Glutamatergic Input From Arcuate Nucleus Kiss1 Neurons to Preoptic Kiss1 Neurons Is Required for LH Surge in Female Mice

Abstract

Hypothalamic kisspeptin (Kiss1) neurons are vital for maintaining fertility in the mammal. In the female rodent, Kiss1 neurons populate the anteroventral periventricular/periventricular nuclei (Kiss1^AVPV/PeN) and the arcuate nucleus (Kiss1^ARH). Kiss1^ARH neurons (also known as KNDy neurons since they coexpress neurokinin B and dynorphin) are considered the “pulse-generator” neurons that presynaptically excite gonadotropin-releasing hormone (GnRH) axons in the median eminence, whereas the Kiss1^AVPV/PeN neurons are the “surge-generator” neurons that depolarize preoptic GnRH neurons directly to drive ovulation. Traditionally, it is believed that Kiss1^ARH neurons are relatively quiet during the late follicular, preovulatory stage of the reproductive cycle due to the 17β-estradiol (E2)-mediated downregulation of the expression of the KNDy peptides. However, based on our single-cell, quantitative polymerase chain reaction and whole-cell electrophysiological recordings, we found that the messenger RNA (mRNA) expression of vesicular glutamate transporter 2 (Vglut2) mRNA and excitatory cation channels in Kiss1^ARH neurons were significantly upregulated by E2, which increased the excitability and glutamate release from these “pulse-generator” neurons. Presently, we demonstrate that optogenetic stimulation of Kiss1^ARH neurons releases glutamate to excite Kiss1^AVPV/PeN neurons via activation of both ionotropic and metabotropic glutamate receptors. CRISPR mutagenesis of Vglut2 in Kiss1^ARH neurons abolished glutamatergic neurotransmission, which significantly reduced the overall glutamatergic input to Kiss1^AVPV/PeN neurons. The mutagenesis of Vglut2 in Kiss1^ARH neurons abrogated the E2-induced luteinizing hormone surge and reduced the formation of corpus lutea, indicative of a reduced ovulatory drive in these Vglut2-mutated Kiss1^ARH mice. Therefore, Kiss1^ARH neurons appear to play a critical role in augmenting the GnRH surge through glutamatergic neurotransmission to Kiss1^AVPV/PeN neurons.

Reproduction is a quintessential homeostatic function, and hypothalamic kisspeptin (Kiss1) neurons are vital for the initiation and maintenance of fertility. In the female rodent neurons expressing kisspeptin (Kiss1) are located primarily in two distinct areas of the forebrain: the preoptic area and the mediobasal hypothalamus. The preoptic Kiss1 populations in rodents are situated in the anteroventral periventricular (AVPV; Kiss1^AVPV) and adjacent periventricular nuclei (PeN; Kiss1^PeN), and the peptide expression in these neurons is increased by 17β-estradiol (E2) (1, 2). The basal hypothalamic Kiss1 population is located in the arcuate nucleus of the hypothalamus (ARH; Kiss1^ARH) with scattered neurons also in the hypothalamic dorsomedial nucleus (3, 4). The expression of the neuropeptides, including Kiss1, neurokinin B (NKB and dynorphin, within the Kiss1^ARH neurons are all inhibited by E2 (5).

High-frequency optogenetic activation of Kiss1^ARH neurons releases NKB and dynorphin, which act to synchronize the Kiss1^ARH neuronal firing through an autoexcitatory mechanism (6). High-frequency activation of Kiss1^ARH neurons also releases kisspeptin, which controls gonadotropin-releasing hormone (GnRH) and luteinizing hormone (LH) release at the level of the median eminence (7, 8). The Kiss1^AVPV/PeN neurons send direct projections onto GnRH neurons and are essential for positive feedback regulation of GnRH and LH secretion driving ovulation in females (6, 9-11). Kiss1^ARH neurons in mice do not appear to contact GnRH cell bodies but form close anatomical contacts with distal GnRH nerve processes and exhibit neurophysiological (functional) interaction with Kiss1^AVPV/PeN neurons (6, 8, 10, 12, 13).

Since the discovery of kisspeptin neurons in the early part of this century, neuroendocrinologists have focused on the peptide expression of the Kiss1^ARH neurons and the coexpression of NKB and dynorphin in these neurons. The expression of all these peptides is downregulated by E2, and hence physiologists thought that these neurons were in the state of negative E2 feedback and therefore less excitable. However, we discovered several years ago that the expression of excitatory cation channels (eg, voltage-gated calcium channels) and the vesicular glutamate transporter 2 (Vglut2) and glutamatergic neurotransmission were increased in Kiss1^ARH neurons by E2 (12). Hence, the excitability of Kiss1^ARH neurons actually increases rather than decreases in the presence of E2. Therefore, the question became what are the downstream targets of this increased excitatory activity of Kiss1^ARH neurons and what are the physiological consequences? We previously identified that the Kiss1^AVPV/PeN neurons were a downstream target of this Kiss1^ARH excitatory output since low-frequency optogenetic stimulation of Kiss1^ARH neurons released glutamate, which depolarized and excited Kiss1^AVPV/PeN neurons (6). Therefore, we hypothesized that this excitation of Kiss1^AVPV/PeN neurons could facilitate the generation of the LH surge, mediated by Kiss1^AVPV/PeN direct excitatory kisspeptin input to GnRH neurons. We initially used conditional knockout of Vglut2 in all Kiss1 neurons to test this hypothesis, but we found that deletion of Vglut2 in all kisspeptin neurons did not alter the estrous cycle or affect fertility in these female mice (12). However, the absence of any effects of knockout of Vglut2 in Kiss1 neurons on the reproduction phenotype could have been due to the deletion of Vglut2 in all Kiss1 neurons (and nonneural cells) or some developmental compensation following embryonic deletion. Therefore, to overcome these limitations, we elected to use a viral-based CRISPR/SaCas9 approach to selectively mutate Vglut2 in Kiss1^ARH neurons in adult females. This strategy has been enhanced by creating a single–adeno-associated viral vector to induce loss of function of Vglut2 in multiple brain regions (14, 15). Indeed, we found that CRISPR/SaCas9 mutagenesis of Vglut2 dramatically reduced mRNA levels in ARH Kiss1^Cre/+ neurons (Kiss1^Cre/+Mut) in adult female mice and greatly attenuated the glutamatergic output of these neurons. With this strategy, we did not observe any effects on estrous cyclicity or LH pulsatile secretion, but we did observe a dramatic reduction in the preovulatory LH surge and a reduced number of ovarian corpora lutea. Therefore, these experiments establish that Kiss1^ARH glutamatergic transmission is vital for driving the LH surge via its glutamatergic excitatory input to Kiss1^AVPV/PeN neurons and consequently ovulation in the female.

Materials and Methods

Animals

All the animal procedures described in this study were performed in accordance with institutional guidelines based on National Institutes of Health standards and approved by the Institutional Animal Care and Use Committee at Oregon Health and Science University (OHSU) or the Brigham and Women's Hospital.

Mice

Kiss1 ^Cre/+ transgenic female mice version 2 (RRID: IMSR_JAX:033169, https://www.jax.org/strain/033169) were obtained from Dr Richard Palmiter (University of Washington, Seattle, Washington) (16). Mice were selectively bred at OHSU and Harvard Medical School. All animals were maintained under controlled temperature and photoperiod (lights on at 06:00 hours and off at 18:00 hours at OHSU, or on at 07:00 hours and off at 19:00 hours at Harvard Medical School) and given free access to food (Lab Diets 5L0D) and water under constant conditions of temperature (22-24 °C). Where specified, Kiss1^Cre/+ mice received viral injections to express channelrhodopsin 2 (ChR2) in Kiss1^ARH neurons, 14 to 21 days prior to each experiment as described (6). In addition, in a few experiments Kiss1^Cre/+ mice were crossed with heterozygous Ai32 mice (RRID: IMSR_JAX:024109, https://www.jax.org/strain/024109, C57BL/6 background), which carry the ChR2 (H134R)–EYFP gene in their Gt(ROSA)26Sor locus (17). When stated, some of the females were ovariectomized (OVX) 7 days prior to an experiment. Each animal was injected on day 5 following OVX with 0.25 μg E2, followed on day 6 with 1.50 μg E2 and used for experiments on day 7 (18).

