Diurnal and photoperiodic changes in thyrotrophin‐stimulating hormone β expression and associated regulation of deiodinase enzymes (DIO2, DIO3) in the female juvenile chicken hypothalamus

Abstract Increased thyrotrophin‐stimulating hormone β (TSHβ) expression in the pars tuberalis is assumed to be an early step in the neuroendocrine mechanism transducing photoperiodic information. The present study aimed to determine the relationship between long‐photoperiod (LP) and diurnal TSHβ gene expression in the juvenile chicken by comparing LP‐photostimulated birds with groups kept on a short photoperiod (SP) for 1 or 12 days. TSHβ expression increased by 3‐ and 23‐fold after 1 and 12 days of LP‐photostimulation both during the day and at night. Under both SP and LP conditions, TSHβ expression was between 3‐ and 14‐fold higher at night than in the day, suggesting that TSHβ expression cycles in a diurnal pattern irrespective of photoperiod. The ratio of DIO2/3 was decreased on LPs, consequent to changes in DIO3 expression, although there was no evidence of any diurnal effect on DIO2 or DIO3 expression. Plasma prolactin concentrations revealed both an effect of LPs and time‐of‐day. Thus, TSHβ expression changes in a dynamic fashion both diurnally and in response to photoperiod.


| INTRODUCTION
Increased thyrotrophin-stimulating hormone β (TSHβ) expression in the pars tuberalis is assumed to be an early step in the neuroendocrine mechanism transducing photoperiodic information to the reproductive axis, driving thyroid hormone conversion in adjacent ependymal tanycyte cells of the ventral hypothalamus. 1 In mammals, photoperiodic information is considered to be interpreted via the nocturnal melatonin signal, as secreted by the pineal acting via the melatonin 1 receptor (MT1) on TSHβ expressing cells, 2 and this sets the phase of transcription factor Eya3 expression. 3 In avian species, despite a functional pineal system with several biological effects, 4,5 the photoperiodic information for reproduction is assumed to be translated by deep brain photoreceptors, 6,7 although there is evidence indicating that the importance of melatonin may depend on the output under investigation. 8 In these cases, a circadian clock mechanism for measuring daylength situated in the medial basal hypothalamus has been demonstrated. 9 In quail and sheep, EYA3 expression is induced on long days and, at least in mammals, is inhibited by melatonin. EYA3 is a transcriptional activator of the TSHβ promoter and fulfills the requirements to be part of the internal coincidence model proposed to explain the interaction of light and the circadian system. 10  in birds, an alternative mechanism involving changes in clock genes Per2 and Cry1 in the pars tuberalis acting on vasoactive intestinal polypeptide neurones has been suggested. 12 The initial discovery of the response of TSHβ to photostimulation in the Japanese quail 1 was preceded by observations of what we now consider to be downstream effectors of this system, including the demonstration that exogenous thyroxine (T4) activates reproduction 13 consequent upon photostimulated conversion of T4 to triodothyronine. 14 Observations in a closelyrelated Galliform, the chicken, have been sparse, although T4 injection results in the stimulation of gonadotrophin-releasing hormone (GnRH) expression and gonadotrophin release. 15 An increase in TSHβ expression has been reported to occur during the first long day and DIO2 after 2 weeks of long-photoperiod (LP) exposure. 16 If TSHβ expression is driven by an internal circadian clock mechanism, we might expect to see differences in the expression of TSHβ at different times of the day. We therefore assessed the diurnal expression of TSHβ under a short photoperiod (SP) and following long-term photostimulation, as well as downstream genes DIO2 and DIO3, in domesticated chickens. Earlier studies have defined photo-inductive effects of LP on gonadotrophin secretion, with a critical daylength of around 11 hours 17 and an increase in GNRH1 expression postphotostimulation. 18 Chickens also exhibit an acute day 1 response to LP of luteinising hormone (LH) release, 15 similar to that observed in quail. In the present study, we investigated whether the dynamic photoperiodic inductive effects observed in the quail are replicated in domesticated chickens and also whether the differences in response of these 2 closely-related species might underlie domestication events or their evolution.

| Experiment 1
Female 1-day-old Isa Brown chicks, a commercial hybrid layer chicken (Gallus gallus domesticus) were maintained for 2 days on LP (14:10 hour light/dark cycle) prior to transfer to a climate controlled chamber on SP (8:16 hour light/dark cycle). They remained there until 5 weeks of age when they were randomised between 2 chambers at the same time as remaining on a SP (8:16 hour light/dark cycle). At 7 weeks of age, the photoperiod for 1 chamber was changed to LP (16:8 hour light/dark cycle) by extending dusk. The next day, hens (n = 7) were killed 4 hours after lights on from both chambers. Additional hens (n = 7) were killed 4 hours after lights off from both chambers. After 12 days, further hens (n = 7) were killed 4 hours after lights on from both chambers and more hens (n = 7) were killed 4 hours after lights off from both chambers. The age and sex of the birds used was based on previous experiments where the most robust responses to photostimulation in prepubertal hens had been observed [17][18][19][20] Prior to killing by an overdose of barbiturate, a blood sample was taken from the brachial vein and the birds were weighed. Immediately after death, the birds were dissected to remove the pituitary and the basal hypothalamus including the pars tuberalis and median eminence. The mean ± SD mass of the basal hypothalamus dissection was 4.1 ± 0.1 mg and the pituitary mass is reported in the results. These tissues were immediately frozen in liquid nitrogen before being stored at −80°C prior to RNA extraction. The ovary and oviduct were also removed and weighed.

