Report. A Circannual Clock Drives Expression of Genes Central for Seasonal Reproduction

Transcription

1 Current Biology 24, , July 7, 2014 ª2014 Elsevier Ltd All rights reserved A Circannual Clock Drives Expression of Genes Central for Seasonal Reproduction Report Cristina Sáenz de Miera, 1,2,4 Stefanie Monecke, 1,4 Julien Bartzen-Sprauer, 1 Marie-Pierre Laran-Chich, 1 Paul Pévet, 1 David G. Hazlerigg, 2,3 and Valérie Simonneaux 1, * 1 Neurobiologie des Rythmes, Institut des Neurosciences Cellulaires et Intégratives, Université de Strasbourg, Strasbourg, France 2 School of Biological Sciences, University of Aberdeen, Aberdeen AB24 2TZ, Scotland, UK 3 Department of Arctic and Marine Biology, Faculty of Biosciences, Fisheries and Economy, University of Tromsø, 9037 Tromsø, Norway Summary Animals living in temperate zones anticipate seasonal environmental changes to adapt their biological functions, especially reproduction and metabolism. Two main physiological mechanisms have evolved for this adaptation: intrinsic long-term timing mechanisms with an oscillating period of approximately 1 year, driven by a circannual clock [1], and synchronization of biological rhythms to the sidereal year using day length (photoperiod) [2]. In mammals, the pineal hormone melatonin relays photoperiodic information to the hypothalamus to control seasonal physiology through well-defined mechanisms [3 6]. In contrast, little is known about how the circannual clock drives endogenous changes in seasonal functions. The aim of this study was to determine whether genes involved in photoperiodic time measurement (TSHb and Dio2) and central control of reproduction (Rfrp and Kiss1) display circannual rhythms in expression under constant conditions. Male European hamsters, deprived of seasonal time cues by pinealectomy and maintenance in constant photoperiod, were selected when expressing a subjective summer or subjective winter state in their circannual cycle of body weight, temperature, and testicular size. TSHb expression in the pars tuberalis (PT) displayed a robust circannual variation with highest level in the subjective summer state, which was positively correlated with hypothalamic Dio2 and Rfrp expression. The negative sex steroid feedback was found to act specifically on arcuate Kiss1 expression. Our findings reveal TSH as a circannual output of the PT, which in turn regulates hypothalamic neurons controlling reproductive activity. Therefore, both the circannual and the melatonin signals converge on PT TSHb expression to synchronize seasonal biological activity. Results and Discussion European Hamster s Seasonal Physiology Circannual species exposed to a constant photoperiod show circannual variations in reproduction, food intake and body 4 Co-first author *Correspondence: simonneaux@inci-cnrs.unistra.fr weight, moult, or hibernation [1, 7, 8], although the robustness of the free-running rhythms may be influenced by the photoperiod employed [9, 10]. Under constant photoperiod, the melatonin profile continues to reflect the prevailing photoperiod [11]. Nevertheless, melatonin received at a particular phase of the circannual cycle entrains the endogenous rhythm in reproduction [12], demonstrating that responsiveness to melatonin is maintained in these conditions. In this study, we used the European hamster, a long-day breeder known to display circannual rhythms [8, 13]. Male European hamsters kept in constant long photoperiod (LP; 16 hr light [L] and 8 hr dark [D]) and pinealectomized in order to remove any potential influence of melatonin showed robust cycles in body weight, testicular size, and body temperature (Figure 1). To avoid the cycling feedback of testosterone on the circannual clock, a subgroup of animals was castrated, and these animals displayed similar strong annual cycles in body weight and temperature (Figure S1 available online). Animals were sampled at the appearance of a summer phenotype (an increase in body weight to a stable level and also large testis size in the gonadally intact group) for the subjective summer (S-Sum) group (Figures 1A and S1A) and the appearance of a winter phenotype (a decrease in body weight for three consecutive points and testis regression in the gonadally intact animals) for the subjective winter (S-Win) group (Figures 1B and S1B). The circulating level of luteinizing hormone (LH), measured as an a posteriori reproductive index, was higher in S-Sum as compared to S-Win but was strongly upregulated in both castrated S-Sum and S-Win groups as a consequence of the removal of the negative testosterone feedback (Figure 1C). For comparative purposes and to exclude any misinterpretation related to putative daily variation in gene expression, we also used two additional groups of intact male European hamsters adapted to LP or short photoperiod (SP; 10L:14D) and sacrificed throughout 24 hr. Their reproductive hormone profiles confirmed the appropriate photoperiodic reproductive state, with higher LH and testosterone levels in LP compared to SP conditions at all day times investigated (Figures S2A and S2B). In view of the similarities between LH level in the circannual gonadally intact animals and the photoperiodic groups, we are confident that sampled hamsters in our circannual experiment were indeed in S-Sum or S-Win states. Our results and previous observations in this species and in the sheep [12, 13] demonstrate the independence of melatonin for the occurrence of circannual rhythms in reproduction, and they demonstrate that the circannual clock is not affected by the sex steroid feedback. TSHb and Dio2 Are Regulated by the Circannual Clock The photoperiodic regulation of seasonal physiology is initiated by circulating melatonin control of the b subunit of thyroid-stimulating hormone (TSHb) in the melatonin-sensitive pars tuberalis (PT) of the pituitary gland. TSHb synthesis is inhibited in SP, when the nocturnal melatonin production is long [14 16]. In LP, the melatonin signal duration is reduced, and TSH is released from the PT and binds to TSH receptors in the neighboring tanycyte cells located in the paraventricular

