Orexin A Suppresses Gonadotropin-Releasing Hormone (GnRH) Neuron Activity in the Mouse

Abstract

GnRH neurons are critical for the central regulation of fertility, integrating steroidal, metabolic and other cues. GnRH neurons appear to lack receptors for many of these cues, suggesting involvement of afferent systems to convey information. Orexin A (orexin) is of interest in this regard as a neuromodulator that up-regulates metabolic activity, increases wakefulness, and affects GnRH/LH release. We examined the electrophysiological response of GnRH neurons to orexin application and how this response changes with estradiol and time of day in a defined animal model. Mice were either ovariectomized (OVX) or OVX and implanted with estradiol capsules (OVX+E). GnRH neurons from OVX+E mice exhibit low firing rates in the morning, due to estradiol-negative feedback, and high firing rates in the evening, due to positive feedback. Orexin inhibited activity of GnRH neurons from OVX mice independent of time of day. In GnRH neurons from OVX+E mice, orexin was inhibitory during the evening, suggesting orexin inhibition is not altered by estradiol. No effect of orexin was observed in OVX+E morning recordings, due to low basal GnRH activity. Inhibitory effects of orexin were mediated by the type 1 orexin receptor, but antagonism of this receptor did not increase GnRH neuron activity during estradiol-negative feedback. Spike pattern analysis revealed orexin increases interevent interval by reducing the number of single spikes and bursts. Orexin reduced spikes/burst and burst duration but did not affect intraburst interval. This suggests orexin may reduce overall firing rate by suppressing spike initiation and burst maintenance in GnRH neurons.

GnRH neurons play a critical role in the central regulation of fertility. GnRH acts upon the anterior pituitary to stimulate release of the gonadotropins LH and FSH, which in turn initiate steroidogenesis and gametogenesis.

Frequency and amplitude of GnRH/LH release are modulated by changes in steroid feedback throughout the female reproductive cycle, ultimately leading to the surges in GnRH and LH near the end of the follicular phase that are required for ovulation (1–11). In addition to steroid feedback, GnRH neurons are regulated by a wealth of factors that may interact with the steroid milieu. Among the most intriguing of these are diurnal signals associated with the day-night/sleep-wake cycle and metabolic cues.

Diurnal changes in GnRH activity are first observed during the pubertal sleep-associated initiation of GnRH pulses, and continue into adulthood (12). Disruptions in this diurnal regulation may be one factor in developing infertility, including progression toward polycystic-ovarian syndrome (12, 13). In rodents, the feedback actions of estradiol are modulated in a diurnal manner, such that negative feedback (defined as an overall reduction in reproductive neuroendocrine output in response to estradiol, measured by serum LH or GnRH, or GnRH neuron activity) transitions into positive feedback (an overall increase in reproductive neuroendocrine output in response to estradiol) to induce the GnRH/LH surge near the time of activity onset (14–16). Blockade of neural activity through barbiturate administration before activity onset delays the LH surge not just for the duration of drug action but for approximately 24 h, suggesting involvement of circadian timing (17). With regard to metabolism, negative energy balance inhibits pulsatile LH and GnRH release and blocks the LH surge (18–21). Furthermore, GnRH neural firing activity is reduced under conditions of glucoprivation (21, 22), and estradiol mitigates this effect (23).

Interestingly, GnRH neurons appear to lack receptors for most steroids and some metabolic cues, including hormonal factors such as leptin, ghrelin, and insulin (24–26). This suggests additional mediators help convey these cues to GnRH neurons. The hypothalamic neuropeptide orexin A (orexin) is of interest in this regard. Orexin neurons, located in the lateral hypothalamus (27, 28), receive inputs from and project to the sleep-wake centers (ventrolatero preoptic nucleus, posterior hypothalamus) and circadian pacemakers in the suprachiasmatic nucleus (29–31) as well as the food intake centers in the lateral hypothalamus and arcuate nucleus (30, 32, 33). Orexin has also been shown to affect GnRH and/or LH release (34–40). Orexin neurons contain a variety of receptors, including estrogen receptor-β (26), serotonin (41), adenosine (42), and neuropeptide Y (Y1, Y4) (32), thereby providing a potential mechanism to integrate and subsequently convey metabolic and sleep-wake information in a steroid-modulated manner. Dual-immunofluorescence labeling in rats revealed that approximately 80% of GnRH neurons express the orexin-1 receptor (OX1R), and orexin immunopositive fibers make synaptic contacts with GnRH neurons as indicated by confocal microscopy, providing an anatomical basis for this regulation (43). Physiological studies, however, have not produced consistent results, with orexin both increasing and decreasing GnRH and/or LH, depending on the experimental paradigm and/or steroid milieu (34–40).

