Evidence That the LH Surge in Ewes Involves Both Neurokinin B–Dependent and –Independent Actions of Kisspeptin

Abstract

Recent evidence has implicated neurokinin B (NKB) signaling in the retrochiasmatic area (RCh) of the ewe in the LH surge. To test this hypothesis, we first lesioned NK3R neurons in this area by using a saporin conjugate (NK3-SAP). Three weeks after bilateral injection of NK3-SAP or a blank control (BLK-SAP) into the RCh, an LH surge was induced by using an artificial follicular-phase model in ovariectomized ewes. NK3-SAP lesioned approximately 88% of RCh NK3R-containing neurons and reduced the amplitude of the estrogen-induced LH surge by 58%, an inhibition similar to that seen previously with intracerebroventricular (icv) infusion of a KISS1R antagonist (p271). We next tested the hypothesis that NKB signaling in the RCh acts via kisspeptin by determining whether the combined effects of NK3R-SAP lesions and icv infusion of p271 were additive. Experiment 1 was replicated except that ewes received two sequential artificial follicular phases with infusions of p271 or vehicle using a crossover design. The combination of the two treatments decreased the peak of the LH surge by 59%, which was similar to that seen with NK3-SAP (52%) or p271 (54%) alone. In contrast, p271 infusion delayed the onset and peak of the LH surge in both NK3-SAP– and BLK-SAP–injected ewes. Based on these data, we propose that NKB signaling in the RCh increases kisspeptin levels critical for the full amplitude of the LH surge in the ewe but that kisspeptin release occurs independently of RCh input at the onset of the surge to initiate GnRH secretion.

Because ovulation is a key event in the female reproductive cycle and is essential for fertility, many studies have explored the neuroendocrine mechanisms underlying the surge in GnRH and LH responsible for follicular rupture. It has been recognized for 60 years that sustained secretion of estradiol (E₂) from the preovulatory follicle is the primary endocrine signal triggering the GnRH and LH surge (1). Moreover, it is also clear that in many species, this positive feedback action of E₂ occurs in both the brain and the pituitary (1). However, there appear to be substantial species differences in the sites of action of estrogen-positive feedback in the brain (1, 2). The neural systems responsible for the GnRH surge have been well-described in rodents and include (i) estrogen-positive feedback in a group of kisspeptin neurons located in the rostral periventricular area of the third ventricle (3) and (ii) a time-of-day signal from the suprachiasmatic nucleus to both these kisspeptin neurons and GnRH cell bodies in the preoptic area (POA) (4). Much less is known about these systems in sheep and primates, but it is generally accepted that estrogen-positive feedback occurs in the medial basal hypothalamus (and pituitary) in both (5, 6).

Early work in sheep with use of local administration of E₂ (6) and Fos expression (7) pointed to the ventromedial hypothalamus and arcuate nucleus (ARC) as sites of estrogen-positive feedback. Not surprisingly, most recent work since the discovery of the role of kisspeptin in control of GnRH (8, 9) has focused on neurons containing this peptide. Lateral ventricle (LV) infusion of a KISS1R antagonist (p271), similar to one (p234) that completely blocked the LH surge in rats (10), decreased the amplitude of the LH surge in ewes by only ∼50% (11), indicating that kisspeptin contributes to, but is not solely responsible for, GnRH release during the surge. Kisspeptin neurons in sheep are found scattered in the POA and concentrated in the ARC (12); the latter are known as KNDy neurons because they also contain neurokinin B (NKB) and dynorphin (13, 14). Based on changes in Fos and kisspeptin expression, it appears that both kisspeptin populations participate in control of the GnRH/LH surge in this species (15). Thus, Fos expression increases only in KNDy neurons at the beginning of estrogen-positive feedback (16), in both populations during the GnRH/LH surge (17), but not in either population late in the follicular phase (16, 17). Exogenous E₂ also increases expression of kisspeptin mRNA and protein in the POA population (18). In contrast, estrogen inhibits kisspeptin expression in KNDy neurons (18, 19), but levels of kisspeptin protein and mRNA increase in the middle/caudal ARC during the follicular phase (16, 17, 20), suggesting that a subpopulation of KNDy neurons may participate in estrogen-positive feedback. Finally, there is a marked sexual dimorphism in kisspeptin expression, with higher levels in females, in both populations in sheep (14). This is similar to that seen in rodent RVRP kisspeptin neurons (4) and may contribute to the sexual dimorphism in estrogen-positive feedback in these species.

