Photoperiod-driven concurrent changes in hypothalamic and brainstem transcription of sleep and immune genes in migratory redheaded bunting

Abstract

The molecular regulation of sleep in avian migrants is still obscure. We thus investigated this in migratory redheaded buntings, where four life-history states (LHS; i.e. non-migratory, pre-migratory, migratory and refractory states) were induced. There was increased night-time activity (i.e. Zugunruhe) during the migratory state with reduced daytime activity. The recordings of the sleep–wake cycle in buntings showed increased night-time active wakefulness coupled with drastically reduced front and back sleep during migratory phase. Interestingly, we found the buntings to feed and drink even after lights-off during migration. Gene expression studies revealed increased hypothalamic expression of glucocorticoid receptor (nr3c1), and pro-inflammatory cytokines (il1b and il6) in pre-migratory and migratory states, respectively, whereas in brainstem Ca²⁺/calmodulin-dependent protein kinase 2 (camk2) was upregulated during the migratory state. This suggested a heightened pro-inflammatory state during migration which is a feature of chronic sleep loss, and a possible role of Ca²⁺ signalling in promoting wakefulness. In both the hypothalamus and brainstem, the expression of melatonin receptors (mel1a and mel1b) was increased in the pre-migratory state, and growth hormone-releasing hormone (ghrh, known to induce sleep) was reduced during the migratory state. The current results demonstrate key molecules involved in the regulation of sleep–wake cycle across LHS in migratory songbirds.

1. Introduction

Sleep is ubiquitous in the animal kingdom, albeit present in radically different amounts and patterns. It is defined as a highly organized phenomenon associated with decreased muscular activity and response to environmental stimuli. A vast array of literature suggests that there is a cross-talk between sleep and the immune system [1]. Krueger et al. [2] were one of the pioneers who first demonstrated this association, and established Interleukin 1 (IL-1) as a key player in sleep regulation. Since then, the race to discover the encephalic sleep centres and sleep-regulatory molecules has begun.

It is now known that interactions between the neurons of the brainstem, midbrain, hypothalamus, thalamus and cortex drive sleep–wake behaviour and physiology [3]; regulatory molecules synthesized in these centres modulate arousal and sleep. IL-1b is one such molecule, involved in sleep promotion. IL-1b administration increases the non-rapid eye movement (NREM) sleep, while its reduction causes sleep disturbances [4]. Similarly, another pro-inflammatory cytokine IL-6 was found to be associated with sleep disturbances in major depressive disorder patients [5]. The expression of these inflammatory cytokines is in turn regulated by nuclear factor-kB (NF-kB), which is activated in children with obstructive sleep apnea syndrome [6]. Even in Drosophila, immune responses gated by NF-kB, increased sleep in the flies [7]. Nitric oxide (NO) further plays a key role in sleep homeostasis; the ratio of NO/NOS2 (nitric oxide synthase 2) maintains adequate REM sleep [8]. Another discrete pathway involved in sleep–wake regulation involves melatonin. Melatonin essentially reflects the nighttime and acts through its receptors Mel1a, Mel1b and Mel1c which have been found in different brain areas of vertebrates [9]. Sleep deprivation has also been associated with an elevated HPA axis, which causes increased glucocorticoid levels, creating a vicious cycle [10]. Both REM and NREM sleep is promoted by the growth hormone-releasing hormone (GHRH) [11], and the reduction of hypothalamic GHRH levels has been associated with elevated wakefulness and reduced slow-wave sleep (SWS). In rats, brain-derived neurotrophic factor (BDNF), which promotes neuronal survival and growth, influenced the sleep–wake activity and homeostatic regulation of REM sleep [12]. This interaction is two-way as sleep disturbance can also lead to decreased BDNF levels [13]. Wakefulness on the other hand is promoted by Ca²⁺ in the dorsal raphe nucleus, and its modulation is in turn maintained by calmodulin-dependent kinase 2 (CaMK2) [14,15].

In birds, sleep has been well characterized by behavioural and electroencephalogram (EEG) recordings. Based on behavioural states, Fuchs et al. [16], categorized awake behaviours into active wakefulness, alert wakefulness and drowsiness, whereas, sleep behaviours comprised front sleep. Back sleep and unilateral eye closure in Swainson's thrush (Catharus ustulatus). This was further modified to classify the sleep postures in redheaded buntings (Emberiza bruniceps) [17]. In both studies, the authors found an increase in night-time alert wakefulness and a decrease in front and back sleep in the migratory state [16,17]. Similar results have been obtained in EEG studies in captive white-crowned sparrows (Zonotrichia leucophrys gambelii), [18] and thrushes [19]. In sparrows, EEG recordings revealed SWS and REM sleep, and in thrushes, unihemispheric sleep (UHS) was also recorded [18,19]. Jones et al. [20], did a similar study on sparrows and found that sleep deprivation caused an increase in slow-wave activity in both SWS and nREM sleep. Further, in the migratory season, they found an upregulation of glucose transporter (glut1) and heat-shock protein (hsp70, 90) mRNA expression in the brain [21].