To better characterize the role of orexin in the reproductive system, we directly examined the electrophysiological response of GnRH neurons to orexin application and how this response changes with estradiol milieu and time of day in a defined animal model.

Materials and Methods

Animals

Studies were performed in adult (2–4 months old) transgenic female GnRH-green fluorescent protein (GFP)-identified mice (44). Mice were housed on a 14-h light, 10-h dark cycle, with lights on at 0200 h EST. Animals were provided with Harlan 2916 rodent chow (Harlan, Bartonsville, IL) and water ad libitum. To control for effects of ovarian steroids in adult females, mice were ovariectomized (OVX) under isoflurane anesthesia (MWI Veterinary Supply Co., Meridian, ID), and received postoperative analgesia via a long-acting local anesthetic (0.25% bupivacaine; 7 μl per site; APP Pharmaceuticals, Schaumburg, IL) directly at the surgical site. At the time of surgery, some mice were implanted with a SILASTIC (Dow Corning, Midland, MI) capsule containing 0.625 μg of estradiol suspended in sesame oil (OVX+E). Recordings were performed 3–5 d after surgery, at which point estradiol has a diurnal effect on GnRH firing activity; specifically estradiol induces negative feedback in the morning and positive feedback in the evening (14). All procedures were approved by the University Committee on Use and Care of Animals from the University of Michigan and were conducted within the guidelines of the National Research Council's Guide for the Care and Use of Laboratory Animals.

Brain slice preparation and recording

All chemicals were purchased from Sigma Chemical Co. (St. Louis, MO) unless otherwise noted. Brain slices were prepared using a previously described method (45, 46). Solutions were continuously bubbled with a 95% O₂-5% CO₂ mixture throughout the experiment and for at least 15 min before exposure to tissues. The brain was rapidly removed and placed in an ice-cold, sucrose-saline solution containing in millimolar: 250 sucrose, 3.5 KCl, 26 NaHCO₃, 10 glucose, 1.25 Na₂HPO₄, 1.2 MgSO₄, and 2.5 MgCl₂. Coronal 300-μm brain slices were then cut using a Vibratome 3000 (Ted Pella, Redding, CA). Slices were incubated in a 1:1 mixture of sucrose-saline and artificial cerebrospinal fluid (ACSF) containing in millimolar 125 NaCl, 3.5 KCl, 26 NaHCO₃, 1.25 NaH₂PO₄, 2.5 CaCl₂, 1.2 MgSO₄, and 10 d-glucose (pH 7.4) for 30 min at 30–32 C and then transferred to 100% ACSF and incubated for at least an additional 30 min at room temperature (22–24 C) before recording.

All recordings were performed from GnRH neurons located in the preoptic area and anterior hypothalamic regions and recorded cells were mapped to an atlas (47). No differences in initial activity or effect of orexin were noted based on location of cells. Individual brain slices were transferred to a recording chamber mounted onto the stage of an upright fluorescent microscope (Olympus BX50WI; Opelco, Dulles, VA). The chamber was perfused continuously with oxygenated ACSF and maintained at 28–30 C using an in-line heater (Warner Instruments, Hamden, CT). Recording micropipettes were pulled from borosilicate capillary glass (type 7052, 1.65 mm outer diameter; 1.12 mm inner diameter; World Precision Instruments, Inc., Sarasota, FL) using a Flaming/Brown P-97 (Sutter Instrument, Novato, CA) to obtain pipettes with a resistance between 1.8 and 3.0 MΩ when filled with HEPES-buffered solution, containing in millimolar: 150 NaCl, 3.5 KCl, 10 HEPES, 10 d-glucose, 2.5 CaCl₂, and 1.3 MgCl₂. GnRH neurons were identified by brief fluorescent illumination at 470 nm to visualize the GFP signal. Pipettes were placed in contact with a GnRH neuron using an MP-200 micromanipulator (Sutter Instruments), creating a 10- to 50-MΩ seal. Recordings were obtained through an EPC-10 amplifier under voltage-clamp mode with a holding potential of 0 mV using PatchMaster software (HEKA Elektronik, Lambrecht/Pfalz, Germany). In this mode, action currents that underlie action potential firing are detected; their frequency is monitored and reflects changes in action potential firing rate.