2 Gene Expression Underlying Circannual Reproduction 1501 Figure 1. Circannual Rhythms in European Hamster s Physiology (A and B) Graphs representing the expression of endogenous rhythms in body temperature (black line), body weight (,), and testis size (:) in two different pinealectomized gonadally intact European hamsters kept in constant conditions during the whole experiment (16L:8D, 20 C 6 2 C). Body weight (g) and scrotal testis size (mm) were measured every two weeks, from the beginning of the experiment until the sampling date, when the animals expressed subjective summer (S-Sum) (A) or subjective winter (S-Win) (B) physiologies. Body temperature ( C) was constantly recorded during the experiment, and the data were assessed postmortem. (C) Blood LH level of circannual animals at the S-Sum or S-Win states, in the gonadally intact and castrated groups. Data in (C) are n = 6 12 animals per group. ***p < indicates a difference between S-Sum and S-Win. zone (PVZ) of the third ventricle to promote the expression of type 2 thyroid hormone deiodinase gene (Dio2) [5, 17]. Dio2 locally converts thyroid hormone from its tetraiodinated circulating form, thyroxine (T4), into the most active form, triiodothyronine (T3) [18]. Exogenous supply of central T3 or TSH in the basal part of the hypothalamus mimics photoperiodic activation of summer physiology [6, 19]. We first studied the photoperiodic variation in TSHb gene expression in the PT and Dio2 in the tanycytes (Figure 2G) in the early morning, at zeitgeber time 2 (ZT02), and during the night, at ZT22 (ZT0: lights-on time), based on previous results in female European hamsters reporting a modest day-to-night difference in Dio2 expression [20]. TSHb expression in the PT was strongly regulated by photoperiod, with high mrna level in LP and nearly undetectable mrna level in SP (Figure 2A; p < ), with no variation detected between ZT02 and ZT22 in either photoperiod (Figure 2A; p = 0.14). Dio2 gene expression was quantified in two regions corresponding to the area where the tanycyte cell bodies and projections are located: in the PVZ and along the border of the median eminence, in the cell-free area called the tuberoinfundibular sulcus (TIS). Dio2 mrna level was markedly increased in LP as compared to SP in the two tanycyte areas examined (Figures 2B and 2C; p < for PVZ and for TIS). In addition, Dio2 expression in LP was significantly higher at ZT02 than at ZT22 in the PVZ region (Figure 2B; p < 0.01), but not in the TIS region (Figure 2C; p > 0.05). This pattern matches the previously demonstrated photoperiodic melatonin regulation of TSHb/Dio2 [3, 5, 14, 15, 21]. In circannual conditions, TSHb mrna level was significantly higher in S-Sum compared to S-Win, and castration did not affect the circannual variation in TSHb expression (Figure 2D; p < for gonadally intact; p < 0.05 for castrated). Strikingly, Dio2 expression showed a similar strong circannual variation in the gonadally intact and castrated groups, with Dio2 mrna level being significantly higher in S-Sum than in S-Win in both tanycyte areas (Figures 2E and 2F; p < for PVZ and TIS). PT TSHb and PVZ Dio2 mrna levels were strongly positively correlated both among the photoperiod adapted (Figure S3A; p < 0.01; Pearson r = 0.65) and the animals in circannual conditions (Figure 2H; p < 0.001; Pearson r = 0.57). These results support the hypothesis that the circannual clock drives TSHb expression in the PT, which in turns drives Dio2 expression in tanycytes. This regulation was identical whether the animals were castrated or not, showing that at this level, the circannual signal is independent of the testosterone feedback. Our results indicate that both the circannual and the photoperiodic signals converge at the level of TSHb in the PT, which in turn, via the regulation of Dio2 expression in

3 Current Biology Vol 24 No Figure 2. Photoperiodic and Circannual Variations in TSHb and Dio2 Gene Expression in European Hamster (A C) Photoperiodic and day-to-night variation in TSHb and Dio2 expression. Photoperiodic and day-to-night variation in TSHb expression in the PT is shown in (A). Mean TSHb labeling intensity was measured at ZT02 (plain bars) and ZT22 (stripes) time points under LP or SP conditions. Representative images of TSHb mrna labeling in the PT in LP and SP at ZT02 are shown. Mean Dio2 labeling intensity in the PVZ (B) and in the TIS (C) was measured at ZT02 (plain bars) and ZT22 (stripes) time points under LP or SP conditions. Representative images of Dio2 mrna labeling in the PVZ and TIS in LP and SP at ZT02 are shown. (D F) Circannual variation in TSHb and Dio2 expression. Circannual variation in TSHb expression in the PT is shown in (D). Mean TSHb labeling intensity was measured in gonadally intact and castrated animals in S-Sum (white bars) and S-Win (black bars). Representative images of TSHb mrna labeling in the PT of a gonadally intact S-Sum and an S-Win animal are shown. Mean Dio2 labeling intensity in the PVZ (E) and in the TIS (F) was measured in gonadally intact and castrated animals in S-Sum (white bars) and S-Win (black bars). Representative images of Dio2 mrna labeling in the PVZ and TIS of a gonadally intact S-Sum and an S-Win animal are shown. All data show mean 6 SEM (n = 5 or n = 6 animals per sampling point in the photoperiodic groups; n = 6 12 animals in the circannual groups). (G) Schematic drawing of a coronal section at the level of the hypothalamus, depicting the areas where gene expression was studied, as shown in (A) (F). (H) Scatterplot showing the positive correlation between TSHb expression in the PT and Dio2 expression in the PVZ. r = 0.57, r 2 = 0.33, p < The following abbreviations were used: au, arbitrary units; 3V, third ventricle; LP, long photoperiod; ME, median eminence; PT, pars tuberalis; PVZ, paraventricular zone; SP, short photoperiod; S Sum, subjective summer; S-Win, subjective winter; TIS, tuberoinfundibular zone; ZT, zeitgeber time. Scale bars of (A) and (D) represent 100 mm; scale bars of (B), (C), (E), and (F) represent 50 mm. In (B), the arrow indicates the PVZ, and the arrowhead points to the TIS. ## p < 0.01 for post hoc analysis in LP ZT02 versus ZT22; ***p < and *p < 0.05 for differences between groups.