While much research has been done to understand the molecular regulation of the sleep–wake cycle in mammals, there is insufficient literature on birds. Our experimental non-model species, redheaded buntings (Emberiza bruniceps) provide a unique opportunity to understand the regulation of the sleep–wake cycle in migratory birds across their life-history states (LHS). We selected redheaded buntings as our experimental animal as it is photoperiodic and shows distinct LHS under captive conditions [22]. The buntings are obligate passerine songbird migrants with predictable annual migrations between breeding (Central Asia and Southeast Europe) and overwintering (mainly India) areas [23]. They arrive in India during October–November to overwinter and start their return journey to their breeding grounds (spring or vernal migration) from late March to early April, when the length of the day (sunrise to sunset) is approximately 12.5 h [23]. The development of vernal migratory phenotype (body fattening, weight gain and nocturnal Zugunruhe) and departure timing can be reproduced under semi-natural photoperiod and ambient temperature in buntings [24]. Similarly, photostimulated vernal migratory and breeding states can be induced in buntings exposed to artificial stimulatory photoperiodic conditions [25–27]. Thus, we induced four different LHS by changing the photoperiodic conditions, and monitored the activity, sleep–wake cycle, general physiology, and measured the mRNA expression of candidate genes involved in sleep homeostasis.

2. Material and methods

(a) . Animals and maintenance

The experiment was carried out on male redheaded buntings (n = 40), which are Palearctic-Indian migrants that overwinter in India. The birds were procured from the overwintering flocks in mid-February 2019. Thereafter, these birds were brought to the laboratory and kept in an outdoor aviary (size = 3.0 × 2.5 × 2.5 m) under the natural day length (approximately11L : 14D; L = light, D = dark). After a week's acclimation, the birds were transferred to an indoor aviary (size = 2.2 × 1.8 × 2.8 m) under either a short photoperiod (n = 30, 8L : 16D) to maintain a photosensitive state, or they were transferred to a long photoperiod (n = 10; 16L : 8D) for 40 weeks to induce photorefractory state [28]. Food and water were given to the birds throughout the 24 h (ad libitum). The overall maintenance and general care were similar, as described in Singh et al. [29].

(b) . Experiment

After the completion of 40 weeks, male redheaded buntings were transferred to separate activity cages either under short photoperiod (n = 30; 8L : 16D) or long photoperiod (n = 10; 16L : 8D). These birds were individually housed in cages (one bird per cage, size = 60 × 45 × 35 cm) that were placed in an independent wooden box (size = 75 × 50 × 70 cm) during the experiment. The cages were equipped with two perches, an IR motion sensor (Haustier PI-meldor; C and K Systems [Intellisense XJ-413T] Conrad Electronic, Hirschau, Germany) [30] and an IR night-vision dome-shaped camera (model no- CP-GTC-D24L2-V3). The Chronobiology Kit program from Stanford Software System, USA was used for the calculation and analysis of the activity counts of the birds. The night-vision camera was wall-mounted at the corner of the box to cover the whole cage and monitor the sleep behaviours of the birds throughout the experiment. The birds were acclimated to this facility for one week before starting the experiment. Throughout the experiment, the temperature (23 ± 2°C) and light intensity (L = 200 ± 10; D < 0.1 l × ) were kept constant.

Before the start of the experiment, the birds were weighed on a top-pan electronic balance to an accuracy of 0.01 g, and they weighed 23.17 ± 0.2 (mean ± s.e.m.) g on average with no subcutaneous fat deposit. Additionally, the birds were laparotomized to check the testes’ size. All the testes were small in size with an average volume of 1.10 ± 0.18 mm³. Laparotomy is a surgical procedure to assess the gonadal volume, that we routinely perform in our laboratory under local anaesthesia [31,32].

The experiment was carried out in the following manner: The birds under short photoperiod (8L : 16D, n = 30) were divided into three groups (i.e. group 1, 2 and 3), whereas the birds under long photoperiod (16L : 8D, n = 10) constituted the last group (i.e. group 4); these discrete groups were created to represent different LHS of the birds. Group 1 (non-migratory, N = 10) was maintained under short photoperiod (8L : 16D) for one more week. Group 2 (pre-migratory, n = 10) and group 3 (migratory, n = 10) were transferred from short to long photoperiod (13L : 11D) for one week or till the development of seven nights of Zugunruhe (i.e. night-time restlessness, a marker of migratory state) respectively. Group 4 (refractory, n = 10) was continued under long days (16L : 8D) for a week. After the completion of the experiment, the birds were sampled.

