Neural and neuroendocrine processing of a non-photic cue in an opportunistically breeding songbird

The physiological mechanisms allowing animals to time breeding appropriately in response to environmental cues have been the subject of research for almost a century. Recent studies of the onset of breeding in long-day photoperiodic breeders have focused on the role of type 2 iodothyronine deiodinase (DIO2) in the conversion of thyroxine (T4) to triiodothyronine (T3) and subsequent activation of the reproductive axis. In seasonally breeding Japanese quail, Coturnix japonica, long day lengths induce an increase in the expression of the thyroid stimulating hormone β-subunit (TSHβ) followed by an increase in local expression of DIO2 and a decrease in type 3 iodothyronine deiodinase (DIO3) in the hypothalamus (Yoshimura et al., 2003; Nakao et al., 2008). In this DIO model, the local photo-induced increase in expression of DIO2 and decrease in DIO3 is viewed as a ‘reciprocal switch’, causing an increase in locally available T3 in the hypothalamus. While it has not been demonstrated directly, this model suggests the increase in hypothalamic T3 causes a release of gonadotropin-releasing hormone (GnRH) from the hypothalamus, thus increasing expression and release of the gonadotropins follicle stimulating hormone (FSH) and luteinizing hormone (LH) from the pituitary and activating the gonads.

The reciprocal switching of DIO2 and DIO3 is hypothesized to be a conserved mechanism of reproductive activation in all vertebrates (Nakane and Yoshimura, 2014), though this has only been tested in photoperiodic breeders (Yoshimura et al., 2003; Revel et al., 2006; Hanon et al., 2008). If the DIO system is an evolutionarily conserved mechanism that regulates GnRH release then it is likely that it responds to cues other than day length, given that multiple cues can influence GnRH release. Species that do not respond reproductively to changes in day length and those that live in areas where day lengths do not differ throughout the year could instead activate the DIO system to cause a release in GnRH in response to stimulatory cues such as availability of mates, food or nesting sites.

Non-photoperiodic breeders that breed whenever conditions are favorable are termed opportunistic breeders (Dawson et al., 2001). To time breeding appropriately, these species – such as the zebra finch, Taeniopygia guttata – respond to and integrate proximate cues of food and water availability, mating opportunities and other resources to time reproduction (Perfito et al., 2007, 2008; Zann et al., 1995; Hahn, 1995). The mechanism(s) underlying reproductive activation in response to food, water or mate availability remain(s) unknown; thus, we sought to investigate a potential role for the DIO system in reproductive activation in this species.

The hypothalamic neuropeptides GnRH and gonadotropin-inhibitory hormone (GnIH) are two likely key integrators of environmental stimuli for reproduction. As such, they are also likely to interact with the DIO system to control breeding. While breeding zebra finches do not differ from non-breeders in hypothalamic immunoreactivity of GnRH and GnIH (Perfito et al., 2006, 2011), the DIO system could be acting to regulate GnRH release when breeding commences. This, together with the proposed role of DIO in the activation of breeding in photoperiodic species, led us to test whether the DIO, GnRH and GnIH systems respond to a social stimulus in zebra finches. Specifically, we sought to test whether there is a reciprocal switch between DIO2 and DIO3 expression following social stimulation of an opportunistic breeder, similar to that seen in response to long days in a photoperiodic species (Yasuo et al., 2005). When males were housed in isolation and deprived of social cues, we expected to see a decrease in DIO2 expression, an increase in DIO3 expression and decreases in GnRH, LH and testosterone. Following presentation of a potential mate to these isolated males, we expected to observe an increase in DIO2 expression, a decrease in DIO3 and GnIH expression and increases in GnRH, LH and testosterone. We expected to see these changes in association with an increase in reproductive behaviors in response to a new female and sought to characterize activity of the DIO system in a social paradigm.

Adult male zebra finches were bred and housed in mixed-sex free-flight aviaries at the Field Station for the Study of Behavior, Ecology, and Reproduction at the University of California, Berkeley. Birds are considered adult at 90 days post-hatch. Birds were exposed to natural changes in day length that were supplemented with full-spectrum artificial light to create a minimum 12 h light:12 h dark photoperiod. Birds were supplied with water and German millet mixed with canary seed ad libitum. Food was supplemented with cuttlebone, grit and lettuce weekly. All animal care and procedures were approved by the University of California Office of Laboratory Animal Care and conducted in accordance with local animal welfare laws and policies.