In light of the report that mutations in NKB signaling caused infertility in humans (21), considerable attention during the last decade has focused on the role of this tachykinin in control of GnRH (22). Most of this work has focused on episodic GnRH secretion (23–25), but we have also obtained evidence that the actions of NKB in the ovine retrochiasmatic area (RCh) are part of the neural circuitry responsible for the GnRH/LH surge. Specifically, local administration of senktide, an agonist to the NKB receptor (NK3R), to the RCh produced surge-like secretion of LH (26–28). More importantly, RCh microimplants containing an antagonist to NK3R (SB222200) reduced the amplitude of the estrogen-induced LH surge by ∼50% (27). This effect was similar to that seen with a KISS1R antagonist (11), so we proposed that NKB actions in the RCh are mediated by kisspeptin release from KNDy neurons. This hypothesis was supported by evidence (28) that (i) RCh neurons project to KNDy neurons, (ii) administration of senktide to the RCh increased Fos expression in KNDy neurons, and (iii) intracerebroventricular (icv) infusion of a KISS1R antagonist (p271) blocked the LH increase in response to senktide microimplants in the RCh.

Despite considerable evidence for this hypothesis, there are two important caveats. First, the effective spread of SB222200 from the RCh microimplants is not known. Second, the data supporting a role for KNDy neurons in mediating NKB actions come solely from pharmacological activation of NK3R neurons in the RCh. These experiments addressed these weaknesses by (i) lesioning RCh NK3R-containing neurons with NK3-SAP, an NK3R agonist conjugated to saporin (29), and (ii) determining whether the combination of NK3-SAP lesions and p271 infusions produce greater, or similar, inhibition of an estrogen-induced LH surge compared with either treatment alone. If NK3R neurons in the RCh act via kisspeptin, there should be no additive effect of blocking both systems.

Materials and Methods

Animals

Adult (age 3 to 7 years), multiparous mixed-breed black-faced ewes were obtained from a local producer, quarantined in an outdoor paddock for a minimum of 2 weeks before use, and then moved into an indoor facility at least 6 days before any experimental work. They were fed timothy grass pellets (Standlee Hay Company, Inc., Kimberly, ID) and Triple Crown Complete Pellets (Triple Crown, Wayzata, MN) twice daily to maintain body weight and had free access to water and mineral blocks. Indoors, animals were maintained two per pen with controlled temperature and lighting that was automatically adjusted to mimic natural changes in day length. All experiments were done in the late breeding season (January to February). Blood samples (3 to 4 mL) were taken by jugular venipuncture into heparinzed tubes; serum was collected and stored at −20°C until assayed for LH concentrations.

Drugs

NK3-SAP ([MePhe7]-NKB conjugated to saporin) and BLK-SAP (a nonspecific peptide conjugated to saporin) were obtained from Advanced Targeting Systems (San Diego, CA) at concentrations of 0.7 mg/mL and 1.0 mg/mL, respectively. The latter was diluted to 0.7 mg/mL with sterile PBS (0.1 M phosphate buffer with 0.9% sodium chloride), both divided into aliquots in 7.5-µL volumes into sterile vials, and stored at −20 °C (if used within 1 month) or −80 °C (if stored >1 month before use). This concentration was based on preliminary work indicating that 0.7 mg/mL, but not 0.5 mg/mL, eliminated almost all NK3R immunoreactivity in the ovine RCh (data not shown). Aliquots were thawed the day of injection. The KISS1R antagonist (p271) (11) was purchased from EZBiolabs (Carmel, IN) and diluted to 3.33 mg/mL with sterile saline; 1-mL aliquots were stored at −20 °C until the day of use. This dose was chosen because we wanted to replicate as closely as possible previous work in which this antagonist decreased the amplitude of the LH surge in ewes (11).

Surgical procedures

Ewes were ovariectomized (OVX) and given injections of saporin conjugates on the same day under gas anesthesia (2% to 5% isoflurane in O₂) using sterile techniques. Ovariectomies were performed first via midventral laparotomy as previously described (30), followed by neurosurgery. Briefly [details can be found in Goodman et al. (31)], a small portion of the dorsal surface of the brain was exposed, the sagittal sinus was ligated, and an 18-gauge needle was lowered into one LV. Radio-opaque dye was injected into the LV, and lateral and frontal radiography was used to lower bilateral guide tubes (5 mm apart) to a position 1.0 to 1.5 mm posterior to the optic chiasm. A 1-µL Hamilton syringe with a fixed needle containing NK3-SAP or BLK-SAP was lowered 3 mm beyond one guide tube and 0.5 to 1.0 mm above the base of the third ventricle. The 1-µL contents were then slowly injected over 2 minutes, and the needle was left in place for an additional 3 minutes. This injection protocol was then repeated on the other side of the hypothalamus, and the needle and guide tubes were removed. For experiment 1, the brain was protected with gel foam and the skin sutured. For experiment 2, the brain was protected with gel foam and a wire mesh, and the LV cannula was cemented in place with dental acrylic, plugged, and protected with a plastic cap (28). All ewes received dexamethasone, analgesics, and antibiotics pre- and postoperatively as previously described in Goodman et al. (31). All animal work was approved by the West Virginia University Animal Care and Use Committee and followed National Institutes of Health guidelines for the use of animals in research.