(c) . Data collection and sampling

On day 1 of the experiment, the body mass and fat score of individual buntings were measured. On days 3 and 4, food intake was quantified, and sleep recording data was taken. Videos were recorded for two consecutive days for each bird across the four LHS. All the video recordings were analysed manually and converted into the binary format of ‘0’ and ‘1’ based on the absence or presence of a particular behaviour (electronic supplementary material, table S1) [17]. Further details of the data collection are given in the electronic supplementary material and methods section.

In all groups, birds (N = 5) were sacrificed by decapitation at 1 h after lights-OFF (i.e. Hour 09 for non-migratory; Hour 14 for pre-migratory, and migratory; Hour 17 for photorefractory, where Hour 0 = lights-on). We chose decapitation, which is a fast and unanticipated procedure, over anaesthesia usage to prevent its probable effects on the mRNA expression [33]. The brains were harvested and stored at −80°C till further analysis.

(d) . Measurement of mRNA expression of candidate genes

To perform the mRNA expression study of our candidate genes, we meticulously excised both the hypothalamus and brainstem out of the brain [26,34]. Briefly, after decapitation, the head was put on ice and the tissues were harvested in a cold room (4°C). The skull was opened carefully, and the meninges were cut open. Then, the brain was placed ventral side up, and two coronal incisions were made to separate the diencephalon, from which the hypothalamus was excised roughly in the shape of an inverted V (Δ) by longitudinal incisions placed at a 45° angle on either side of the third ventricle [34]. Thereafter, the posterior stalk-like brainstem was carefully cut out and snap-frozen in dry ice (electronic supplementary material, figure S2). We wanted to assess the differential effects of LHS on sleep-immune health; therefore, we measured the expression of genes involved in the cytokine (il1b, il6 and nfkb) and RNS (nos2) pathway, genes coding for melatonin (mel1a and mel1b) and glucocorticoid receptors (GRs) (nr3c1), and genes directly involved in sleep (ghrh and bdnf) and arousal (camk2) promoting pathway. We used the partial gene sequences from previous studies [26,35], whereas degenerate primers were used to clone sequence of genes not available to us (see electronic supplementary material and methods for details).

For qPCR, specific primers were designed using the online software Primer Quest program (https://eu.idtdna.com/PrimerQuest). All the samples were analysed in duplicates using SYBR green (Applied Biosystems, A25742) run on Agilent real-time PCR system. In each PCR plate, beta-actin was used as a housekeeping gene or reference gene. The relative mRNA expression of each candidate gene was determined by the ΔΔCt method [36]. The values of cycle threshold (Ct) were determined by fluorescence surpassing the background noise which will give us ΔCt value (Ct [gene of interest] – Ct [reference gene]). The Ct values were then normalized against the Ct value of pooled cDNA from all samples; this gave ΔΔCt. The negative value of ΔΔCt powered to 2 (2−ΔΔCt) was plotted.

(e) . Statistics

All statistical analyses were performed using GraphPad PRISM software (version 5.0), San Diego, CA, USA. We used two-way repeated measures of analysis of variance (two-way RM ANOVA) to compare the activity counts/h, and sleep behaviours among the groups, where factor 1 was ‘Time’, and factor 2 was ‘LHS’. Whereas, to assess the effect of LHS on total activity/day, daytime activity/h, night-time activity/h, physiological parameters and gene expressions we performed one-way ANOVA followed by Bonferroni post-test. Effect sizes were calculated (Eta squared (η²) for one-way ANOVA, and partial Eta squared (η²_p) for two-way ANOVA; 0.01 = small effect, 0.06 = medium effect, ≥ 0.14 = large effect, Tomczak and Tomczak, 2014) [37] to evaluate the magnitude of the responses. Additionally, Pearson's correlation was performed between gene expression values and total sleep after lights-off (L_OFF0-1), prior to the sampling (i.e. sleep duration 1 h before sampling) to assess its link with the variation in transcript level. For significance, the alpha was set at 0.05.

