Effects of Melatonin on Peripheral Reproductive Function: Regulation of Testicular GnIH and Testosterone

Study of seasonal reproduction has focused on the brain. Here, we show that the inhibition of sex steroid secretion can be seasonally mediated at the level of the gonad. We investigate the direct effects of melatonin on sex steroid secretion and gonadal neuropeptide expression in European starlings (Sturnus vulgaris). PCR reveals starling gonads express mRNA for gonadotropin inhibitory hormone (GnIH) and its receptor (GnIHR) and melatonin receptors 1B (Mel 1B) and 1C (Mel 1C). We demonstrate that the gonadal GnIH system is regulated seasonally, possibly via a mechanism involving melatonin. GnIH/ GnIHR expression in the testes is relatively low during breeding compared with outside the breeding season. The expression patterns of Mel 1B and Mel 1C are correlated with this expression, and melatonin up-regulates the expression of GnIH mRNA in starling gonads before breeding. In vitro, GnIH and melatonin significantly decrease testosterone secretion from LH/FSH-stimulated testes before, but not during, breeding. Thus local inhibition of sex steroid secretion appears to be regulated seasonally at the level of the gonad, by a mechanism involving melatonin and the gonadal GnIH system.

In photoperiodic vertebrates, the annual cycle of reproduction is controlled principally by changing day length, a highly predictive cue of seasonal resource abundance. In all diurnal vertebrates studied, melatonin is synthesized in and secreted from the pineal gland during night (1) such that the timing and duration of melatonin secretion provide an endocrine proxy of changes in day length (2).

In mammals, melatonin stimulates reproductive physiology and activity in short day breeders and inhibits reproductive physiology and activity in long-day breeders. Consequently, manipulating the duration of melatonin secretion experimentally causes changes in reproductive physiology in photoperiodic mammals. For example, in pinealectomized (Px) sheep, which are short-day breeders, long durations of melatonin administration caused early onset of estrus in ewes and significantly higher FSH, testosterone, and testes volume in rams. In contrast, short-duration melatonin infusion to Px sheep inhibited reproduction (3). In Px Syrian and Djungarian hamsters, which are long-day breeding species, the opposite effects are seen. Long-duration melatonin infusions caused involution of the gonads, whereas short-duration melatonin infusions stimulated reproductive development (4).

In photoperiodic birds, the relationship between melatonin signal and reproductive timing is more complex than in photoperiodic mammals. In seasonally breeding birds, a prolonged exposure to short days induces a physiological state termed “photosensitivity” (5). When photosensitive birds experience a critical day length (for European starlings, 11.5 h), the hypothalamic-pituitary-gonadal axis becomes stimulated, and gonadal recrudescence and initiation of reproductive behaviors occurs. They are thus photostimulated by long days (6). After several weeks of photostimulation by long days, these birds become photorefractory (6), which means that the reproductive system becomes inactive, the gonads regress, and reproductive behaviors cease even while days are still increasing in length (7). Because melatonin provides a measure of changing day length, and day length is a primary cue in mediating the transition to photostimulation (from photosensitivity), it seems likely that melatonin could be involved in this event. For example, male Baya weavers, a seasonally breeding species, had an increased rate of photoperiodically induced testicular growth after pinealectomy (8). In another study, injection of antimelatonin just before night caused gonadal development in photosensitive quail (9). However, seasonally breeding birds also respond to photoperiodic cues via encephalic photoreceptors (10, 11), thus circumventing the need for absolute control of photoperiod detection by melatonin. In fact, removing the sources of melatonin (pinealectomy with bilateral enucleation) caused no appreciable effects on the photoperiodic response in American tree sparrows (10). In another study, injecting melatonin into quail exposed to long day lengths (before dusk or at dawn, to extend subjective night) did not inhibit long day-induced gonadal growth (12), providing further evidence that the relationship between melatonin and reproductive timing is complex and not direct. Furthermore, the transition to photorefractoriness occurs while days are increasing (and thus while melatonin duration is continuing to decrease). Thus it is unlikely that melatonin is directly involved in mediating the regression of the reproductive system that occurs in the summer.

More recent evidence suggests that although melatonin is not the sole regulator of photo-induced reproductive cycles in birds, it can exert inhibitory effects on avian reproductive cycles via the hypothalamic neuropeptide gonadotropin-inhibitory hormone (GnIH). GnIH inhibits the synthesis and secretion of LH and FSH from the pituitary gland and is thought to act on GnRH neurons directly (13–15). Melatonin receptor 1C (Mel 1C) mRNA is expressed by GnIH-ir neurons (16), and melatonin causes hypothalamic GnIH release in quail (17). Pinealectomy combined with orbital enucleation (Px+Ex), which removes the major sources of melatonin in quail, decreases the expression of GnIH in the diencephalons, whereas subsequent melatonin administration rescues these effects (16). Thus in birds, melatonin appears to act directly on GnIH neurons via its receptor to induce hypothalamic GnIH expression and release. The potential exists, therefore, for photoperiodic regulation of the avian reproductive axis via action of the changing melatonin signal on the GnIH system.