Experimental Protocols

Animal model

The positive feedback actions of estrogen were examined by using well-established steroid treatments in OVX animals designed to mimic concentrations and patterns of E₂ and progesterone during the ovine estrous cycle (30, 32). At the time of ovariectomy, a 1-cm-long Silastic implant (inner diameter, 0.34 cm; outer diameter, 0.46 cm; Dow Corning Corp., Midland, MI) containing E₂ was placed subcutaneously and two progesterone-containing controlled intravaginal drug release devices (CIDRs; Zoetis, Parsippany, NJ) were inserted intravaginally. The CIDRs were then removed 8 to 9 days later to simulate a luteal phase followed by luteolysis. The short ovine follicular phase was mimicked by subcutaneous insertion of four 3-cm-long E₂ implants 24 hours after CIDR removal, which induced an LH surge about 24 hours later.

Experiment 1: Determining effects of lesioning NK3R-containing neurons in RCh

After ovariectomy, ewes were given an E₂ implant and two CIDRs, and then injected bilaterally in the RCh with NK3-SAP (n = 5) or BLK-SAP (n = 3). The CIDRs were removed 8 to 10 days later, and two new CIDRs were inserted 3 days later to simulate two luteal phases and allow saporin time to lesion NK3R-containing neurons. CIDRs were removed 9 days later, and four long E₂ implants were inserted 24 hours after CIDR removal. Blood samples were collected for analysis of LH pulses shortly before CIDR removal (every 12 minutes for 4 hours) and just before insertion of E₂ implants (every 10 minutes for 3 hours). LH surges were monitored in samples collected every 4 hours from 0 to 12 hours after start of E₂ treatments, then every 2 hours for the next 24 hours, and every 4 hours from 36 to 48 hours. Tissue was collected the next day for immunohistochemical analysis.

Experiment 2: Determining whether the effects of NK3-SAP lesions and KISS1R antagonist are additive

The protocol described in experiment 1 was replicated by using ewes receiving NK3-SAP (n = 7) or BLK-SAP (n = 5) injections, but with three alterations. The three changes were as follows: (i) Ewes received p271 or saline (SAL) infusions during the artificial follicular phase, (ii) these infusions were done in two consecutive follicular phases using a crossover design, and (iii) LH pulses were not monitored. Specifically, during the first artificial follicular phase, animals received p271 antagonist (200-µg loading dose, 300 µg/h) or SAL (60-µL loading dose, 90 µL/h) from 16 to 24 hours after insertion of E₂ implants (n = 6/treatment). These infusions replicated previous icv treatments with p271 that decreased the amplitude of the LH surge in ewes (11). Blood samples were collected every 2 to 4 hours, as described in experiment 1. At the end of these infusions, the long E₂ implants were removed, two new CIDRs were inserted, and this protocol was repeated with a crossover of p271 and SAL treatments. Tissue was then collected for immunohistochemical analysis. Because of technical difficulties with lateral ventricle infusions, one of the BLK-SAP–injected ewes was omitted from this study.

Tissue collection

Animals were treated with heparin (20,000 U intravenously) 10 minutes before and immediately before euthanasia via overdose (8 to 16 mL intravenously) of Euthasol (Patterson Veterinary, Bessemer, AL). When breathing stopped and there was no eye reflex, the carotids were cut and the head removed and perfused with 6 L of 4% paraformaldehyde in 0.1M phosphate buffer with 0.1% sodium nitrite. A block of tissue including the hypothalamus and POA was removed and stored in paraformaldehyde solution overnight at 4°C. Tissue was then stored in 30% sucrose in phosphate buffer at 4°C for at least 1 week. Tissue was frozen and 45-μm-thick coronal sections were cut with a microtome equipped with a freezing stage. Sections were collected in 10 series (450 μm apart) and stored at −20°C in cryoprotectant solution.