4 Gene Expression Underlying Circannual Reproduction 1503 the tanycytes, most likely controls hypothalamic level of T3. Recently, we reported on the Soay sheep, an endogenous variation in PT TSHb and hypothalamic deiodinases associated with the endogenous changes in reproductive physiology [22]. When maintained under constant nonstimulatory LP conditions, sheep exhibited an endogenous reactivation toward S-Win physiology, accompanied by a decrease in TSHb and Dio2 levels. In the opposite conditions, in constant SP, sheep exhibited an endogenous switch toward S-Sum physiology, together with a reversion in deiodinases gene expression, although independent of TSHb change [22]. Therefore, it appears that in both the sheep (a short-day breeder) and the European hamster (a long-day breeder), the circannual clock is using comparable mechanisms relaying on the PT to regulate hypothalamic T3 level. TSH and Dio2 Convey the Circannual Information to Hypothalamic RF-Amides In seasonal rodents, the TSH-induced control of seasonal reproduction is associated with effects on two neuropeptides of the RF-amide family, RF-amide-related peptide (RFRP-3) and kisspeptin [6]. In seasonal breeders, Rfrp expression in the dorsomedial hypothalamic (DMH) and ventromedial hypothalamic (VMH) nuclei is higher in LP as compared to SP [4, 23, 24]. In male hamsters, RFRP-3 stimulates gonadotrophin and testosterone production [24, 25], whereas in ewes, the peptide may inhibit [26] or may not inhibit [27] gonadotrophin secretion. Kisspeptin, notably expressed in the arcuate nucleus (ARC), displays a potent and well-conserved stimulatory effect on gonadotrophin-releasing hormone (GnRH) neurons [28]. ARC kisspeptin neurons are the central site for the negative steroid feedback occurring in the breeding season, but the peptide expression is also inhibited by the SP melatonin signal [29 31]. As a consequence of this dual testosterone and melatonin inhibitory regulation, ARC kisspeptin expression displays seasonal variations but with species-specific differences [31, 32]. A central infusion of kisspeptin in hamsters and sheep [29, 33] or RFRP-3 in hamsters [25], kept in photoinhibitory conditions, is able to restore the reproductive function. Thus, RF-amide peptides, highly regulated by the melatonin-driven TSH, are critical for the control of seasonal breeding [34]. Next, we investigated how Rfrp in the DMH and VMH area and Kiss1 in the ARC (Figure 3E) are regulated in European hamsters kept in photoperiodic conditions and whether their expression is subjected to circannual variation. The number of Rfrp-expressing cells in the DMH and VMH area was found to be markedly larger in LP than in SP (p < 0.001), but, in LP, the number of cells appeared larger in the morning than at night (Figure S2C; p < 0.05 for overall time effect). Therefore, we further measured the level of total Rfrp mrna expressed in the DMH and VMH area at ZT02 and ZT22 under both photoperiods. At these times, total hypothalamic Rfrp mrna was markedly higher in LP than in SP (p < 0.001) but with no variation for time of sampling (Figure 3A; p = 0.08). In circannual conditions, Rfrp displayed a significantly higher expression in S-Sum as compared to S-Win, with no difference in Rfrp expression whether animals were castrated or not (Figure 3B; gonadally intact: p < 0.001; castrated: p < 0.01; p = 0.55 for overall testosterone effect), indicating that the circannual control of Rfrp expression is independent of testosterone feedback, as observed in its photoperiodic regulation in other species [4, 24]. The number of cells expressing Kiss1 in the medial region of the ARC was lower in LP than in SP (p < 0.001), with an additional daily variation with a higher value observed at ZT18 in LP (Figure S3D; p < 0.01). The analysis of total Kiss1 mrna levels in the medial ARC confirmed the photoperiodic difference, but no variation was found between ZT02 and ZT22 in either photoperiod (Figure 3C; p < for photoperiod; p = 0.57 for time of day). Similar to animals in photoperiodic conditions, total ARC Kiss1 mrna levels were significantly lower in the S-Sum compared to the S-Win state in gonadally intact animals (Figure 3D; p < 0.05). In castrated animals, the overall level of ARC Kiss1 mrna was increased compared to intact hamsters (Figure 3D). Moreover, the circannual variation in Kiss1 expression was opposite, with higher level in S-Sum than in S-Win (Figure 3D; p < 0.01). In the absence of testosterone negative feedback, this circannual variation was found to display a significant positive correlation with that of Rfrp mrna (p < 0.05; Pearson r = 0.57). The circannual variation in Kiss1 mrna observed in pinealectomized castrated hamsters suggests that the circannual signal, like the photoperiodic signal, regulates Kiss1 gene expression, but this effect is strongly altered by the negative testosterone feedback. Dio2 and Rfrp mrna levels were strongly correlated in the photoperiodic groups (Figure S3B; p < 0.001; Pearson r = 0.76). Similarly, there is a strong positive correlation between the circannual variation in Dio2 and Rfrp mrna levels (Figure 3F; p < ; Pearson r = 0.67), suggesting that the circannual clock signal is transmitted to Rfrp neurons in the DMH and VMH area via the changes in TSHb/Dio2. Similar mechanisms were previously described for the photoperiodic regulation of Rfrp expression in Syrian and Siberian hamsters [4, 6]. Altogether, our findings indicate that the circannual clock signal, acting through TSHb/Dio2, is integrated into the reproductive axis at the level of the hypothalamic RF-amide neurons. Previous studies have demonstrated that both kisspeptin and RFRP-3 increase GnRH neuron activity and gonadotrophin secretion in seasonal rodents [24, 25, 29]. Our observation of increased expression of Rfrp in sexually active LP-adapted and S-Sum European hamsters supports the idea that in this species, as well as in other hamster species [29, 30], RFRP-3 may serve as a seasonally regulated stimulatory neuropeptide acting upstream of the GnRH neurons to synchronize reproductive activity. The mode of action of RFRP-3 on GnRH neurons in European hamsters might be direct and/or via kisspeptin neurons as observed in other species [24, 25, 35]. Conclusions Our study demonstrates that in the European hamster, the circannual clock regulates TSHb expression in the PT. It is also well established that melatonin binds directly to its receptors on the PT to regulate TSHb expression [15]. Altogether, these data situate the PT in the spotlight for the interaction of different cues to time seasonal functions (Figure 4). The level of TSHb is strongly correlated to that of Dio2 in the tanycytes and, hence, probably to the local concentration of T3 (Figure 4). Our data indicate that changes in T3 may be relayed via RFRP and kisspeptin neurons to influence GnRH stimulation of pituitary gonadotrophins and gonadal activity (Figure 4). Additionally, we show that the seasonal change in testosterone production feeds back specifically at the kisspeptin neurons and not upstream on the circannual clock signaling (Figure 4). Furthermore, we recently reported that photoperiod, independently of melatonin, entrains circannual rhythms in European hamsters [36], and, therefore, the photoperiodic signal can