Melatonin may also act peripherally to control reproductive physiology. The concentration, duration, and phase of plasma melatonin follows that of the pineal (4). Thus peripheral organs are also provided with a hormonal measurement of day length. Melatonin appears to mediate the observed seasonal changes in immune function in the thymus, spleen, and bursa of Fabricius in conjunction with gonadal steroids in birds and mammals (18). Cell-mediated immunity can also be modulated by melatonin administration in birds (19). Thus, peripheral tissues in birds are able to detect and respond to changes in the melatonin signal.

In mammals, the testes respond directly to melatonin. Intratesticular injection of melatonin decreased androgen receptor and androgen-binding protein expression, 3β-hydroxysteroid dehydrogenase (3β-HSD) activity, testosterone release, spermatogenesis, and testes weight, while increasing melatonin 1a receptor immunoreactivity on Leydig cells in Indian palm squirrels (20). Expression of mRNA for the Mel1a receptor subtype was detected in isolated Syrian hamster Leydig cells (21). In addition, melatonin administration to long-day (breeding) hamsters significantly reduced human chorionic gonadotropin-stimulated testosterone and cAMP production and decreased the expression of steroidogenic acute regulatory protein, P450 side chain cleavage, 3β-HSD, and 17β-HSD (21). Similarly, melatonin reduced LH-stimulated testosterone secretion from Djungarian hamster Leydig cells (22).

Avian gonads also have the potential to respond to plasma melatonin, but the effect on gonadal activity is not well studied. In galliform birds (chicken, duck, quail), the testes and ovary bind 2-[¹²⁵I]iodomelatonin, and the melatonin-binding site appears to be coupled to a guanine nucleotide-binding protein second messenger system (23). The membrane fraction of hen granulosa cells binds 2-[¹²⁵I]iodomelatonin, and LH-stimulated progesterone production is markedly reduced with melatonin coincubation (24).

Along with melatonin receptor, GnIH and its receptor (GnIHR) are expressed in avian testes and ovaries in addition to the brain (25). GnIH reduces testosterone secreted from cultured testes in a dose-dependent manner (25). As already mentioned, there is a strong relationship between melatonin and GnIH release in the brain. The regulation of the gonadal GnIH system is, as yet, unknown, and evidence for the influence of melatonin and/or GnIH on the annual reproductive cycle is sparse. Due to the inhibitory actions of melatonin on the gonads of long-day breeding mammals, the expression of melatonin receptors in the gonads of birds and the interaction of melatonin and GnIH in the avian brain, we predicted that melatonin could directly inhibit gonadal function in long-day breeding birds. Because these birds have seasonal cycles of reproduction and the melatonin signal differs with season, we hypothesized that this inhibition was mediated seasonally, through the expression and action of the gonadal GnIH system.

Materials and Methods

All procedures were performed in accordance with federal and state laws and with appropriate agreements from the University of California Berkeley Office of Laboratory Animal Care. European starlings (Sturnus vulgaris) housed under natural photoperiods in outdoor aviaries in Berkeley were captured in mist nets and euthanized on six separate days from 2008–2011. All birds were terminally anesthetized using isoflurane within 3 min of capture. Tissues were removed after decapitation. A summary of photoperiod lengths and starling breeding states for each date can be found in Fig. 1. Birds were classified as photosensitive, photostimulated, or photorefractory following previously established parameters for European starlings (5, 6).

Fig. 1.

Summary of photoperiod lengths and starling breeding states for the six sampling dates of European starlings. Starlings were collected from outdoor aviaries at 0011 h on each date. Average volume of testes is reported in mm³ (black line) Photoperiod length is reported in hours of daylight (gray line). On October 1, 2008, starlings were exposed to a prolonged period of short days after photorefreactoriness and were thus characterized as photosensitive [day length (DL): 11h, 46 min, average testes volume (ATV): 8.6 ± 2.3 mm³]. On April 1, 2009, day length had exceeded 11.5 h (DL: 12 h, 38 min), so starlings were characterized as photostimulated (ATV: 445.9 ± 48.0 mm³). On June 19, 2009, starlings had been exposed to prolonged long days (via natural photoperiod, DL: 14 h, 46 min) and their gonads were regressing (ATV: 65.2 ± 13.8 mm³). On February 5, 2010, starlings were exposed to a prolonged period of short days after photorefractoriness and had not yet experienced critical day length for the onset of photostimulation (6,7). They were thus characterized as photosensitive (DL: 10 h, 27 min; ATV: 76.0 ± 22.5 mm³). On April 7, 2010, day length had exceeded 11.5 h (DL:12 h, 52 min), so starlings were characterized as photostimulated (ATV: 502.0 ± 168.9 mm³). Similar to February 2010, starlings were photosensitive on February 4, 2011 (DL): 10 h, 25 min; ATV: 158 ± 64.3 mm³). On this date, one male had a testis volume of 807.9 mm³. The ATV excluding this male (n = 9) was 70.4 ± 18.9 mm³. Day lengths plotted were recorded on the first of each month. The photosensitive period is delineated by the light gray boxes. The photostimulated period is delineated by the dark gray boxes. The photorefractory period is delineated by the white boxes.