Adeno-associated Virus Delivery to Kiss1^cre/+ Mice

Fourteen to 28 days prior to each experiment, the Kiss1^Cre/+ mice (aged >60 days) received bilateral ARH or AVPV/PeN injections of AAV1-Ef1a-DIO-ChR2:YFP (University of Washington) or AAV1-Ef1a-DIO-ChR2:mCherry (University of Washington) or AAV1 vectors designed to encode SaCas9 and single-guide RNA (sgRNA) (see the SaCas9 section for specifics on the sgRNA design). Using aseptic techniques, anesthetized (1%-1.5% isoflurane/O₂) mice received a medial skin incision to expose the surface of the skull. A glass pipette (No. 3-000-203-G/X, Drummond Scientific) with a beveled tip (diameter = 45 μm) was filled with mineral oil and loaded with an aliquot of AAV using a Nanoject II (Drummond Scientific). Arcuate nucleus (ARH) injection coordinates were anteroposterior (AP): −1.10 mm, mediolateral (ML): ± 0.33 mm, and dorsoventral (DV): −5.80 and −5.70 (surface of the brain z = 0.0 mm); 250 nL of the AAV was injected (100 nL/min) at each position, and the pipette was left in place for 10 minutes post injection and then slowly retracted from the brain. For AVPV/PeN injections, the coordinates were x: ± 0.33, y: 0.55, and z: 5.1 and −4.7. To better cover the longer and thinner shape of the AVPV/PeN, the virus was injected at 2 sites separated by 0.5 mm (300 nL/site, 100 nL/min). The pipette was left in place for 10 minutes after injection at each site to allow for even distribution of the virus and to limit viral reflux as the pipette was withdrawn from the brain. The skin incision was closed using Vetbond (3 M), and each mouse received analgesia (Rimadyl, 4-5 mg/kg, subcutaneously).

Generation and Injection of AAV1-FLEX Sacas9-sgSlc17a6 Mice for In Vitro Studies

The generation of AAV1-FLEX Sacas9-sgSlc17a6 virus was performed at the University of Washington using published methods (14, 19). The constructs of sgRNA for Vglut2 (AAV1-FLEX Sacas9-sgSlc17a6) was designed to target exon 4 (Fig. 1A) as described previously (14). Therefore, to visualize the quality and location of the injection/infection, the CRISPR/SaCas9 vector was spiked with a high-titer virus encoding YFP. The resulting mixture allowed a single injection of both viruses at a similar titer. Coinjected viruses were always the same serotype (AAV1) so as not to affect the infection efficiency. Four weeks prior to each experiment, the Kiss1^Cre/+ mice (aged >60 days) received bilateral ARH injections of 10% AAV1-FLEX-YFP (University of Washington) coinjected with 90% AAV1-FLEX Sacas9-sgSlc17a6 or 90% AAV1-FLEX Sacas9-sgROSA26 control virus.

Figure 1.

CRISPR mutagenesis of Vglut2 in Kiss1^ARH neurons. A, Outline of Cre-dependent AAV for targeting Vglut2 (Slc17a6). Targeted exon 4 with guide sequence is highlighted in red. The PAM is underlined. B1, Schematic of viral injection into the arcuate nucleus (ARH). B2, Photomicrographs showing coronal sections confirming targeted bilateral injections of AAV1-sFLEX-SaCas9-U6-sgVglut2 or control virus into the arcuate of adult female Kiss1^Cre/+ mice. The scale bar is 200 μm. C, Quantitative polymerase chain reaction (qPCR) amplification assay illustrating the cycle threshold (CT) for the Vglut2 for Kiss1^Cre vs Vglut2 mutated (Vglut2^Mut) Kiss1^ARH neurons. Cycle number is plotted against the normalized fluorescence intensity (ΔRn) to visualize the PCR amplification of Vglut2 and the reference gene Gapdh in 10 cell Kiss1^ARH neuronal pools. D, qPCR assay measuring Vglut2 expression in Kiss1^ARH neuronal pools from Kiss1^Cre:Ai32 (n = 4), Kiss1^{Cre:control vector} (n = 14), and Vglut2^Mut (n = 21) mice (10 cells per pool, 3-4 pools/animal) using Vglut2 primers (product length 116 bp) that span the single-guide RNA (sgRNA) and PAM sites of the Vglut2 sequence. Bar graphs represent mean ± SEM, with data points representing individual animals. Unpaired t tests: Kiss1^Cre:Ai32 vs Vglut2 ^Mut: t₍₂₃₎ = 18.17; ****P < .0001; Kiss1^{Cre:control vector} vs Vglut2 ^Mut: t₍₃₃₎ = 17.72; ****P < .0001.

Visualized Whole-Cell Patch Recording

Electrophysiological and optogenetic studies were made in coronal brain slices (250 μm) containing the ARH or AVPV/PeN from AAV1-EF1α-DIO-mCherry or AAV1-Ef1a-DIO-ChR2:YFP–injected Kiss1^Cre/+ mice, which were E2-treated OVX females aged 10 weeks or older as previously described (6, 12). Whole-cell patch recordings were performed in voltage-clamp and current-clamp as previously described (12) using an Olympus BX51 W1 fixed-stage scope outfitted with epifluorescence and IR-DIC video microscopy. Patch pipettes (A-M Systems; 1.5 μm outer diameter borosilicate glass) were pulled on a Brown/Flaming puller (Sutter Instrument, model P-97) and filled with the following solution: 128 mM potassium gluconate, 10 mM NaCl, 1 mM MgCl₂, 11 mM EGTA, 10 mM HEPES, 2 mM ATP, and 0.25 mM GTP adjusted to pH 7.3 with KOH; 295 mOsm. Pipette resistances ranged from 3.5 to 4 MΩ. In whole-cell configuration, the access resistance was less than 30 MΩ and was 80% compensated. The input resistance was calculated by measuring the slope of the I-V relationship curve between −70 and −50 mV. Standard whole-cell patch recording procedures and pharmacological testing were performed as previously described (20, 21). Electrophysiological signals were digitized with a Digidata 1322A (Axon Instruments), and the data were analyzed using p-Clamp software (Molecular Devices). The liquid junction potential was corrected for all data analysis.

To characterize electrophysiological properties, current-clamp recordings were obtained in the presence of antagonists of ionotropic glutamate and γ-aminobutyric acid (GABA) receptors (50 μM D-APV, 10 μM CNQX, and 100 μM picrotoxin). The membrane potential of neurons was subsequently maintained at −70 mV by direct current injection. The excitability of neurons was assessed by injecting depolarizing currents (1-second steps, 5-pA increments). The first current step to display an AP was defined as the rheobase. Miniature excitatory postsynaptic currents (mEPSCs) were recorded in the presence of tetrodotoxin (TTX) (1 μM) and picrotoxin (100 μM). Cell membrane potentials were voltage-clamped at −60 mV. EPSC traces were filtered at 10 kHz and acquired at a sampling rate of 5 kHz (22). The duration of the recordings for the amplitude measurements was 3 minutes, with a threshold set to 10 pA for the amplitude analysis using Mini Analysis 6.0.7 software (Synaptosoft).

For optogenetic stimulation, a light-induced response was evoked using a light-emitting diode (LED) 470 nm blue light source controlled by a variable 2A driver (ThorLabs) with the light path directly delivered through an Olympus 40× water-immersion lens. For high-frequency (20 Hz) stimulation, the length of stimulation was 10 seconds (light intensity 660 μW and pulse duration, 10 ms) (6).