5 Current Biology Vol 24 No Figure 3. Photoperiodic and Circannual Variations in Rfrp and Kiss1 Gene Expression in European Hamster (A) Photoperiodic and day-to-night variation in Rfrp expression in the DMH and VMH area. Total Rfrp mrna level was measured at ZT02 (plain bars) and ZT22 (stripes) time points under LP or SP conditions. Representative images of Rfrp mrna labeling in LP and SP at ZT02 are shown. (B) Circannual variation in Rfrp expression in the DMH and VMH area. Total Rfrp mrna level measured in gonadally intact and castrated animals in S-Sum (white bars) and S-Win (black bars). Representative images of Rfrp mrna labeling in a gonadally intact S-Sum and a S-Win animal are shown. (C) Photoperiodic and day-to-night variation in Kiss1 expression in the mid-arc. Total Kiss1 mrna level was measured at ZT02 (plain bars) and ZT22 (stripes) time points under LP or SP conditions. Representative images of Kiss1 mrna labeling in LP and SP at ZT02 are shown. (D) Circannual variation in Kiss1 expression in the mid-arc. Total Kiss1 mrna level was measured in gonadally intact and castrated animals in S-Sum (white bars) and S-Win (black bars). Representative images of Kiss1 mrna labeling in gonadally intact (left) and castrated (right) S-Sum and S-Win animals are shown. All data show mean 6 SEM (n = 5 or n = 6 animals in the photoperiodic groups; n = 6 12 animals in the circannual groups). (E) Schematic drawing of a coronal section at the level of the mid-arc, depicting the areas where gene expression was studied, as shown in (A) (D). (F) Scatterplot showing the positive correlation between Dio2 expression in the PVZ and Rfrp expression in the DMH and VMH area in circannual conditions. r = 0.67, r 2 = 0.45, p < The following abbreviations were used: ARC, arcuate nucleus; au, arbitrary units; DMH, dorsomedial hypothalamic nuclei; VMH, ventromedial hypothalamic nuclei; LP, long photoperiod; SP, short photoperiod; S-Sum, subjective summer; S-Win, subjective winter; ZT, zeitgeber time. Scale bars of (A) (D) represent 100 mm; ***p < and **p < 0.01 for differences between groups. also directly synchronize the circannual clock through yetunknown mechanisms (Figure 4). Our study confirms that other biological functions, in addition to reproduction, are regulated by the circannual clock, like body weight [13] and temperature [37]. In chipmunks, a central circannual clock drives the hormonal signal for hibernation [38]. Our hypothesis is that TSHb/Dio2 represents a conserved signaling pathway through which the circannual clock regulates these functions too. Indeed, hypothalamic T3 implants [19] or a chronic infusion of TSH [6] restores the LP body weight phenotype of SP-adapted Siberian hamsters. Although the circannual control of biological functions is clearly established in some seasonal species, the anatomical localization of the circannual clock remains to be established. Our study on the European hamster has disclosed TSHb in the PT as the most-upstream neuroendocrine element regulated by the circannual clock so far investigated. TSH serves as a PT circannual output signal to the hypothalamic sites controlling seasonal reproduction and body weight. The PT has also been involved in the circannual control of prolactin secretion in sheep, although here, the mechanism is independent of the hypothalamus [39] and of thyroid hormones [40]. Thus, the PT emerges as the strongest candidate site for a circannual pacemaker so far identified, and future work should be aimed at understanding the molecular mechanisms behind circannual rhythm generation. A current hypothesis suggests that cyclical histogenesis may be important [41], whereas another hypothesis focuses attention on epigenetic mechanisms [42], with a combination of these processes being a third plausible scenario.