3. Results

(a) . Activity behaviour and physiology

Buntings in non-migratory (NM) and pre-migratory (PM) states showed a diurnal pattern of activity rest, with activity being confined to the 8 h and 13 h light periods, respectively. In the migratory (MIG) state, there was a phase transition, where the birds exhibited intense night-time activity (i.e. Zugunruhe), with reduced daytime activity. In refractory (REF) state, the activity was again diurnal and limited to the 16 h light period (figure 1). Two-way RM ANOVA revealed a significant effect of time (F_23,805 = 12.08, p < 0.0001, η²_p = 0.156), LHS (F_{3, 805} = 11.86, p < 0.0001, η²_p = 0.282) and time × LHS interaction (F_69,805 = 35.69, p < 0.0001, η²_p = 0.626; figure 1a–d) on hourly activity counts. There was a significant difference in total activity counts (F_3,35 = 11.87, p < 0.0001, η² = 0.537), activity in the daytime (F_3,35 = 6.781, p < 0.0010, η² = 0.461), activity in nighttime (F_3,35 = 55.11, p < 0.0001, η² = 0.808). The total and night-time activity was high whereas daytime activity was low in MIG state. (p < 0.05, Bonferroni post-test, figure 1e–g).

Figure 1.

Activity data and physiological changes. (a–d) Diurnal activity heatmap of redheaded buntings in four LHS (i.e. (a) non-migratory, NM; (b) pre-migratory, PM; (c) migratory, MIG; (d) refractory, REF) with overlayed mean ± (s.e.m.) values of 24 h activity profile. (e–k) Mean ± (s.e.m.) of total activity counts (e), activity counts in day (f) and night (g), body mass (h), food intake (i), body fattening (j) and testis volume (k) in the same four LHS of redheaded buntings. Different letters on bars represent a significant difference. Alpha was set at 0.05.

There was a significant difference in body mass (F_3,36 = 21.12, p < 0.0001, η² = 0. 638), food intake (F_3,36 = 9.80, p < 0.0001, η² = 0.450), fat score (F_3,36 = 58.08, p < 0.0001, η² = 0.829) and testis volume (F_3,36 = 31.61, p < 0.0001, η² = 0.725 one-way ANOVA). All four physiological parameters were found to be increased under MIG state (p < 0.05, Bonferroni post-test; figure 1h–k).

(b) . Active and sleep behaviours

In the four different physiological states, redheaded buntings showed differential active (i.e. active wakefulness, alert wakefulness, quiet wakefulness, feeding and drinking) and sleep (i.e. drowsiness, front sleep, back sleep and UHS) behaviour, which were distributed throughout the 24 h. In non-migratory, pre-migratory and refractory states active wakefulness (ACW), alert wakefulness (ALW) and quiet wakefulness (QW) were distributed during day hours whereas in the migratory state these behaviours were distinctly high at the nighttime.

Two-way RM ANOVA revealed a significant effect of time (F_23,828 = 34.15, p < 0.0001, η²_p = 0.440), LHS (F_3,828 = 74.69, p < 0.0001, η²_p = 0.516) and time × LHS interaction (F_69,828 = 33.19, p < 0.0001, η²_p = 0.696) (figure 2a–d) on active wakefulness. Likewise, we found a significant effect of time (ALW: F_23,828 = 24.82, p < 0.0001, η²_p = 0.270; QW: F_23,828 = 31.56, p < 0.0001, η²_p = 0.417), LHS (ALW: F_3,828 = 13.05, p < 0.0001, η²_p = 0.580; QW: F_3,828 = 5.059, p = 0.0050, η²_p = 0.072) and time × LHS interaction (ALW: F_69,828 = 6.834, p < 0.0001, η²_p = 0.696; QW: F_69,828 = 14.61, p < 0.0001, η²_p = 0.498) on both alert and quiet wakefulness. Interestingly we observed prolonged night-time feeding in the migratory state as compared to other states. Two-way RM ANOVA revealed a significant effect of time (F_23,828 = 40.85, p < 0.0001, η²_p = 0.449), LHS (F_3,828 = 4.006, p = 0.0147, η²_p = 0.086) and time × LHS interaction (F_69,828 = 7.548, p < 0.0001, η²_p = 0.312) on feeding behaviour. On the other hand, we found the effect of time (F_23,828 = 4.866, p < 0.0001, η²_p = 0.089) and time × LHS interaction (F_{69, 828} = 1.716, p = 0.0004, η²_p = 0.083) on drinking behaviour (figure 2a–d).

Figure 2.

Active behaviours. Polar plots of the 24 h profile of active wakefulness, alert wakefulness, quiet wakefulness, feeding and drinking in four LHS (i.e. non-migratory (a), pre-migratory (b), migratory (c) and refractory (d)) of redheaded buntings. The unit of the radial axis is ‘minutes’ (range: 0–60 min h⁻¹) and the distance between the spokes is 1 h.