| Experiment 2
In Experiment 1, we observed no change in DIO2 expression related to photoperiod, unlike in the quail. 1 To confirm the ability of our DIO2 assay to detect dynamic changes in expression, an experiment was performed using quail bred at Roslin Institute derived from a population originally kept by Professor Brian Follett. 21 The protocol was similar in terms of lighting to Experiment 1, except only 12 days of photostimulation was used and dissections were performed only 4 hours after lights off (n = 8). The quail were photostimulated at 3 weeks of age to match their earlier sexual maturity.
Experiments were carried out under the Animals (Scientific Procedures) Act 1986, project license 70/7909 and individual experiments were approved by the institutional ethics committee.

| Measurement of gene expression
Total RNA was extracted from the tissue samples using Ultraspec Quantitative PCR for estimation of levels of gene expression were made using specific primers (Table 1)  For primers, the chromosomal position and sequence are provided in Table 1. PCR products from the amplification of cDNA with the gene specific primers were purified for standards using a QIAEX II gel ex- Standards were diluted to produce top standards detectable after approximately 15 PCR cycles. Assays were analysed using mxpro (Agilent Technologies) and expression was normalised using a weighted average of lamin B receptor (LBR), β-actin and glyceraldehyde-3-phosphate dehydrogenase expression. In the case of the quail experiment, only LBR was used as a control gene. There was no evidence of any treatment effects for the control gene expression in an ANOVA (for primers, chromosomal position and sequence, see Table 1).

| Hormone assays
Plasma prolactin and LH were measured in 1 assay using homologous radioimmunoassays as described previously. 23,24 The intra-assay coefficient of variance was 8.4% and 10.2%, respectively.

| Statistical analysis
The results were analysed by ANOVA in genstat, version 13 (VSN International, Hemel Hempstead, UK) using log-transformed data to approximate to a normal distribution where appropriate. The significance of differences between means was calculated by least significant differences where indicated as appropriate from the ANOVA.

| Body mass and reproductive organs are increased by photostimulation
Body mass or organ mass did not show any diurnal variation.
Therefore, for these traits, only photostimulation (LP vs SP) and the duration of treatment (1 vs 12 days) were used as variables for ANOVA. As predicted, body mass and organ mass increased after 12 days (F 1,55 = 26.92-158.26, P < .001) because animals were still in the juvenile growth period. There was an effect of photostimulation on body mass (F 1,55 = 7.29, P = .009) and, after 12 days of exposure, LP animals were heavier than those on SP (858 g vs 772 g, P < .05) ( Table 2). Pituitary and ovary mass were significantly higher in LP animals (F 1,53 = 11.65, P = .001 and F 1,54 = 7.62, P = .008, respectively, when body mass was fitted as a covariate). Specifically, at 1 and 12 days, pituitary mass was greater in photostimulated birds (P < .01 Significance of differences between the means for SP-LP contrasts; *P < .05, **P < .01, ***P < .001. a Estimated mean from ANOVA using body mass as a covariate. b Observed values but ANOVA performed on log-transformed values. c Back-transformed estimated means from ANOVA using body mass as a covariate. and P < .001, respectively) ( Table 2), as was also the case for the ovary (P < .01 and P < .05, respectively) ( Table 2). In the case of the oviduct mass, the ANOVA was also significant for photostimulation (F 1,55 = 12.23, P < .001) but oviduct mass was increased significantly (P < .01) only after 12 days of exposure to LP when the difference between the treatment means was tested (Table 2).

| Plasma LH varies diurnally
Plasma LH concentrations were not significantly elevated after photostimulation; however, there was a clear diurnal effect on LH concentrations (ANOVA, F 1,55 = 9.13, P = .004) across the entire experiment ( Figure 1).   Figure 3). The magnitude of the differences between LP and SP groups after 12 days was a 23-fold increase in expression during the day (P < .001), which is greater than the 11-fold difference observed during the night (P < .01).

