17β-estradiol increases arcuate KNDy neuronal sensitivity to ghrelin inhibition of the M-current in female mice

Abstract

Obesity and anorexia result in dysregulation of the hypothalamic-pituitary-gonadal axis, negatively impacting reproduction. Ghrelin, secreted from the stomach, potentially mediates negative energy states and neuroendocrine control of reproduction by acting through the growth hormone secretagogue receptor (GHSR). GHSR is expressed in hypothalamic arcuate (ARC) Kisspeptin/Neurokinin B (Tac2)/Dynorphin (KNDy) neurons. Ghrelin is known to inhibit the M-current produced by KCNQ channels in other ARC neurons. In addition, we have shown 17β-estradiol (E2) increases Ghsr expression in KNDy neurons 6-fold and increases the M-current in NPY neurons. We hypothesize that E2 increases GHSR expression in KNDy neurons to increase ghrelin sensitivity during negative energy states. Furthermore, we suspect ghrelin targets the M-current in KNDy neurons to control reproduction and energy homeostasis. We utilized ovariectomized (OVX) Tac2-EGFP adult female mice, pre-treated with estradiol benzoate (EB) or oil vehicle and performed whole-cell-patch-clamp recordings to elicit the M-current in KNDy neurons using standard activation protocols in voltage-clamp. Using the selective KCNQ channel blocker XE-991 (40 μM) to target the M-current, oil- and EB-treated mice showed a decrease in the maximum peak current by 75.7 ± 13.8 pA (n=10) and 68.0 ± 14.7 pA (n=11), respectively. To determine the actions of ghrelin on the M-current, ghrelin was perfused (100 nM) in oil- and EB-treated mice resulting in the suppression of the maximum peak current by 58.5 ± 15.8 pA (n=9) and 59.2 ± 11.9 pA (n=9), respectively. KNDy neurons appeared more sensitive to ghrelin when pre-treated with EB, revealing that ARC KNDy neurons are more sensitive to ghrelin during states of high E2.

Introduction

Periods of undernutrition caused by food scarcity as well as overnutrition caused by an abundance of food are both associated with an imbalance in energy status that leads to a suppressed hypothalamic-pituitary-gonadal (HPG) axis, resulting in cessation of reproduction. While obesity has been linked to coronary heart disease, and type 2 diabetes mellitus (1), it also negatively impacts reproduction. In women, obesity leads to irregular menses, infertility, and miscarriages, among other complications (2, 3). These problems in reproductive health extend to individuals suffering from undernutrition. Amenorrhea is prominent in athletes and anorexia nervosa patients, largely due to their decreased body mass index (4). These perturbations in energy balance can lead to problems in the HPG axis, though the precise mechanisms remain unclear.

Energy balance and reproduction are centrally regulated processes that are controlled, in part, by neurons in the arcuate nucleus of the hypothalamus (ARC). Regulation of reproduction is controlled by negative and positive feedback of 17β-estradiol (E2) on the HPG axis mediated, to some extent though not entirely, by neurons expressing kisspeptin. There are two main groups of kisspeptin neurons located in the rodent brain (5). Kisspeptin is expressed in neurons in the anteroventral periventricular nucleus (AVPV) and ARC. ARC kisspeptin neurons co-express neurokinin B (Tac2), and dynorphin (Pdyn), and are therefore termed KNDy neurons. KNDy neurons mediate the negative feedback of E2 and progesterone on the HPG axis in females (6, 7). Neurokinin B and dynorphin in KNDy neurons act as positive and negative autoregulators of KNDy neuronal excitability, respectively. Their combined actions produce the pulse generator that controls gonadotropin releasing hormone (GnRH) release into the median eminence and subsequent release of Luteinizing Hormone (LH) and Follicle Stimulating Hormone (FSH) from the anterior pituitary (8–12). Both AVPV and ARC kisspeptin regulate the HPG axis by binding to its receptor Kiss1r expressed on the soma and axons of GnRH neurons (7–9, 13). KNDy neurons also contribute to the control of energy homeostasis, especially in females. Ablation of KNDy neurons abrogates the post-ovariectomy weight gain associated with E2 in rats, suggesting that KNDy neurons mediate, in part, the anorectic effects of E2 (14). One pathway for KNDy neurons to control energy balance is by directly depolarizing POMC neurons via kisspeptin and/or glutamate (15, 16).