Expression of GnIH, GnIHR, and Mel 1B and 1C in the testis cDNA to determine the expression of GnIH, GnIHR, and Mel 1B and Mel 1C was synthesized from the testes of four male European starlings collected on October 1, 2008 (photoperiod, 11 h, 46 min; Fig 1). Tissues were immediately removed, placed on dry ice, and stored at −80 C. Partial precursors to GnIH, GnIHR, Mel 1B, and Mel 1C were amplified from 1 μg cDNA by PCR using primers based on European starling (S. vulgaris) and chicken (Gallus gallus) cDNA sequences [GenBank accession nos. EF486798 (GnIH), EF212891 (GnIHR), NM001048258 (Mel 1B), U31821 (Mel 1C)]. Primers are as follows: GnIH forward, 5′-GGAAGAAAAGCAGAGGAGTCTC-3′; reverse, 5′-TGGAGATCTCCCAAGCCTGT-3′; GnIHR forward, 5′-TCCTGGCCTACACCTTCATCT-3′; reverse, 5′-AGATGATGGCGATGGTCAGCA-3′; Mel 1B forward, primer 5′-TACAATGTTGTATGTCTCCTTAGTC-3′; reverse, 5′-CTTTTGGTGCCATTTCCGTA-3′; Mel 1C forward, 5′-TCAACCTGAAGAACACCTGC-3′; reverse, 5′-AGAAATTCCGTATGTCAGCA-3′.

PCR products were run on a 1.5% agarose gel, and bands of the appropriate size were selected for identity verification. PCR products were subcloned into a pGEM-T Easy vector (Promega Corp., Madison, WI; catalog no. A1360), and the DNA inserts of positive clones were amplified by PCR with universal M13 primers. The amplified products were sequenced at the University of California Berkeley DNA sequencing facility, and their homologies to European starling GnIH and GnIHR and chicken Mel 1B and Mel 1C were assessed in GenBank (accession nos. EF486798, EF212891, NM001048258, and U31821).

Seasonal GnIH/ GnIHR and Mel 1B/ Mel 1C expression

Testes from European starlings were collected on April 1, 2009 [n = 10, (photostimulated), day length, 12 h, 38 min], June 19, 2009 [n = 10, (photorefractory), in molt, day length, 14 h 46 min], and February 4, 2011 [n = 10, (photosensitive), day length, 10 h 25 min] (Fig. 1). Testes were measured, removed, frozen on dry ice, and then stored at −80C until use.

RNA was isolated from the left testis of each bird using TRIzol (Invitrogen, Carlsbad, CA; catalog no. 15596018) according to the manufacturer's protocol. For maximum extraction efficiency, at least 1 ml of TRIzol per 100 mg of tissue was used. RNA concentration and quality were assessed using a NanoDrop and NanoDrop-1000 3.3.0 accessory software. RNA samples with 10 mm absorbance at wavelengths from 220–350 nm or 260:280 ratio of more than 1.6 were not used for further analysis.

For cDNA, 1 μg total RNA for each testis was reverse transcribed using oligo(deoxythymidine)15 primer and M-MLV Reverse Transcriptase (Promega catalog nos: C1101, M1701). Concentration of cDNA was assessed using a NanoDrop and NanoDrop-1000 3.3.0 accessory software. Partial European starling GnIH, GnIHR, 1B, and 1C precursors were amplified from 500 ng cDNA by PCR using primers based on European starling GnIH and GnIHR and chicken (G. gallus) Mel 1B and Mel 1C precursor cDNA sequences (respective GenBank accession nos. EF486798, EF212891, NM001048258, and U31821): GnIH forward primer, 5′-GGAAGAAAAGCAGAGGAGTCTC-3′; GnIH reverse primer, 5′-TGGAGATCTCCCAAGCCTGT-3′; GnIHR forward primer, 5′-TCCTGGCCTACACCTTCATCT-3′; GnIHR reverse primer, 5′-AGATGATGGCGATGGTCAGCA-3′; Mel 1B forward primer, 5′-TACAATGTTGTATGTCTCCTTAGTC-3′; Mel 1B reverse primer, 5′-CTTTTGGTGCCATTTCCGTA-3′; Mel 1C forward primer, 5′-TCAACCTGAAGAACACCTGC-3′; Mel 1C reverse primer, 5′-AGAAATTCCGTATGTCAGCA-3′. β-Actin was also amplified from the cDNA as the endogenous control. All PCR amplifications were performed using a Taq polymerase kit (TaKaRa Ex Taq; Takara Bio, Inc., Shiga, Japan).

Samples for quantitative PCR must be assayed for concentration within the exponential phase before amplification efficiency decreases and relative concentration of coamplicons begins to vary (26). To establish an optimized and standardized cycle number, a trial using European starling testis cDNA was run and a 4 μl aliquot of each PCR was removed every five cycles during a 40-cycle amplification. The aliquots were run on an agarose gel and assessed using UV Transilluminator and National Institutes of Health software package. Another parameter of PCR that can affect amplification efficiency is the extension time. However, this is a set value of 1 min/1-kb template with a minimum of 1 min; thus a 1-min extension time was used. Annealing temperature, 55 C, was based on our established laboratory protocol for these primers and published data on GnIH and GnIHR sequences using this temperature.

