17β-Estradiol Increases Persistent Na⁺ Current and Excitability of AVPV/PeN Kiss1 Neurons in Female Mice

Abstract

In vitro slice studies have revealed that there are significant differences in the spontaneous firing activity between anteroventral periventricular/periventricular preoptic nucleus (AVPV/PeN) and arcuate nucleus (ARC) kisspeptin (Kiss1) neurons in females. Although both populations express similar endogenous conductances, we have discovered that AVPV/PeN Kiss1 neurons express a subthreshold, persistent sodium current (I_NaP) that dramatically alters their firing activity. Based on whole-cell recording of Kiss1-Cre-green fluorescent protein (GFP) neurons, I_NaP was 4-fold greater in AVPV/PeN vs ARC Kiss1 neurons. An LH surge-producing dose of 17β-estradiol (E2) that increased Kiss1 mRNA expression in the AVPV/PeN, also augmented I_NaP in AVPV/PeN neurons by 2-fold. Because the activation threshold for I_NaP was close to the resting membrane potential (RMP) of AVPV/PeN Kiss1 neurons (−54 mV), it rendered them much more excitable and spontaneously active vs ARC Kiss1 neurons (RMP = −66 mV). Single-cell RT-PCR revealed that AVPV/PeN Kiss1 neurons expressed the requisite sodium channel α-subunit transcripts, Na_V1.1, Na_V1.2, and Na_V1.6 and β subunits, β2 and β4. Importantly, Na_V1.1α and -β2 transcripts in AVPV/PeN, but not ARC, were up-regulated 2- to 3-fold by a surge-producing dose of E2, similar to the transient calcium current channel subunit Ca_v3.1. The transient calcium current collaborates with I_NaP to generate burst firing, and selective blockade of I_NaP by riluzole significantly attenuated rebound burst firing and spontaneous activity. Therefore, I_NaP appears to play a prominent role in AVPV/PeN Kiss1 neurons to generate spontaneous, repetitive burst firing, which is required for the high-frequency-stimulated release of kisspeptin for exciting GnRH neurons and potentially generating the GnRH surge.

Kisspeptin (Kiss1) neurons in the anteroventral and more caudal periventricular preoptic nucleus (AVPV/PeN) are thought to be vital presynaptic neurons that drive GnRH neuronal excitability (1, 2), and it has been shown that high-frequency stimulation of these neurons (>5 Hz) is required to induce kisspeptin release and hence excitation of GnRH neurons (3). Both a low threshold, transient calcium current (I_T) and a hyperpolarization-activated, cyclic nucleotide-gated current (I_h) are prominently expressed in both AVPV/PeN and arcuate nucleus Kiss1 neurons (4, 5). Kiss1 neurons also express the corresponding critical channel transcripts, Ca_v3 and HCN, which underlie these currents (5, 6). In addition, AVPV/PeN Kiss1 neurons express a prominent persistent Na⁺ current (I_NaP) that is activated approximately 10 mV negative to the threshold for the transient, rapidly inactivating, sodium current (I_NaT), which underlies the action potential generation (5). Although 17β-estradiol (E2) treatment increases both I_h and I_T and the excitability of AVPV/PeN Kiss1 neurons (5, 7), nothing is known about the physiological regulation of I_NaP in Kiss1 neurons during the female reproductive cycle.

Persistent sodium currents help sculpt firing patterns of central nervous system (CNS) neurons (8–10). Na⁺ channel proteins that underlie the persistent sodium current are complexes of a 260-kDa α-subunit in association with one or two auxiliary β-subunits (β1–4) of 33–34 kDa (11). Three different α-subunits are expressed in the adult mammalian brain, Na_V1.1, Na_V1.2, and Na_V1.6, all of which can contribute to I_NaP. The α-subunits associate with one or more β-subunits, which regulate not only the trafficking of α-subunits but also their gating properties. Although it is cellular environmental dependent, β1-, β2-, and β3-subunits all promote persistent sodium currents in CNS neurons (8–10). Therefore, Na_V1.1-, Na_V1.2-, and Na_V1.6-α subunits not only generate I_NaT but also associate with β-subunits to produce a regenerative depolarizing current in the subthreshold voltage range, ie, the persistent sodium current. Although the magnitude of I_NaP is only about 2%–5% of the transient sodium current, it is activated approximately 10 mV negative to I_NaT, in which few voltage-gated channels (eg, potassium channels) are active and the input resistance is high (10), so the depolarizing effects of this current are greatly amplified. Therefore, I_NaP shapes repetitive firing, generates rhythmicity, and amplifies excitatory postsynaptic potentials (12). I_NaP modulates pacemaking activity in numerous CNS neurons including cerebellar Purkinje neurons (13, 14), CA1 pyramidal neurons (15), ventral tegmental dopamine neurons (16), hypothalamic suprachiasmatic neurons (17), tuberomammilary histamine neurons (18), and GnRH neurons (19), all of which depend on the interplay between I_NaP, I_T, and I_h (20). Also, neurotransmitters that affect any of these conductances greatly facilitate burst firing activity. Therefore, we hypothesized that similar to its effects on I_T and I_h, E2 increases I_NaP in AVPV/PeN Kiss1 neurons and consequently the burst firing capability of these vital reproductive neurons.