Experimental design

OVX+E mice exhibit a diurnal rhythm in GnRH neuronal firing activity. More specifically, the frequency of firing varies over the course of a day, with low firing frequency in the morning due to estradiol negative feedback and high firing frequency in the evening due to positive feedback (14). This defined model was used to examine the interactions of estradiol and time of day in modulating the response of GnRH neurons to orexin. To do this, brain slices were prepared and recordings made during both the morning and evening. For morning recordings, brain slices were prepared between 0700 and 0830 h EST and recordings were made from 0930 to 1230 h EST; for evening recordings, mice were euthanized between 1300 and 1400 h EST, and recordings were made from 1430 to 1900 h EST. Brain slices were prepared at the same times from OVX mice, in which no diurnal phase in activity or other biophysical parameters has yet been observed (10, 14). All recordings from an individual mouse were conducted during either negative or positive feedback and did not continue into the next phase. This allowed standardization of the time from slice preparation to recording.

To examine the effect of orexin application on GnRH neuron activity, targeted extracellular recordings were made from GFP-identified GnRH neurons. This type of recording does not modify intracellular milieu and is minimally invasive; thus the intrinsic nature of the cellular response to synaptic input (within the brain slice) is left intact during recording (46). After establishing the recording configuration, cells were acclimated for 5–20 min to ensure a stable baseline. Firing rate was then monitored for 10–15 min under control conditions. Orexin (1–100 nm; Tocris Bioscience, Ellisville, MO) was then bath applied for 7–10 min, followed by a washout period (10–30 min). If no firing was observed during the washout period, ACSF with increased potassium (+15 mm KCl) was applied to verify the cell's ability to produce action currents and the integrity of the recording. If no action currents were observed in response to increased potassium, only the recording period up to the last clearly verifiable action current was used for analysis. Firing amplitude of action currents was not taken into account because extracellular recordings do not allow for precise characterization of firing mechanisms based on amplitude. To test whether orexin was acting via the OX1R receptor previously reported to be expressed in GnRH neurons (43), cells were recorded under control conditions for 5–10 min, followed by a pretreatment with the OX1R antagonist SB-334867 (5 μm; Tocris Bioscience) for 5–10 min. Cells were then treated with orexin plus SB-334867 (5–10 min). The effects of SB-334867 alone were also examined.

Data collection and analysis

Data were analyzed using custom programs developed in Matlab (Mathworks, Inc., Natick, MA) to identify action currents, the membrane currents associated with action potential firing. Firing (action current) rate was averaged during the last five minutes of each control and treatment period. Spontaneous events within those timeframes were automatically detected using custom programs and confirmed by eye. Both false-positive and false-negative detection errors were corrected manually. Because GnRH neurons undergo spontaneous changes in firing frequency, percent change in firing frequency was calculated as a measure of effect, with a minimum of a 25% change in firing frequency being defined as a response for calculation of percent responding cells (both responding and nonresponding cells were included in statistical analyses to be more rigorous). Two-tailed paired parametric or nonparametric analyses, as appropriate for number of groups and data distribution, were used to determine whether overall firing rates differed before and during treatment.