6 Gene Expression Underlying Circannual Reproduction 1505 Figure 4. Working Model Showing the Neuroendocrine Pathway through which Circannual and Photoperiodic Cues Regulate the Reproductive Axis in the European Hamster The circannual clock, like the photoperiodic melatonin signal, drives TSHb gene expression in the PT. The photoperiodic melatonin signal also entrains the circannual clock, but photoperiodic entrainment of the circannual clock can also happen via melatonin-independent pathways [36]. In turn, TSH, through TSH receptor (TSH-R), activates Dio2 expression in the tanycytes surrounding the third ventricle, leading to a local increase in T3 level. T3 is suggested to control the expression of Rfrp and, eventually, Kiss1, which in turn regulate GnRH neuronal activity. GnRH release represents the final step in the neural regulation of reproduction. The seasonal changes in circulatory sex steroids feed back specifically on kisspeptin neurons. This neuroendocrine pathway is likely regulating the seasonal variation in body weight, but the level at which this is taking place remains to be determined. Direct inputs are indicated with black arrows, and indirect or uncharacterized inputs are indicated with light gray arrows. ARC, arcuate nucleus; DMH, dorsomedial hypothalamic nuclei; VMH, ventromedial hypothalamic nuclei. Experimental Procedures Animals and Experimental Design Male European hamsters (Cricetus cricetus) were born and raised in our animal facilities. The animals were aged 11 months to 2 years and weighed g at the day of sacrifice. Animals were kept in individual type 3 Macrolon cages, in rooms with constant temperature at 20 C 6 2 C and humidity at 55% 6 5%. Food (Safe005 diet for rodents; SAFE) and water access was ad libitum. Animals were maintained in dim red light during the night phase. The study was conducted at the Chronobiotron (CNRS- UMS 3415) in accordance with the European Communities Council Directive of November 24, 1986 (86/609/EEC) and the French laws. Photoperiodic Investigation Seventy-two male European hamsters were kept in a controlled photoperiod, with 6 months of LP followed by 6 months of SP. The LP group was sampled after 3 months, under 16L:8D at six different time points along 24 hr. The sampling times in this group were ZT02, ZT06, ZT10, ZT14, ZT18, and ZT22, where ZT0 corresponds to lights-on time (04:00 CET). The SP group was sampled after 3 months under 10L:14D along 24 hr. The sampling times for this group were ZT02, ZT08, ZT12, ZT15, ZT18, and ZT22, where ZT0 corresponds to lights-on time (08:00 CET). These time points were chosen to be equally distributed around lights-on and lights-off times in both photoperiods (i.e., a time point 2 hr before and another 2 hr after each light change). Circannual Experiment A total of 54 animals, born in March 2010, were kept in LP conditions. All animals were pinealectomized, at 3 months of age, under anesthesia (intraperitoneal injections using 0.2 ml per 100 g body weight of a 4:1 mixture of Zoletil 20 [Virbac] and Rompun [Bayer HealthCare]) according to [36]. Additionally, an ibutton (Maxim) was implanted in the abdominal cavity for body temperature recordings. Thirty-four of these animals were also castrated. Body weight was assessed every 2 weeks, and, in gonadally intact animals, the reproductive state was assessed by measuring scrotal testicular length under short isofluorane anesthesia. The subjective seasonal state of the animals was determined according to the following parameters: (1) animals showing increasing body weight, which then stabilized after at least three consecutives measurements, and also presenting testes over 18 mm in length in the gonadally intact animals [36] were considered to be in S-Sum (Figure 1A), and (2) animals showing decreasing body weight for at least three consecutive measurements, accompanied by testes that regressed into the abdomen in the gonadally intact animals, were considered to be in S-Win (Figure 1B). Temperature data were used a posteriori to corroborate the subjective state of the animals: a stable temperature above the year s mean being summer-like temperatures (Figure 1A) and a stable temperature below the year s mean and/or phases of hypothermia being winter-like temperatures. Nonetheless, all the S-Win animals were sampled during euthermic phases (Figure 1B). In order to avoid any daily variation in the data, all animals were sampled between ZT07 and ZT08. Only the animals fitting these criteria without doubt and showing strong rhythms were selected for the study. Some animals were not selected for the study because they presented long rhythms with low amplitude, maintained constant testicular activity and high body weight, or never developed descended testes during the study (as was the case with one animal). Others were eliminated for developing agerelated diseases. We selected 35 out of 54 animals whose circannual state undoubtedly matched our criteria. Of these, all the S-Sum animals were in their first breeding season, whereas seven were taken during their first S-Win and eight during their second S-Win. No differences were found in hormone level or gene expression between animals during their first or second summer or winter; therefore, animals of both ages were included in the study, with confidence that the results of this study are not influenced by the age of the animals. Supplemental Information Supplemental Information includes Supplemental Experimental Procedures and three figures and can be found with this article online at org/ /j.cub Author contributions S.M., P.P., and V.S. designed this experiment; C.S.d.M., S.M., and V.S. performed animal experiments; C.S.d.M., J.B.-S., and M.-P.L.-C. performed the gene expression studies; C.S.d.M. analyzed data and made figures; C.S.d.M., S.M., V.S., and D.G.H. discussed the data and wrote the manuscript; and V.S. supervised this study. Acknowledgments The authors wish to thank Daniel Bonn and Olivier Arnaud for their expert animal care; Paul Klosen for providing probes for in situ hybridization and technical help; Manuel Tena-Sempere for the LH measurements; Jens Mikkelsen for the plasma testosterone assay; and Patrick Vuillez, Beatrice Bothorel, Dominique Sage-Ciocca, Christiane Calgary, and Sylviane Gourmelen for their help with the surgeries. This work was supported by grants from the CNRS, the Région Alsace, DREAL Alsace, the German Research Foundation (Mo 1742/1-1), and the German Wildlife Foundation. Received: March 26, 2014 Revised: May 2, 2014 Accepted: May 9, 2014 Published: June 26, 2014 References 1. Zucker, I. (2001). Circannual rhythms: mammals. In Circadian Clocks: Handbook of Behavioral Neurobiology, Volume 12, Takahashi, J., ed. (New York: Academic Press), pp Goldman, B.D. (2001). Mammalian photoperiodic system: formal properties and neuroendocrine mechanisms of photoperiodic time measurement. J. Biol. Rhythms 16,