| DIO3 but not DIO2 expression responds to photostimulation but shows no diurnal variation
Measurement of DIO2 did not reveal any discernible pattern of expression attributable to time of day or photostimulation ( Figure 4A) but did increase with age (ANOVA, F 1,55 = 16.92, P < .001). The samples taken after 12 days of photostimulation had approximately 3.7 times the level of expression observed at 1 day, which corresponds to approximately 7 and 9 weeks of age.
By contrast, DIO3 expression ( Figure 4B) was clearly reduced after photostimulation (ANOVA, F 1,55 = 37.76, P < .001) and DIO3 was expressed at a higher level in the 12-day exposure group than in the 1-day group (ANOVA, F 1,55 = 27.64, P < .001) as was the case for DIO2. There was also a tendency for the photostimulation effect to be stronger after 12 days of exposure to LP than after 1 day, as indicated by the day × photostimulation interaction (ANOVA, F 1,55 = 3.96, P = .052). A significant effect of day length was observed after only 1 day of exposure to LP in the night samples (P < .001), although the difference was not significant in the day samples.
F I G U R E 1 Plasma luteinising hormone concentrations in samples taken prior to death. Hens were exposed to either 1 or 12 days of a long photoperiod (LP) (16:8 hour light/dark cycle) or maintained on a short photoperiod (SP) (8:16 hour light/dark cycle). Open bars represent the samples taken during the light period (4 hours after lights on), whereas the solid bars represent the night samples (4 hours after lights off)

| DIO2 expression is increased in quail after photostimulation
To confirm the lack of effect observed in DIO2 expression in the chicken, we aimed to confirm that our assay can detect differences in the quail hypothalamus after photostimulation. The DIO2 primers were selected to work on both species and an increase in expression in the quail after 12 days of photostimulation was observed, as well as an increase in TSHβ expression ( Figure 5).

| DISCUSSION
We have demonstrated that juvenile domesticated chickens have as robust a photoperiodic response in terms of TSHβ expression and changes in the ratio of DIO2/3 expression as that previously observed in quail. 1 However, in the present study, there is no evidence that DIO2 is affected by photostimulation in juvenile chickens as it is in quail. We have observed considerably higher TSHβ expression in LPs compared to the respective SP birds. Equally striking was that differences in TSHβ expression between day and night samples on both LP and SPs.
There is a 3-and 11-fold difference in TSHβ expression between the night and day sample in 1 day and 12 day birds on SP, respectively.
This was proportionally as large as the fold difference produced by photostimulation. Thus, TSHβ expression in the pars tuberalis cycles daily, regardless of whether the bird was photostimulated or not.
In domesticated chickens, the region of the genome containing TSHR and DIO2 and DIO3 has been under intense selection 25,26 and a genetic association at the locus was observed in a study on incubation behaviour and persistency. 27 Therefore, there may be a link between the locus, photoperiod and reproduction that may have increased fecundity during domestication. Although there is evidence for differences in TSHβ expression with selection, it is only in the context of declining photoperiod, 28 and so, currently, we can only speculate on the involvement of these genes and domestication, although it does provide a testable hypothesis. The fact that the response of DIO2 expression does not appear to respond to photoperiod in juvenile chickens but does change with age might be a suitable place to start. This is especially the case given that hens bred for egg production, although influenced by light, do start reproductive activity on short days. 29 Studies describing the link between TSHβ in the pars tuberalis and the photoperiodic response in quail, 1 revealed that expression of TSHβ was increased 14 hours after lights on, returning to baseline 6 hours later, although only in photostimulated conditions. There was no evidence of an increase in TSHβ prior to or during the nocturnal phase on SPs. 1 In the mouse, a species not classically considered as photoperiodic, the results clearly show the levels of TSHβ mRNA and protein are higher during the dark phase than the light phase 30  F I G U R E 4 (A), DIO2, (B), DIO3 expression and (C) their ratio in basal hypothalamus. Hens were exposed to either either 1 or 12 days of a long photoperiod (LP) (16:8 hour light/dark cycle) or maintained on a short photoperiod (SP) (8:16 hour light/dark cycle). Open bars represent the samples taken during the light period (4 hours after lights on), whereas the solid bars represent the night samples (4 hours after lights off). Specific comparisons between means are indicated for SP-LP comparisons, *P < .05; ***P < .001. There were no significant night day effects response to photostimulation is more robust 34 and, in the present study, we observed a persistent and significantly higher level at night after a single long day on LP. Prolactin concentrations were significantly different between night and day after 1 or 12 days on LP. Both a diurnal pattern in circulating LH and prolactin have previously been observed in chickens. 35,36 We could speculate that, in birds, prolactin might be ultimately controlled by the TSHβ system in the pars tuberalis because of the correlation between their release and expression and, similarly, the diurnal variation of LH. The major problem with this hypothesis is the belief that photoperiod induced LH secretion is driven through GnRH release by changes in the DIO2/3 enzyme ratio. Although we observed a change in the ratio of DIO2/3 expression in the night sample after a single long day, there is no evidence for a diurnal effect on the DIO2/3 expression ratio. Of course, this may be because the fluctuating TSHβ signal is integrated into a linear DIO2/3 response.
Whatever the link between TSHβ and reproduction we should not think of TSHβ expression as a static entity but, instead, as a dynamic system that changes diurnally and in response to photoperiod.
In conclusion, we have confirmed changes in the expression of TSHβ and the ratio of DIO2/3 in response to increased photoperiod in juvenile domestic chickens and we have also shown that TSHβ but not DIO2/3 shows a very large diurnal variation of similar magnitude to that of the response to photostimulation.