Products were run on a 1.5% agarose gel and quantified by fluorescence of ethidium bromide under a UV Transilluminator (UVP, Inc., Upland, CA) using a two-dimensional image analysis of the gel in Adobe Photoshop CS2. The intensity of each of the signals was normalized to the β-actin internal control intensity of the same bird. Expressing the values as relative, rather than absolute, amounts controls for a reasonable margin of error in the estimation of original RNA concentration (26). The relative mRNA levels of GnIH and GnIHR in each bird (expressed as 0–100% of B-actin) were then compared. This procedure is a minor modification of that described in Refs. 26–29. It is advantageous over quantitative PCR in this protocol as it allows for a greater number of samples to be run concurrently and allows comparison of RNA in tissues of highly disparate volumes (i.e. regressed and recrudesced testes).

To verify the identity of the bands, a random sample of PCR products was subcloned into a pGEM-T Easy vector (Promega catalog no. A1360), and the DNA inserts of positive clones were amplified by PCR with universal M13 primers. The amplified products were sequenced at the University of California Berkeley DNA sequencing facility and their homologies to European starling GnIH, GnIHR, Mel 1B, and Mel 1C were assessed in GenBank (accession nos. EF486798, EF212891, NM001048258, and U31821), translated by Translate (Molecular Toolkit, Colorado State University) and compared using CLUSTAL W.

Gonad culture

Testes were collected from European starling males on February 5, 2010 [six males, (photosensitive), day length, 10 h, 27 min] and April 7, 2010 [six males, (photostimulated), day length, 12 h 52 min] (Fig 1). Gonad cultures were performed on both days with an identical procedure, except where noted. Length and width of each testis were measured after which testes were removed and placed in individual tubes of DMEM/Nutrient Mixture F12 (Sigma catalog no. D8437). Gonads were incubated in culture medium for 4 h at 4 C before experimentation to establish basal steroidogenesis/steroid secretion/protein transcription.

Each right testis (T_R) was snipped into four pieces using clean dissection scissors. The mass of each piece was recorded and then placed randomly into one of four fresh culture media containing: 1) 500 μl DMEM alone, 2) 0.25 μg/ml LH/FSH (0.244 μg/ml LH and 0.006 μg/ml FSH; National Hormone and Peptide Program, Torrance, CA) in DMEM, 3) 1 μm melatonin (Sigma catalog no. M5250) and 0.25 μg/ml LH/FSH in DMEM or 4) 1 μm GnIH and 0.25 μg/ml LH/FSH in DMEM. Hereafter, cultures are referred to as: 1) basal, 2) LH/FSH, 3) LH/FSH + melatonin, and 4) LH/FSH + GnIH. LH and FSH concentration used is deemed to be physiological as defined by Ref. 30. Because the concentration of GnIH within the testis requires further elucidation, the doses of GnIH used were based on physiological concentrations of GnRH and CRH in the testis (31, 32). The concentration of melatonin used was based on a previous study involving Leydig cell culture (22). Because of differences in functional equivalence of culture medium vs. blood plasma (passive diffusion vs. vascular delivery), it is hard to identify a perfectly biologically relevant hormonal concentration in culture. Hormones used were the best match to passerine LH, FSH, and GnIH commercially available and have been used successfully in the past (25). Culture followed the technique of Refs. 29 and 30.

All cultures were placed in a sealed incubator for a 4 h, high humidity incubation at 37 C immediately after placement of tissue into media. Oxygen (100%) was pumped into the chamber at 4 liters/min to allow for maximum respiration of cells. Cultures were removed to ice after 4 h and then centrifuged at 1500 × g at 4 C. The supernatant was removed to a clean tube and stored at −20 C. The tissues were washed once with clean incubation medium and then stored at −20 C before analysis.

The culture media supernatants from all cultures were assayed for testosterone using ELISA (Cayman Chemical catalog no. 582701) according to the manufacturer's instructions. Briefly, standards and samples were incubated with testosterone antiserum and testosterone acetylcholinesterase tracer in mouse antirabbit IgG-coated microplate wells at room temperature for 2 h at 45 rpm. Blank (no reagents) and maximum binding (buffer only) wells were also prepared at this time. After washing, all wells were developed with Ellman's reagent, 5,5′-dithiobis-(2-nitrobenzoic acid), in the dark for 60 min. Assays were performed in duplicate and were read at 415 nm on a microplate reader (Bio-Rad, model 680XR) at +0, +5, +15, and +30 min after development. Data were collected when maximum binding wells reached an absorbance within 0.3–1 absorbance units. All readings were corrected to the absorbance of the blank wells, after which standards and samples were converted to percentages of maximum binding. A standard curve was drawn using log (agonist) vs. response variable slope four parameter curve fit in GraphPad Prism 5.0 software. The concentrations of testosterone in the culture media were interpolated from the appropriate standard curves.

To control for potential differences in individual culture tissue masses, testosterone data were corrected to testes mass (picograms testosterone/ml per mg tissue).