Materials and Methods

Animals

All procedures performed with animals were in accordance with the National Institutes of Health guidelines for the care and use of laboratory animals and were approved by our local committee on animal care and use. Kiss1-Cre-green fluorescent protein (Kiss1-Cre-GFP; C57BL6/J and S129 background) mice were produced by Dr Robert Steiner and his colleagues at the University of Washington (4). Offspring that carried the transgene were identified by PCR on the genomic DNA extracted from tail biopsies. Transgenic animals were maintained as heterozygous by breeding with wild-type C57BL6/J mice. Animals were housed under constant temperature and light in a 12-hour light, 12-hour dark cycle, with lights on at 6:00 am [zeitgeber time (ZT) 0] and lights off at 6:00 pm (ZT12). Food and water were provided ad libitum. Heterozygous female animals between 8 and 16 weeks of age were used for all experiments. Bilateral ovariectomies (OVX) were performed under inhalant isofluorane anesthesia. Carprofen (Rimadyl; Pfizer) was given immediately after a surgery at a dose of 4 mg/kg for analgesia. After OVX, animals were treated with E2, which has been shown to induce an LH surge in Kiss1-Cre-GFP mice (5). Five days (d 5) after ovariectomy, animals were treated with a sc injection of 0.25 μg 17β-estradiol benzoate in sesame oil (50 μl) at ZT4. On the following day (d 6), animals were given a sc injection of 1.5–2.0 μg 17β-estradiol benzoate in sesame oil. On the day of the induced surge (d 7), E2-treated animals (high-E2 group; uterine weights 90–130 mg) were killed at approximately ZT4.

An additional group of animals was treated as above but received less E2 as a model for the diestrous stage of the estrous cycle. These animals received injections of 0.0625 μg 17β-estradiol benzoate in sesame oil on day 5 and/or day 6 after OVX and were used for experiments on day 7 (low-E2 group; uterine weights 30–44 mg). Other animals used for cell harvesting of arcuate nucleus (ARC) Kiss1 neurons were OVX and received oil vehicle treatment (uterine weights 13–18 mg).

Slice preparation

On the day of each experiment, the animal was killed by decapitation and the brain was removed from the skull. The brainstem was removed, and the resulting block was mounted on a cutting vibratome and submerged in ice-cold oxygenated (95% O₂-5% CO₂) high-sucrose artificial cerebral spinal fluid (aCSF) (in millimoles: 208 sucrose, 2 KCl, 1 MgCl₂, 1.25 NaH₂PO₄, 10 HEPES, 26 NaHCO₃, 10 dextrose, 2 MgSO₄, and 1 CaCl₂). Slices of 220–250 μm thickness were cut through the AVPV/PeN area and the ARC. Slices were transferred to an auxiliary chamber containing oxygenated aCSF at room temperature (in millimoles: 124 NaCl, 5 KCl, 1.44 NaH₂PO₄, 5 HEPES, 10 dextrose, 26 NaHCO₃, 2 MgCl₂, and 2 CaCl₂) and allowed to recover for at least 1 hour. Also, the uterus was removed, trimmed of fat, blotted dry, and weighed. Uterine weights served as an indicator of circulating E2 levels (5, 21).

Cell harvesting and RT-PCR

Tissue slices (250 μm) containing the AVPV/PeN were microdissected and used for neuronal dispersion and harvesting as described previously (5, 22). Briefly, the tissue was treated with protease (from Streptomyces griseus; Sigma) at 37°C for 15 minutes and then washed three times with low Ca²⁺ aCSF (1 mM Ca²⁺) and two times in aCSF. Using flame-polished glass Pasteur pipettes of decreasing sizes, tissues were titrated and dispersed neurons were plated onto a 60-mm cell culture dish with a glass bottom. Oxygenated aCSF at room temperature was constantly perfused into the dish at a rate of 2 ml/min. Individual GFP-positive and several GFP-negative neurons were harvested into a standard glass pipette (1.5 mm outer diameter/0.83 mm inner diameter; World Precision Instruments) using the XenoWorks microinjector system (Sutter Instruments). Importantly, GFP-negative cells were always negative to Kiss1 mRNA, whereas GFP-expressing cells were positive to Kiss1.

Only fully intact, healthy neurons that showed uniform GFP fluorescence (or no fluorescence) and were anchored to the glass plate were harvested.