Electrophysiological Solutions/Drugs

A standard artificial cerebrospinal fluid (aCSF) was used (12, 20, 23). All drugs were purchased from Tocris Bioscience unless otherwise specified. A total of 1 mM TTX (Alomone Labs), 50 mM DL-2-amino-5-phosphonopentanoic acid sodium salt (AP5), 10 mM 6-Cyano-7-nitroquinoxaline-2,3-dione disodium (CNQX), 100 mM picrotoxin and 50 mM (S)-3,5-dihydroxyphenylglycine (DHPG), stocks (1000×) were prepared in dimethyl sulfoxide (picrotoxin) or water (TTX, AP5, CNQX and DHPG) and stored at −20 °C. Aliquots of the stock solutions were stored as appropriate until needed.

Cell Harvesting of Recorded and Dispersed Kiss1^Cre/+ Neurons and Conventional Polymerase Chain Reaction Real-time Quantitative Polymerase Chain Reaction

Some cells were harvested following whole-cell patch recordings. Complementary DNA (cDNA) synthesis and polymerase chain reaction (PCR) were performed for Kiss1 as previously described (12). Cell harvesting and quantitative PCR (qPCR) was conducted as previously described (18). The ARH was microdissected from basal hypothalamic coronal slices obtained from female Kiss1^Cre/+ mice injected with AAV1-FLEX solution as described earlier. The dispersed cells were visualized, patched, and then harvested (10 cells/tube) as described previously (18). Briefly, ARH tissue was incubated in papain (7 mg/mL in oxygenated aCSF) for 50 minutes at 37 °C then washed 4 times in low Ca²⁺ aCSF and 2 times in aCSF. For cell dispersion, Pasteur pipettes were flame-polished to decreasing tip sizes and gentle trituration used to disperse the neurons onto a glass-bottom dish. The plated cells were bathed in oxygenated aCSF using a peristaltic pump to keep the cells viable and clear of debris. Healthy cells with processes and a smooth cell membrane were harvested. Pipettes (World Precision Instruments; 1.5 μm outer diameter borosilicate glass) were pulled on a Brown/Flaming puller (Sutter Instrument, model P-87) to a 10-µm diameter tip. The cells were harvested using the XenoWorks Microinjector System (Sutter Instruments), which provides negative pressure in the pipette and fine control to draw the cell up into the pipette. Cell pools were harvested and stored at −80 °C. All cell pools were DNAse-treated using DNase1 (Invitrogen). cDNA synthesis was performed as previously described (18).

Primers for the genes that encode for Kiss1, Vglut2, Trpc5, and glyceraldehyde-3-phosphate dehydrogenase (Gapdh) were designed using Clone Manager software (Sci Ed Software) to cross at least one intron-exon boundary and optimized as previously described using Power Sybr Green method (18). Real-time qPCR controls included neuronal pools without reverse transcriptase (-RT), hypothalamic RNA with RT (+) and without RT (−), as well as water blanks. Standard curves using ARH cDNA were used to determine the real-time PCR efficiency (E = 10^(−1/m) – 1) (24-26). Only primers resulting in efficiencies of 90% to 100% were used for analysis. Primer sequences, qPCR parameters, and efficiency calculations are provided in Table 1.

Table 1.

Primer table

Gene name (encodes for)	Accession No.	Primer location, nt	Product length, bp	Annealing temp, °C	Efficiency
					Slope	r 2	%
Kiss1 (Kiss1)	NM_178260	64-80	120	60	−3.410	0.989	97
		167-183
Slc17a6 (VGLUT2)a	NM_080853	1275-1296	116	60	−3.374	0.997	98
		1371-1390
Gapdh (GAPDH)	NM_008084	689-706	93	60	−3.352	0.998	99
		764-781
Trpc5 (TRPC5)	NM_009428	734-753	118	60	−3.161	0.952	100
		832-851

Abbreviations: bp, base pair; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; Kiss, kisspeptin; VGLUT2, vesicular glutamate transporter 2.

^a Primers designed to cross single-guide RNA and PAM sites that cross intron-exon boundary between exons 4 and 5.

Messenger RNA Expression Analysis

qPCR was performed on a Quantstudio 7 Flex Real-Time PCR System (Applied Biosystems) using Power SYBR Green PCR Master Mix (Applied Biosystems) according to established protocols (18). The comparative ΔΔC_T method (24-26) was used to determine values from duplicate samples of 4 µL for the target genes and 2 µL for the reference gene GAPDH in a 20-μL reaction volume containing 1× Power SYBR Green PCR Master Mix and 0.5-μM forward and reverse primers. Four 10-cell pools per animal were analyzed in a total of 18 Kiss1^Cre controls and 21 Vglut2^Mut animals. Trpc5 was measured in three 10-cell pools per animal in Kiss1^Cre OVX and OVX + E2 animals (n = 5 animals). The relative linear quantity was determined using the 2^−ΔΔCT equation (24-26). Relative mRNA expression level of target genes in Kiss1^Cre/+ neurons was obtained by comparing OVX E2-treated Kiss1^Cre/+ controls to OVX E2-treated Vglut2^Mut animals. The mean Δ CT for the target genes from the OVX E2-treated Kiss1^Cre/+ control samples was used as the calibrator. The data were expressed as n-fold change in gene expression normalized to the reference gene GAPDH and relative to the calibrator.

CRISPR Mutagenesis of Vglut2 in Kiss1^ARH Neurons (Vglut2^Mut) for In Vivo Studies

Adult female Kiss1^Cre/+ (heterozygous) mice were deeply anesthetized with isoflurane and placed into the stereotaxic apparatus (Kopf Instruments, model 940). Loss of consciousness was confirmed with a toe pinch test before transferring mice to the stereotaxic apparatus. Mice were administered preoperative subcutaneous injections of buprenorphine (0.6 mg/kg) and meloxicam (5 mg/kg), in addition to protective eye lubricant. The skull was exposed via incision and a small hole was drilled above the ARH for bilateral injections. ARH coordinates used were AP −1.6 mm, ML ± 0.25 mm, and DV −5.85 mm. A 5-μL Hamilton syringe (model 175 RN SYR, catalog No. 80016) was lowered into the brain at the appropriate DV coordinates. Each infusion (200 nL) was slowly delivered (50 nL/minute). The Kiss1^Cre/+ female mice received bilateral ARH injections of the same viruses used for the in vitro electrophysiology and molecular biology experiments described earlier. Animals received meloxicam 24 hours after surgery to control postoperative pain and inflammation and were allowed a 3-week recovery before onset of experiments. Viral injections were validated in postfixated perfused brains via immunohistochemistry for YFP.

Immunohistochemistry

Mice were terminally anesthetized with a ketamine/xylazine in saline (0.9% NaCl) cocktail and transcardially perfused with 0.1-M phosphate buffer (0.1-M PB) followed by 4% paraformaldehyde diluted in 0.1 M PB (PFA; Electron Microscopy Sciences). Brains were removed, stored in the same fixative overnight, and then transferred into sucrose solution (Thermo Fisher Scientific; 20% sucrose in 0.1-M PB containing 0.01% sodium azide (Sigma-Aldrich)) at 4 ˚C. After sucrose infiltration, tissue was cut into 30-µm coronal sections using a freezing stage microtome (Fisher HM440E). Free-floating sections were washed in phosphate-buffered saline (PBS) then transferred to blocking buffer (PBS + 0.4% triton + 0.1% bovine serum albumin [BSA]) for 1 hour at room temperature on an orbital mixer. Sections were then incubated in a rabbit green fluorescent protein (GFP) antibody (Life Technologies No. A6455, 1/5000 dilution in blocking buffer) overnight at 4 °C. The following day, sections were washed in PBS then incubated in goat anti-rabbit DyLight 488 (ThermoFisher Scientific 1:200 dilution in blocking buffer: PRID: AB_844398) in the dark for 1 hour at room temperature. After incubation and a final washing, sections were mounted on glass slides and cover-slipped with Vectashield + DAPI (4′,6-diamidino-2-phenylindole) solution (Vector Laboratories).