In the case of sleep behaviours, drowsiness was prevalent during the daytime in all states except migratory, where it was also visibly high during the first quarter of the night. Two-way RM ANOVA showed a significant effect of time (F_23,828 = 16.26, p < 0.0001, η²_p = 0.269), LHS (F_3,828 = 4.757, p = 0.0068, η²_p = 0.068) and time × LHS interaction (F_69,828 = 3.612, p < 0.0001, η²_p = 0.197) on drowsiness. In non-migratory, pre-migratory and refractory states, both front (FS) and back sleep (BS) were found to be distributed throughout the night hours but in the migratory state, they were distributed during the day hours and were visibly absent from the night-time. Two-way RM ANOVA revealed the effect of time (FS: F_23,828 = 6.972, p < 0.0001, η²_p = 0.123; BS: F_23,828 = 101.7, p < 0.0001, η²_p = 0.687), LHS (FS: F_3,828 = 3.160, p = 0.0363, η²_p = 0.068; BS: F_3,828 = 36.85, p = 0.0363, η²_p = 0.405) and time × LHS interaction (FS: F_69,828 = 7.767, p < 0.0001, η²_p = 0.319; BS: F_69,828 = 21.87, p < 0.0001, η²_p = 0.587) (figure 3a–d) on both front and back sleep. UHS was distributed in small bouts during the night-time in non-migratory, pre-migratory and refractory states, whereas, in the migratory state, it was present throughout the 24 h. We found the effect of only LHS (F_3,828 = 5.262, p = 0.0041 η²_p = 0.042, two-way RM ANOVA) on UHS.

Figure 3.

Sleep behaviours. (a–d) Polar plots of 24 h profile of drowsiness, front sleep, back sleep and UHS in four LHS (i.e. non-migratory, NM (a); pre-migratory, PM (b); migratory, MIG (c); refractory, REF (d)) of redheaded buntings. The unit of the radial axis is ‘minutes’ (range: 0–60 min h⁻¹) and the distance between the spokes is 1 h. (e–g) Mean ± (s.e.m.) of total sleep in 24 h (e), daytime sleep (f), night-time sleep (g) is plotted in four LHS. Different letters on bars indicate a significant difference between the LHS (Bonferroni post-test following one-way ANOVA). For statistical significance, alpha was set at 0.05.

We also measured the total sleep (FS + BS + UHS) in 24 h, daytime sleep and night-time sleep in all four LHS. We found a significant difference in total sleep among the groups as revealed by one-way ANOVA (F_3,36 = 93.13, p < 0.0001, η² = 0.869). In particular, we found a significant reduction in total sleep in the migratory state (p < 0.05; Bonferroni post-test, figure 3e). The daytime sleep (F_3,36 = 32.29, p < 0.0001, η² = 0.718) and night-time sleep (F_3,36 = 331.3, p < 0.0001, η² = 0.966, one-way ANOVA) was also found to be significantly different, where daytime sleep was higher, while night-time sleep was lower in the migratory state, compared to the other states (p < 0.05, Bonferroni post-test; figure 3f–g). We also found a significant difference in total sleep (L_OFF0-1) prior to sampling (F_3,36 = 26.80, p < 0.0001, η² = 0.834), with a reduction of sleep in the migratory state (p < 0.05, Bonferroni post-test, electronic supplementary material, table S3.2).

The total, daytime and night-time feeding and drinking were measured in the four LHS, and we found a significant difference in night-time feeding (F_3,36 = 9.998, p < 0.0001, η² = 0.402) and drinking (F_3,36 = 11.33, p < 0.0001, η² = 0.262, one-way ANOVA) among the groups, where both were increased in the migratory state (p < 0.05, Bonferroni post-test; electronic supplementary material, figure S1c, f). However, we didn't find any difference in total and daytime feeding and drinking between the LHS.

(c) . Gene expressions

We measured the mRNA expression of genes involved in cytokine (il1b, il6 and nfkb) and RNS (nos2) pathways. In the hypothalamus, we found a significant effect of LHS on il1b (F_3,16 = 12.37, p = 0.0002, η² = 0.699; figure 4a), il6 (F_3,16 = 5.08, p = 0.01, η² = 0.488 figure 4c) and nos2 (F_3,16 = 3.25, p = 0.04, η² = 0. 379, one-way ANOVA; figure 4g). Whereas, in the brainstem, we found a significant effect on nfkb (F_3,16 = 3.35, p = 0.04, η² = 0.386; figure 4f), and nos2 (F_3,16 = 3.95, p = 0.02, η² = 0.426, one-way ANOVA; figure 4h). The il1b and il6 mRNA levels were found to be elevated in MIG (p < 0.05, Bonferroni post-test; figure 4a,c) state in the hypothalamus. In the brainstem, levels of nkfb in PM > MIG, and nos2 in NM > PM (p < 0.05, Bonferroni post-test; figure 4f,h).