Neurons were ejected into a siliconized 0.5-ml tube containing a solution of 1× Invitrogen Superscript III buffer, 15 U RNasin (Promega), 10 mM of dithiothreitol, and diethylpyrocarbonate-treated water in a total of 5 μl (single cells) or 8 μl (pools of cells). Cells were harvested individually or as pools (five cells/pool) and were frozen on dry ice and stored at −80°C until further processing.

Harvested cells were reverse transcribed according to established protocols (22). The cDNA was stored at −20°C. The positive and negative controls included preoptic or arcuate hypothalamic RNA samples with reverse transcriptase added (+RT) or without reverse transcriptase (−RT), single cells without reverse transcriptase, aCSF from the surrounding area, and water blank. Each cell or pool of cells was then evaluated by using single-cell PCR or quantitative real-time PCR (qPCR), respectively.

Primer design

PCR primers were designed using Clone Manager Software (Sci Ed Software). Primers were designed to cross at least one intron/exon boundary. Only primers that demonstrated a single-peak melting curve and had efficiency between 95% and 100% were used. Primer sequences for single-cell PCR were as follows: Na_V1.1α (accession number NM_018733), forward primer 3393–3414 nt, reverse primer 3587–3608 nt, 216 bp; Na_V1.2α (accession number NM_001099298), forward primer 894–911 nt, reverse primer 1066–1083 nt, 190 bp; Na_V1.6α (accession number NM_001077499), forward primer 3448–3468 nt, reverse primer 3561–3581 nt, 134 bp; Na_Vβ2 (accession number NM_001014761), forward primer 85–104 nt; reverse primer 313–332 nt, 248 bp; Na_Vβ4 (accession number NM_001013390), forward primer 216–237 nt, reverse primer 351–372 nt, 157 bp. Primer sequences for qPCR were as follows: Na_V1.1α (118 bp product), forward primer 3501–3519 nt, reverse primer 3598–3618 nt; Na_V1.2α (130 bp product), forward primer 3925–3946 nt, reverse primer 4033–4054 nt; Na_V1.6α (82 bp product), forward primer 4290–4309 nt, reverse primer 4352–4371 nt; Na_Vβ2 (105 bp product), forward primer 187–206 nt, reverse primer 272–291 nt; Na_Vβ4 (117 bp product), forward primer 24–41 nt, reverse primer 123–140 nt; glyceraldehyde-3-phosphate dehydrogenase (GAPDH; accession number NM_008084), forward primer 689–706 nt, reverse primer 764–781 nt, 93 bp. Primers for Ca_V3.1, Kiss1, and β-actin were as described previously (5, 22).

Single-cell PCR and qPCR

Analysis of single cells was performed on 3 μl of cDNA as described previously (22). Conditions for the PCR were as follows: initial denaturation for 2 minutes at 95°C; 50 cycles of amplification at 63°C for Na_V1.1α, Na_V1.2α, and Na_V1.6α; 50 cycles of amplification at 61°C for Na_Vβ2 and Na_Vβ4, followed by a final extension at 72°C for 5 minutes. PCR products were visualized with ethidium bromide on a 2% agarose gel. Analysis by qPCR was performed using the Power SybrGreen master mix method as described previously (22, 23) on a Life Technologies Quant Studio 7 Flex System (Life Technologies). cDNA samples of AVPV/PeN neurons were run in 3- to 4-μl duplicates for the target genes (Kiss1, Na_V1.1α, Na_V1.2α, Na_V1.6α, Na_Vβ2, Na_Vβ4, Ca_V3.1) and 2 μl duplicates for the reference genes (β-actin, GAPDH), and the mean Δcycle threshold of the low E2-treated pool values was used as a calibrator when comparing the mRNA quantities with the high E2-treated samples. The mRNA expression of Kiss1, Na_V1.1α, Na_V1.2α, Na_V1.6α, and Na_Vβ2 (Na_Vβ4 was not detected in ARC Kiss1 neurons) was also quantified in harvested arcuate kisspeptin neurons from OVX high E2, as compared with OVX oil-treated animals as described above with the exception that GAPDH was used as the reference gene (see below). Data were analyzed following the comparative ΔΔcycle threshold method described previously (22, 23) and are reported as relative amounts of mRNA expression to the calibrator.

Electrophysiological recording

Whole-cell patch clamp recordings were conducted from GFP-labeled Kiss1 neurons with the use of an Olympus BX51 (Olympus) or a Zeiss Axioskop FS upright microscope (Carl Zeiss Microscopy) equipped with fluorescence and infrared differential interference contrast imaging devices (1, 4, 24, 25). Patch pipettes (1.5 mm outer diameter borosilicate glass; A-M Systems) were pulled on a Flaming/Brown puller (model P-97; Sutter Instrument Co) and filled with the following internal solution (in millimoles): 128 potassium gluconate, 10 NaCl, 1 MgCl₂, 11 EGTA, 10 HEPES, 2 or 4 ATP, and 0.25 GTP; adjusted to pH 7.3 with KOH; 290 mOsm. Pipettes filled with the above internal solution had a resistance of 2–3 MΩ. In whole-cell configuration, and access resistance was less than 25 MΩ and was 80% compensated for the I_NaT. Generally, I_NaP is measured from the current activated by a slow voltage ramp (<100 mV/sec) from −80 to −20 mV (17, 18).