Characterization of the Estrous Cyclicity and the Estradiol-induced Luteinizing Hormone Surge

Estrous cycle tracking started 4 weeks before the viral injections to establish a baseline estrous cyclicity in the Vglut2^Mut (n = 10) and Kiss1^Cre (n = 7) females and continued 10 weeks after injections. Cycle stages were determined by vaginal cytology and were obtained every morning (8-10 Am) and placed on a glass slide for determination of the estrous cycle stage under the microscope as previously described (27).

To characterize the E2-induced LH surge, vaginal smears were taken at 8 Am every day from intact females, as described earlier. The day of diestrus for each of the females, 4 μL of tail-tip blood samples were collected to establish baseline LH, then LH surge was induced with a subcutaneous injection of 50 µg/kg of estradiol benzoate (EB) at 10 Am based on a protocol designed and optimized in the Navarro laboratory. This produced elevated proestrus-like E2 levels to trigger an LH surge the following day at the time of lights out (7 Pm). The next day, 4-µL tail-tip blood samples were collected at 9 Am, 7 Pm, and 7:30 Pm. Each sample was diluted in 116 µL of 0.05% PBST (PBS [Boston Bio Products, catalog No. BM220] containing Tween-20 [Sigma, catalog No. P2287]), vortexed, and frozen on dry ice. LH was measured with the high-sensitivity LH enzyme-linked immunosorbent assay (ELISA) sandwich to evaluate the surge of Kiss1^Cre and Vglut2^Mut females (28).

Body Weight

Body weight measurements of Kiss1^Cre (n = 7) and Vglut2^Mut (n = 10) females were taken biweekly at the time of vaginal smear throughout the duration of the experiment.

Luteinizing Hormone Pulsatile Secretion Profile

To assess the profile of LH pulses secretion, Kiss1^Cre and Vglut2^Mut females were trained to the procedure daily for 3 weeks preceding the experiment to allow acclimation to sampling conditions prior to the experiment. Pulses were assessed by repeated blood collection through a single incision at the tip of the tail and collection of 4 µL of whole blood every 10 minutes for 180 minutes. Each sample was diluted in 116 µL of 0.05% PBST (phosphate buffer saline [Boston Bio Products, catalog No. BM220] containing Tween-20 [Sigma, catalog No. P2287]), vortexed, and frozen on dry ice. Samples were stored at −80 °C until analyzed with ELISA as described later.

Ovarian Histology

Bilateral OVX of Kiss1^Cre and Vglut2^Mut females was performed under light isoflurane anesthesia. Briefly, the ventral skin was shaved and cleaned, and one small abdominal incision was made. Once the ovaries were identified and excised, the muscle incision was sutured, and the skin was closed with surgical clips. Ovaries were stored in Bouin's fixative, sectioned, and stained with hematoxylin and eosin at the Harvard Histopathology Core. Corporal lutea were counted in the middle section of each ovary.

Hormone Measurements

LH was measured by an ultra-sensitive ELISA sandwich for assessment of whole-blood LH concentrations as previously described (28). A 96-well microplate (Corning; catalog No. 0720039) was coated with 50 µL of capture antibody (bovine LHβ 518B7 Mab at 1:1000 dilution in 1× PBS. PRID:_AB_2665514). Plates were incubated in a humidified chamber at 4 °C overnight. The following morning, plates were incubated with 200 µL blocking buffer (5% skim milk powder in 1× PBST) for 2 hours in a humidified chamber at room temperature. A standard curve was generated using 2-fold serial dilution of LH in 0.2% BSA-PBST. A total of 50 µL of standards and blood samples were incubated in humidified chamber at room temperature for 2 hours. Next, 50 µL of detecting antibody (LH antibody 5303 SPRN at 1:1000 dilution in blocking buffer; PRID:_AB_2784503) was incubated in duplicate in a room temperature, humidified chamber at room temperature for 1.5 hours. The plate was washed with PBST then loaded with 50 µL of secondary antibody (polyclonal goat antirabbit horseradish peroxidase-conjugated antibody: Poly-HRP Streptavidin at 1:8000 dilution in PBST), which was incubated a further 1.5 hours at room temperature. To visualize the secondary antibody, 100 µL o-phenylenediamine dihydrochloride (OPD) (1 tablet OPD catalog No. PI34006, Fisher, per 10 mL citrate buffer: 1 mL 10× citric acid + 1 mL 10× sodium phosphate + 8 mL distilled H₂O) was added to each well and the plate was covered in foil to prevent light degradation of signal before incubating at room temperature for 30 minutes. The reaction was stopped with 50 µL of 3 M HCl, and absorbance at 490 nm was read. The concentration of LH in whole-blood samples was determined in comparison to the standard curve.

Luteinizing Hormone Pulse Analysis

LH pulsatility was assessed by measuring (1) the total number of pulses throughout the 180-minute sampling period, (2) the LH pulse amplitude, calculated by averaging the 4 highest LH values in the samples collection period for each animal, (3) the basal LH, calculated by averaging the 4 lowest LH values in the sample collection period for each animal, and (4) the total secretory mass, assessed by area under the curve (AUC). LH pulsatility was analyzed using a reformulation of the PULSAR software (29).

Microscopy and Image Analysis

Validation of the AAV injections was documented in Kiss1^Cre control and Vglut2^Mut females. Sections from animals injected with AAV vectors in the ARH were examined, and the location of viral expression was confirmed in YFP-positive neurons. Analysis of positive neurons in the ARH was carried out in images taken at × 20 magnification. A digital camera (CoolSnap EZ, Photometrics) attached to a microscope (Nikon Eclipse 90i), with the appropriate excitation for DyLight 488 (green fluorescence), and NIS-Elements Viewer AR 310 software were used to examine tissue sections. Montages of images and adjustments of brightness and contrast levels were made in Image J. Only animals with accurate and restricted injection sites were included in the analysis.

Experimental Design and Statistical Analysis

For the pharmacological experiments, only one cell was recorded per slice. Two to 3 arcuate slices and 1 AVPV/PeN slice were analyzed from each mouse, with at least 3 to 5 mice contributing to each group. For cell harvesting of dispersed Kiss1^Cre/+-YFP or -mCherry neurons and qPCR measurements, 10 cells per pool and 4 pools from each animal were used, unless otherwise specified. Statistical comparisons between 2 groups were performed using an unpaired, 2-tailed t test. Comparisons between more than 2 groups were performed using the repeated-measures, multifactorial analysis of variance (ANOVA). If a significant interaction was encountered, we then moved to the one-way ANOVA, followed by the multiple range tests as specified in the appropriate figure legends. All data were analyzed using GraphPad Prism version 6. All data are presented as mean ± SEM. Differences were considered statistically significant if the probability of error was less than 5%.

Results

CRISPR Editing of Vglut2 Significantly Decreases Vglut2 Expression

To specifically inactivate Vglut2 in Kiss1^ARH neurons in adult animals while preventing developmental compensation, we employed a CRISPR approach to mutate Vglut2 in Kiss1^ARH neurons, following protocols established in our prior studies (14, 30). Hunker and colleagues (14) developed a single viral vector for conditional expression of the smaller Staphylococcus aureus (SaCas9) and sgRNA, which demonstrated high-efficiency mutagenesis in specific cell types. To selectively mutate Vglut2 in Kiss1^ARH neurons, a guide RNA was targeted to exon 4, which is conserved across all splice variants (Fig. 1A). A cohort of Kiss1^Cre/+ (heterozygous) mice received bilateral stereotaxic injections into the ARH of AAV1-FLEX-SaCas9-sgSlc17a6 or a control virus containing the Rosa26 guide. Additionally, a Cre-dependent AAV of the same serotype (AAV; serotype 1), driving expression of AAV1-Ef1a-YFP or AAV1-Ef1a-mCherry, was coadministered to visualize injection quality and facilitate cell harvesting (Fig. 1B). Three weeks later, mice underwent OVX and were treated with E2 to mimic the proestrous stage of the cycle (18). Brain slices were prepared, evaluated for quality of viral injections, and used for electrophysiology or cell harvesting. Cells were harvested as previously described (12) and analyzed with qPCR. Based on the cycle threshold (CT) values (Fig. 1C) and quantitative analysis (Fig. 1D), we observed a highly significant 4-5-fold reduction in relative expression of Vglut2 in Kiss1^ARH neurons in the Vglut2 mutagenesis group compared to the control viral–injected Kiss1^Cre mice or the Kiss1^Cre:Ai32 control group. Thus, the qPCR data confirmed selective reduction of Vglut2 gene expression in targeted Kiss1^ARH neurons of sgVglut2-targeted mice.