Figure 4.

mRNA expression of cytokine and RNS pathway genes in hypothalamus (a,c,e,g) and brainstem (b,d,f,h). Mean ± (s.e.m.) mRNA levels of il1b (a,b), il6 (c,d), nfkb (e,f) and nos2 (g,h) in redheaded bunting in four LHS (i.e. non-migratory, NM; pre-migratory, PM; migratory, MIG; refractory, REF). Different letters on bars indicate a significant difference between the LHS (Bonferroni post-test following one-way ANOVA). For statistical significance, alpha was set at 0.05.

Next, we measured mRNA expressions of melatonin (mel1a and mel1b) and GR (nr3c1) genes. In the hypothalamus, there was a significant effect of LHS on mel1a (F_3,16 = 3.49, p = 0.04, η² = 0.396; figure 5a), mel1b (F_3,16 = 4.46, p = 0.01, η² = 0.456; figure 5c) and nr3c1 (F_3,16 = 8.26, p = 0.001, η² = 0.608, one-way ANOVA; figure 5e). Similarly, in brainstem, we found a significant effect on mel1a (F_3,16 = 4.37, p = 0.01, η² = 0.451; figure 5b) and mel1b (F_3,16 = 6.86, p = 0.003, η² = 0.563, one-way ANOVA; figure 5d). The mRNA levels of mel1a in PM > REF, mel1b in PM > NM and nr3c1 PM > NM, MIG and REF (p < 0.05, Bonferroni post-test, figure 5a,c,e) in the hypothalamus. In brainstem levels of mel1a in PM > MIG, and mel1b PM > NM, MIG and REF (p < 0.05, Bonferroni post-test; figure 5b,d).

Figure 5.

mRNA expression of melatonin and GR (a–f), and sleep and arousal promoting (g–l) genes in hypothalamus (a,c,e,g,i,k) and brainstem (b,d,f,h,j,l). Mean ± (s.e.m.) mRNA levels of mel1a (a,b), mel1b (c,d), nr3c1 (e,f), ghrh (g,h), bdnf (i,j) and camk2 (k,l) in redheaded bunting in four LHS (i.e. non-migratory, NM; pre-migratory, PM; migratory, MIG; refractory, REF). Different alphabets on bars indicate a significant difference between the LHS (Bonferroni post-test following one-way ANOVA). For statistical significance, alpha was set at 0.05.

Finally, we measured mRNA expressions of pro-somnogenic (ghrh and bdnf) and anti-somnogenic (camk2) genes. In the hypothalamus, we found a significant effect of LHS on ghrh (F_3,16 = 4.66, p = 0.01, η² = 0.647; figure 5g). Likewise, in the brainstem, we found a significant effect on ghrh (F_3,16 = 3.45, p = 0.04, η² = 0.393; figure 5h) and camk2 (F_3,16 = 6.78, p = 0.003, η² = 0.560, one-way ANOVA; figure 5l). The mRNA levels of ghrh in PM > MIG in the hypothalamus (p < 0.05, Bonferroni post-test; figure 5g). In brainstem, the mRNA levels of ghrh in NM > MIG, and camk2 in MIG > NM and REF (p < 0.05, Bonferroni post-test; figure 5h,l).

Pearson's correlation test revealed a negative correlation between total sleep (L_OFF0-1) and il1b (r = −0.7668, p < 0.0001), il6 (r = −0.5899, p = 0.0062) and nos2 (r = −0.5163, p = 0.0198) in the hypothalamus, whereas in the brainstem we found a positive relation of total sleep (L_OFF0-1) with nr3c1 (r = 0.5170, p = 0.0196) and negative with camk2 (r = −0.5274, p = 0.0169; electronic supplementary material, table S3.5).

4. Discussion

The present study demonstrates the LHS-dependent changes in activity-rest pattern, general physiology and awake/sleep behaviours. Furthermore, we show the concomitant changes in the hypothalamic and brainstem molecular machinery which directly or consequentially influence the sleep–wake cycle.

(a) . Changes in activity counts and physiological parameters

The changes in the activity profile across the LHS are supported by previous studies performed on both red and blackheaded buntings [28,38]. The total activity of the birds was significantly increased during the migratory phase, which was exclusively due to the increased night-time activity (i.e. vernal Zugunruhe), while the daytime activity was contrastingly low, compared to the non- and pre-migratory birds. This was accompanied by parallel changes in the body mass, body fattening and testicular volume, where all the physiological parameters were found to be elevated during the migratory phase [9]. The increased body mass and fattening provides essential fuel to the migrants, to sustain their long arduous flights [39]. This occurs due to increase in food intake i.e. hyperphagia, as found in our study, and associated changes in the metabolism [40].