7 Current Biology Vol 24 No Revel, F.G., Saboureau, M., Pévet, P., Mikkelsen, J.D., and Simonneaux, V. (2006). Melatonin regulates type 2 deiodinase gene expression in the Syrian hamster. Endocrinology 147, Revel, F.G., Saboureau, M., Pévet, P., Simonneaux, V., and Mikkelsen, J.D. (2008). RFamide-related peptide gene is a melatonin-driven photoperiodic gene. Endocrinology 149, Hanon, E.A., Lincoln, G.A., Fustin, J.-M.M., Dardente, H., Masson- Pévet, M., Morgan, P.J., and Hazlerigg, D.G. (2008). Ancestral TSH mechanism signals summer in a photoperiodic mammal. Curr. Biol. 18, Klosen, P., Sébert, M.E., Rasri, K., Laran-Chich, M.P., and Simonneaux, V. (2013). TSH restores a summer phenotype in photoinhibited mammals via the RF-amides RFRP3 and kisspeptin. FASEB J. 27, Pengelley, E.T., and Asmundson, S.J. (1974). Circannual rhythmicity in hibernating mammals. In Circannual Clocks: Annual Biological Rhythms, E.T. Pengelley, ed. (London: Academic Press), pp Canguilhem, B. (1989). External and endogenous control of body weight rhythm in the European hamster, Cricetus cricetus. In Living in the Cold, A. Malan and B. Canguilhem, eds. (London: John Libbey Eurotext), pp Karsch, F.J., Robinson, J.E., Woodfill, C.J.I., and Brown, M.B. (1989). Circannual cycles of luteinizing hormone and prolactin secretion in ewes during prolonged exposure to a fixed photoperiod: evidence for an endogenous reproductive rhythm. Biol. Reprod. 41, Gwinner, E., and Wozniak, J. (1982). Circannual rhythms in European starlings: why do they stop under long photoperiods? J. Comp. Physiol. 146, Lincoln, G.A., Johnston, J.D., Andersson, H., Wagner, G., and Hazlerigg, D.G. (2005). Photorefractoriness in mammals: dissociating a seasonal timer from the circadian-based photoperiod response. Endocrinology 146, Woodfill, C.J., Wayne, N.L., Moenter, S.M., and Karsch, F.J. (1994). Photoperiodic synchronization of a circannual reproductive rhythm in sheep: identification of season-specific time cues. Biol. Reprod. 50, Masson-Pévet, M., Naimi, F., Canguilhem, B., Saboureau, M., Bonn, D., and Pévet, P. (1994). Are the annual reproductive and body weight rhythms in the male European hamster (Cricetus cricetus) dependent upon a photoperiodically entrained circannual clock? J. Pineal Res. 17, Böckers, T.M., Niklowitz, P., Bockmann, J., Fauteck, J.D., Wittkowski, W., and Kreutz, M.R. (1995). Daily melatonin injections induce cytological changes in pars tuberalis-specific cells similar to short photoperiod. J. Neuroendocrinol. 7, Dardente, H., Klosen, P., Pévet, P., and Masson-Pévet, M. (2003). MT1 melatonin receptor mrna expressing cells in the pars tuberalis of the European hamster: effect of photoperiod. J. Neuroendocrinol. 15, Dardente, H., Wyse, C.A., Birnie, M.J., Dupré, S.M., Loudon, A.S.I., Lincoln, G.A., and Hazlerigg, D.G. (2010). A molecular switch for photoperiod responsiveness in mammals. Curr. Biol. 20, Ono, H., Hoshino, Y., Yasuo, S., Watanabe, M., Nakane, Y., Murai, A., Ebihara, S., Korf, H.-W., and Yoshimura, T. (2008). Involvement of thyrotropin in photoperiodic signal transduction in mice. Proc. Natl. Acad. Sci. USA 105, Lechan, R.M., and Fekete, C. (2005). Role of thyroid hormone deiodination in the hypothalamus. Thyroid 15, Barrett, P., Ebling, F.J.P., Schuhler, S., Wilson, D., Ross, A.W., Warner, A., Jethwa, P., Boelen, A., Visser, T.J., Ozanne, D.M., et al. (2007). Hypothalamic thyroid hormone catabolism acts as a gatekeeper for the seasonal control of body weight and reproduction. Endocrinology 148, Hanon, E.A., Routledge, K., Dardente, H., Masson-Pévet, M., Morgan, P.J., and Hazlerigg, D.G. (2010). Effect of photoperiod on the thyroidstimulating hormone neuroendocrine system in the European hamster (Cricetus cricetus). J. Neuroendocrinol. 22, Nakao, N., Ono, H., Yamamura, T., Anraku, T., Takagi, T., Higashi, K., Yasuo, S., Katou, Y., Kageyama, S., Uno, Y., et al. (2008). Thyrotrophin in the pars tuberalis triggers photoperiodic response. Nature 452, Sáenz de Miera, C., Hanon, E.A., Dardente, H., Birnie, M.J., Simonneaux, V., Lincoln, G.A., and Hazlerigg, D.G. (2013). Circannual variation in thyroid hormone deiodinases in a short-day breeder. J. Neuroendocrinol. 25, Smith, J.T., Coolen, L.M., Kriegsfeld, L.J., Sari, I.P., Jaafarzadehshirazi, M.R., Maltby, M., Bateman, K., Goodman, R.L., Tilbrook, A.J., Ubuka, T., et al. (2008). Variation in kisspeptin and RFamide-related peptide (RFRP) expression and terminal connections to gonadotropin-releasing hormone neurons in the brain: a novel medium for seasonal breeding in the sheep. Endocrinology 149, Ubuka, T., Inoue, K., Fukuda, Y., Mizuno, T., Ukena, K., Kriegsfeld, L.J., and Tsutsui, K. (2012). Identification, expression, and physiological functions of Siberian hamster gonadotropin-inhibitory hormone. Endocrinology 153, Ancel, C., Bentsen, A.H., Sébert, M.E., Tena-Sempere, M., Mikkelsen, J.D., and Simonneaux, V. (2012). Stimulatory effect of RFRP-3 on the gonadotrophic axis in the male Syrian hamster: the exception proves the rule. Endocrinology 153, Clarke, I.J., Sari, I.P., Qi, Y., Smith, J.T., Parkington, H.C., Ubuka, T., Iqbal, J., Li, Q., Tilbrook, A., Morgan, K., et al. (2008). Potent action of RFamide-related peptide-3 on pituitary gonadotropes indicative of a hypophysiotropic role in the negative regulation of gonadotropin secretion. Endocrinology 149, Caraty, A., Blomenröhr, M., Vogel, G.M.T., Lomet, D., Briant, C., and Beltramo, M. (2012). RF9 powerfully stimulates gonadotrophin secretion in the ewe: evidence for a seasonal threshold of sensitivity. J. Neuroendocrinol. 24, Pinilla, L., Aguilar, E., Dieguez, C., Millar, R.P., and Tena-Sempere, M. (2012). Kisspeptins and reproduction: physiological roles and regulatory mechanisms. Physiol. Rev. 92, Revel, F.G., Saboureau, M., Masson-Pévet, M., Pévet, P., Mikkelsen, J.D., and Simonneaux, V. (2006). Kisspeptin mediates the photoperiodic control of reproduction in hamsters. Curr. Biol. 16, Smith, J.T. (2009). Sex steroid control of hypothalamic Kiss1 expression in sheep and rodents: comparative aspects. Peptides 30, Ansel, L., Bolborea, M., Bentsen, A.H., Klosen, P., Mikkelsen, J.D., and Simonneaux, V. (2010). Differential regulation of kiss1 expression by melatonin and gonadal hormones in male and female Syrian hamsters. J. Biol. Rhythms 25, Wagner, G.C., Johnston, J.D., Clarke, I.J., Lincoln, G.A., and Hazlerigg, D.G. (2008). Redefining the limits of day length responsiveness in a seasonal mammal. Endocrinology 149, Caraty, A., Smith, J.T., Lomet, D., Ben Saïd, S., Morrissey, A., Cognie, J., Doughton, B., Baril, G., Briant, C., and Clarke, I.J. (2007). Kisspeptin synchronizes preovulatory surges in cyclical ewes and causes ovulation in seasonally acyclic ewes. Endocrinology 148, Simonneaux, V., Ancel, C., Poirel, V.J., and Gauer, F. (2013). Kisspeptins and RFRP-3 Act in Concert to Synchronize Rodent Reproduction with Seasons. Front Neurosci 7, Rizwan, M.Z., Poling, M.C., Corr, M., Cornes, P.A., Augustine, R.A., Quennell, J.H., Kauffman, A.S., and Anderson, G.M. (2012). RFamiderelated peptide-3 receptor gene expression in GnRH and kisspeptin neurons and GnRH-dependent mechanism of action. Endocrinology 153, Monecke, S., Sage-Ciocca, D., Wollnik, F., and Pévet, P. (2013). Photoperiod can entrain circannual rhythms in pinealectomized European hamsters. J. Biol. Rhythms 28, Monecke, S., Malan, A., and Pévet, P. (2011). Longterm temperature recordings in European hamsters. S. Monecke and P. Pévet, eds. Proceedings of the 18 th Meeting of the International Hamster Workgroup, Kondo, N., Sekijima, T., Kondo, J., Takamatsu, N., Tohya, K., and Ohtsu, T. (2006). Circannual control of hibernation by HP complex in the brain. Cell 125, Lincoln, G.A., Clarke, I.J., Hut, R.A., and Hazlerigg, D.G. (2006). Characterizing a mammalian circannual pacemaker. Science 314, Billings, H.J., Viguié, C., Karsch, F.J., Goodman, R.L., Connors, J.M., and Anderson, G.M. (2002). Temporal requirements of thyroid hormones for seasonal changes in LH secretion. Endocrinology 143, Hazlerigg, D.G., and Lincoln, G.A. (2011). Hypothesis: cyclical histogenesis is the basis of circannual timing. J. Biol. Rhythms 26, Stevenson, T.J., and Prendergast, B.J. (2013). Reversible DNA methylation regulates seasonal photoperiodic time measurement. Proc. Natl. Acad. Sci. USA 110,