Immunocytochemistry

General

All immunostaining procedures were performed on free-floating sections with gentle agitation on a horizontal orbital shaker. For each analysis, a series of sections from all animals were processed simultaneously. Sections were washed with PBS between all incubations. Before incubation with primary antibodies, sections were treated with 1% H₂O₂ (catalog no. H323-500, Fisher Scientific, Pittsburgh, PA) for 10 minutes and with the antibody incubation solution (PBS containing 4% normal goat serum (catalog no. 005-000-121, Jackson ImmunoResearch, West Grove, PA) and 0.4% Triton X-100 (catalog no. BP151-100, Fisher Scientific). This antibody incubation solution was used for all primary and secondary antibody incubations. Following immunostaining, sections were mounted on SuperFrost Plus glass slides (catalog no. 22-037-246, Fisher Scientific) and coverslipped with DPX Mountant (Electron Microscopy Sciences, Hatfield, PA). Antibodies used for kisspeptin (13), NK3R (33), and tyrosine hydroxylase (TH) (28) have been previously characterized for use in the sheep brain.

Immunohistochemistry of NK3R and NeuN.

A series of sections from all animals in experiments 1 and 2 encompassing the hypothalamus were immunoprocessed for NK3R and for dual labeling of NK3R and NeuN, a commonly used neuronal marker. Sections were incubated successively with rabbit anti-NK3R (dilution 1:20,000, catalog no. NB300-102; Novus Biological, Littleton, CO; RRID: AB_350431, 17 hours) (34) biotinylated goat anti-rabbit secondary antibody (1:500, 1 hour, catalog no. BA-1000, Vector Laboratories, Burlingame CA), Vectastain ABC-elite (1:500, 1 hour; catalog no. PK6100, Vector Laboratories), and 3,3′-diaminobenzidine tetrahydrochloride (DAB, 0.2 mg/mL, catalog no. D5905, Sigma-Aldrich, St. Louis, MO) with 0.08% nickel sulfate hexahydrate (catalog no. AC211085000, Acros Organics, Morris Plains, NJ), and 0.012% hydrogen peroxide for 10 minutes to create a black reaction product. Next, for sections dual-labeled for NeuN, sections were sequentially incubated in mouse anti-NeuN (dilution 1:10,000, 17 hours, catalog no. MAB377, Millipore, Billerica, MA; RRID: AB_2298772) (35), biotinylated goat anti-mouse secondary antibody (1:500, 1 hour, catalog no. BA-1000, Vector Laboratories), ABC-elite (1:500; 1 hour), and DAB (0.2 mg/mL) with 0.012% hydrogen peroxide (10 minutes) to create a brown reaction product.

Immunohistochemistry of kisspeptin or TH.

A series of sections from experiment 2 containing the RCh or ARC were processed for visualization of TH or kisspeptin, respectively. Sections were incubated with polyclonal rabbit anti-kisspeptin (dilution 1:25,000, catalog no. AB-9754, Millipore, Darmstadt, Germany, RRID: AB_2296529) (36) or monoclonal mouse anti-TH (dilution 1:15,000, catalog no. MAB5280, EMD Millipore, Billerica, MA, RRID: AB_2201526) (37) for 17 hours. Next, sections were incubated with biotinylated goat anti-rabbit antibody (dilution 1:500, catalog no. BA-1000) or biotinylated goat anti-mouse antibody (dilution 1:500, catalog no. BA-1000, Vector Laboratories) for 1 hour and Vectastain ABC-elite (1:500) for 1 hour. Antigen was visualized using DAB (0.2 mg/mL) with 0.08% nickel sulfate hexahydrate (Acros Organics), and 0.012% hydrogen peroxide for 10 minutes.

Analyses

LH assays

Plasma samples (50- to 200-μL aliquots) were assayed in duplicate for LH by using a double liquid phase radioimmunoassay that has been previously validated for use in sheep (28). Assay reagents from the National Hormone and Peptide Program were used, and LH concentrations are expressed in terms of NIH S24. The sensitivity of the LH assays averaged 0.10 ± 0.02 ng/mL, and the inter- and intra-assay coefficients of variation were 6.9% and 5.1%, respectively.

Analysis of immunoreactive neurons

In each of the two experiments, for each ewe, the number of immunoreactive (ir) cells single-labeled for NK3R or all cells ir for NeuN (including those also labeled for NK3R) in the RCh were counted under brightfield illumination using a Leica DMR microscope with a photo tube attached (Leica Microsystems GmbH, Wetzlar, Germany). Cell counts were performed by two independent experimenters blinded to the experimental group. Cells were counted in a standard analysis area (1200 × 1400 µm) placed ventral to the fornix in the RCh (for sections dual-labeled for NK3R and NeuN) in four sections in both hemispheres, resulting in eight data points per animal. Counts were averaged first for each hemisphere separately to confirm that there were no differences within animals between hemispheres (left: 0.66 ± 0.6 cells/mm²; right: 0.64 ± 0.39 cells/mm²) and thus the lesions were bilateral. Cell numbers were then averaged over all eight hemisections for each animal and thus represent an average number of ir cells in one hemisphere through the RCh. In addition, in experiment 2, neurons ir for TH in the RCh were analyzed as described above, except that a standard area of analysis was not required because of the restricted distribution of TH cells in this area. Finally, in experiment 2, cells ir for kisspeptin were counted in both hemispheres of two sections for each of the rostral, middle, and caudal areas within the ARC. For each animal, hemispheric cell counts were averaged for each rostrocaudal level of the ARC to obtain a single cell count for one hemisphere.