(b) . Changes in active and sleep behaviours

The present study demonstrates the diurnal distribution of different active and sleep postures in redheaded buntings and validates the previous observations by Yadav et al. [17]. However, while the previous study [17] had made observations at specific time points during the day, we have video monitored the birds throughout the 24 h which gave us a better resolution of their daily behaviour. Additionally, we have monitored the active/sleep behaviours in the photorefractory state, which provided further insight to their life history.

The birds in the migratory group showed abounding active wakefulness during the night-time, while the birds in the other three groups showed the same only during the daytime. Thus, we found an effect of ‘time × LHS’ interaction, which was expected as the activity-rest data had already revealed increased night-time activity in the migratory birds, which occurs largely due to perch hopping. Similarly, we found an effect of LHS on alert and quiet wakefulness, which showed short night-time bursts during the migratory phase, where both peaked during the early night. Interestingly, for the first time, we recorded the redheaded buntings to feed and drink during the night-time in the migratory phase, which probably resulted in increased food intake. Although there is no field data available on buntings, which show that the birds feed at night-time during their migratory journey, our study suggests that if food is available nearby the birds can potentially feed during the dark hours.

Both front and back sleep was drastically reduced during the migratory phase and showed an effect of LHS which was expected, as the migratory birds are largely active during the night. The birds, however, were drowsy during the first couple of night hours. Similar data were obtained in white-crowned sparrows, where in comparison to the non-migratory state, the migratory sparrows spent almost two-thirds less time sleeping with an advancement in the REM sleep timing [18]. Interestingly, the cognitive function of the sparrows did not decline in the migratory phase, despite the chronic sleep loss. However, in another study, it was found that during the migratory state, the ability to perform ‘differential reinforcement of low rate behaviour’ (DRL, a marker for operant performance) was significantly impaired in sparrows [41]. They compensated for this loss of sleep, by increasing the front sleep, back sleep and UHS in the daytime during migration, which was practically negligible during the other LHS. Fuchs et al. [16,17] reported similar daytime micro naps in Swainson's thrush from EEG and behavioural recordings, where brief episodes of daytime sleep compensated for the extended periods of sleep loss. Future studies on migratory buntings can greatly benefit from additional EEG recordings alongside video monitoring data, which is a shortcoming of the current study.

(c) . Changes in the regulatory gene expression

We found tissue-specific changes in mRNA expression in the hypothalamus and brainstem, which provides key insights into the molecular regulation of the sleep–wake cycle in redheaded buntings. First, in the cytokine pathway, we looked into IL1 and IL6, which are pro-inflammatory cytokines, in which IL1 is well known to promote sleep [4], whereas IL6 is involved in sleep disturbances [5]. We found increased hypothalamic expression of both il1b and il6 during the migratory state, where night-time sleep was severely compromised. This may appear to be counterintuitive as IL1b is supposed to induce sleep, but it does suggest a heightened encephalic inflammatory state possibly owing to the drastic sleep loss. This theory is supported by studies performed on zebra finches (Taeniopygia guttata), where sleep loss was accompanied by an increased expression of il6 in the hippocampus, and a peripheral increase of il1b and il6 in the spleen and adipose tissue, respectively [42]. Similarly, in mice (Mus musculus), sleep fragmentation has been shown to increase il1b expression in both the hypothalamus and adipose tissue [43]. In the brainstem, we found increased expression of nfkb in the pre-migratory state, but we didn't find any changes in the downstream genes (i.e. il1b and il6). It is to be noted that apart from inducing cytokine expression, NFkB is involved in multiple pathways including the induction of anti-apoptotic factors, cell cycle regulators, chemokines, etc. [44]. Jones et al. [21] have found similar changes in the gene expression of glucose transporter and heat-shock proteins in the migratory state, which suggest dynamic changes in brain cellular stress and energetic demands. In the RNS pathway, we found tissue-specific differences in the nos2 expression. Studies have shown that nos2-induced neuronal NO is a key homeostatic factor to promote recovery sleep following sleep deprivation [45]. This transitory NO/nos2 production is accompanied by inflammatory events ensured by the glial cells [8]. The hypothalamic expression of nos2 in the migratory state may be a direct reaction to night-time sleep loss and associated inflammation. This, however, does not explain the increased nos2 expression in the non-migratory state. While interpreting the mRNA expression data, we have to be cautious as many of these molecules tend to be pleiotropic, and NO is no exception. For example, it acts as a retrograde neurotransmitter, which helps in the brain's blood flow, and it also regulates the dendritic spine growth [46].