In addition to mean firing rate, characteristic features of the firing pattern of GnRH neurons were also determined during the last 5 min before and during treatment with orexin. The number of single spikes, bursts, burst duration, and spikes per burst were analyzed in each cell for both OVX combined and OVX+E evening groups (OVX+E morning data were excluded due to paucity of events). A burst was defined as any sequence of action currents occurring with an interval of 210 msec or less from one spike to the next; this value was chosen based on previous studies (48, 49) and examination of the inter-action current interval distribution of present data, which showed a cluster from approximately 120 to 210 msec. For interevent analysis, each burst was considered as one event. Any spike detected after an interval greater than 210 msec was considered to be the start of a new event. Intervals within bursts (intraburst intervals) as well as between defined events (interevent intervals) were analyzed. A schematic of how data were organized for analysis is shown elsewhere (see Fig. 5). Two-way ANOVA was used to compare intraburst interval, interevent interval, burst duration, and spikes per burst during control and orexin treatment in cells from OVX and OVX+E mice. Interevent interval data were formatted to a histogram plot to better visualize changes in the distribution of interval lengths. Kolmogorov-Smirnov tests were used to compare the interevent interval distribution between OVX and OVX+E groups and between control and orexin treatments. All statistical analyses were performed using Prism (GraphPad Software, Inc., La Jolla, CA) or Matlab (Mathworks). Significance was set at P < 0.05.

Fig. 5.

Schematic showing how bursts (B) and single spikes (S) were identified, and how intraburst and interevent intervals were defined. A, Bursts consist of all action currents occurring within 210 msec or less of each other (four action currents in this example burst). Intraburst intervals are the intervals between each action current within a burst. (solid double headed arrows). Single spikes are action currents greater than 210 msec from other spikes or bursts (white double-headed arrow). B, Interevent intervals were determined between both single spikes and bursts but did not include intraburst intervals (A, left). Note change in scale between A and B.

Results

Orexin A suppresses GnRH neuron firing frequency in a dose-dependent manner

We first examined whether orexin modulates GnRH neuron firing frequency. An initial dose-response test was conducted, in which 1, 10, or 100 nm orexin was bath applied to GnRH neurons in brain slices from OVX mice in the evening. Extracellular recordings were performed to monitor firing activity. Orexin elicited a dose-dependent suppression of GnRH neuron firing rate (Fig. 1, n = 3–4 cells per dose). This dose-response curve was obtained from separate cells because the orexin-induced suppression was long lasting, and washout (>30 min) did not always return cells to control firing levels. Based on these observations, an orexin dose of 10 nm was used for the remainder of studies.

Fig. 1.

Dose-response curve of orexin effects on activity of GnRH neurons from OVX mice during the evening (n = 3–4 cells per dose, one dose per cell). Orexin was bath applied to coronal slices and targeted extracellular recordings monitoring action currents were made of GnRH neurons.

Orexin A suppresses GnRH neuron firing frequency in OVX mice in both morning and evening

To study whether the effect of orexin depends on time of day in OVX mice, we performed additional extracellular recordings in both the morning and evening. Spontaneous firing rate was recorded from GnRH neurons during an initial control period (≥5 min) followed by bath application of 10 nm orexin (7–10 min). Representative traces from OVX mice illustrate firing levels before and during treatment with orexin in both the morning and evening (Fig. 2, A and B). As previously observed, GnRH neuron control firing rates in OVX mice were similar at both times of day (14). Orexin inhibited mean firing rate in the morning (n = 6 cells from four mice, P < 0.05), with 100% of cells responding with an average percent inhibition of 60 ± 18%, and in the evening (n = 7 cells from five mice, P < 0.05), with 71% of cells responding and an average percent inhibition of 62 ± 22% from all cells (Fig. 2C). Together these data suggest that both the degree and percentage of GnRH neurons inhibited by orexin is independent of the time of day in the brain slices from OVX mice (P = 0.37). Data from OVX mice in the morning and evening were thus combined for later comparisons.

Fig. 2.

Orexin reduces the overall firing frequency of GnRH neurons from OVX mice, and the effect is independent of time of day. A and B, Representative raw data traces showing targeted extracellular recordings during the morning (AM) (A) and evening (PM) (B); control conditions are on the left and treatment with 10 nm orexin on the right. C, Mean ± sem firing frequency of GnRH neurons before and during orexin application (n = 6–7 cells/group). *, P < 0.05, orexin vs. control.