For publication, images of immunostained sections were captured at magnification of ×10 or ×40 using a digital microscope camera (MBF Bioscience, Williston, VT) attached to a Leica DMR microscope and using identical camera settings for sections of NK3-SAP–treated and BLK-SAP–treated animals. Images were opened in Adobe Photoshop (Adobe, San Jose, CA) for preparation of figures and were not altered in any way.

Statistical analyses

LH pulses were defined as previously described (38), and LH pulse amplitudes and mean LH were compared between groups by a Student t test. LH pulse frequencies were compared by using the Wilcoxon-Mann-Whitney test. An LH surge was defined as a sustained rise in LH that lasted at least 8 hours with a maximum point at least four times the baseline LH value (average of samples at time 4 to 12 hours). Starting time was defined as the first point of a continuous increase in concentrations that was >2 SDs above baseline values, and the ending time was the last point in this continuous set that was >2 SDs above baseline values. Statistical comparisons of LH surge parameters in experiment 1 were done by using Student t tests, whereas the data in experiment 2 were analyzed by two-way ANOVA with repeated measures (main effects: NK3R-SAP vs BLK-SAP and SAL vs p271). If significant effects were observed, we performed pairwise comparisons using the Holm-Sidak method. Differences in number of cells ir for NK3R, TH, NeuN, or kisspeptin were compared between ewes injected with NK3-SAP and BLK-SAP in both experiments using t tests. A P value <0.05 was considered to indicate statistical significance.

Results

Cell-specific lesions of NK3R-containing neurons in RCh

Experiment 1

Four of the five ewes administered NK3-SAP had markedly fewer NK3R-ir cell bodies in the RCh and numbers of NK3R-containing cells were reduced to 10.8% of BLK-SAP–treated ewes (range, 1.5% to 30.6%; P = 0.004) (Fig. 1). The number of NK3R-ir cell bodies in the fifth ewe was similar to that in controls (108.7% of BLK-SAP controls and 8 SDs from the mean of the other NK3-SAP–treated ewes), so data from this animal were not used in analyses. Quantitative analysis of NK3R in parallel sections that were dual-stained for NK3R and NeuN further confirmed these observations, and a significant decrease in RCh NK3R-ir neurons was noted (to 5% of BLK-SAP controls; P < 0.0001; Fig. 1). In contrast, the overall neuronal number based on NeuN staining did not differ between groups (NK3-SAP ewes had 94.2% compared with BLK-SAP controls; P = 0.53; Fig. 1), confirming the specificity of the NK3R neuron lesions.

Figure 1.

Effect of NK3-SAP lesions in the RCh of the ewe. Representative photomicrographs of RCh sections from ewes injected with (A and C) BLK-SAP or (B and D) NK3-SAP are shown. Immunolabeling for NK3-ir cells in (A and B) sections stained for the receptor showed a marked reduction of RCh NK3R-ir cells (indicated by arrows) in ewes (B) with NK3-SAP injections compared with (A) BLK-SAP. (C and D) In contrast, immunolabeling for NeuN in sections processed for NueN (brown) and NK3R (black) show no qualitative change in NeuN immunolabeled cells between BLK-SAP and NK3-SAP. Boxed inserts show representative NeuN cells at higher magnification for each group; arrows indicate NeuN cells that colocalize NK3R-ir in control, but not in NK3-SAP, animal. (A–D) Scale bars, 100 μm and [boxed inserts in (C) and (D)] 40 μm. (E and F) Mean ± SEM number of NK3R-ir cells (left) and NeuN-ir (right) per square millimeter in the RCh from BLK-SAP–injected (open bars) and NK3-SAP–injected (solid bars) ewes for experiments 1 and 2. *P < 0.05 vs BLK-SAP–injected ewes. fx, fornix.