Effects of culture on expression of GnIH, Mel 1B, and Mel 1C in the testes

Testis tissue recovered from each culture on February 5, 2010 and April 7, 2010 were subjected to RNA isolation, cDNA synthesis, and semiquantitative PCR using an endogenous control according to the methods outlined in Seasonal GnIH/ GnIHR and Mel 1B/ Mel 1C expression. Briefly, RNA was extracted using TRIzol (Invitrogen catalog 15596018) according to the manufacturer's protocol and assessed using a NanoDrop and NanoDrop-1000 3.3.0 accessory software. cDNA synthesis proceeded using 1 μg RNA, oligo(deoxythymidine)15 primer and M-MLV Reverse Transcriptase (Promega catalog nos. C1101 and M1701). Partial European starling GnIH, GnIHR, Mel 1B, and Mel 1C precursors were amplified from 500 ng cDNA by PCR using primers based on European starling GnIH and GnIHR and chicken (G. gallus) Mel 1B and Mel 1C precursor cDNA sequences (respective GenBank accession nos. EF486798, EF212891, NM001048258, and U31821). Products were quantified by fluorescence of ethidium bromide bands on a 1.5% agarose gel under a UV Transilluminator (UVP, Inc., Upland, CA) using a two-dimensional image analysis of the gel in Adobe Photoshop CS2. The intensity of each band was normalized to the β-actin internal control intensity of the same bird. The relative mRNA levels of GnIH, GnIHR, Mel 1B, or Mel 1C in each bird (expressed as 0–100% of B-actin) were then compared.

Statistical analysis

For seasonal GnIH and GnIHR expression and for seasonal Mel 1B and Mel 1C expression in which three independent groups were compared, one-way ANOVA was used. Where P < 0.05, Tukey's post hoc tests were used to analyze differences between each group.

Differences in testosterone secretion from cultured testes in which four groups of data collected from each individual were compared, data were pretested for normality and then repeated measures ANOVA were used. If P < 0.05, Bonferroni post hoc tests were used to analyze differences between LH/FSH-treated and each other culture treatment.

Differences in GnIH expression in cultured testes in which three groups of data collected from each individual were compared, data were pretested for normality after which a repeated measures ANOVA was used. Bonferroni post hoc tests analyzed the differences between LH/FSH treatment vs. basal and LH/FSH + melatonin.

Differences in Mel 1B and Mel 1C expressions between LH/FSH-treated and basal cultured testes collected from each individual were compared using paired t tests.

Correlations between expression levels of GnIH, GnIHR, and melatonin receptors were tested using Nonparametric Pearson correlation.

All statistical analyses were performed using GraphPad Prism 5.0 software.

Results

European starling testes express mRNA for GnIH, GnIHR, Mel 1B, and Mel 1C

European starling testes express mRNA for GnIH, GnIH receptor, and two forms of melatonin receptor, 1B and 1C. RT-PCR yielded electrophoretic bands of the appropriate size, which were then confirmed by sequencing. Mel 1A could not be isolated from European starling testes cDNA using RT-PCR, nor could it be sequenced.

Expression of testicular GnIH and GnIHR are season dependent

There is a statistically significant seasonal expression pattern of GnIH (ANOVA: P = 0.033, F = 3.87, r² = 0.22, df = 28) and its receptor (ANOVA: P = 0.0001, F = 12.73, r² = 0.48, df = 28) in the testes of European starlings (Fig. 2). Relative GnIH mRNA expression was significantly higher in photosensitive birds on February 4 compared with photostimulated birds on April 1 (Tukey's: P < 0.05). GnIHR mRNA expression was significantly higher in photorefractory birds on June 19 compared with photosensitive birds on February 4 (Tukey's: P < 0.005) and to photostimulated birds on April 1 (Tukey's: P < 0.01).

Fig. 2.

Seasonal status affects expression of GnIH and GnIHR in European starling testes. There is a significant effect of season on both (A) (ANOVA: P =0.033, F = 3.87, r² = 0.22, df = 28) and (B) GnIHR mRNA expression (ANOVA: P = 0.0001, F = 12.7.3, r² = 0.48, df = 28) in the testes. GnIH mRNA expression is significantly higher in photosensitive birds on February 4 compared with photostimulated birds on April 1 (Tukey's: P < 0.05, represented by asterisk in panel A). GnIHR mRNA expression is significantly higher in photorefractory birds on June 19 compared with photosensitive birds on February 4 (Tukey's: P < 0.005, represented by asterisk in panel B) and to photostimulated birds on April 1 (Tukeys's: P <0.01, represented by asterisk in panel B).

Effects of melatonin and GnIH in the testes are season dependent

The testes of photosensitive and photostimulated birds responded to 0.25 μg/ml LH/FSH stimulation in culture by significantly increasing testosterone secretion above basal levels (Fig 3; Bonferroni: February 5, P < 0.01, t = 4.68; April 7, P < 0.05, t = 3.19). The effects of melatonin and GnIH on LH/FSH-stimulated testosterone secretion from the testes are season dependent. On February 5, before breeding when starlings were photosensitive, the testes exhibited a reduced release of testosterone in response to melatonin and GnIH in culture [Fig 3A; repeated-measures ANOVA (RMANOVA): P = 0.003, F = 8.04, r² = 0.67, df = 5; Bonferroni: LH/FSH + melatonin P < 0.05, t = 3.02; LH/FSH + GnIH P < 0.05, t = 3.60]. On April 7, during breeding while starlings were photostimulated, the testes showed no reduction in testosterone release (relative to LH/FSH-stimulated controls) in response to melatonin and GnIH in culture (Fig 3B; RMANOVA: P = 0.12, F = 2.42, r² = 0.38, df = 5; Bonferroni: LH/FSH + melatonin P > 0.05, t = 0.01; LH/FSH + GnIH P > 0.05, t = 0.01).