Orexin A suppresses GnRH neuron firing frequency in OVX+E mice during evening

OVX+E mice exhibit diurnal changes in GnRH neuron activity due to a shift in action of estradiol feedback, with low firing activity in the morning as a result of negative feedback and high firing activity due to positive feedback in the evening. To test whether estradiol influences the response to orexin, extracellular recordings were made during either the morning or evening from GnRH neurons in brain slices prepared from OVX+E mice. Representative traces displaying firing rates of GnRH neurons from OVX+E mice before and during treatment with 10 nm orexin are shown in Fig. 3, A and B. The firing rate of GnRH neurons from OVX+E mice during the pretreatment period was increased in the evening compared with the morning (n = 9 cells from five mice, P < 0.01) and also compared with OVX mice at both times of day (Fig. 3C, P = 0.05,), as previously reported (14). Similar to the above studies in the OVX mice, application of 10 nm orexin inhibited the firing frequency of GnRH neurons from OVX+E mice during the evening (P < 0.05) with 90% of cells responding with an average inhibition of 61 ± 8% (Fig. 3C).

Fig. 3.

Orexin reduces the overall firing frequency of GnRH neurons in the presence of estradiol. A and B, Representative raw data traces showing targeted extracellular recordings during the morning (AM) (A) and evening (PM) (B); control conditions are on the left and treatment with 10 nm orexin on the right. Cells were almost entirely quiescent during OVX+E morning recordings. C, Mean ± sem firing frequency of GnRH neurons before and during orexin application (OVX, n = 13; OVX+E morning, n = 7; OVX+E evening, n = 9 cells). OVX morning and evening recordings were combined for comparison. *, P < 0.05 orexin vs. control.

As a result of the impact of estradiol-negative feedback, control firing rate of GnRH neurons from OVX+E mice recorded during the morning (n = 6 cells from five mice) was lower than that of GnRH neurons from OVX mice in the morning and evening as well as that of OVX+E mice during the evening (Fig. 3C, all P < 0.01). Representative data traces show little to no firing occurs either before or during bath application of 10 nm orexin (Fig. 3A). Due to the quiescence of cells during OVX+E morning recordings, we were unable to detect any effect of orexin on firing rate (control, 0.05 ± 0.02 Hz; 10 nm orexin, 0.03 ± 0.02 Hz, P = 0.24). These data suggest orexin is either unable to alter GnRH firing frequency further in the morning, when GnRH neurons are relatively inactive due to estradiol negative feedback or, alternatively, that endogenous orexin exerts a suppressive effect on GnRH neurons in addition to estradiol-negative feedback.

Orexin acts via the OX1R to inhibit GnRH neurons

To determine whether the OX1R, previously reported to be expressed by GnRH neurons (43), is responsible for the inhibitory effect on GnRH firing frequency, we examined the effect of the selective OX1R antagonist SB-334867. Extracellular recordings were made of GnRH neurons in brain slices from OVX+E mice in the evening to ensure a high rate of firing during the control period and allow for any potential suppressive effect of orexin to be detected. Spontaneous firing rate was recorded for 10 min. SB-334867 (5 μm) was then bath applied for 5–10 min, followed by SB-334867 plus 10 nm orexin for 5–7 min. Figure 4A shows raw data traces from OVX+E evening recordings before and during treatment with orexin and the OX1R antagonist; there was no effect of antagonist alone on firing rate (n = 8 cells from five mice, P = 0.74), but SB-334867 blocked the suppressive effect of orexin on GnRH neuron activity (Fig. 4B, P = 0.16 vs. pretreatment control).

Fig. 4.

The inhibitory effect of orexin on overall GnRH neuron firing rate is blocked in the presence of the selective OX1R antagonist SB-334867. A, Representative raw data traces showing targeted extracellular recordings during the control period (left) and during treatment with SB-334867 + 10 nm orexin (right). PM. Evening. B, Mean ± sem firing frequency of GnRH neurons during control, SB-334867, and SB-334867 with 10 nm orexin periods. C, Representative raw data trace showing SB-334867 did not increase firing in GnRH neurons from OVX+E mice recorded in the morning (AM; same scale as in A).