8 Current Biology, Volume 24 Supplemental Information A Circannual Clock Drives Expression of Genes Central for Seasonal Reproduction Cristina Sáenz de Miera, Stefanie Monecke, Julien Bartzen-Sprauer, Marie-Pierre Laran- Chich, Paul Pévet, David G. Hazlerigg, and Valérie Simonneaux

9 Supplemental Figures Figure S1. Related to Figure 1. Circannual rhythms in pinealectomized castrated European hamsters (A, B) Graphs representing the expression of endogenous rhythms in body temperature (black line) and body weight ( ) in two different pinealectomized castrated European hamsters kept in constant conditions during the whole experiment (16L:8D, 20 ± 2ºC). Body weight (g) was measured every two weeks from the beginning of the experiment until the sampling date, when the animals expressed subjective summer (S Sum; A) or subjective winter (S-Win; B) physiologies. Body temperature ( C) was constantly recorded during the experiment and the data were assessed post-mortem.

10 Figure S2. Related to Figure 3. Photoperiodic and daily variation in reproductive hormones and RFamides in European hamster (A) Blood LH and (B) testosterone levels along 24h in LP (open squares) and SP (filled squares) conditions. Photoperiod effect, p< (A, B); Time of day effect, p<0.05 (B). (C) Number of Rfrp expressing neurons in DMH/VMH and (D) Kiss1 expressing neurons in the ARC along 24h in LP and SP conditions. Photoperiod effect, p< (C, D); Time of day effect, p<0.05 (C), p<0.01 (D). All data are mean ± SEM of 5 or 6 animals per sampling point. Zeitgeber time 0 (ZT0) is the lights on. Black horizontal bars represent the night phase for each photoperiodic regime.

11 Figure S3. Related to Figure 2. Correlation between photoperiodic variation in Rfrp, Dio2, and TSHß in European hamster Scatterplots showing the positive correlation found in photoperiodic conditions (A) between TSHß and Dio2 mrna levels; r= 0.65, r 2 = 0.43, p < 0.01; and (B) between Dio2 and Rfrp mrna levels; r= 0.76, r 2 = 0.58, p <