In preliminary experiments we examined the I_NaP with voltage ramps from 20 mV/sec to 100 mV/sec and found that the amplitude of I_NaP decreased with the decrease in the ramp speed and reached a state at 20 mV/sec when the I_NaP was relatively stable and action currents (from the unclamped axon) were eliminated (Figure 1A). Therefore, we measured the I_NaP at a ramp speed of 20 mV/sec in the absence and presence of tetrodotoxin (TTX; 1 μM), which blocks the persistent (and transient) sodium current. To record I_NaT, a step voltage command of 100 milliseconds from −60 mV to −20 mV with a step size of 2.5 mV was applied to cells. Postinhibitory rebound burst firing was examined at resting membrane potentials after a series of hyperpolarizing steps of 1 second duration from −70 mV to −120 mV to recruit active T-type calcium channels (23). The spontaneous firing of AVPV/PeN and arcuate Kiss1 neurons were examined in the presence of ionotropic glutamate and γ-aminobutyric acid (GABA)_A receptor blockers [10 μM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), 50 μM DL-2-amino-5-phosphonopentanoic acid (AP-5), and 100 μM picrotoxin]. The liquid junction potential of −10 mV was corrected for in all analyses. Consecutive current traces were filtered at 2–10 kHz and acquired at a sampling rate of 0.5–5 kHz. Electrophysiological signals were amplified with an Axopatch 200B amplifier and digitized with a Digidata 1440A (Molecular Devices), or with an Axopatch 1D amplifier and digitized with a Digidata 1322 (Axon Instruments). Data were analyzed with Clampfit software (version 9.0 or 10.0; Molecular Devices).

Figure 1.

I_NaP in AVPV/PeN and ARC Kiss1 neurons. A, Representative recordings showing the relationship between I_NaP and voltage ramps (speeds from 20 mV/sec to 100 mV/sec) from an AVPV/PeN Kiss1 neuron. B and C, Voltage clamp recordings showing I_NaP generated by a voltage ramp from −80 to −20 mV at 20 mV/sec before (black trace; control) and after (red trace) application of TTX (1 μM) in representative AVPV/PeN (B) and ARC (C) Kiss1 neurons. The black straight line indicates the linear part of the whole-cell current in the presence of TTX. The horizontal stippled line indicates the zero current level. The vertical stippled line marks the resting membrane potential, and the arrow points to activation threshold for the I_NaP. D, Comparison of I_NaP density measured at −45 mV in AVPV/PeN Kiss1 neurons and ARC Kiss1 neurons from high-E2 animals. **, P < .01, n = 9–14; Student's t test.

Drugs or chemicals

Riluzole, CNQX, AP-5, and picrotoxin were purchased from Tocris Bioscience and were made up as a stock of 10, 10, 50, and 100 mM, respectively, in DMSO. TTX was purchased from Alomone Laboratories and was made up as a 1-mM stock in Milli-Q water.

Data analysis

PCR/qPCR

For quantification of mRNA expression differences between low-E2 (n = 5) and high-E2 (n = 5–6) animals, Kiss1 neuronal pools were analyzed using real-time qPCR as described above. Three pools of five cells each were analyzed from each animal, and the average mRNA expression relative to β-actin was determined for each animal and used for further analysis of mean ± SEM, and differences in gene expression were analyzed by a two-tailed Student's t test or one-way ANOVA. Note that β-actin was used as a reference gene for quantification of mRNA expression in AVPV/PeN Kiss1 neurons because its mRNA was not regulated by E2 in these neurons. For quantification of mRNA expression in arcuate Kiss1 neurons we used GAPDH rather than β-actin because β-actin, but not GAPDH, was regulated by E2 in these neurons.

Electrophysiology

The spontaneous firing frequencies of cells were measured at their resting membrane potential. The I_NaP activation threshold was defined as the voltage in which 5% of the maximum I_NaP was activated (15). Data were analyzed with Mini Analysis (Synaptosoft Inc) and Clampfit 9.2 (Molecular Devices) software. Graphs were plotted using GraphPad Prism 4, Sigma Plot 8.0, and Macromedia Freehand 10 software. Comparisons between different treatments were performed using a paired or unpaired Student's t test. Differences were considered significant if the probability of error was less than 5%. All data are presented as mean ± SEM.