Vglut2 Mutagenesis Does not Affect Endogenous Electrophysiological Properties of Kiss1^ARH Neurons

To assess whether Vglut2 mutagenesis affected Kiss1^ARH neuronal excitability, we measured the endogenous electrophysiological properties and the optogenetically generated slow excitatory postsynaptic potentials (EPSPs) (Fig. 2), which is a measure of the combined peptide (NKB) excitatory output of all Kiss1^ARH neurons (6). With the Vglut2 mutagenesis there was no difference in the resting membrane potential (see Fig. 2A), cell capacitance (Fig. 2B), input resistance (Fig. 2C), or rheobase (Fig. 2D). Therefore, the Vglut2 mutagenesis in Kiss1^ARH neurons did not affect the endogenous electrophysiological properties, which was further proof that Vglut2 sgRNA was not targeting any other channel or transporter (14). High-frequency optogenetic stimulation (20 Hz, 10 seconds) reliably induced a slow EPSP in Kiss1^ARH neurons from the Kiss1^Cre/+ mice, which received an injection of a ChR2-expressing viral vector into the ARH (Fig. 2E). This slow EPSP plays a crucial role in synchronizing Kiss1^ARH neurons and stimulating GnRH release (6, 7). Comparing the slow EPSPs between Kiss1^Cre/+ and Vglut2^Mut mice, mutagenesis attenuated but did not abrogate the slow EPSP (Figs. 2F-2H), which is what we would have predicted based on our previous studies (6, 12).

Figure 2.

Vglut2 mutagenesis attenuates the slow EPSPs in Kiss1^ARH neurons. A to D, The endogenous electrophysiological properties of Kiss1^ARH neurons from Kiss1^Cre and Vglut2^Mut E2-treated female mice showed no statistically significant differences. Specifically, there were no differences in A, resting membrane potential; B, cell capacitance; C, input resistance; or D, rheobase. Unpaired t tests for A, t₍₅₆₎ = 0.2545; P = .8001; for B, t₍₅₆₎ = 0.3662; P = .77156; for C, t₍₅₆₎ = 0.1197; P = .9065; and for D, t₍₅₆₎ = 0.3759; P = .7084. Kiss1^Cre, n = 28, Vglut2^Mut, n = 30 cells. E, Experimental protocol: AAV1-DIO-ChR2:YFP was bilaterally injected into the ARH of Kiss1^Cre/+:GFP control and Vglut2^Mut mice. In slices, recording was made in Kiss1^ARH neurons with high-frequency (20 Hz) photostimulation of Kiss1^ARH neurons/terminals. F, Example of high-frequency optogenetic stimulation (20 Hz, 10 seconds) generating a slow EPSP in a Kiss1^ARH neuron from E2-treated, OVX Kiss1^Cre/+ control mouse. The amplitude of the response was similar to our previous published findings (6). G, High-frequency optogenetic stimulation (20 Hz, 10 seconds) generated a reduced slow EPSP in a Kiss1^ARH neuron from an E2-treated, OVX Vglut2^Mut mouse. H, The efficacy of high-frequency photostimulation to generate a slow EPSP in Kiss1^ARH neurons was significantly reduced in Vglut2^Mut mice (EPSP amplitude: unpaired t test, t₍₄₈₎ = 3.214; ** P = .0023). The bar graphs represent mean ± SEM, with data points representing the number of cells. Kiss1^Cre, n = 20, Vglut2^Mut, n = 30 cells. However, Vglut2 mutagenesis did not completely abrogate the slow EPSP in Kiss1^ARH neurons, which is mostly dependent on neurokinin B release and TACR3 signaling in Kiss1^ARH neurons.

Vglut2 Mutagenesis Abrogates Glutamatergic Input to Kiss1^AVPV/PeN Neurons From Kiss1^ARH Neurons

Consistent with previous findings demonstrating that Kiss1^ARH neurons can excite Kiss1^AVPV/PeN neurons via glutamate release (6), light-evoked fast EPSCs were elicited in Kiss1^AVPV/PeN neurons from Kiss1^Cre/+ female mice that had received bilateral ARH injections of a Cre-dependent AAV (serotype 1) vector encoding ChR2-mCherry (AAV1-Ef1a-DIO-ChR2:mCherry) (Fig. 3). On patching 32 Kiss1-^AVPV/PeN neurons from 3 animals with uniform bilateral ChR2 injections into the ARH, we observed glutamatergic EPSC responses in 8 cells, corresponding to a ratio of 25% (Figs. 3B and 3C). However, when patching 36 cells from 5 animals with uniform bilateral ChR2 injections in combination with the sgSlc17a6 (sgVglut2), into the ARH, we observed only 1 Kiss1^AVPV/PeN neuron with a clear glutamatergic EPSC response (see Fig. 3C). Kiss1^AVPV/PeN neurons were identified by tell-tale biophysical properties (31) and post hoc RT-PCR expression of Kiss1 mRNA (Fig. 3D).

Figure 3.

Photoactivation of Kiss1^ARH fibers excites Kiss1^AVPV/PeN neurons via glutamate release. A, Schematic of experimental design; whole-cell, voltage-clamp (V_hold = −60 mV) recordings were made in Kiss1^AVPV/PeN neurons from an OVX + E2 Kiss1^ARH-Cre:ChR2:mCherry mouse, and a single pulse (intensity, 660 μW; 10 ms duration, 0.5 Hz) of blue light (470 nm) was delivered to the AVPV. ARH, hypothalamic arcuate nucleus; AVPV/PeN, anteroventral periventricular and periventricular nuclei. B, Whole-cell, voltage clamp recording of Kiss1^AVPV/PeN neuron exhibiting light-evoked fast EPSCs (V_hold = −60 mV). Overlay of individual (gray) whole-cell recordings where photostimulation of Kiss1^ARH-ChR2:mCherry neuronal terminals induced EPSCs in a Kiss1^AVPV/PeN neuron. The averaged responses before (red trace) and after (blue trace) the application of ionotropic glutamate receptor antagonists CNQX (10 μM) and AP5 (50 M). Solid blue bar above recordings indicates light stimulus. Following recording, cells were harvested for single-cell reverse-transcription polymerase chain reaction (PCR) determination of Kiss1 messenger RNA (mRNA) expression. All responsive AVPV/PeN cells (mean response: 17.3 ± 2.8 pA, n = 8) expressed Kiss1 mRNA. C, A total of 25% (8 cells) of Kiss1^AVPV/PeN neurons responded to optogenetic stimulation of Kiss1^ARH-Cre:ChR2:mCherry fibers in the Kiss1^Cre mice, whereas only 1 of 36 cells from 5 animals responded in the Vglut2^Mut mice. D, Examples of single-cell PCR identification of the recorded Kiss1^AVPV/PeN neurons from both groups of animals.

Additionally, we recorded mEPSCs in the presence of TTX to block the fast sodium channel and picrotoxin to block GABA_A receptors in Kiss1^AVPV/PeN neurons from animals that received the sgRosa26 or sgSlc17a6 virus (Fig. 4A). We found no difference in the EPSC amplitude between the two groups (Fig. 4B and 4C), but the EPSC frequency was reduced in the Vglut2^Mut group compared to the Kiss1^Cre control group (Fig. 4B and 4D). Therefore, Vglut2 mutagenesis in Kiss1^ARH neurons significantly reduces glutamatergic input to Kiss1^AVPV/PeN neurons, which is congruent with previous studies showing that Vglut2 expression controls the efficacy of glutamatergic neurotransmission by thalamic neurons (32).

Figure 4.