Adult male zebra finches were caught from the colony between 10:00 h and 11:00 h and transferred to an isolation cage (18×8×12 in) inside a custom-made temperature- and light-controlled insulated box. Within each box, males did not receive any visual or auditory input from the colony or adjacent isolation boxes. For the 2 days of isolation, males were maintained on the 12 h light:12 h dark photoperiod they had experienced in the colony setting and had ad libitum access to water and German millet mixed with canary seed.

For experiment 1, after the 2 days of isolation, birds (N=8 per group) were randomly assigned to one of three treatments: (1) maintained in isolation (isolated), (2) presented with a novel female for 30 min (female stimulus) or (3) presented with a novel male for 30 min (male stimulus). Behavior was recorded for 30 min before and 30 min after stimulus presentation for all treatments. The novel stimulus birds came from separate colony rooms, ensuring that the experimental males had had no previous experience with these stimulus animals.

After treatment, experimental males were deeply anesthetized with isoflurane (Phoenix Pharmaceuticals Inc., Burlingame, CA, USA) and rapidly decapitated, at which time trunk blood was collected. Brain and testis tissue were frozen immediately on dry ice. Heads with pituitary tissue intact were frozen on dry ice and the pituitaries were subsequently extracted under a dissection microscope.

For experiment 2, the above protocol was followed with N=6 birds in each group. The brains were fixed in a 4% paraformaldehyde solution (PFA) for 3 days before being cryoprotected in 30% sucrose in 0.1 mol l⁻¹ phosphate-buffered saline (PBS) and frozen.

Trunk blood was spun in a centrifuge at 4°C at 1500 g for 10 min to separate blood plasma. Plasma was stored at −80°C before assay for LH, testosterone and corticosterone. Pituitaries were extracted from the skull and placed into 1 ml of TRIzol reagent (Invitrogen by Life Technologies, Grand Island, NY, USA) and homogenized. One testis from each bird was placed into 1 ml of TRIzol and homogenized.

Brains were cut into 40 µm-thick slices on a cryostat (cm3050s, Leica Microsystems, Buffalo Grove, IL, USA) and mounted directly onto slides. Brain tissue was mounted onto slides beginning with the appearance of the tractus septomesencephalicus, a neuroanatomical landmark anterior to the hypothalamus, and all slices were mounted until the appearance of the cerebellum. Then, 3 mm punches were taken through the hypothalamus from alternating sections and placed in TRIzol. All sections were mounted on slides for subsequent immunocytochemistry (ICC). The brains put in PFA immediately after collection were also cut into 40 µm-thick slices, but were stored in antifreeze prior to ICC. Prior to freezing, the length and width of the testes were measured to assess fresh testis volume. Volume was calculated as V=4/3πa²b, where a is half the width and b is half the length (long axis) of the testis.

RNA was extracted from hypothalamus, pituitary and gonadal tissue using chloroform (as described in Perfito et al., 2012). RNA extracts were treated with a DNAse (Invitrogen by Life Technologies) to digest any single- and double-stranded DNA and subsequently reverse transcribed using M-MLV reverse transcriptase (Promega Corporation, Madison, WI, USA) to create cDNA from each tissue.

Real-time quantitative PCR (qPCR) on cDNA obtained from hypothalamic punches was run for a number of genes: DIO2, DIO3, GnRH, GnIH, and EGR-1, an immediate-early gene (IEG). This IEG was taken as an indicator of general activation of neurons (Hoffman et al., 1993; Morgan and Curran, 1989, 1995). The reference gene 18S was used as a control in hypothalamic, pituitary and gonadal tissue as its expression did not change with treatment in any tissue. cDNA was diluted 1:25 with water treated with diethylpyrocarbonate. Primers were designed for each gene from previously published sequences in the zebra finch genome (see Table 1 for accession numbers) using Primer3 software (simgene.com). qPCR was performed following the manufacturer's instructions for SYBR green reagent (Applied Biosystems by Life Technologies). Post-qPCR products were cloned, sequenced, and compared with GenBank to confirm identification of target genes. Raw fluorescence data were analyzed with the RT-PCR Miner program (Zhao and Fernald, 2005) and cycle thresholds (C_t) were obtained from this program. Expression values were calculated as 1/(1+E)^C_t, where E is the average PCR efficiency for the gene of interest as calculated by a standard curve. Data are shown as fold-change, which was calculated by dividing the expression of the gene of interest (corrected for the reference gene 18S) by the average expression for the isolated group.

