was left untreated during the entire pregnancy. In experiment 2, the dams underwent surgery to remove their pineal glands before they were mated with males and exposed to LL (for the surgical procedure, see below). During the pregnancy, melatonin or vehicle was administered in the same way as described for Experiment 1. In both experiments, the pups were sacrificed by decapitation at 3-h intervals during the 24 h period of the 1st–2nd postnatal day and samples of tissues (brains and livers in Experiment 1 and only brains in Experiment 2) were collected. The brains were immediately frozen on dry ice and stored at 280 C. The samples of the livers were immersed into RNAlater stabilization reagent (Qiagen, Valencia) and stored at 280 C. A series of control experiments was performed as follows:
i. An effect of exposure to LL on the endogenous melatonin serum levels in dams was determined. Dams were maintained either in the same LL conditions as in Experiments 1 and 2 (n 5 6) or in LD12:12 (n 5 8). After 17 days in LL, the dams were sacrificed at ZT8 (corresponding to the previous daytime), or at ZT16 and ZT20 (corresponding to the previous first and second halves of the nighttime, respectively), and blood samples were collected and processed to measure the melatonin levels by radioimmunoassay as described below. ii. An effect of the injection procedure per se (vehicle injection and handling) to pregnant rats on the serum corticosterone levels was assessed. The pregnant dams (n 5 9) were maintained in LL from the beginning of pregnancy until the gestational day 17. Then, one group of pregnant rats (n 5 5) received a single vehicle injection (i.p.) between ZT11:30 and ZT12:30, that is, at approximately the same time as melatonin or vehicle was administered in Experiments 1 and 2) and was sacrificed 30 min afterward. The other group of dams (n 5 4) was left untreated and was sacrificed at the same time as the rats of the first group. The collected blood samples were processed to determine the levels of corticosterone by radioimmunoassay as described below.
Experimental Protocol
Locomotor Activity Monitoring
Adult female rats were inspected by vaginal smears every day to detect the phases of their estrous cycles. On the night of proestrus, the dams were mated with males, and the day when the vaginal smears were found to be sperm-positive was considered embryonal day (E) 0. From E0 until the end of the experiment, the pregnant rats were maintained in LL
Pregnant rats were maintained individually in cages equipped with infrared movement detectors attached above the center of the top of the cage, thus enabling detection of their locomotor activity across the entire cage. The activity was detected every minute using a circadian activity monitoring system (Dr. H.M. Cooper, INSERM, France). Developmental Neurobiology
134
Houdek et al.
Double-plotted actograms were generated for visualization of data.
Pinealectomy Pinealectomy (PINX) of adult female rats for Experiment 2 was performed using deep anesthesia (ketamine 85 mg/kg b.w., xylazine 10 mg/kg b.w., and atropine 0.1 mg/kg b.w.). The hair on the head was shaved, the skin was incised, and a hole in the skull was bored at the lambda position by a micro drill. The pineal gland was removed by fine forceps. The surgical wound was closed with the bone fragment being returned back, and the skin was sutured. The threads were removed 2 weeks after the surgery. The recovery period took approximately 1 month until the sutures were no longer visible. Thereafter, mating with male rats was realized as described above. The completeness of the pineal removal was assessed after the Experiment 2 was terminated via (i) visual inspection post mortem and (ii) detection of the plasma melatonin levels using radioimmunoassay. Before the blood sampling, the PINX dams were exposed to LD12:12 for 10 days and then they were sacrificed, and the blood samples were collected around midnight, that is, between ZT17 and ZT19, when the highest endogenous melatonin levels were expected.