Experiment 2

Similar to experiment 1, expression of NK3R in the RCh was decreased by NK3-SAP injection in all but one of the seven ewes; data from that animal were 7.4 SDs above the mean of the rest of the NK3-SAP group and similar to the BLK-SAP control group (106% of BLK-SAP controls); they were therefore excluded from subsequent analyses. Statistical analysis demonstrated a significant decrease in NK3R-containing cells (to 22.2% of controls), but not NeuN-positive cells (to 88% of controls) in the RCh (Fig. 1F). Similarly, the number of A15 dopaminergic neurons, which do not contain NK3R (26), was unaffected by NK3-SAP injections (100.8% of controls; average numbers BLK-SAP: 6.0 ± 0.4; NK3-SAP: 6.1 ± 0.37; P = 0.846). Finally, we examined whether NK3-SAP injections affected ARC KNDy cells, as these also express NK3R (33, 39) and are adjacent to the injection sites. There was no significant difference between BLK-SAP and NK3-SAP treatments in the total number of ARC KNDy neurons (average number BLK-SAP: 36.1 ± 5.3; NK3-SAP: 27.7 ± 9; P = 0.133), but there was a tendency (P = 0.052) toward fewer KNDy neurons in the rostral ARC after NK3-SAP injections (Fig. 2), probably because these neurons are near the injection site. However, this difference did not reach statistical significance, and there was no correlation between kisspeptin cell number in the rostral ARC and the peak of the LH surge (R = 0.57, R² 0.326).

Figure 2.

Effects of BLK-SAP and NK3-SAP injections in RCh on ARC kisspeptin expression. Top panels depict representative images showing examples of kisspeptin-ir (Kiss-ir) cells in the middle ARC of (A) BLK-SAP–treated and (B) NK3-SAP–treated ewes. (C) No differences were detected in Kiss-ir cell numbers between BLK-SAP–treated (white bars) and NK3-SAP–treated (black bars) ewes. Values represent means ± SEM. Scale bars, 100 μm.

Effects of NK3-SAP lesions on LH secretion

Experiment 1

The peak of the estrogen-induced LH surge in all four NK3R-lesioned ewes was clearly less than that in controls (Fig. 3), and the mean peak concentration in this group was 42% of the mean peak in controls (20.5 ± 2.5 ng/mL vs 48.7 ± 2.8 mg/mL; P < 0.0001). The ewe without decreased NK3R expression in the RCh (ewe 103) had a normal LH surge (Fig. 3). In contrast, both the start (23.0 ± 2.4 hours after E₂ vs 21.3 ± 1.8 hours after E₂) and peak (28.5 ± 1.3 hours after E₂ vs 26.0 ± 2.0 hours after E₂) occurred at similar times in lesioned and control groups. Lesions of NK3R-containing neurons in the RCh also had no effect on episodic LH secretion (Table 1) in the presence of either progesterone and E₂ (before CIDR removal) or E₂ alone (before insertion of the four long E₂ implants).

Figure 3.

LH surges in five individual ewes that were injected with NK3-SAP in the RCh (white circles) compared with mean (±SEM) LH concentrations in control ewes normalized to the peak of the LH surge (black circles). Bottom-right panel depicts mean (±SEM) LH concentrations in both groups. The number after ewe identification in each panel is the mean NK3R-ir neurons/section in the RCh of that animal. The number of these neurons in ewe 103 (top right panel) was the same as in control ewes (16.3 ± 3.2 NK3R cells/section), so data from this animal were not included in mean LH values for the NK3-SAP–treated ewes.

Table 1.

Effect of Lesions of NK3R-ir Neurons in the RCh on Episodic LH Secretion

Variable	OVX + E2 + P Ewes		OVX + E2 Ewes
	BLK-SAP	NK3-SAP	BLK-SAP	NK3-SAP
Mean LH, ng/mL	1.65 ± 0.15	1.59 ± 0.18	3.83 ± 0.50	3.06 ± 0.44
Frequency, n/bleed	1.67 ± 0.33	1.50 ± 0.29	3.67 ± 0.67	2.50 ± 0.96
AMPL, ng/mL	2.45 ± 0.51	1.20 ± 0.37	3.36 ± 0.73	2.72 ± 0.11

Values are the mean ± SEM. There were no significant differences between treatment groups, but LH pulse amplitude in OVX + E₂ + P ewes (samples collected before CUDR removal) tended (P = 0.098) to be lower in the NK3-SAP–treated ewes, but the lack of significance could have been due to the low number of animals used.

Abbreviation: P, progesterone.

Experiment 2

As shown in Fig. 4, NK3-SAP treatment decreased the peak of the estrogen-induced LH surge in SAL-infused ewes by the same amount as in experiment 1 (NK3-SAP: 26.2 ± 5.4 ng/mL vs BLK-SAP: 61.2 ± 7.6 ng/mL), without affecting the timing of this surge. A similar decrease in the peak of the LH surge (28.2 ± 4.4 ng/mL) was produced by the KISS1R antagonist infusion in ewes in which RCh NK3R neurons were not lesioned (Fig. 4), confirming previous data in ewes (11). However, in contrast to this earlier report (11), p271 infusion delayed both the start (SAL: 22.5 ± 0.5 vs p271: 29.5 ± 1.0 hours after E₂) and peak (SAL: 26.5 ± 0.5 vs p271: 34.2 ± 1.0 hours after E₂) of the LH surge.