ICC was performed on a series of brain tissue to label GnRH and EGR-1 proteins. One series of brains from experiment 1 (mounted on slides) was single-labeled for GnRH and another series was double-labeled for GnRH and EGR-1. EGR-1 single-label ICC was conducted on one series of brain tissue from experiment 2 (free-floating brain sections) that had been fixed at the time of collection. For GnRH single-labeling and GnRH/EGR-1 double-labeling, slides were fixed in 4% PFA for 20 min and washed in 0.1 mol l⁻¹ PBS (pH 7.4) three times. Slides were then incubated in 0.03% hydrogen peroxide in methanol for 30 min followed by three PBS washes and 1 h of incubation in 2% normal goat serum (Vector Laboratories, Burlingame, CA, USA) in 0.2% PBS with Triton X-100 (PBS-T). Slides were incubated in GnRH primary antibody (HU60, a gift from Dr Henryk Urbanski, Portland, OR, USA; 1:5000 in PBS-T) for 1 h at room temperature and subsequently for 48 h at 4°C. After incubation in the primary antibody, slides were washed with PBS, incubated in the secondary antibody biotinylated goat anti-rabbit (1:250 in 0.2% PBS-T, Vector Laboratories) for 1 h, washed in PBS-T, and incubated in avidin/biotinylated enzyme complex [PK-6100, Vectastain Elite ABC Kit (Standard), Vector Laboratories] as per the manufacturer's instructions. Slides were washed in PBS and incubated with DAB peroxidase substrate kit (SK-4100, Vector Laboratories). For double-labeling, slides were then incubated in the primary antibody for EGR-1 (c-19, sc-189, Santa Cruz Biotechnology, Santa Cruz, CA, USA; 1:1000 in 0.2% PBS-T) for 24 h. Slides were then washed and incubated as described above with biotinylated goat anti-rabbit secondary antibody, then ABC, then DAB with nickel sulfate. For EGR-1 single-labeling, all steps were the same as described above on free-floating tissue, then brain tissue was mounted onto slides. All slides were then dehydrated with graded ethanol, cleared in xylenes and coverslipped with Permount (SP 15-100, Fisher Scientific Inc., Waltham, MA, USA).

Cell counts for all immunocytochemistry were obtained by a researcher unaware as to the experimental treatment. All cells immunoreactive (-ir) for GnRH were counted for each animal and assessed for co-localization with EGR-1. Cells single-labeled for EGR-1 were counted at 20× magnification centered on the midline, with the base of the brain just visible in the microscope field from the pre-optic area through the appearance of the cerebellum. All EGR-1-ir cells in this space were counted for each individual and averaged over the total number of sections counted for analysis.

Concentrations of testosterone and corticosterone in plasma were measured using enzyme immunoassay kits (EIA, Enzo Life Sciences, Farmingdale, NY, USA). These kits have been validated and optimized for zebra finches (Lynn et al., 2007, 2010). Protocols for both hormones were followed as specified by Lynn et al. (2010). Plasma was diluted 1:40 for testosterone and corticosterone EIA. All samples were run in duplicate on one plate per hormone, with a standard curve also in duplicate for comparison.

Plasma was assayed in 15 μl duplicate samples in a single assay using a micro-modification of the radioimmunoassay (RIA) originally devised by Follett and colleagues (1975). Intra-assay coefficient of variation was 3.4% and the lower detection limit was 0.07 ng ml⁻¹.

Video recordings were taken before and after presentation of the stimulus animal with security cameras (SWADS-120CAM-US, Swann Communications, Santa Fe Springs, CA, USA) and recorded with a 4-channel DVR (Q-See, Anaheim, CA, USA). Animals were recorded for 30 min prior to stimulus presentation and 30 min during stimulus presentation. Individuals receiving no stimulus animal had their cages opened and handled in a similar fashion to those receiving a stimulus animal. Behavior was recorded for each individual animal by counting copulation attempts and number of songs and by timing the duration of total time resting. Activity levels were calculated by subtracting the time resting from the total time of recording. The number of songs was determined by counting song bouts with five or more seconds between each song. Because of the low number of songs, directed (towards the stimulus animal) and undirected (with no obvious intended recipient) songs were combined and taken as a measure of total song activity. The first 5 min of each video were not included in the analysis to allow the animals time to adjust to the experimenter opening the cage. Video analysis was conducted by researchers with no explicit knowledge of the experimental treatment of the animal or experimental design. Video analysis could not be completed entirely blind, as the researcher analyzing the video could also see the stimulus animal. The researcher was given 30 min videos to watch, and was not told whether it was a video of the animal prior to stimulus or a control animal that received no stimulus.