Although the neuroendocrine mechanisms that link reproduction and energy balance are not well understood, it is known that ghrelin suppresses the frequency of pulsatile LH release in male and female rats as well as humans, in part, due to activation of β-endorphin (POMC) signaling (17–19). Ghrelin is a brain-gut peptide hormone secreted from the stomach to stimulate food intake by acting on its receptor, growth hormone secretagogue receptor (GHSR) expressed in the hypothalamus (20, 21). GHSR is expressed throughout the brain particularly in NPY/AgRP neurons in the ARC, while only 10% of POMC neurons express GHSR (22, 23). Furthermore, ghrelin activation of GHSR excites NPY neurons and controls calcium homeostasis (23, 24) by activating calcium channels (25, 26) and inhibiting the M-current, a hyperpolarizing, inwardly-rectifying potassium channel current (26, 27). Recently we reported that E2 increases Ghsr expression in the ARC, but not in NPY neurons (26, 28). While ghrelin actions on NPY neurons are well characterized, only a few studies have examined the actions of ghrelin on KNDy neurons. In one study, central ghrelin administration had no effect on ARC Kiss1 expression in a fed or fasted state (29) and, in second study, ghrelin depolarized more KNDy neurons in E2-treated ovariectomized (OVX) females than in oil-treated females (29, 30). These data suggest that direct effects of ghrelin on KNDy neuronal excitability, and not kisspeptin transcription, mediate the suppression of LH pulse frequency. Furthermore, we recently demonstrated that E2 increases Ghsr expression by 6-fold in KNDy (Tac2-GFP) neurons (30), which can potentially increase KNDy neuronal sensitivity to ghrelin.

The M-current is a subthreshold voltage gated potassium current generated by KCNQ channels which modulate firing frequency in neurons. KCNQ channels are expressed in numerous brain regions including: the hippocampus, the cortex, and the hypothalamus (27). Additionally, KCNQ channels are negatively modulated by Gq-coupled G-protein coupled receptors, including GHSR. The binding of ghrelin to GHSR activates the alpha subunit which activates the phospholipase C (PLC) signaling cascade. Once PLC becomes activated, it hydrolyzes the membrane phospholipid, phosphatidylinositol 4, 5-bisphosphate (PIP₂) for inositol triphosphate (IP₃) and diacylglycerol (DAG). IP3 is released into the cytosol which goes on to increase intracellular calcium. Additionally, PLC activation of the membrane associated DAG can activate protein kinase C (PKC) which then decreases KCNQ channel (or M-current) conductance (27, 31). Our previous work has shown that the M-current can be modulated by fasting and 17β-estradiol in NPY neurons of the hypothalamus (32). In addition, activation of the 5-HT_2c receptors inhibits the M-current in POMC neurons of the hypothalamus (33). Because KNDy neurons can directly depolarize POMC neurons via kisspeptin and glutamate, (15, 16) and because fasting (increasing ghrelin) suppresses the M-current in NPY neurons (32), we speculate that KNDy neuronal ghrelin sensitivity and the M-current play a significant role in the control of energy balance and reproduction.

Because E2, ghrelin, and KNDy neurons all control reproduction and energy homeostasis (14, 34–36), we hypothesize one potential mechanism could be that steroids (E2) increase ghrelin sensitivity in KNDy neurons by augmenting KCNQ channel conductance and Ghsr expression to modulate the control of reproduction and other hypothalamic pathways. During states of elevated E2 (proestrus or ovariectomy with E2 replacement), GHSR expression is increased in KNDy neurons, augmenting KNDy sensitivity to ghrelin. These transcriptional effects produce KNDy neurons more sensitive to states of elevated ghrelin (fasting, caloric restriction) increasing KNDy neuronal output, kisspeptin or glutamate release onto ARC POMC and NPY neurons (37). POMC activation would increase the release of β-endorphin on GnRH neurons to suppress LH pulsatility (23). To begin to elucidate this circuit and the interactions of E2 and ghrelin in KNDy neurons, we used whole-cell patch-clamp electrophysiology in a Tac2-GFP mouse model and determined the impact of E2 and ghrelin on M-current activity and excitatory postsynaptic currents.