In situ Hybridization The pup heads, which had been maintained at 280 C, were sectioned in a cryostat into five series of 12-mm-thick slices throughout the entire rostrocaudal extent of the SCN, and the sections were organized in an alternating order on slides. The sections were further processed for in situ hybridization to determine the profiles of c-fos mRNA and Avp hnRNA in the SCN. The cDNA fragments of rat c-fos (1160 bp; corresponding to nucleotides 141–1300 of the sequence in GenBank accession number X06769) and Avp (506 bp; identical to nucleotides 796–1302 of the intronic sequence in GenBank accession no. X01637) were used as templates for in vitro transcription of complementary RNA probes. The c-fos fragment–containing vector was generously donated by Dr. Tom Curran (Children’s Hospital of Philadelphia, Pennsylvania). The Avp cDNA was cloned in our laboratory. Probes were labeled using 35S-UTP, and the in situ hybridizations were performed as previously described (El-Hennamy et al., 2008). The sections were hybridized for 20 h at 60 C. Following a posthybridization wash, the sections were dehydrated in ethanol and dried. Finally, the slides were exposed to BIOMAX MR film (Kodak) for 10–14 days and developed using the ADEFOMIX-S developer and ADEFOFIX fixer (ADEFO-CHEMIE Gmbh, Germany). Brain sections from all experimental groups were processed simultaneously under identical conditions. Autoradiographs of the sections were analyzed using an image analysis system (Image Pro, Olympus, New Hyde Park, NY) and the relative optical density (OD) of the speDevelopmental Neurobiology
cific hybridization signal was detected. For each animal, the mRNA or hnRNA was quantified bilaterally, at the midcaudal SCN section containing the strongest hybridization signal. Each measurement was corrected for nonspecific background by subtracting the OD values from the same adjacent area in the hypothalamus. The background signal of that adjacent area served as an internal standard; it was consistently low and did not exhibit marked changes with the time of day. Finally, the slides were counterstained with cresyl violet to check for the presence and midcaudal position of the SCN in each section. The OD for each animal was calculated as a mean of values for the left and right SCN.
RNA Isolation and Real-Time RT-PCR Total RNA was isolated using an RNeasy Mini kit (Qiagen, Valencia) according to the manufacturer’s instructions. RNA concentrations were determined using spectrophotometry at 260 nm, and the RNA quality was assessed by electrophoresis on a 1.5% agarose gel. Moreover, the integrity of randomly selected samples of total RNA was tested using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara). The real-time RT-PCR method, which was used to detect mRNA levels of Per1, Per2, Rev-erba, Bmal1, Dbp, and Wee1, was described previously (Polidarova et al., 2011; Sladek et al., 2007a). Briefly, 1 mg of total RNA was reverse transcribed using a SuperScript VILO cDNA Synthesis kit (Life Technologies, USA). Diluted cDNA was then amplified on a LightCycler (Roche, Basel, Switzerland) using a Express SYBR Green qPCR Supermix Universal kit (Life Technologies, USA) and the corresponding primers. Relative quantification was achieved using a standard curve and subsequently normalizing the measured gene expression to b2-microglobulin. The housekeeping gene b2-microglobulin has been used for normalization previously (Sladek et al., 2007a; Sladek et al., 2007b), and its expression was stable throughout the 24 h and did not vary markedly between the analyzed tissues.
Radioimmunoassay The collected blood samples (5–10 mL per animal) were left to clot for 45 min at room temperature while protected from light. After centrifugation (1800 g, 30 min), the serum was separated and stored at 220 C until the RIA was performed. For assessment of the serum melatonin and corticosterone levels, the Melatonin Research RIA kit (LDN, Germany) and Rat Corticosterone H3 kit (MP Biomedicals, CA), respectively were used according to the guidelines of the manufacturers.
Statistical Analysis For analysis of the gene expression profiles, data were tested for rhythmic and nonrhythmic expression by fitting to two alternative regression models, that is, either a horizontal straight line (null hypothesis) or a single cosine curve (alternative hypothesis), which was defined by the
Melatonin Entrains Fetal Circadian Clocks equation Y 5 mesor 1 [amplitude 3 cos(2 3 p 3 (X-acrophase)/wavelength)] with a constant wavelength of 24 h. The extra sum-of-squares F test was used for comparison, and when the p value exceeded 0.05, the cosine curve parameters were calculated. The amplitude (i.e., the difference between the peak or trough and the mean value of a cosine curve), acrophase (i.e., the phase angle of the peak of a cosine curve), mesor (i.e., the average value around which the variable oscillates), and coefficient of determination R2 (i.e., goodness of fit) were detected. The leastsquares regression method implemented in the Prism 5 software (GraphPad, La Jolla) was applied. The mean daily levels of melatonin were compared between two experimental groups of rats by Student’s t test, and the results of the Student’s t-test are expressed as tx values (with x in the lower index representing degrees of freedom) and as p values (level of significance), with p < 0.05 required for significance.