One-way ANOVA tests were performed for immunoreactive cell counts, gene expression as measured by qPCR, and hormone concentrations from RIA and EIA tests. One-way ANOVA tests were also performed for behavioral measures of total time active, copulation attempts and number of songs. Significance was set at P<0.05. In cases of statistical significance, Tukey's test of multiple comparisons was used to assess which groups were significantly different. All statistical analyses were performed using Prism 6 (GraphPad, La Jolla, CA, USA).

Experiment 1 had 8 animals per group, though there are some data sets where all animals could not be used for analysis. For circulating plasma hormones, RIA for LH was performed first, followed by EIAs for testosterone and corticosterone. Thus, there was not enough plasma collected from some birds to run all three assays. For qPCR analysis, samples with expression levels 3 or more standard deviations above the mean were considered outliers and were excluded from analysis. This occurred for two samples for expression of DIO3, one from the male stimulus group and one from the female stimulus group, giving a sample size of 8 for the isolated group and 7 each for the male stimulus and female stimulus groups.

Hypothalamic expression of DIO2 and DIO3 did not differ across groups (Fig. 1A,B; ANOVA, DIO2: F_2,21=0.264, P=0.599, DIO3: F_2,19=0.624, P=0.547). Expression of GnRH and GnIH was also similar across groups (Fig. 1C,D; ANOVA, GnRH: F_2,21=1.907, P=0.173, GnIH: F_2,21=2.321, P=0.123). Expression of the housekeeping gene 18S did not differ across groups (F_2,21=0.117, P=0.256). EGR-1 expression was greater in the group receiving a female stimulus as compared with the isolated group (ANOVA, F_2,21=3.507, P<0.049), while the group exposed to a male stimulus had an intermediate level of EGR-1 expression (Fig. 2A). Tukey's multiple comparisons test produced a mean difference of −1.035 between the isolated group and the group receiving a female stimulus, indicating a significant difference between these two groups. Pituitary expression of the β-subunit of FSH (FSHβ) did not differ across groups (ANOVA, F_2,21=1.080, P=0.358). GnIH expression in the gonads was also not significantly different between groups (ANOVA, F_2,21=1.528, P=0.240).

Analysis of immunocytochemistry revealed no significant differences in the number of GnRH-immunoreactive (ir) cells between groups (F_2,21=0.606, P=0.555; Fig. 4). In sets of tissue double-labeled for EGR-1 and GnRH, EGR-1 was not found to be co-localized with GnRH-ir cells. Analysis of single-label EGR-1 ICC showed that number of EGR-1 labeled cells was significantly greater in the hypothalamus in the female stimulus group as compared with the isolated group and the male stimulus group (F_2,21=6.686, P<0.04, ANOVA; Figs 2B and 3).

EIA revealed circulating testosterone was significantly higher in males receiving a male stimulus and in those receiving a female stimulus as compared with isolated males (Fig. 5A; multiple comparison t-tests with Sidak–Bonferroni correction, P<0.012 for isolated versus male stimulus, P<0.021 for isolated versus female stimulus). The male stimulus group did not differ in circulating testosterone level from the female stimulus group (multiple comparison t-tests with Sidak–Bonferroni correction, P=0.848).

LH did not differ across groups (Fig. 5B; F_2,21=1.958, P=0.166), nor did circulating corticosterone (Fig. 5C; F_2,14=0.555, P=0.586). The RIA was run first followed by the EIAs for testosterone and corticosterone; thus, plasma was not available from all birds for measurement of testosterone (total of 8 samples in the isolated and female stimulus groups and 7 in the male stimulus group) and corticosterone (total of 6 samples for isolated and female stimulus groups and 5 for male stimulus group). Testis volume was not significantly different across groups (F_2,21=0.985, P=0.211).