Secondly, in the hormone receptor pathway, we found increased expression of mel1a and mel1b in the pre-migratory state in both tissues. Night-time melatonin levels reflect the day length and thus the seasons [9], and studies have shown melatonin levels to be significantly reduced during the migratory phase in garden warblers (Sylvia borin) [47] and blackcaps (Sylvia atricapilla) [48]. The upregulation of the MEL receptor level may be a compensatory mechanism to counteract the reduction in plasma melatonin concentration, but this theory needs to be verified by further studies in buntings. Interestingly, changes in MEL receptors can directly affect the sleep cycles in vertebrates. For example, MEL1a is known to regulate REM sleep, whereas MEL1b increases non-REM sleep [49]. We also found increased hypothalamic expression of nr3c1 in the pre-migratory state, which encodes GR. GR in locus coeruleus has been shown to play a significant role in stress-associated sleep disorders [50]. Further, previous studies on buntings indicate an increase in baseline plasma corticosterone levels during the pre-migratory state [26,34]. This elevation in corticosterone can play a crucial role in the development of migratory phenotype.

Thirdly, we found a reduction in ghrh expression in both hypothalamus and brainstem in the migratory state. GHRH is neuropeptide which stimulates the synthesis and secretion of growth hormone (GH). It is well established that GHRH promotes both REM and nREM sleep in animals, including humans [11]; for example, when GHRH is administered exogenously it enhances nREM, but endogenous inhibition of it reduces sleep depth and duration [51], Additionally, during sleep, the secretion of GH is profoundly increased which is associated with the appearance of SWS [52]. The reduction of ghrh expression can thus be directly correlated with the reduction in the night-time as well as total sleep in buntings. On the other hand, camk2 expression was found to be increased in the migratory state. CaMK2 is a key signal-transducing molecule that is activated by the presence of Ca²⁺. Interestingly, in rodents Ca²⁺ in the dorsal raphe nucleus promotes wakefulness and suppresses REM and nREM sleep [14]. Furthermore, the activation of Ca²⁺/CaMK2 signalling in the pedunculopontine tegmentum nucleus promotes wakefulness, while suppressing sleep [53]. Thus, CaMK2 in buntings is a potentially key molecule that promotes wakefulness during night-time in the migratory phase and reduces sleep.

It is well established that the encephalic gene expression levels change dramatically between wakefulness and sleep in both mammals and birds [20,21]. Hence, we wanted to assess how the sleep timing before the sampling might affect the expression values of our target genes. The negative correlation of pro-inflammatory cytokines and nos2 with total sleep (L_OFF0-1), suggests that the acute reduction of sleep duration (L_OFF0-1) can give rise to an inflammatory state in the hypothalamus. Further, in the brainstem, the negative relation between camk2 and total sleep (L_OFF0-1) suggests that arousal before sampling can also influence mRNA expression levels. Furthermore, it is to be noted that the gene expression levels in the brain can also vary as a function of the circadian rhythm which might influence fold changes. In the current study, we examined the mRNA expressions at a single time point that varied across LHS relative to the lights ON/OFF, which might be a limitation. Future studies need to be performed to evaluate the tissue-specific changes in diurnal mRNA expression profiles under different states.

5. Conclusion

There was an overall decrease in night-time sleep in the migratory phase, which was moderately compensated by the increased daytime UHS, front and back sleep. The intense Zugunruhe exhibited by the buntings corresponded to the increased active wakefulness. The migratory sleep deprivation paralleled an augmented pro-inflammatory state coupled with nitrosative stress. Melatonin potentially acts through receptors to signal seasonal changes, and there is strong evidence for a pre-migratory CORT-signalling within the hypothalamus. Changes in the calcium and GH genes reveal a putative machinery that regulates the sleep–wake cycle in migratory redheaded buntings.

Ethics

The experiments were performed as per approval of Institutional Animal Ethics Committee of University of Lucknow, Lucknow, India. Protocol number: LU/ZOOL/SR-SM/DBT/11-2014.

Data accessibility

The mRNA sequences with their partial CDS can be accessed using GenBank accession numbers as provided in the electronic supplementary material, table S2.

The data are provided in the electronic supplementary material [54].

Authors' contributions

J.T.: conceptualization, data curation, formal analysis, investigation, methodology, validation, visualization and writing—original draft; S.S.: formal analysis, validation, visualization and writing—review and editing; A.Y.: investigation; R.K.: investigation; N.R.: resources; S.R.: conceptualization, resources and supervision; S.M.: conceptualization, funding acquisition, investigation, project administration, resources, supervision and writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

The authors declare that they have no competing interests.

Funding

The financial support by R&D (Research and Development) state grant to S.M. and fellowship (ID:108/2021/2585/Seventy-4-2021-4(28)/2021) to J.T., and SERB (Science and Engineering Research Board) Research Project [grant no. CRG/2019/000669] to S.R. Department of Zoology, University of Lucknow is highly acknowledged.