Figure 4.

Mean (±SEM) LH concentrations during the LH surge in ewes receiving BLK-SAP injections and icv SAL infusions (black circles), BLK-SAP injections and icv p271 infusions (red circles), NK3-SAP injections and icv SAL infusions (white circles), and NK3-SAP and icv p271 infusions (green circles). LH concentrations were normalized to the peak of the LH surge in each group and plotted based on the mean time of the LH peak in each group. The top panel compares each individual treatment with controls (BLK-SAP and SAL infusions), and the bottom panel compares the combined treatment with each individual treatment.

Infusion of the KISS1R antagonist in ewes injected with NK3-SAP produced no further decrease in the peak of the estrogen-induced LH surge but delayed the timing of the surge as it did in ewes treated with BLK-SAP (Fig. 4). Two-way ANOVA of peak LH concentrations indicated significant effects of NK3-SAP (P = 0.022) and p271 (P = 0.004) and an interaction (P = 0.005) of the two. Similar analysis of the timing of the surge indicated a significant effect of p271 on the start (P = 0.018) and peak (P = 0.007) of the LH surge but no effect of NK3-SAP and no interaction. Although p271 delayed the LH surge, it had no significant effect on its duration (SAL: 15.2 ± 1.2 vs p271: 14.6 ± 1.4 hours).

Discussion

These results confirm and extend previous data that NKB signaling in the RCh (27) and kisspeptin release in the hypothalamus (11) are each necessary for the full LH surge in the ewe. However, in contrast to its role in the ARC (24), NKB signaling in the RCh likely plays little role in the control of episodic LH secretion. These data also strongly support the hypothesis that RCh NK3R-containing neurons and kisspeptin cells function in series during the estrogen-induced LH surge. Finally, we observed that the KISS1R antagonist, p271, delayed the timing of the LH surge, an effect that was not seen in previous work using this antagonist in ewes (11).

Our earlier work reported that stimulation of NK3R signaling in the RCh with senktide produced surge-like LH secretion (26, 27), and local administration of the NK3R antagonist, SB222200, to this area produced a 42% decrease in the peak of an estrogen-induced LH surge, without affecting the timing of the surge (27). In contrast, infusion of a different NK3R antagonist into the LV failed to affect an LH surge induced by estradiol benzoate in OVX ewes (40). Thus, the first aim of this study was to test the role of RCh NK3R-containing neurons in the LH surge by using a different approach that allowed us to assess the number of these neurons affected by treatment. The effects of NK3-SAP injections that reduced the number of NK3R-ir cells in the RCh by >78% to 88% were qualitatively identical to those of SB222200 (27). Although the results of these two studies cannot be statistically compared, the degree of inhibition of the LH peak produced by NK3-SAP lesions in this study (42.1% of controls in experiment 1; 42.8% in experiment 2) is greater than that observed with SB222200 (58.6% of controls). Thus, SB222200 may not have affected as many NK3R-containing RCh neurons as the NK3-SAP injections in this study. Similarly, the most likely explanation for the lack of effect of infusion of a NK3R antagonist into the LV is that it failed to reach a sufficient number of RCh NK3R-containing neurons, which are >2 mm lateral to the third ventricle (33). Finally, NK3-SAP produced a modest, but not significant, decrease in KNDy cell number in the rostral ARC. It is unlikely, however, that this contributed to the inhibition of the amplitude of the LH surge because there are few KNDy neurons in this region of the ARC, and there is no evidence that they play a major role in generation of the LH surge in ewes (16, 17).

The effects of the KISS1R antagonist on the peak of the estrogen-induced LH surge in ewes injected with BLK-SAP confirms previous work with the same antagonist (11): there was a 56% decrease in surge amplitude in that study compared with a 52% decrease in this report. However, we observed a marked delay in the time of the LH surge in ewes give p271 icv, which was not observed in the previous study. This was surprising because we had attempted to replicate the previous p271 treatment as closely as possible in experiment 2. We used the same dose and duration of antagonist (11) but delayed the start of the infusion relative to estrogen treatment by 4 hours to take into account that our ewes were pretreated with progesterone, which delays the onset of the LH surge (32), whereas their animals were not. The timing of the LH surge relative to icv infusions occurred about 2 hours earlier in our experiment (i.e., the surge started at 22 hours with SAL infusions, which ended at 24 hours) compared with the previous work (the surge started at 16 hours, and infusions ended at 20 hours), but it seems unlikely that this can account for the dramatic differences in effects of p271 on the timing of the LH surge in these two studies. The most likely explanation is that our animals had an artificial luteal phase before estrogen treatment because progesterone pretreatment augments the amount of GnRH released (41) and the activation of non-GnRH neurons in the POA (42) by estrogen-positive feedback in the ewe.