Vglut2 mutagenesis attenuates miniature excitatory postsynaptic current (mEPSC) frequency in Kiss1^AVPV/PeN neurons from 17β-estradiol (E2)-treated ovariectomized (OVX) mice. A, Schematic of experimental design; whole-cell, voltage-clamp (V_hold = −60 mV) recordings were made in Kiss1^AVPV/PeN neurons from OVX + E2 Kiss1^Cre/+:Ai32 mice that received an ARH viral injection of sgVglut2 and mCherry or sgRosa26 and mCherry. ARH, hypothalamic arcuate nucleus; AVPV/PeN, anteroventral periventricular and periventricular nuclei. B, The recordings were performed in the presence of TTX (1 μM), a fast sodium channel blocker, and picrotoxin (100 μM), a GABA-A channel blocker. Miniature postsynaptic currents in Kiss1^AVPV/PeN neurons from E2-treated Kiss1^Cre/+ mice are depicted in the upper trace, while those from Vglut2^Mut mice are shown in the lower trace. The displayed traces cover a duration of 40 seconds, representing a segment of the total 3-minute recording. C, The mean amplitude of the mEPSCs of Kiss1^AVPV neurons in Vglut2^Mut mice did not differ from the Kiss1^Cre/+ control mice. The data are expressed as mean ± SEM, with the data points representing the number of cells (unpaired 2-tailed t test comparing Kiss1^Cre (5 animals, 39 cells total recorded) vs Vglut2^Mut (5 animals, 38 cells total recorded: t₍₇₅₎ = 0.5312; P = .5968). D, However, the mean frequency of the mEPSCs of Kiss1^AVPV/PeN neurons in Vglut2^Mut mice was significantly reduced compared to Kiss1^Cre/+ control mice (unpaired 2-tailed t test comparing Kiss1^Cre (5 animals, 39 cells total recorded) vs Vglut2^Mut (5 animals, 38 cells total recorded: t₍₇₅₎ = 2.510; *P = .0142). This indicates that Vglut2 mutagenesis in Kiss1^ARH neurons reduces the glutamatergic synaptic input to Kiss1^AVPV/PeN neurons.

Glutamate Excites Kiss1^AVPV/PeN Neurons via Ionotropic and Metabotropic Glutamate Receptors

In addition to the fast glutamate response with low-frequency optogenetic stimulation (0.5 Hz, 10 ms duration) (Fig. 5A), high-frequency optogenetic stimulation (20 Hz, 10 seconds) of Kiss1^ARH ChR2 terminals surrounding Kiss1^AVPV/PeN neurons induced a slow EPSP response that increased action potential firing (Fig. 5B); we also measured the slow metabotropic response in 3 Kiss1^AVPV/PeN neurons that were depolarized by 4.7 ± 0.1 mV (n = 3) from a baseline of −60 mV. Additionally, there was an increase in firing activity with 95 ± 45 spikes recorded over 5 minutes starting from a previously silent state, suggesting the involvement of both fast ionotropic and metabotropic glutamatergic responses. Kiss1^AVPV/PeN neurons may undergo modulation by group I metabotropic glutamate receptors. Dong and Ennis (33) showed that the activation of these receptors not only triggers a nonselective cation current but also enhances a persistent sodium current, thereby inducing rhythmic bursting in main olfactory bulb external tufted cells. Therefore, we explored the effects of the mGluR group I agonist dihydroxyphenylglycine (DHPG) on Kiss1^AVPV/PeN neurons, identified by their biophysical properties (31) and post hoc RT-PCR. DHPG depolarized and increased firing in current clamp (Fig. 5C) and induced an inward current in voltage clamp (Fig. 5D). The I/V relationship indicated that a nonselective cation current mediated the response (Fig. 5E). Based on the fact that mGluR1 is coupled to activation of TRPC4/5 channels in many central nervous system neurons (34-36), we surmised that DHPG activated these channels in Kiss1^AVPV/PeN neurons. Indeed, we measured the expression of Trpc 5 mRNA in Kiss1^AVPV/PeN by quantitative single cell PCR and found that not only was it expressed, but was upregulated with E2 treatment, consistent with our present model (Fig. 5F). Therefore, our findings demonstrate that the activation of group I metabotropic glutamate receptors contribute to the Kiss1^ARH excitation of Kiss1^AVPV/PeN neurons.

Figure 5.

The metabotropic glutamate receptor 1/5 agonist DHPG excites a Kiss1^AVPV/PeN neuron. A, Fast glutamatergic currents were evoked in voltage clamp (V_hold = −60 mV) recordings of Kiss1^AVPV/PeN neurons following optogenetic stimulation of Kiss1^ARH ChR2 terminals (intensity: 660 µW, 10 ms duration, 0.5 Hz). Gray traces represent individual recordings, while the black trace shows the averaged response. The blue bar above the recordings indicates the timing of the LED stimulus. B, High-frequency optogenetic stimulation (20 Hz, 10 seconds) of Kiss1^ARH ChR2 terminals induced membrane depolarization and increased action potential firing in the same Kiss1^AVPV/PeN neuron, suggesting the involvement of both fast ionotropic and metabotropic glutamatergic responses. Two other Kiss1^AVPV/PeN neurons showed a similar response profile. C, Representative whole-cell, current clamp recording of another Kiss1^AVPV/PeN neuron from an ovariectomized (OVX) + 17β-estradiol (E2) Kiss1^Cre/+:Ai32 mouse showed that the mGluR1/5 agonist DHPG (50 µM) depolarized and induced robust firing (n = 3). The cell was silent at its resting membrane potential of −60 mV. D, Whole-cell, voltage clamp recording of a Kiss1^AVPV/PeN neuron from an OVX + E2 Kiss1^Cre/+:Ai32 mouse in the presence of the fast sodium channel blocker TTX and ionotropic glutamate receptor blockers (CNQX and AP5), which electrically isolated the cell from synaptic input. Rapid bath application of DHPG induced inward currents in Kiss1^AVPV/PeN neurons. V_hold = −60 mV. E, Example of current/voltage (I/V) relationships of a Kiss1^AVPV/PeN neuron before and after rapid application of DHPG. Voltage ramps from 0 to −100 mV were applied (over 2 seconds) before and after treatment with DHPG. V_hold = −60 mV. The crossing point of the 2 I/V plots (reversal potential) was at −20 mV, which indicates that DHPG activated a nonselective cation channel like TPRC5, which is also expressed in Kiss1^AVPV/PeN neurons. F, Comparison of Trpc5 messenger RNA expression between E2-treated and oil-treated OVX females was performed using the comparative 2^−ΔΔCT method. The E2-treated group was used as the calibrator. Bar graphs represent the mean ± SEM, with data points indicating the number of animals (unpaired t test t₍₈₎ = 4.478; **P = .0021).