Fig. 3.

Effects of melatonin and GnIH on testosterone secretion in European starling testes are dependent upon season. A, On February 5, in photosensitive birds just before breeding, testes coculture with melatonin or GnIH significantly reduces LH/FSH-stimulated testosterone secretion (RMANOVA: P = 0.003, F = 8.04, r² = 0.67, df = 5, Bonferroni: LH/FSH + melatonin p < 0.05, t = 3.02; LH/FSH + GnIH p < 0.05, t = 3.60) B, On April 7, in photostimulated birds during breeding, testes coculture with melatonin ot GnIH does not significantly reduce LH/FSH-stimulated testosterone secretion (RMANOVA: P = 0.12, F= 2.42, r² = 0.38, df = 5, Bonferroni: LH/FSH + melatonin, P < 0.05, t = 0.01; LH/FSH + GnIH P > 0.05, t = 0.01). On February 5 and April 7, LH/FSH stimulation alone significantly increases testosterone secrecretion from testes cultures relative to basal (Bonferroni: February 5, P < 0.01, t = 4.68; April 7, P < 0.05, t = 3.19). Asterisks indicate statistical significance from LH/FSH-treated group.

Gonadal GnIH is regulated by LH/FSH and melatonin in photosensitive birds

The expression of gonadal GnIH in cultured testes from photosensitive European starlings was significantly altered by LH/FSH) and melatonin on February 5, 2010 (Fig 4; RMANOVA: P = 0.01, F = 7.46, r² = 0.60, df = 5). Testes cultured with LH/FSH had significantly reduced GnIH mRNA expression compared with testes cultured without LH/FSH, in basal medium alone (Bonferroni: LH/FSH vs. basal, P < 0.05, t = 3.13). However, melatonin treatment was able to restore GnIH expression levels in the presence of LH/FSH, significantly increasing GnIH mRNA expression above LH/FSH levels (Bonferroni: LH/FSH alone vs. LH/FSH + melatonin, P < 0.05, t = 3.53).

Fig. 4.

LH/FSH and melatonin (mel) significantly affect GnIH expression in cultured photosensitive (Feb 5, before breeding) European starling testes (RMANOVA: P = 0.01, F = 7.46, r² = 0.60, df = 5). Testes cultured with LH/FSH have significantly reduced GnIH mRNA expression compared with untreated testes cultured in basal medium alone (Bonferroni: LH/FSH vs. basal, P < 0.05, t = 3.13, represented by asterisk). Melatonin significantly increased GnIH mRNA expression in testes cultured with LH/FSH (Bonferroni: LH/FSH alone vs. LH/FSH + melatonin, P < 0.05, t = 3.53; represented by asterisk).

Expression of Mel 1B and 1C in the testes is season dependent

There was a statistically significant effect of seasonal status on both Mel 1B and Mel 1C expression in the testes of European starlings (1B ANOVA: P = 0.009, F = 5.77, r² = 0.32, df = 28; 1C ANOVA: P = 0.002, F = 12.13, r² = 0.50, df = 28) (Fig. 5). Mel 1B mRNA was significantly up-regulated in the testis of photorefractory birds on June 19 compared with photosensitive birds on February 4 (Tukey's P < 0.05) and compared with photostimulated birds on April 1 (Tukey's: P < 0.01). There was no statistically significant difference in Mel 1B expression between photosensitive and photostimulated birds. Mel 1C mRNA expression was significantly down-regulated in the testis of photostimulated birds (April 1) compared with photosensitive birds (February 4) (Tukey's: February vs. April, P < 0.05) and compared with photorefractory birds (June 19) (Tukey's: June vs. April, P < 0.005).

Fig. 5.

Seasonal status affects expression of melatonin receptors in European starling testes. There is a significant effect of season on both (A) Mel 1B mRNA expression (ANOVA: P = 0.009, F = 5.77, r² = 0.32, df = 28) and (B) Mel 1C mRNA expression (ANOVA: P = 0.002, F = 12.13, r² = 0.50, df =28) in the testes. Mel 1B mRNA expression is significantly higher in the testes of photorefractory birds on June 19 compared with photosensitive birds on February 4 (Tukey's: P < 0.05) and compared with photostimulated birds on April 1 (Tukey's: P < 0.01). Mel 1C mRNA expression is significantly up-regulated in the testis of photosensitive birds on February 4 and photorefractory birds on June 19 compared with photostimulated birds on April 1 (Tukey's: Feb vs. April, P < 0.05; June vs. April, P < 0.005). The highest expression of Mel 1C mRNA is seen in the testis of photorefractory birds (Tukey's: February vs. June, P < 0.01). a, P < 0.05; b, P <0.01; c, P < 0.005.