It is unclear whether the KISS1R antagonist reset the neural mechanisms responsible for the GnRH surge or simply blocked GnRH secretion during treatment without altering subsequent release. In 8 of 10 ewes, the surge did not start until the end of p271 infusion, which argues for the latter explanation. On the other hand, the similar duration of the LH surge with and without p271 treatment supports resetting of the neural mechanisms. However, since the duration of the GnRH surge is much longer than the LH surge (41, 43), the normal neural mechanisms could have produced sufficient GnRH release after the end of p271 infusion to result in an LH surge of normal duration. Clearly, more work is needed to resolve this question.

Comparison of the combined effects of NK3-SAP lesions and p271 infusions was designed to test the hypothesis that NKB signaling in the RCh and kisspeptin release represent systems that function in series, rather than in parallel, during the generation of the LH surge. Thus, the observation that the peak of the estrogen-induced LH surge seen with the combined treatments was essentially identical to that observed with either treatment alone strongly supports this hypothesis. These data do not address whether kisspeptin acts via NKB-responsive RCh neurons or NKB-responsive neurons act via kisspeptin, but there is strong evidence for the latter, including that there are few, if any, kisspeptin-ir fibers in the RCh (12) and that p271 blocked the stimulatory effects of senktide in the RCh (28). Moreover, the RCh NK3R-ir neurons likely act via KNDy neurons because RCh neurons project to the mediobasal hypothalamus, not the POA (44), and administration of senktide to the RCh produced a marked increase in Fos expression in KNDy neurons (28). In contrast to the control of the amplitude of the LH surge, the ability of p271 to delay the LH surge was not affected at all by the NK3-SAP lesions, indicating that the initiation of the LH surge by kisspeptin is independent of NKB signaling in the RCh.

The physiological significance of low-amplitude LH surges, like those seen with the KISS1R antagonist or disruption of NKB-NK3R signaling in the RCh, on ovarian function is unclear because no studies have systematically evaluated the effects of different LH surge amplitudes in ewes. However, prenatal androgen treatment of ewes for 60 days in utero produced a delay and a similar-magnitude decrease in the amplitude of estrogen-induced LH surges (37%) (45) and resulted in marked defects in the luteal phase (46) when these animals were adults. Defects in follicular function were also observed in these ewes, but these could not account for the luteal deficits (46). The continuation of low-amplitude LH surges after disruption of RCh NK3R-KNDy neural circuitry also illustrates another important physiological characteristic: There is considerable redundancy in the neural systems responsible for the GnRH/LH surge in ewes. Specifically, convincing data implicate noradrenergic neurons in the brain stem (47, 48) and possibly somatostatin-containing neurons in the ventromedial nucleus (49, 50) in estrogen-positive feedback. These systems are likely responsible for the LH surges, albeit of low amplitude, in these experiments.

These observations can be integrated with other data on the hypothalamic neural systems mediating estrogen-positive feedback in the ewe using the conceptual model that divides this process into three steps: activation, transmission, and secretion (51). Activation most likely occurs in the ventromedial hypothalamus and ARC (6, 7) and includes increased activity of KNDy neurons (16). NKB-responsive neurons in the RCh then participate in the transmission phase, possibly by increasing kisspeptin expression in KNDy neurons in the middle and caudal ARC (16, 17). Finally, kisspeptin release from both KNDy (17) and POA kisspeptin (17, 52) neurons initiates GnRH secretion during the surge.

In summary, the results of this study strongly support the hypothesis that NKB-responsive neurons in the RCh act via kisspeptin release from KNDy neurons to augment the amplitude of the estrogen-induced LH surge in ewes. However, the ability of the KISS1R antagonist to delay the onset of the LH surge independent of NK3R-ir cells in the RCh indicates that kisspeptin release initiates the LH surge and that the timing of this release is independent of input from the RCh.

Glossary

Abbreviations:

ARC

arcuate nucleus

BLK

blank

CIDR

controlled intravaginal drug release device

DAB

3,3′-diaminobenzidine tetrahydrochloride

E₂

estradiol

icv

intracerebroventricular

ir

immunoreactive

LV

lateral ventricle

NKB

neurokinin B

OVX

ovariectomized

POA

preoptic area

RCh

retrochiasmatic area

SAL

saline

SAP

saporin

TH

tyrosine hydroxylase

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.