Results

I_NaP in AVPV/PeN and ARC Kiss1 neurons

Kiss1 neurons in the AVPV/PeN and ARC of females exhibit dramatically different firing activities (4, 5, 7, 26, 27). Because both AVPV/PeN and ARC Kiss1 neurons express I_T and I_h but exhibit these differences in firing activity (4, 5), we hypothesized that the expression of I_NaP in these two distinct populations of Kiss1 neurons would account for these differences. First, we isolated I_NaP in AVPV/PeN and ARC Kiss1 neurons using a ramp voltage protocol in the presence or absence of sodium channel blocker TTX (see Materials and Methods). Similar to other hypothalamic neurons (17, 18), I_NaP in Kiss1 neurons was evoked by slowly ramping the membrane voltage from −80 to −20 mV at 20 mV/sec. In this way we were able to isolate I_NaP without contamination of I_NaT. Figure 1A illustrates that I_NaP decreased with the ramp speed from 100 mV/s to 40 mV/s, but stabilized between 40 mV/sec and 20 mV/sec. Also, we observed that at 20 mV/sec the I_NaT was eliminated in most cells. Therefore, a ramp speed of 20 mV/sec was chosen for later experiments, and cells that displayed I_NaT in the ramp were excluded from the analysis (17, 18). In AVPV/PeN and ARC Kiss1 neurons, the I_NaT was not activated until −43.1 ± 2.4 mV and −41.4 ± 1.8 mV (n = 14) mV, respectively, and was inactivated with the slow voltage ramp speed of 20 mV/sec (Figure 1A). Therefore, we measured I_NaP at −45 mV by subtracting the linear background current obtained in the presence of TTX. I_NaP was approximately 4-fold greater in AVPV/PeN Kiss1 neurons vs ARC Kiss1 neurons (1.16 ± 0.18 pA/pF, n = 14 vs 0.34 ± 0.11 pA/pF, n = 9, respectively; Figure 1).

Also, the activation threshold for I_NaP was significantly different between the two population of neurons (AVPV/PeN: −54.2 ± 1.3 mV, n = 14 vs ARC: −50.1 ± 1.5 mV, n = 9) (Figure 1, B and C). Furthermore, the whole-cell current between the resting membrane potential and the I_NaP threshold was inward (depolarizing) in the majority (88%) of AVPV/PeN Kiss1 neurons (Figure 1B) but outward (hyperpolarizing) in the majority (70%) of ARC Kiss1 neurons (Figure 1C), which is consistent with the fact that the resting membrane potential of AVPV/PeN Kiss1 neurons was depolarized (−54.6 ± 0.8 mV, n = 15) relative to the resting membrane potential of ARC Kiss1 neurons (−65.9 ± 1.2 mV, n = 13). However, the activation threshold for I_NaT was not significantly different between the two groups of Kiss1 neurons (AVPV/PeN neurons: 43.1 ± 2.4 mV, n = 14 vs ARC neurons: −41.4 ± 1.8 mV, n = 9).

E2 increases persistent sodium current in AVPV/PeN Kiss1 neurons

Previous studies showed that the excitability of Kiss1 neurons was regulated by circulating levels of E2 (4, 5, 7, 26, 27). Therefore, we examined whether E2 regulates the expression of I_NaP in AVPV/PeN and ARC Kiss1 neurons by comparing the whole-cell current density from both low-E2 and high-E2 animals. GFP neurons in the AVPV/PeN from low-E2 animals harvested in the slice during whole-cell recording (or after dispersion) were found consistently to be positive to Kiss1 mRNA (Figure 2, A and B) as described previously in high-E2 animals (5). Importantly, as shown in Figure 2, A and C, E2 increased the current density of I_NaP by greater than 2-fold in AVPV/PeN Kiss1 (1.25 ± 0.18 pA/pF for high-E2 vs 0.55 ± 0.14 pA/pF for low-E2, P < .01) but not in ARC Kiss1 neurons (0.38 ± 0.11 pA/pF for high-E2 vs 0.49 ± 0.23 pA/pF for low-E2).

Figure 2.

High estradiol treatment increased I_NaP in AVPV/PeN Kiss1 neurons but not in ARC Kiss1 neurons. A, I_NaP current in AVPV/PeN Kiss1 neurons is increased in high-E2 as compared with low-E2 animals. **, P < .01, Student's t test. B, qPCR measurements of Na_V1.1α and Na_Vβ2 mRNAs in AVPV/PeN Kiss1 neuronal pools (three pools per animal) from low-E2 and high-E2 OVX mice (n = 5 animals in each group). **, P < .01, n = 10–14; Student's t test. The gel figures in panels A and B illustrate that AVPV/PeN GFP neurons from low-E2 Kiss1 animals harvested during recording in the slice (A) or after dispersion (B) are Kiss1 positive. C, There was no difference in the magnitude of I_NaP in ARC Kiss1 neurons between high-E2 and low-E2 animals. D, In ARC Kiss1 neurons, Kiss1 mRNA expression was decreased with high-E2 treatment vs vehicle treatment; however, Na_V1.1α and Na_Vβ2 mRNA expression from high-E2 groups were not different from vehicle-treated OVX females. ***, P < .001, n = 5; Student's t test.