Mutagenesis of Vglut2 in Kiss1^ARH Neurons Does not Impair Estrous Cyclicity

To address the role of Vglut2 within Kiss1^ARH neurons in regulating reproductive function in females, we used CRISPR technology to mutate Vglut2 mRNA within Kiss1^ARH neurons in Kiss1^Cre/+ females (see Fig. 1). We verified that only the ARH showed infected AAV1-FLEX Sacas9-SgSlc17a6 cells (for Vglut2^Mut females) or AAV1-FLEX Sacas9-SgROSA 26 (for Kiss1^Cre control mice), while there were no infected cells in the AVPV/PeN area. Therefore, we confirmed that only Kiss1^ARH neurons were targeted with the Vglut2 mutagenesis. We then assessed estrous cyclicity in the Vglut2^Mut females (n = 10) and their Kiss1^Cre controls (n = 7) for 10 weeks following AAV injections. The Vglut2^Mut females displayed regular cycles, similar to those found in Kiss1^Cre controls (Fig. 6A and 6B). There was no difference in the average number of days spent in each stage of the cycle between the two groups (2-way ANOVA; P > .05). Next, we assessed the LH pulsatile secretion pattern in these females in the gonad-intact state. We found that gonad-intact Vglut2^Mut females (n = 8) had a small but statistically significant decrease in the total number of pulses compared to Kiss1^Cre controls (n = 6) (Vglut2^Mut = 2.25 ± 0.16 vs Kiss1^Cre = 3.16 ± 0.30; P = .015). However, these females did not exhibit any impairment in the total secretory mass measured by AUC representing total LH release (Vglut2^Mut AUC = 161.32 ± 8.704 vs Kiss1^CreAUC = 171.9 ± 18.88; P = .58), basal LH corresponding to the lowest LH values for each animal (Vglut2^Mut basal = 0.616 ± 0.04140, Kiss1^Cre basal = 0.570 ± 0.08597; P = .5975), and LH amplitude, corresponding to highest LH values for each animal (Vglut2^Mut amplitude = 1.23 ± 0.05457 vs Kiss1^Cre amplitude = 1.38 ± 0.1342; P = .2731) that were similar to those found in their Kiss1^Cre controls (Fig. 6C and 6D). We also investigated LH pulses in the Vglut2^Mut and Kiss1^Cre females in the OVX state and found that Vglut2^Mut females (n = 10) had similar LH pulses as Kiss1^Cre controls (n = 6) (total number of LH pulses: Vglut2^Mut = 3.3 ± 0.36 vs Kiss1^Cre = 3.66 ± 0.760; P = .63; AUC: Vglut2^Mut: 647 ± 17.57 vs Kiss1^Cre = 690.7 ± 34.96; P = .23; basal LH: Vglut2^Mut = 2.76 ± 0.094 vs Kiss1^Cre: 2.928 ± 0.153; P = .35; and LH amplitude: Vglut2^Mut = 4.576 ± 0.132 vs Kiss1^Cre: 4.944 ± 0.363; P = .27) (Fig. 6E and 6F). We also investigated if the loss of glutamate signaling in Kiss1^ARH neurons impairs body weight. Body weights of Vglut2^Mut (n = 10) and Kiss1^Cre control (n = 7) females were assessed for 10 weeks following AAV injections. There was no statistically significant difference in the weight gain between the two groups of females (Kiss1^Cre: 1.757 ± 0.1702 g, n = 7 vs Vglut2^Mut: 1.790 ± 0.1894 g, n = 10; P > .05).

Figure 6.

Estrous cycles are not impaired by the loss of Vglut2 in Kiss1^ARH neurons. A, Representative examples of estrous cycles in Kiss1^Cre (n = 7) and Vglut2^Mut (n = 10) females assessed by daily vaginal cytology for 15 days. B, Vglut2^Mut females display regular estrous cycles, with similar times spent in diestrus, estrus, and proestrus as Kiss1^Cre controls. Pattern of luteinizing hormone (LH) pulsatility was analyzed in C and D, gonad-intact and E and F, ovariectomized Kiss1^Cre and Vglut2^Mut females. LH samples were collected every 10 minutes for 180 minutes. *Represents LH pulses. Total number of LH pulses show a small but statistically significant decrease in the Vglut2^Mut females compared to Kiss1^Cre females in the intact state (unpaired t test, t₍₁₂₎ = 2.825; *P = .0153; Vglut2^Mut = 2.3 ± 0.2 vs Kiss1^Cre = 3.2 ± 0.3). Total number of pulses/180 minutes, LH total secretory mass assessed by area under the curve. Basal LH and LH pulse amplitude were assessed.

Vglut2 Expression in Kiss1^ARH Neurons Is Required for the Preovulatory Luteinizing Hormone Surge Driving Ovulation

To assess the contribution of Vglut2 signaling in Kiss1^ARH neurons in the generation of the preovulatory LH surge driving ovulation, Vglut2^Mut (n = 9) and Kiss1^Cre (n = 6) females underwent an LH surge induction protocol. The LH surge in the Vglut2^Mut females was significantly blunted in the afternoon/evening of proestrus (Fig. 7A) (8 Am: Vglut2^Mut vs 7:30 Pm: Vglut2^Mut; P = .9969) compared to Kiss1^Cre controls, which displayed a significantly higher LH surge compared to morning values (8 Am: Kiss1^Cre vs 7:30 Pm: Kiss1^Cre; P = .0087), and compared to afternoon/evening values in the Vglut2^Mut group (7:30 Pm: Kiss1^Cre vs 7:30 Pm: Vglut2^Mut; P = .0119). Congruent with these findings, the ovaries of Vglut2^Mut females displayed fewer corpora lutea (Fig. 7B). These findings suggest a role of Vglut2 in Kiss1^ARH neurons in the generation of the LH surge driving ovulation in females. These findings are perhaps not surprising based on the fact that there is a direct glutamatergic excitatory input to Kiss1^AVPV/PeN neurons from Kiss1^ARH neurons, which was significantly attenuated (absent) in Vglut2^Mut females (Fig. 3).

Figure 7.

Vglut2^Mut females display impaired LH surge and fewer corpora lutea in ovaries. A, Kiss1^Cre control (n = 6) and Vglut2^Mut (n = 9) females were subjected to an LH surge induction protocol. LH samples were collected in the morning (AM [8 Am]) and evening (PM [7:30 Pm]) after lights off. Two-way analysis of variance followed by Sidak post hoc test: main effect of treatment F_{(1, 26)} = 5.854; P = .0228, main effect of time F_{(1, 26)} = 9.472; P = .0049, and interaction F_{(1, 26)} = 5.959; P = .0218. **P < .01, *P < .05. B, Ovarian histology shows a decrease in the number of corpora lutea (CL) in the Vglut2^Mut females (unpaired t test, t₍₇₎ = 2.628; *P = .0340; Vglut2^Mut = 2.0 ± 0.8 vs Kiss1^Cre = 5.6 ± 1.0).

Discussion

We have shown for the first time that mutagenesis of Vglut2 in Kiss1^ARH neurons abrogates glutamatergic synaptic transmission between Kiss1^ARH and Kiss1^AVPV/PeN neurons. This is both a direct, fast ionotropic glutamate response but also possibly a slower metabotropic (mGluR1) glutamate response that excites Kiss1^AVPV/PeN neurons. Moreover, mutating Vglut2 in Kiss1^ARH neurons abrogated the LH surge and significantly attenuated the ovulatory drive. This is consistent with earlier findings from the O’Byrne laboratory (37, 38) showing that low-frequency (5 Hz) optogenetic stimulation of Kiss1^ARH neurons, which releases only glutamate (6), generates a surge-like increase in LH release following optical stimulation (37, 38). Therefore, there appears to be a critical role for glutamatergic transmission from the Kiss1^ARH to Kiss1^AVPV/PeN neurons in generating the LH surge driving ovulation in the female mouse.

Based on early immunocytochemical and in situ hybridization studies, Kiss1^ARH neurons were believed to be under “inhibitory” control of E2 feedback based on the fact that the expression of the peptide (kisspeptin, NKB and dynorphin) neurotransmitters in Kiss1^ARH neurons were all downregulated by E2 (1, 5, 39, 40). However, in terms of Kiss1^ARH cellular excitability, voltage-gated calcium channels, hyperpolarization-activated cyclic nucleotide-gated (HCN) channels and Vglut2 expression are all upregulated by E2 in Kiss1^ARH neurons (12). At the synaptic level, E2 also increases glutamatergic transmission from Kiss1^ARH neurons to pro-opiomelanocortin, NPY/AgRP, but most importantly to Kiss1^AVPV/PeN neurons (12). Kiss1^ARH neurons are also excited by coreleased glutamate from neighboring Kiss1^ARH neurons in a “paracrine” manner (6). Interestingly, Vglut2 mRNA expression in Kiss1^ARH neurons and the efficacy of glutamate neurotransmission, along with the neuropeptides, are increased with castration in male mice (41), which reveals a significant sex difference in the glutamatergic signaling by Kiss1^ARH neurons (12, 41). However, in males there is no preovulatory LH surge so one could argue that there is no need for excitatory glutamatergic input to the few Kiss1^AVPV/PeN neurons expressed in males.