Seasonal up-regulation of GnIH and GnIHR is correlated with up-regulation of Mel 1C and Mel 1B, respectively

There was a significant correlation between the up-regulation of gonadal GnIH expression and gonadal Mel 1C expression in birds from all time points (Pearson r = 0.43, r² = 0.18, P = 0.022). There was also a significant correlation between the up-regulation of gonadal GnIHR expression and gonadal Mel 1B expression in birds from all time points (Pearson r = 0.46, r² = 0.21, P = 0.019). The linear regression lines are shown in Fig. 6.

Fig. 6.

Seasonal up-regulation of GnIH and GnIHR is correlated with seasonal up-regulation of Mel 1C and Mel 1B, respectively. There is a significant correlation between gonadal GnIH expression and gonadal Mel 1C expression in European starling testes (Pearson: r = 0.43, r² = 0.18, P = 0.022). This relationship is described by the linear regression line y = 0.15 + 0.51x. There is a significant correlation between gonadal GnIHR expression and gonadal Mel 1B expression in European starling testes (Pearson: r = 0.46, r² = 0.21, P = 0.019). This relationship is described by the linear regression line y = 0.07 + 0.66x.

Gonadal expression of Mel 1C is regulated by LH/FSH

The expression of Mel 1C mRNA in cultured European starling testes was significantly attenuated by LH/FSH compared with testes cultured without LH/FSH, in basal media alone (paired t test: P = 0.006, T = 4.53, r = 0.80, df = 5) (Fig. 7B). There was no statistically significant difference in the expression of Mel 1B mRNA in testes cultured with and without LH/FSH (paired t test: P = 0.09, T = 2.10, r = 0.47, df = 5) (Fig. 7A).

Fig. 7.

LH/FSH treatment has no effect on Mel 1B mRNA expression (A, paired t test: P = 0.09, T = 2.10, r = 0.47, df = 5), but significantly down-regulates Mel 1C expression (B, paired t test: P = 0.006, T = 4.53, r = 0.80, df = 5) in cultured photosensitive European starling testes.

Discussion

Seasonally-breeding European starlings exhibit a localized seasonal pattern of responsiveness to melatonin and to GnIH in the gonads. The response to melatonin and GnIH consists of modulation (inhibition) of testosterone release in photosensitive birds. Melatonin increased gonadal GnIH expression at this time; thus the action of melatonin appears to involve regulation of the endogenous gonadal GnIH/GnIHR system. Neither melatonin nor exogenous GnIH decreased testosterone secretion in culture during the breeding stage, a time when endogenous GnIH and GnIHR expression are down-regulated. Because gonadal melatonin receptors are responsive to LH and Mel 1C is up-regulated in vivo and ex vivo (in culture) in photosensitive birds (before breeding), it is likely that the direct inhibitory action of melatonin on the gonads is season dependent. The longer duration of melatonin secretion in February (where night is 13 h, 35 min) compared with April (where night is 11 h, 22 min) may be responsible for modulating testosterone secretion before breeding in the wild. This hypothesis remains to be tested. Alternatively, this responsiveness could be mediated by plasma LH/FSH or gonadal melatonin receptors.

The expression and action of the gonadal GnIH system in European starlings also depend on season. We showed previously that GnIH acts directly on the testes to reduce LH/FSH-stimulated testosterone secretion (25). Here we show that the testes of European starlings respond to the inhibitory effects of gonadal GnIH on testosterone secretion during the photosensitive period, just before breeding, but not in the photostimulated period, during breeding. In vivo, this may be physiologically mediated by up- and down-regulation of gonadal GnIH expression. Evidence supporting this comes from our data indicating that gonadal GnIH expression is down-regulated in April, when the birds are in full reproductive condition, compared with February, before the onset of breeding. This is in contrast to hypothalamic GnIH expression, which is up-regulated in April during breeding in European starlings (34, 35). The action and regulation of gonadal GnIH therefore cannot be inferred from the action and regulation of hypothalamic GnIH. Future studies are needed to determine whether and how these systems interact to time reproductive activity. It is possible that starling testes use the gonadal GnIH system to modulate testosterone secretion and regulate the exact timing of the onset of full reproductive growth. During photosensitivity, before the critical day length threshold is reached, starlings exhibit greater responsiveness to supplementary cues (such as food availability, presence of conspecifics, and stress) than at other stages of the photo-induced breeding cycle. These supplementary signals can delay or advance the onset of full gonadal recrudescence (36–39). Endocrine responses (e.g. glucocorticoid release) to many supplementary cues are detectable in the plasma (40–42); thus gonadal GnIH could act as a mediator of supplemental information in the testes if the gonadal GnIH system is responsive to such endocrine signals. During photostimulation, when the starling reproductive axis is already maximally stimulated, supplementary cues are less effective (36–39, 44, 45). A physiological commitment to reproduction has already been made: full reproductive growth has occurred and high levels of testosterone secretion are required (7, 37). Thus gonadal, GnIH expression down-regulation would be required, as our data indicate.

The seasonal inhibitory effects of gonadal GnIH do not appear to be mediated by seasonal changes in the gonadal GnIH receptor. There does not appear to be a significant down-regulation of GnIHR receptor expression in April compared with February. The effects of gonadal GnIH in February vs. April are thus not mediated by up- or down-regulation of the receptor, but by another mechanism, possibly an interaction of the changing melatonin signal and its effect upon gonadal GnIH expression, as described above.