Low GnRH neuron activity during negative feedback does not appear to be due to endogenous orexin tone

Although a suppressive effect of orexin in the OVX+E morning studies could not be observed due to quiescence, an interesting hypothesis is that the low activity of GnRH neurons during negative feedback is due to endogenous inhibition by orexin. To test whether endogenous orexin was exerting a suppressive effect on GnRH neurons from OVX+E mice during the time of negative feedback, SB-334867 was bath applied to GnRH neurons during the morning. Raw data traces show spontaneous firing before and during treatment with the SB-334867 (Fig. 4C). Contrary to our hypothesis, application of SB-334867 did not increase firing frequency in the morning and cells remained almost entirely quiescent during recordings (control, 0.04 ± 0.05 Hz; 5 μm SB-334867, 0.06 ± 0.10 Hz, n = 7 cells from four mice, P = 0.29). This was also true when treatment with SB-334867 was extended to up to 25 min to reflect the long duration of inhibition in response to exogenous orexin. Viability of cells and ability to record action currents was confirmed through the administration of increased potassium (+20 mm KCl) at the end of the recording. These results suggest the low spontaneous firing rate of GnRH neurons in coronal brain slices from OVX+E mice in the morning is unlikely due to a suppressive effect of endogenous orexin under these experimental conditions.

Orexin reduces both bursts and single spikes in GnRH neurons

Firing dynamics and how action potentials are clustered relative to one another can affect neuroendocrine secretion (50). We examined the effects of estradiol and orexin on GnRH neuron firing dynamics; recordings made from OVX+E mice during the morning were excluded due to the paucity of spikes. Figure 5 illustrates how intervals and events were grouped for analysis (see Materials and Methods for details). We first examined whether the number of bursts and single spikes during the control periods were different between OVX morning (n = 6) and OVX evening (n = 7) recordings. There was no effect of time of day on these parameters (bursts per 5 min: OVX morning, 26 ± 11, OVX evening, 27 ± 17, P > 0.90; single spikes per 5 min: OVX morning, 68 ± 17, OVX evening, 94 ± 27, P > 0.25). Furthermore, orexin suppressed both number of bursts and single spikes in OVX recordings at both times of day (Fig. 6, A and B; P < 0.05). These groups were thus combined for analysis.

Fig. 6.

Mean ± sem firing dynamics of GnRH neurons from OVX (morning-evening combined, n = 13) and OVX+E evening (n = 9) recordings, during the last 5 min before (gray bars) and during treatment with orexin (black bars). A, Single spikes. B, Bursts. C, Intraburst interval. D, Burst duration. E, Spikes per burst. F, Interevent intervals. *, P < 0.05 orexin vs. control; **, P ≤ 0.05 OVX vs. OVX+E.

We next determined how estradiol itself altered spike and burst patterning because previous work quantified only the changes in mean firing rate and quiescence in this model (14). Estradiol increased both the number of bursts and single spikes (P = 0.01, unpaired t test), reducing the interevent interval (Fig. 6F, P = 0.05); the difference in number of bursts was not evident (P = 0.09) when two-way ANOVA was used to coevaluate the effects of orexin and estradiol (Fig. 6, A and B). There was no interaction between estradiol and orexin for either spike or burst number (spikes, F = 0.71; bursts, F = 0.23), but both of these treatments altered the distribution of interevent intervals, with estradiol shifting the distribution toward smaller intervals and orexin shifting the distribution toward larger intervals (Fig. 7, both P < 0.05 Kolmogorov-Smirnov test).

Fig. 7.

Histograms showing distribution of interevent intervals as a function of interevent interval duration. Data are plotted on double-log scales. A and C, Distribution of interevent intervals of all cells from OVX (morning-evening) recordings during the last 5 min of control period (A) and orexin administration (C). B and D, Distribution of interevent intervals of all cells from OVX+E evening recordings during the last 5 min of control period (B) and orexin administration (D).

We then examined how intraburst and interevent dynamics were altered by estradiol and orexin. Intraburst interval was not altered in the presence of estradiol or orexin (Fig. 6C, P > 0.19). In contrast, estradiol increased burst duration during the control period (Fig. 6D, P < 0.05). Interestingly, there was no effect of estradiol on number of spikes per burst (Fig. 6E, P = 0.11). Given the precision of the intraburst interval data, this suggests either a type 1 error in detecting effects of estradiol on burst duration or a type 2 error in detecting spikes per burst. Based on the mean increase in burst duration, approximately 25% of bursts would have an additional spike in the presence of estradiol, an effect that is likely to be biologically subtle. In contrast to the effects of estradiol during the control period, orexin robustly altered burst dynamics in the opposite direction, reducing both spikes per burst and burst duration, regardless of estradiol milieu (Fig. 6, D and E).