RESULTS LL Abolishes the Rhythms in Behavior, Plasma Melatonin Levels and Gene Expression in the SCN of Pregnant Rats, As Well As the Rhythms in Gene Expression in the SCN of Newborn Pups As expected, exposure to LL abolished the circadian rhythm in behavior in all pregnant rats. At the beginning, the locomotor activity became free running with a long period for approximately the first 2 weeks, and later, during the last week of pregnancy, all rats became completely arrhythmic [Fig. 1(A)]. To detect an effect of the 3-week exposure to LL on endogenous melatonin levels, melatonin was assayed in the serum of rats at three different time points, namely at the times corresponding to the previous daytime (ZT8) and the previous nighttime (ZT16 and 20). The daily melatonin levels were compared with those in control animals that remained in LD12:12 and sampled at the corresponding time points [Fig. 1(B)]. The daily melatonin levels expressed as a mean of the three samples during the 24 h were significantly lower in the animals maintained in LL compared with those maintained in LD12:12 (t-test, F12 5 9.894, p < 0.001). The data confirmed that in the pregnant rats, the endogenous pineal melatonin production was significantly reduced throughout the 24-h period. Additionally, in the SCN of the pregnant rats maintained in LL, no significant rhythms in c-fos and Avp gene expression were detected [Fig. 1(C)]. To determine whether the lack of the maternal rhythmicity also affects the fetal SCN, the daily profiles of c-fos and
135
Avp gene expression were determined in the SCN of newborn pups at P1 [Fig. 1(C)]. Cosinor analysis did not reveal significant circadian rhythms in the expression of any of these genes (c-fos: p 5 0.4473; Avp: p 5 0.3343). These results demonstrate that due to exposure to the LL for 3 weeks, the maternal SCN clock did not drive the circadian rhythm in locomotor activity, pineal melatonin production and the SCN c-fos and Avp gene expression. The SCN clocks of the newborn pups did not exhibit circadian rhythms in c-fos and Avp gene expression.
The Injection Procedure Per Se May Affect the Behavior and Corticosterone Levels in Some But Not All Pregnant Rats Maintained in LL Inspection of the individual locomotor activity records revealed that repeated injections (vehicle and/or melatonin) to pregnant rats at E17–21 affected the behavior of some of the animals. To analyze whether the injections entrained locomotor activity, the actograms for the interval of the entire pregnancy and v2 periodograms for the interval of the last 5 days when injections were administered to each pregnant rat were evaluated. In most of the animals, the behavior was not affected by administration of vehicle [Fig. 2(A,C)] or by melatonin injections [Fig. 2(B,D)], and the rats exhibited arrhythmic behavior that was undisturbed by the injections. In a minority of the rats, the effect of the injections was manifested by slightly increased activity immediately after each injection (2 out of 8 vehicle-treated and in 3 out of 9 melatonin-treated animals). Only one out of eight vehicle-treated and one out of nine melatonin-treated animals exhibited a significant 24-h period in locomotor activity during the 5 days of injections. Obviously, the proportion of animals affected by the injections was approximately the same in the vehicle- and melatonin-treated groups. Therefore, the repeated melatonin injections did not significantly entrain the behavioral rhythms of the pregnant rats. The vehicle contained ethanolic saline because of melatonin solubility, and thus, we could not exclude a possibility that these injections were slightly painful to the pregnant rats. Therefore, in a separate experiment, we tested whether the i.p. injections represented a mild stressor to the animals. In untreated controls (n 5 4), the serum corticosterone levels measured at ZT11:30 were 490.0 6 62.8 ng/mL. In the vehicle-injected animals (n 5 5), the serum corticosterone levels measured 30 min after the injection Developmental Neurobiology
136
Houdek et al.
Figure 1 Effect of constant light on locomotor activity, melatonin levels in blood and gene expression in the SCN. (A) A representative double-plotted actogram of a pregnant rat maintained in constant light throughout the entire gestation, that is, from embryonic day (E)1 until E22. (B) The melatonin levels in the serum (pg/mL) of pregnant rats maintained in light/dark regime (LD12:12) or constant light (LL) and sampled at 3 time points a day. ZT8 represents the day, ZT16 represents the first half of night, and ZT20 represents the second half of night. (C) Daily profiles of c-fos and Avp expression in the SCN of mothers after delivery that were maintained in LL throughout entire pregnancy (upper panel) and in their 1-day-old pups (lower panel). The data are expressed as the relative OD measured in the individual SCN. The data were tested by cosinor analysis for rhythmic expression, and the horizontal straight lines demonstrate the absence of significant circadian rhythms. Time is expressed in Zeitgeber time (ZT). For more details, see Materials and Methods.
at ZT11 varied among the pregnant rats; in three animals, the levels were approximately the same as in controls (528.7 6 17.9 ng/mL), but in two animals, the levels were remarkably increased (2407.7 and Developmental Neurobiology
1278.7 ng/mL). The data suggested that at least in some (but not in all) pregnant rats, the i.p. injections of vehicle/melatonin affected the behavior and basal serum corticosterone levels.