Melatonin seasonally mediates the expression of the gonadal GnIH system. Gonadal GnIH is up-regulated in photosensitive European starling testes in response to melatonin in culture. Furthermore, gonadal GnIH expression is down-regulated by LH/FSH, but this effect is rescued by melatonin. Additionally, Mel 1C expression is up-regulated in February relative to April, and this expression is significantly correlated with gonadal GnIH expression. It is tempting to assume that gonadal melatonin receptors are directly regulated by the duration of melatonin secretion. Mel 1C could be up-regulated by the longer duration melatonin secretion in the plasma of birds on February 4 (where night is 13 h, 35 min) compared with April 1 (where night is 11 h, 22 min). However, the highest expression of Mel 1C is observed in testes from birds collected June 19 (where night is only 9 h, 14 min). Instead, we propose that Mel 1C is potentially regulated by LH/FSH rather than melatonin duration: Mel 1C expression is significantly reduced in testes cultured with LH/FSH compared with testes cultured in basal media alone. In this way, Mel 1C expression is high during nonbreeding periods (photosensitivity and photorefractoriness) when LH/FSH concentration is relatively low and Mel 1C is, by contrast, minimal during the breeding season (photostimulation) when plasma LH/FSH concentration is high. Thus gonadal GnIH can be directly up-regulated by seasonal status: via seasonally appropriate durations of melatonin secretion acting on a receptor (Mel 1C) down-regulated by hormones (LH and FSH) secreted during the breeding season (Fig. 8A).

Fig. 8.

Schematic representation of relative gonadal expression of GnIH and Mel 1C (A) and GnIHR and Mel 1B (B) according to data collected in this study, together with relative endogenous plasma concentrations of LH/FSH and melatonin in European starlings (33, 43, 50). GnIH, Mel 1C, GnIHR, and Mel 1B data are fit by nonlinear regression with second order polynomial best fit equations using GraphPad Prism 5.0 software.

Interestingly, as in the brain, the direct action of melatonin explains the inhibition of gonadal function in the photosensitive period, but not at the onset of photorefractoriness. Starling gonads begin to regress while days are still long in the summer, and this is neither due to significant up-regulation of gonadal GnIH (testes expression of GnIH is not significantly different in June compared with April) nor duration of melatonin signal. However, both the expression of GnIH receptor and Mel 1B are up-regulated in photorefractory birds compared with photosensitive and photostimulated birds. GnIH receptor does not appear to be regulated by gonadal GnIH expression, melatonin duration, Mel 1C or LH/FSH secretion, because its seasonal pattern of expression is not well matched to these agents (the highest expression of GnIHR is in June, when these factors are relatively low). However, GnIHR expression is correlated with gonadal expression of Mel 1B. Mel 1B is significantly up-regulated in June (compared with February and April), and its expression does not appear to be regulated by LH/FSH (Fig. 8B). Because the agents governing the onset of photorefractoriness in birds are not well understood, gonadal GnIHR and Mel 1B, which are up-regulated at this time, may be an important avenue for future study.

It is also interesting to note that Mel 1B and Mel 1C are differentially regulated in the testes. Mel 1C expression is high during nonbreeding periods (photosensitivity and photorefractoriness) and is reduced to a minimum as plasma LH and FSH become maximal during breeding. Mel 1B, in contrast, is up-regulated only during photorefractoriness. The functional significance of the uncoupling of melatonin receptor subtype expression in the testes remains to be tested, but it may have an important role in allowing for differential actions of melatonin across seasons.

In sum, we propose that in European starlings, melatonin acts via Mel 1C and Mel 1B in the testis to seasonally regulate the expression and action of the gonadal GnIH system, which modulates testosterone secretion.

In a broader context, mammals also express the mammalian form of GnIH, RF-amide related peptide, and its receptors GPR74/GPR147/RFRPR in their gonads (46, 47) Photoperiodic mammals have detectable levels of melatonin in their plasma and seminal plasma that are consistent with pineal secretions (48, 49); they express melatonin subtype 1a receptor (21) in the testes and respond to melatonin challenge with reduced testosterone output from the testes (20–22). Thus the testes of mammals are capable of directly responding to melatonin, and this effect may be mediated through the gonadal GnIH system as well. More study will be needed to confirm this. Nonetheless, the ability of melatonin to directly convey photoperiodic information to the gonads, and the responsiveness of the gonads to this information, appears to be a common property among photoperiodic vertebrates. The timing of breeding of photoperiodic vertebrates thus appears to involve integration of cues at the level of the gonads in addition to the brain.

Abbreviation:

 
• GnIH

  gonadotropin-inhibitory hormone

 
• GnIHR

  GnIH receptor

 
• HSD

  hydroxysteroid dehydrogenase

 
• Mel 1B and 1C

  melatonin receptors 1B and 1C

 
• Px

  pinealectomized

 
• RMANOVA

  repeated-measures ANOVA.

Acknowledgments

This work was supported by National Science Foundation Grants IOS 0641188 and 0920753 (to G.E.B.). N.L.M. was the recipient of a Graduate Research Fellowship.

Disclosure Summary: The authors have nothing to disclose.