Discussion

As the final common pathway for central control of fertility, GnRH neurons integrate multiple inputs to determine whether reproduction is appropriate, given the resources available. Steroids (1–11), metabolic cues (18–25, 51–55), and signaling associated with the day-night/sleep-wake cycle (12, 13) are among the factors determining final GnRH output. Orexin neurons are poised to integrate many of these cues (26, 32, 41, 42) and make connections with GnRH neurons (43). The present study investigated the effect of orexin on GnRH neurons and whether this effect was altered by estradiol or time of day. In spontaneously active GnRH neurons, orexin inhibited activity, independent of the time of day or estradiol. Orexin reduced the number of single spikes and bursts, spikes per burst, and burst duration but did not affect intraburst interval. Orexin may thus inhibit GnRH firing activity by targeting mechanisms involved in both action potential initiation and burst maintenance.

Previous studies have demonstrated both inhibitory and excitatory effects of orexin administration on GnRH and/or LH release. This disparity may be due to the variety of models used and the times of the day at which the effects were investigated. Intracerebroventricular (ICV) administration of orexin to OVX rats inhibited LH pulse frequency and mean LH secretion in some studies (35, 37, 40) but had no effect in another (34). However, in the study in which orexin had no effect in OVX animals, orexin did inhibit LH pulse frequency in rats treated with low-dose estradiol implants that did not affect pulsatile LH release during the control period (34). In contrast, 48 h after a larger dose of estradiol provided by injection of estradiol-benzoate, orexin administered directly into the preoptic area of rats had a transient stimulatory effect on mean LH levels, whereas an orexin injection in the arcuate region had no effect (38). OVX rats primed with estrogen and progesterone exhibited both increases and decreases in LH in response to ICV orexin (34, 35, 38), but samples were obtained at different times of day and relative to steroid treatment. In a similar estrogen-progesterone priming model, orexin antibodies blocked the LH surge in fed animals, suggesting a stimulatory effect of orexin itself (34). However, ICV orexin failed to rescue the LH surge in fasted animals, suggesting additional mediators may be involved (34). Discrepancies in the effect of orexin on LH release in vivo may be in part attributable to additional steroidal effects at the pituitary to affect response to GnRH. In this regard, when GnRH release from hypothalamic explants from OVX-estradiol benzoate-injected rats was studied directly, orexin stimulated GnRH release (38). Orexin treatment acutely increased GnRH release from immortalized GT1–7 cells as well as increased GnRH mRNA expression after several hours, suggesting a stimulatory effect of orexin on GnRH neurons (36).

We examined the effect of orexin by recording directly from GnRH neurons in brain slices from an animal model exhibiting daily switches in estradiol feedback action. Under these conditions, application of orexin consistently inhibited GnRH neuron firing frequency in cells exhibiting spontaneous activity, regardless of estradiol or time of day. These data are thus consistent with previous studies demonstrating orexin decreases LH release when given ICV and would suggest orexin suppresses GnRH release. However, these data conflict with results from previous experiments demonstrating orexin increased GnRH release in hypothalamic explants and immortalized GT1–7 cells (36, 38). One possible explanation for the discrepancy in results is that the brain slice preparation used in the present study is different from both of these other systems, having both advantages and disadvantages. An advantage is to record directly from an identified GnRH neuron in a slice that maintains a portion of the afferent network in a thin-enough preparation to minimize the effects of hypoxia. It is difficult, however, to perform biochemistry or long-term treatments, and some of the afferent network is clearly removed. Another difference between our study and other in vitro studies is the dose at which effects were observed. In the present study, 1–100 nm orexin reduced GnRH neuron activity, but as the plateau of the dose-response curve for this preparation was reached, higher doses, such as 1 μm that were used in the other studies, were not examined. From a physiological standpoint, the role of orexin in homeostatic mechanisms during negative energy balance (27, 56) is consistent with its inhibition of GnRH neuron activity. Negative energy balance strongly inhibits GnRH activity (18–20, 22). Furthermore, orexin immunoreactivity has been shown to increase after paradoxical sleep deprivation (57), a stressor that has previously been shown to adversely affect reproductive cyclicity and metabolism (57, 58). Orexin may help signal reduced metabolic resources or impaired sleep status to GnRH neurons, leading to their inhibition.