To see whether the increase in the amplitude of I_NaP in AVPV/PeN Kiss1 neurons was correlated to an increased mRNA expression of Na_Vα and/or the Na_Vβ subunits, we first determined which α- and β-subunits were expressed in AVPV/PeN Kiss1 neurons, and then we compared the expression levels in high-E2 vs low-E2 animals using quantitative PCR on pooled cells. Based on sc RT-PCR, Na_V1.1α, Na_V1.2α, Na_V1.6α, Na_Vβ2 and Na_Vβ4 were all expressed in AVPV/PeN Kiss1 neurons (n = 50 cells from four animals). Based on qPCR, the mRNA expression of Na_Vα and Na_Vβ subunits in AVPV/PeN Kiss1 neurons were as follows: Na_V1.2α was similar to or greater than Na_V1.6α, which was greater than Na_V1.1α (P < .05; ANOVA, n = 3 animals), whereas Na_Vβ2 was the predominant β-subunit expressed in these cells, and Na_Vβ4 mRNA was barely detectable. Moreover, our qPCR results showed that Na_V1.1α mRNA expression in AVPV/PeN Kiss1 neurons was increased by 2- to 3-fold, and the β2-subunit was increased by 2-fold in high-E2 compared with low-E2 females (Figure 2B). Also, as found previously in tissue, Kiss1 mRNA was significantly increased in Kiss1 neuronal pools in high-E2 vs low-E2 groups (Figure 2B). In addition, Na_V1.2α (high-E2, 1.61 ± 0.22; low-E2, 1.03 ± 0.04) and Na_V1.6α (high-E2, 2.40 ± 0.43; low-E2, 1.22 ± 0.19) were significantly increased in AVPV/PeN Kiss1 neuronal pools from high-E2 females (P < .05; n = 5 animals in each group). In ARC Kiss1 neurons, Na_V1.2α was greater than Na_V1.6α which was greater than Na_V1.1α (P < .05–0.01; ANOVA, n = 3 animals), and Na_Vβ2, but not β4, was also expressed. Comparing Kiss1 neurons from OVX vehicle-treated vs high-E2 animals revealed that Na_V1.1α and Navβ2 mRNA levels were not altered by E2 treatment, whereas Kiss1 mRNA levels were greatly reduced (Figure 2D). Also, similar to Na_V1.1α, the mRNA levels of Na_V1.2α and Na_V1.6α were also not regulated by E2 (data not shown).

Persistent sodium current and firing properties of AVPV/PeN Kiss1 neurons

Given that persistent sodium currents are critical for generating repetitive firing in the face of depolarizing synaptic input, we used the relatively selective I_NaP blocker riluzole (28) to abrogate I_NaP to examine the contribution of I_NaP to the overall excitability of AVPV/PeN Kiss1 neurons. In voltage clamp, riluzole (10 μM) robustly inhibited the I_NaP (74.1% ± 3.3%, n = 6; Figure 3A) but had negligible effects (<10% inhibition) on I_NaT (Figure 3A, inset). In current clamp, riluzole decreased the baseline firing activity of Kiss1 neurons by 88% ± 3.5% (n = 3) (Figure 3B). Moreover, riluzole attenuated the sustained firing induced by a depolarizing current (20 pA) injection by greater than 2-fold (Figure 3, C and D).

Figure 3.

Riluzole blocked I_NaP and inhibited firing in AVPV/PeN Kiss1 neurons. A, Representative recordings showing that treatment with riluzole (10 μM) for 10 minutes inhibited the I_NaP in an AVPV/PeN neuron. The inset shows that the I_NaT was only slightly inhibited (<10%) by riluzole. B, Representative recording showing that riluzole inhibited the spontaneous firing of an AVPV/PeN neuron by 90%. C, The current-induced firing before and after exposure to riluzole. The current injection protocol (20 pA, 1 sec duration) is illustrated at the bottom. D, Summary of the inhibitory effects of riluzole (illustrated in panel C) on the mean firing rate. **, P < .01, n = 6; Student's t test.