Prior to presentation of a stimulus, animals in all three groups had similar levels of activity (defined as time resting subtracted from the total time; F_2,21=0.212, P=0.81). The isolated and male stimulus groups maintained levels of activity similar to those of their baseline recording, while males presented with a female were more active (F_2,21=14.42, P<0.0002; Fig. 6A). Males in the presence of a female also showed significantly more copulation attempts (F_2,21=14.11, P<0.0001; Fig. 6B) and a greater number of songs than isolated males and those in the presence of a male (F_2,21=4.74, P<0.02, ANOVA; Fig. 6C).

In this experiment, we sought to test whether social cues activate the DIO system in male zebra finches. If the mechanism of DIO-induced GnRH release is evolutionarily conserved across vertebrate taxa, and if there is no separate or additional mechanism for regulating GnRH release, we would expect DIO2/DIO3 expression in non-photoperiodic species to respond to stimulatory cues other than photoperiod. We predicted that the hypothalamic-pituitary–gonadal axis would be activated in male zebra finches that were presented with a female after isolation. Thus, we also predicted that there would be reciprocal switching of DIO2 and DIO3 expression in the hypothalamus, precipitating other indicators of reproductive activation (increased GnRH expression, increased circulating LH or testosterone, increased reproductive behaviors). In our non-photoperiodic model and within this experimental paradigm, no changes in the DIO system were observed with a social stimulus.

Males receiving a female stimulus showed increased levels of activity, song behavior and copulation attempts compared with isolated birds or those receiving a male stimulus, suggesting males were responding to females with appropriate reproductive effort. Males with a female stimulus also showed a physiological response to these females in the form of an increase in EGR-1 expression and EGR-1 protein immunoreactivity in the hypothalamus. It is unknown, however, what cell type is being activated in response to the social stimulus. In double-label immunocytochemistry, EGR-1 was not found to be co-localized with GnRH. While this indicates that GnRH-ir cells might not be directly activated with presentation of a female, EGR-1-ir cell count in the hypothalamus did increase, including in the pre-optic area where GnRH cells are located. A similar pattern in IEG-ir cells surrounding but not co-localized with GnRH cells was observed in white-crowned sparrows following treatment with NMDA (Meddle et al., 1999). In that study, increased c-FOS immunoreactivity was associated with an increase in LH release, suggestive of GnRH release. While the activation of hypothalamic cells shown with an increase in EGR-1 indicates a hypothalamic response to the female stimulus, theoretically this increase in expression could be associated with the change in behavior seen with the presence of a female as opposed to an increase in activity in the reproductive axis.

Males exhibited a behavioral response to the presence of a female, but other reproductive parameters usually associated with reproductive activation, including increased GnRH expression and increased circulating LH, did not change. Likewise, no changes in the expression of DIO2 or DIO3 were found. While 30 min is considered sufficient time for changes in expression to be seen, it is possible that dynamic changes of DIO2, DIO3 and GnRH expression were missed because of the analysis of a single time point. Experiments inducing DIO2 expression in photoperiodic breeders with a photic stimulus show great variation in the timing of DIO2 expression after stimulus presentation. Japanese quail show a change in expression a few hours into the first long day (Yoshimura et al., 2003), while Syrian hamsters show increases in DIO2 expression 8 days after transfer from short to long day lengths (Yasuo et al., 2010). DIO2 expression can be stimulated by social cues in photoperiodic breeders several days after stimulus presentation (Perfito et al., 2015), and thus an experiment with many time points after social stimulation would be required to provide a definitive conclusion as to whether the DIO system responds to the presence of a potential mate in zebra finches. Alternately, a time-course study of zebra finches could show an increase in GnRH expression and release independently from changes in DIO2 expression, as was found in European starlings (Bentley et al., 2013). Testosterone levels were significantly higher in males receiving a female stimulus than in isolated males. Testosterone levels were also higher in males receiving a male stimulus, indicating an increase in testosterone with all social stimuli, not just female stimuli.

In summary, this study provides evidence that male zebra finches respond to potential mates with increased reproductive effort and increased activity of the hypothalamus. The lack of changes in DIO2, DIO3, GnIH and GnRH expression across groups indicate that either these components of the reproductive axis are not influenced by social cues in this species or our experimental time course was not sufficient to reveal the influence of social cues on reproductive activation. If the DIO system is truly an evolutionarily conserved mechanism in the initiation of vertebrate reproduction, then further experiments over different time frames will elucidate changes in the DIO response to social cues.

We thank members of the Bentley lab, especially S. Hsia, for assistance in tissue collection and laboratory work. We thank John Wingfield (UC Davis) for the use of his laboratory to perform the LH assay.