Within the Kiss1^ARH neurocircuitry, the lack of glutamate transmission did diminish the amplitude of the slow EPSP, which underlies the synchronization of Kiss1^ARH neurons. The slow EPSP, which is dependent on NKB, is attenuated by E2 treatment (6), and it was further reduced in E2-treated, Vglut2^Mut OVX females (see Fig. 2). Since the slow EPSP underlies NKB-dependent engagement of synchronous activity in Kiss1^ARH neurons (6), glutamate may also play a role in modulating Kiss1^ARH synchronous firing activity and LH pulsatility in the presence of high circulating levels of E2, when the expression of peptides in the Kiss1^ARH neurons are at a nadir but glutamate levels are high in female Kiss1^ARH neurons. However, females with mutagenesis of Vglut2 in Kiss1^ARH neurons appeared to exhibit a normal ovulatory cycle, an indication that glutamatergic neurotransmission from Kiss1^ARH neurons plays only a minor role to support pulse-generator activity. This is in apparent contrast to male mice, in which glutamate via glutamatergic AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) receptors appear to drive the synchronization of Kiss1^ARH neurons (42, 43). However, the authors did not mutate Vglut2 in Kiss1^ARH neurons to support their supposition.

Low-frequency (5 Hz) optogenetic stimulation of Kiss1^ARH neurons, which releases glutamate (6), generates a surge-like increase in LH release during optical stimulation (37, 38), and intra-AVPV infusion of the glutamate ionotropic receptor blockers AP5 (2-amino-5-phosphonopentanoate) and CNQX (6-cyano-7-nitroquinoxaline-2,3-dione) blocks LH secretion induced by optogenetic stimulation of Kiss1^ARH terminals in the AVPV of proestrus females (44). Presently, we show that mutagenesis of Vglut2 in Kiss1^ARH neurons significantly attenuated the estradiol benzoate–induced LH surge and reduced the ovulatory drive (ie, reduced the number of corpus lutea; see Fig. 7).

In females, mutagenesis of Vglut2 in Kiss1 neurons essentially abrogates glutamate release from Kiss1^ARH neurons. However, the lack of glutamate transmission does not abolish the slow EPSP, which is generated by high-frequency synchronous firing of these neurons (see Fig. 2). Indeed, glutamate generates only small “synchronizing” events that are dependent on glutamatergic ionotropic receptors (42), but the fast neurotransmitter is unable to support the sustained firing (ie, slow EPSP) that is necessary for peptide release and synchronization of the Kiss1^ARH network. Rather, as we demonstrate here, glutamate neurotransmission appears to be more important for excitation of Kiss1^AVPV/PeN neurons to facilitate the GnRH (LH) surge in the presence of high circulating levels of E2. Ultimately, the timing of the LH surge is dependent on the vasopressin input from suprachiasmatic nucleus neurons that increases HCN channel activity in Kiss1^AVPV/PeN neurons (45). This HCN channel depolarizes the Kiss1^AVPV/PeN neurons into the voltage range to facilitate the activation of a persistent Na⁺ current that drives the high-frequency firing of these Kiss1^AVPV/PeN neurons (31). Importantly, in the presence of high physiological levels of E2, the mRNA expression of Tac2 is many-fold higher than Kiss1, which is essential for NKB maintaining the synchronous firing of Kiss1^ARH neurons, albeit at a lower frequency, across all physiological states (6). In the ewe there is a progressive change from a strictly pulsatile pattern of GnRH in the hypophyseal portal blood to one containing both pulsatile and “nonpulsatile” components during the GnRH surge (46, 47). Also, physiological and anatomical studies in the ewe support our findings in mice that Kiss1^ARH (KNDy) neurons are active during the preovulatory LH surge (48, 49) and modulate the LH surge (50). Moreover, in the ewe there is an E2-dependent increase in glutamatergic input to Kiss1^ARH neurons (51), which is congruent with the E2-driven increase in Vglut2 expression and glutamate release from mouse Kiss1^ARH neurons (12). Similarly, pulsatile LH secretion is maintained during the preovulatory LH surge in humans (52). Therefore, the concept that E2 only inhibits Kiss1^ARH neuronal activity (ie, “negative feedback”) is outdated. Rather, we believe that there is a transition from a “pulse-generator” mode to a “surge-generating” mode for the Kiss1^ARH neurons so that they switch from a synchronized firing pattern, favoring peptide release, to a burst firing mode favoring glutamate release in mammals (Fig. 8). With the E2-driven increase in the expression of voltage-gated calcium channels and pacemaker HCN channels (12), Kiss1^ARH neurons may also become “pacemaker” neurons that are in synchrony with the faster firing Kiss1^AVPV/PeN neurons and thus provide the crucial input at the time of the preovulatory surge. Therefore, our in vitro and in vivo findings provide a critical conceptual framework of how Kiss1^ARH neurons participate in the preovulatory GnRH (LH) surge, which ultimately is vital for ovulation in females.

Figure 8.

The activation of Kiss1^ARH neurons augments the excitability of Kiss1^AVPV/PeN during the preovulatory surge. At low estradiol levels, high-frequency photostimulation of Kiss1^ARH neurons triggers the release of neurokinin B (NKB), which depolarizes and recruits other Kiss1^ARH neurons. Dynorphin, coreleased by these neurons, acts presynaptically to modulate (inhibit) NKB release. This balance between NKB and dynorphin governs the synchronous activity of Kiss1^ARH neurons, promoting the release of kisspeptin. This kisspeptin release stimulates GnRH release in the median eminence (ME), which results in pulsatile luteinizing hormone (LH) release. At surge levels of estradiol, Kiss1^ARH neurons communicate with Kiss1^AVPV/PeN neurons through the fast neurotransmitter glutamate, which triggers burst firing in Kiss1^AVPV/PeN neurons. The activation of these rostral Kiss1 neurons leads to kisspeptin release, robustly exciting gonadotropin-releasing hormone (GnRH) neurons via the GPR54 signaling cascade, ultimately driving the release of GnRH during the preovulatory LH surge.

Abbreviations

AAV

adeno-associated virus

aCSF

artificial cerebrospinal fluid

ANOVA

analysis of variance

AP

anteroposterior

ARH

arcuate nucleus

AUC

area under the curve

AVPV

anteroventral periventricular

cDNA

complementary DNA

DHPG

dihydroxyphenylglycine

DV

dorsoventral

E2

17β-estradiol

ELISA

enzyme-linked immunosorbent assay

EPSP

excitatory postsynaptic potential

GABA

γ-aminobutyric acid

GnRH

gonadotropin-releasing hormone

HCN

hyperpolarization-activated cyclic nucleotide-gated

LH

luteinizing hormone

ML

mediolateral

mEPSCs

miniature excitatory postsynaptic currents

mRNA

messenger RNA

NKB

neurokinin B

Kiss1

kisspeptin

OHSU

Oregon Health and Science University

OVX

ovariectomized

PB

phosphate buffer

PBS

phosphate-buffered saline

PeN

periventricular nuclei

qPCR

quantitative polymerase chain reaction

sgRNA

single-guide RNA

TTX

tetrodotoxin

Vglut2

vesicular glutamate transporter 2

Funding

This work was supported by the National Institutes of Health (NIH) (grant No. R01-DK68098), a grant from OHSU Center for Women's Health (Circle of Giving; to O.K.R. and M.J.K.), and NIH grants HD090151, HD099084, and DK133760 (to V.M.N.). The generation of the sgRNAs was funded by NIH grants P30-MH048736 and R01-MH104450 (to L.S.Z.).

Author Contributions

O.K.R. conceived the study; J.Q., R.T., M.A.B., L.S.Z., M.J.K., V.M.N., and O.K.R. designed the study; J.Q., R.T., M.A.B., and E.M. performed the experiments. J.Q., R.T., M.A.B., E.M., M.J.K., V.M.N., and O.K.R. analyzed data. J.Q., R.T., M.J.K., V.M.N., and O.K.R. wrote the manuscript with input from all the authors.

Disclosures

The authors declare no competing financial interest.

Data Availability

Original data generated and analyzed during this study are included in this article or in the data repositories listed in Reference (53).