Because I_NaP collaborates with I_T to generate repetitive burst firing and we found that I_T in AVPV/PeN Kiss1 neurons was augmented by 6-fold in high-E2 females (5), we measured Ca_V3.1 mRNA levels in these neurons. Indeed, high-E2 compared with low-E2 increased the mRNA expression of Ca_V3.1 by 2.5-fold (Figure 4A). In high-E2 animals, we also ascertained whether I_NaP could modulate the transient calcium channel-mediated rebound burst firing (see Materials and Methods). AVPV/PeN Kiss1 neurons typically generated two to four action potentials during rebound burst firing under control conditions (Figure 4B). However in the presence of the I_NaP blocker riluzole (10 μM), the rebound burst firing was reduced to one to two spikes, and the interspike interval between the first and second spikes increased from 12 to 18 milliseconds. Riluzole had no effect on the amplitude of T-type calcium current (Figure 4B, inset). The summarized results in Figure 4, C and D, show that inhibiting I_NaP by riluzole significantly reduced the mean number of action potentials (ie, fast sodium spikes) generated during the rebound burst from 3.0 ± 0.37 to 1.5 ± 0.22 spikes and increased the interspike interval from 13.3 ± 1.3 to 19.1 ± 2.0 milliseconds (n = 6). Finally, because both I_T and I_NaP are critical for generating spontaneous firing and both are augmented in AVPV/PeN Kiss1 neurons by E2, we examined the spontaneous firing of these neurons in the presence of ionotropic glutamate and GABA_A receptor blockers (10 μM CNQX, 50 μM AP-5, and 100 μM picrotoxin). As one would predict, the mean firing rate of AVPV/PeN Kiss1 neurons from high-E2 vs low-E2 females was significantly increased (3.7 ± 0.6 Hz in high-E2 vs 1.5 ± 0.2 Hz in low-E2, P < .05; Figure 5). However, ARC Kiss1 neurons from low-E2 and high-E2 animals were not significantly different in the mean firing rate (0.04 ± 0.02 Hz for low-E2 vs 0.06 ± 0.03 Hz for high-E2).

Figure 4.

I_NaP modulates T-type calcium channel-mediated rebound burst firing. A, qPCR measurements of Ca_V3.1 mRNA in AVPV/PeN Kiss1 neuronal pools (three pools per animal) from low-E2 and high-E2 mice (n = 5 animals in each group). **, P < .01, Student's t test. B, Representative recordings showing that in control conditions hyperpolarizing a cell to −95 mV for 1 second from −60 mV induced rebound burst firing with four action potentials (upper trace); however, in the presence of riluzole (10 μM), the same hyperpolarization from −60 mV induced rebound burst firing with only two action potentials, and the interspike interval from the first to the second action potential increased from 12 milliseconds to 18 milliseconds. The inset (middle trace) shows that riluzole had no effect on the T-type calcium current amplitude. C and D, Summary of the effects of riluzole on the number of action potentials (spikes) in the rebound burst and interspike intervals. **, P < .01, n = 6; Student's t test.

Figure 5.

E2 increases the excitability of AVPV/PeN Kiss1 neurons. A and B, The spontaneous firing rate of Kiss1 neurons was increased in high-E2 vs low-E2 animals in the presence of ionic glutamatergic and GABAergic receptor blockers (10 μM CNQX, 50 μM AP-5, and 100 μM picrotoxin). C, A representative recording showing that E2-treated ARC Kiss1 neurons are silent in the presence of ionic glutamatergic and GABAergic receptor blockers. D, Summary of the effects E2 treatment on the spontaneous firing of AVPV/PeN and ARC Kiss1 neurons. *, P < .05; ***, P < .001, Student's t test.

Discussion

We have discovered a fundamental difference between AVPV/PeN and ARC Kiss1 neurons that may account for the dramatic differences in their spontaneous firing activity: AVPV/PeN Kiss1 neurons express a persistent sodium current that is approximately 4-fold greater than that of ARC Kiss1 neurons. The activation threshold and peak voltage were more negative in AVPV/PeN Kiss1 neurons than ARC Kiss1 neurons. Moreover, the resting membrane potential of AVPV/PeN Kiss1 neurons (−54 mV) was close to the threshold potential for activation of I_NaP, whereas the resting membrane potentials of ARC Kiss1 neurons (−64 mV) was significantly hyperpolarized relative to the threshold potential of I_NaP (4, 5). As shown previously, AVPV/PeN Kiss1 neurons tend to fire spontaneously (5, 26, 29), whereas ARC neurons are mostly silent except when they are depolarized to the I_NaP threshold (4, 26). Therefore, the more depolarized resting membrane potential in combination with the expression of the I_NaP may account for the fundamental differences in the spontaneous activity and burst firing capability of AVPV/PeN vs ARC Kiss1 neurons.