Melatonin Entrains Fetal Circadian Clocks
137
Figure 2 Analysis of locomotor activity in pregnant rats maintained in constant light and injected with vehicle (A,C) and melatonin (B,D) during the last 5 days of gestation. Representative doubleplotted actograms demonstrate the locomotor activity of a rat treated with vehicle (A) and of a rat treated with melatonin (B). The vertical bars depict the days and time of the treatment. Individual v2 periodograms for the same animals as in A and B were calculated from the 5 days of injections with vehicle (C) and melatonin (D). The diagonal line depicts the threshold for significant period of the rhythm (p < 0.001).
Repeated Injections of Melatonin to Pregnant Rats Maintained in LL Affect the Gene Expression in the Fetal SCN Daily profiles of c-fos and Avp expression were determined in the SCN of 1-day-old pups that were born to mothers maintained in LL and injected with melatonin/vehicle as described in Experiment 1. The data were fitted with cosine curves (Fig. 3). The results of the cosinor analysis, namely the acrophases, amplitudes, mesors, R2, and p values, are summarized in Table 1. The c-fos and Avp expression profiles exhibited significant circadian rhythms in the SCN of the 1day-old pups born to mothers maintained in LL and
injected with vehicle or melatonin. However, the phases of the rhythms in the c-fos expression (Fig. 3; upper part) differed significantly between the melatonin- and vehicle-treated groups (t-test, F124 5 19.544, p < 0.001). Similarly, the phases of Avp expression profiles (Fig. 3; lower part) also significantly differed in between the pups born to the mothers injected with vehicle and those born to mothers injected with melatonin, (t-test, F124 5 22.559, p < 0.01). Therefore, the repeated vehicle injections to pregnant rats were able to synchronize the c-fos and Avp expression in the fetal SCN and the rhythms peaked approximately 6 and 10 h after the injections, respectively. The repeated melatonin injections synchronized the c-fos and Avp expressions and the rhythms Developmental Neurobiology
138
Houdek et al.
Figure 3 Daily profiles of c-fos (upper panel) and Avp (lower panel) expression in the SCN of 1-day-old pups born to mothers that were maintained in constant light throughout the entire gestation and were treated with vehicle (left) and melatonin (right). The data are expressed as values of the relative OD measured in the individual SCN. The data were fit with cosine curves to determine the presence or absence of circadian rhythms. The sinusoid curves demonstrated significant circadian rhythms. The parameters of the rhythms are summarized in Table 1. Time is expressed in ZT, and the apostrophe in ZT120 means that it is 24 h after the ZT12 when the sampling of the pups started.
peaked approximately 12 and 13 h after the injection, respectively.
Repeated Injections of Melatonin to Pregnant PINX Rats Entrain Gene Expression in the Fetal SCN Only dams whose pineal glands were completely removed and their nighttime serum melatonin levels were reduced to a minimum (23.55 6 3.13 pg/mL) were used in the study. Similarly to intact rats, the PINX pregnant rats maintained in LL also did not exhibit a circadian rhythm in locomotor activity, and the injection procedure had only a minor effect on the behavior. During the 5 days of injections, periodograms detected a significant 24-h period in 0 out of 5 melatonin-treated and 1 out of 6 vehicle-treated PINX pregnant rats. Daily profiles of c-fos and Avp expression were detected in the SCN of 1-day-old pups born to the PINX mothers maintained in LL as described in Experiment 2. The data were fitted with cosine curves Developmental Neurobiology
(Fig. 4), and results of the cosinor analysis are summarized in Table 1. In the SCN of 1-day-old pups born to PINX mothers repeatedly injected with vehicle, the c-fos expression (Fig. 4; upper part) was completely desynchronized. However, in the pups born to mothers that were repeatedly injected with melatonin, the c-fos expression exhibited a significant circadian rhythm. For the Avp expression (Fig. 4; lower part), only a weak rhythm was present in the SCN of the pups born to mothers repeatedly injected with vehicle. In contrast, after repeated melatonin injections, a circadian rhythm of Avp expression with a higher amplitude and R2 value was detected. Moreover, the acrophases of the Avp expression rhythms in the SCN of the pups born to vehicle- and melatonin-treated mothers were significantly different (t-test, F124 5 29.334, p < 0.001). Therefore, in the SCN of the pups born to PINX mothers, the vehicle injections were not able to synchronize the c-fos expression, and only weakly synchronized the Avp expression in the fetal SCN; the Avp expression rhythm peaked approximately
0.063 6 0.003 – – 0.008 0.7611 0.022 6 0.001 23.7 6 0.6 0.014 6 0.002 0.627