12 Supplemental experimental procedures Tissue collection Animals were culled by CO 2 saturation. Immediately after animals death blood was sampled by intracardiac punction and animal tissues were fixed by transcardiac perfusion first with a saline solution and then with 4% paraformaldehyde in 0.1M phosphate buffer (ph 7.4). Both testes were dissected and weighed. Brains were removed carefully from the skull, taking care not to damage the pituitary stalk, and immediately checked for correct pinealectomy, which was confirmed in all the animals. They were postfixed in the same fixative for 24h, dehydrated in serial ethanol and stored in butanol until embebbed in polyethyleneglycol [S1]. Brains were cut into serial 12 m thick coronal sections using a microtome (Leica Microsystems, Rueil- Malmaison, France). Sections were stored in acclimatised rooms until mounted. One in ten sections through the hypothalamus were mounted on SuperFrost ultraplus slides (Menzel-Glaser, Braunschweig, Germany) and stored at -80ºC until processed for in situ hybridization. Hormone Analysis Free testosterone in plasma was measured using a direct radioimmunoassay (RIA) kit [S2]. Serum LH level were determined in a volume of 50µl using a RIA kit, as described previously [S2]. Non-radioactive In situ hybridization Antisense rat probes were used for detection of Kiss1 expression [S3] and TSHß [S4]. Rfrp and Dio2 expression were analysed using antisense Siberian hamster probes [S5]. All probes were digoxigenin (DIG) labelled, according to the manufacturer s instructions (Roche, Meylan, France). On the day of hybridization, sections were postfixed in paraformaldehyde 4% in phosphate buffer for 10 minutes at room temperature, rinsed with PBS, treated with g/ml proteinase K (Roche, Meylan, France) for 30 minutes at 37ºC, rinsed with ice-cold paraformaldehyde 2% in phosphate buffer, rinsed again in PBS, acetylated twice for 10 minutes with 0.25% acetic anhydride in 100 mm triethanolamine and finally equilibrated in 5X salinesodium citrate (SSC) 0.05% Tween-20 twice for 5 minutes at room temperature. Hybridization was performed with 2 g/ml antisense probe in a medium containing 50% formamide, 5X SSC, 5X Denhardts solution, 0.1% Tween20 and 1 mg/ml salmon sperm DNA for 40h at 60ºC. High stringency washes were performed with 0.1X SSC 0.05% Tw20 at 72ºC to reduce nonspecific labelling. Hapten labelled riboprobes

13 were detected using an alkaline phosphatase-labelled antidigoxigenin antibody (Roche, Meylan, France). The alkaline phosphatase activity was detected with nitroblue tetrazolium and bromo-chloro-indolyl phosphate. After detection, slides were premounted using Crystalmount aqueous mounting medium (Sigma-Aldrich, Lyon, France) and mounted with Eukitt (Sigma-Aldrich, Lyon, France). Quantification Cell counting For Kiss1 and Rfrp mrna, labelled neurons were hand-counted on a Leica DMRB microscope (Leica Microsystems) by a person naive of the experimental groups. All Kiss1 expressing neurons were counted in 4 sections per animal, corresponding to the level of the mid ARC region. The number of Rfrp expressing neurons was counted in 12 sections per animal to cover the rostro-caudal hypothalamic expression of Rfrp. The total number of neurons expressing each gene per animal was calculated by multiplying the counted neurons by 10 (1 in 10 serial sections). Semiquantitative analysis All slides and images were processed in the same way to ensure identical conditions during the semiquantitative image analysis. Photos taken at 10X magnification using a Leica DMRB microscope (Leica Microsystems, Rueil-Malmaison, France) with an Olympus DP50 digital camera (Olympus France, Rungis, France) and digital images were captured using the camera software ViewfinderLite, (Olympus France, Rungis, France). Labelling intensity was calculated using IMAGEJ software (NIH Image, Bethesda MD, USA). For each slide, a background value was measured and was subtracted from the labelling measure. For Rfrp and Kiss1 the digital image analysis consisted of measurements of mean pixel grey level of the staining in an elliptical selection for each neuronal cell body. A minimum of 50 neurons per animal were measured. For each animal, the mean integrated density of each measured cells was multiplied by the total number of cells to obtain total mrna expression per animal. Dio2 and TSH quantification was made using segmented line tool of ImageJ. A 140µm long and 7µm wide line was used to cover Dio2 staining along the PVZ from the basal end of the 3V wall; and a 5µm line that covered the well delimited expression of Dio2 in the TIS. A line that covered all along the PT, adapted to the thickness of each section, was used for measuring TSH Three consecutive sections along the rostro-

14 caudal axis were measured per animal. The mean grey value in arbitrary units for each animal was calculated as the average of the three measurements. Statistical analysis All data are represented as mean ± SEM, with n=number of animals analysed. Data were assessed for normality using Kolmogorov-Smirnov test and homogeneity of variance using Bartlett s test. When data were not normally distributed they were square root or log transformed to fit this distribution and reanalysed. For the analysis of photoperiodic effects along the day, data were analyzed by two-way Analysis of Variance (ANOVA) with time of day and photoperiod as independent factors followed by post-hoc Bonferroni s tests when appropriate. To test the differences between subjective seasonal states we used t-tests. Pearson correlation analyses were used to evaluate relations between different genes expression in the circannual experiment. The threshold for statistical significance was set at p<0.05. All analyses and graphs were performed using GRAPHPAD PRISM version 5 (GraphPad software Inc., San Diego, CA, USA). Curves for body weight, testis size and body temperature data were design using SigmaPlot version 12 (Systat Software Inc., San Jose, CA, USA) Supplemental references S1. Klosen, P., Maessen, X., and van den Bosch de Aguilar, P. (1993). PEG embedding for immunocytochemistry: application to the analysis of immunoreactivity loss during histological processing. J. Histochem. Cytochem. 41, S2. Ancel, C., Bentsen, A. H., Sébert, M. E., Tena-Sempere, M., Mikkelsen, J. D., and Simonneaux, V. (2012). Stimulatory Effect of RFRP-3 on the Gonadotrophic Axis in the Male Syrian Hamster: The Exception Proves the Rule. Endocrinology 153, S3. Ansel, L., Bolborea, M., Bentsen, A. H., Klosen, P., Mikkelsen, J. D., and Simonneaux, V. (2010). Differential Regulation of Kiss1 Expression by Melatonin and Gonadal Hormones in Male and Female Syrian Hamsters. J. Biol. Rhythms 25, S4. Dardente, H., Klosen, P., Pévet, P., and Masson-Pévet, M. (2003). MT1 melatonin receptor mrna expressing cells in the pars tuberalis of the European hamster: effect of photoperiod. J. Neuroendocrinol. 15, S5. Klosen, P., Sébert, M. E., Rasri, K., Laran-Chich, M.-P., and Simonneaux, V. (2013). TSH restores a summer phenotype in photoinhibited mammals via the RF-amides RFRP3 and kisspeptin. FASEB J. 27,