Although all of the adult Na_Vα subunits were detected and regulated by E2 in AVPV/PeN Kiss1 neurons, we found that Na_V1.1 was the most highly up-regulated by a surge-inducing dose of E2. The β2 subunit, which was also expressed and highly up-regulated by E2 in AVPV/PeN Kiss1 neurons, in combination with the pore-forming α-subunit will generate the persistent sodium current (8–10). As predicted from the regulation of the subunit mRNA expression, I_NaP was augmented by 2-fold in AVPV/PeN Kiss1 neurons from high-E2 females. Neither Na_V1α nor Na_Vβ mRNA expression was regulated by E2 in ARC Kiss1 neurons, which corresponded to the lack of E2 effects on I_NaP in these neurons. Therefore, it appears that there is a close correlation between ion channel mRNA expression and function in Kiss1 neurons as shown previously in other hypothalamic neurons (23, 30). Moreover, I_NaP in combination with the other endogenous conductances in Kiss1 neurons promoted the burst firing that is required for kisspeptin release (see below). Indeed, selective blockade of I_NaP with riluzole significantly diminished the ability of Kiss1 neurons to fire spontaneously and to burst fire after a hyperpolarizing stimulus.

Persistent sodium currents have been reported to drive spontaneous firing and rhythmic (pacemaking) burst firing in many CNS neurons (15, 17–19, 31, 32). Likewise, I_NaP appears to play a critical role in the control of spontaneous repetitive firing and T-type calcium channel-mediated burst firing of AVPV/PeN Kiss1 neurons. The function of I_NaP in different cells is determined by its interaction with other subthreshold conductances including I_T, I_h, and inwardly rectifying potassium conductances (20, 32, 33). For example, hippocampal CA1 pyramidal neurons are normally quiescent, but they express a robust I_NaP and fire spontaneously when depolarized by muscarinic agonists that bring the membrane potential into the activation voltage range of I_NaP (15). Therefore, to drive spontaneous firing, the resting membrane potential of the cell should be close to or above the activation threshold of I_NaP. For AVPV/PeN Kiss1 neurons, the mean resting membrane potential is −54 mV, which is right at the activation threshold of I_NaP; thus, they can fire spontaneously (5, 7, 27). In contrast, the resting membrane potential of ARC Kiss1 neurons is −64 mV, which is farther from the activation threshold of I_NaP (−50 mV); therefore, they are mostly quiescent except when they are depolarized by glutamate, neurokinin B (4, 34), and other metabotropic receptors (eg, leptin receptor) that are coupled to activation of canonical transient receptor potential channels (TRPCs) (25). With depolarizing stimuli, which can be induced with current injection, I_NaP increases exponentially, and AVPV/PeN neurons fire in a repetitive high frequency manner. Indeed, we found that blockade of I_NaP by riluzole attenuated spontaneous repetitive firing (ie, Figure 3B). Furthermore, Amarillo et al (20) predicted that I_NaP in combination with I_T will enhance burst firing in thalamocortical neurons, due to the fact that the activation of I_T depolarizes the membrane potential into the activation range of I_NaP, which boosts the generation of action potentials during a burst firing event. As predicted, riluzole significantly diminished the number of spikes during rebound burst firing of AVPV/PeN Kiss1 neurons (ie, Figure 4).

Generally, high-frequency burst firing is required for peptide release from neurons (35–37). Studies from Liu et al (3) confirmed that kisspeptin release from AVPV/PeN Kiss1 neurons is frequency dependent, and a minimum of 5 Hz stimulation is required to induce kisspeptin-mediated excitation of GnRH neurons. It is noteworthy that T-type calcium channel-mediated burst firing reaches frequencies of 15–200 Hz in AVPV/PeN Kiss1 neurons, which is an ideal range for triggering the release of kisspeptin (3, 5). On the other hand, low-frequency spontaneous firing would most probably not evoke kisspeptin release but could induce GABA release from AVPV/PeN Kiss1 neuronal terminals (3). Activation of GABA_A receptors causes short-lived excitation of GnRH neurons (38, 39), whereas activation of the Gq-coupled kisspeptin receptor (also known as GPR54) causes a sustained depolarization that lasts tens of minutes in an in vitro slice preparation (1, 40, 41). One caveat is that high-frequency stimulation may also release more GABA, which could spill over and activate GABA_B receptors postsynaptically in GnRH neurons (42). Although activation of GABA_B receptors hyperpolarizes GnRH neurons via activation of G protein-coupled K⁺ channels (43–45), this hyperpolarization could facilitate the subsequent actions of kisspeptin by recruiting more T-type calcium channels through deinactivation of I_T and activation of I_h at more hyperpolarized potentials (5, 23). This scenario needs to be elucidated using optogenetic tools to differentially evoke the release of GABA or kisspeptin from these AVPV/PeN neurons and study the complexity of the postsynaptic responses using select pharmacological ligands. However, previous and present findings clearly show that I_T, I_h, and I_NaP are all up-regulated by surge-inducing doses of E2, and the interaction of I_NaP and I_T enhances the spontaneous activity and high-frequency burst firing that is necessary for the release of kisspeptin from AVPV/PeN neurons.