In contrast to the consistent reduction in overall GnRH firing activity observed in OVX and OVX+E evening recordings, no effect of orexin was observed in OVX+E morning recordings. In the animal model used, estradiol-negative feedback inhibits GnRH neuron activity during the morning. Previous work has suggested that this quiescence is due to reduced stimulatory fast synaptic transmission to GnRH neurons (59–62) and altered intrinsic properties (63, 64). Interestingly, orexin expression is regulated in a diurnal manner (65, 66) and would be highest during morning recordings in our model. This suggested the hypothesis that GnRH neuron quiescence in OVX+E morning recordings is due at least in part to an estradiol-mediated increase in endogenous orexin input to GnRH neurons. Blockade of OX1R, which mediates inhibitory effects of orexin on GnRH neurons, however, failed to increase GnRH neuron activity during negative feedback, causing us to reject this hypothesis. It is important to bear in mind that time of day is important in this system, and it is possible that blocking orexin action at a different time would alter the course of negative feedback.

In spontaneously active GnRH neurons, orexin exhibited a consistent inhibitory effect that was similar in the presence and absence of estradiol. Furthermore, because orexin caused a similar suppression throughout the wide range of control firing frequencies observed among OVX recordings and OVX+E evening recordings, orexin may be acting through mechanisms that directly alter the firing pattern of GnRH neurons, rather than through downstream pathways regulated by cellular activity. Analysis of the firing patterns of GnRH neurons in response to estradiol and orexin could help sculpt future investigations into the underlying mechanisms of orexin action. Orexin decreased the number of single spikes and bursts as well as reducing the within-burst parameters of spikes per burst and burst duration. These observations suggest orexin may act via at least two potential mechanisms: one that reduces spike initiation, based on increased interevent intervals, and another that shortens bursts.

OX1R is coupled to Gq, and in other neuronal populations, it has been shown to act through a variety of mechanisms including alterations to intracellular calcium via activation of L-type calcium channels, activation of protein kinase C and activation of a nonselective cation channel (67–69). These observations fit well with mechanisms that may regulate GnRH spike initiation and burst activity. For example, slow afterhyperpolarization currents have been proposed to modulate intraburst dynamics, such as number of spikes per burst as well as the intervals between bursts in GnRH neurons (70–72). This modulation is dependent upon extracellular Ca²⁺ influx through L-type calcium channels or Ca²⁺-activated K⁺ channels (70). Blocking hyperpolarization-activated currents, a nonselective cation current, reduces overall spontaneous firing activity of GnRH neurons by reducing the number of single spikes and spikes per burst, similar to the current observations with orexin, and may alter the action potential waveform, thereby contributing to changes in intraburst and/or interevent intervals (49, 73). Furthermore, orexin can affect secretion of γ-aminobutyric acid and glutamatergic transmission (67, 74), providing an indirect mechanism to influence GnRH neural activity.

In contrast to the inhibitory action of orexin, the positive feedback action of estradiol increased the number of bursts and single spikes as well as burst duration. Because the effect of orexin is not altered in the presence of estradiol, it is likely these two modulators of GnRH neuronal activity are working through independent mechanisms to alter the intraburst and interevent dynamics of GnRH neurons. Notably, estradiol has been shown to influence intracellular excitability through changes to potassium (transient outward A-type K⁺ current, delayed-rectifier K⁺ current, ATP-sensitive potassium channel) (64, 75, 77) and calcium (high and low voltage activated) conductances (63, 78) to increase spontaneous single spiking and burst firing by affecting slow afterdepolarization currents mediated by sodium flux (80) and to indirectly adjust spontaneous activity through signaling cascades (reviewed in Refs. 73 and 76) and regulation of kisspeptin action (79). These methods of action provide pathways for estradiol regulation of GnRH single spiking and burst firing activity that may be independent from the mechanisms of orexin action.

The present study demonstrates for the first time an electrophysiological response of GnRH neurons to orexin. These data support and extend previous work and establish orexin as an inhibitor of GnRH neuronal activity via the OX1R. Future studies will address the mechanisms by which orexin alters GnRH neural activity.