Materials and Methods

Animals

All animal procedures were completed in compliance with institutional guidelines based on National Institutes of Health standards and were performed with Institutional Animal Care and Use Committee approval at Rutgers University. Adult mice (6–12 weeks of age) were housed under constant photoperiod conditions (12/12 h light/dark cycle), with lights on/off at 700 h and 1900 h and maintained at a controlled temperature (25°C). Animals were given food (LabDiet 5V75) and water ad libitum. Animals were weaned at postnatal day 21 (PD21). Sexually mature Tac2-EGFP females were used for all electrophysiology experiments. Genotype was determined by using PCR products of extracted DNA from ear clippings. Tac2-EGFP mice were generated by mating either positive EGFP males to positive EGFP females or positive EGFP males to WT EGFP females.

Surgical Procedure

To elucidate the interactions of E2 and the M-current, adult females were bilaterally ovariectomized (OVX) under isoflurane anesthesia using sterile no-touch technique according to the NIH Guidelines for Survival Rodent Surgery. Animals were given a dose of analgesic [4 mg/kg carprofen (Rimadyl®)] one day following surgery for pain management. Animals typically lost 1–2 grams of weight one day after surgery. Following OVX, females were separated into 2 treatment groups – oil and estradiol benzoate (EB) (n = 8–10 cells, 7–8 mice per group). An EB injection protocol was used that has previously been shown to alter gene expression in the hypothalamus and mimic a proestrus state (13, 38). Animals were injected subcutaneously (s.c.) at 1000 h on post-OVX day 5 with either 0.25 μg of EB or oil-vehicle. On post-OVX day 6, a 1.5 μg dose of EB or oil was injected at 1000 h. On post-OVX day 7, mice were rapidly decapitated, and the brain was prepared for electrophysiology.

Drugs

All drugs were purchased from Tocris unless otherwise specified. Tetrodotoxin (TTX, 1 mM stock), Ghrelin, a GHSR agonist, (100 mM stock) and the GHSR antagonist [D-Lys3]-GHRP-6 (100 mM stock) were all dissolved in ultrapure H₂O. The PKA inhibitor H-89 dihydrochloride (10 mM stock), the PKC inhibitor rottlerin (10 mM stock), the PLC inhibitor U73122 (20 mM stock), and 10, 10-bis (4-pyridinylmethyl)-9(10h)-anthracenone dihydrochloride (XE-991), a KCNQ channel blocker (40 mM stock) were all dissolved in DMSO. Aliquots of the stock solutions were stored at −20°C until needed. Estradiol benzoate (EB) was purchased from Steraloids (Newport, RI, USA) and dissolved in ethanol (1 mg/ml) prior to mixing in sesame oil (Sigma-Aldrich). Ethanol was allowed to evaporate off for 24h prior to storage. EB was stored at 4°C until needed.

Tissue Preparation

Animals were rapidly decapitated on day 7 following OVX. The brain was quickly removed from the skull, and a block containing the hypothalamus was immediately dissected and submerged in cold (4°C), oxygenated (95% O₂, 5% CO₂), high-sucrose artificial cerebrospinal fluid (aCSF) consisting of 208 mM sucrose, 2 mM KCl, 26 mM NaHCO₃, 10 mM glucose, 1.25 mM NaH₂PO₄, 2 mM MgSO₄, 1 mM MgCl₂, 2 mM CaCl₂, and 1 mM HEPES (pH 7.3; 300 mOsm). Coronal slices (250 μm) were cut on a vibratome at 4°C. The slices were transferred to an auxiliary chamber where they were kept at room temperature (25°C) in aCSF consisting of 124 mM NaCl, 5 mM KCl, 2.6 mM NaH₂PO₄, 2 mM MgCl₂, 2 mM CaCl₂, 26 mM NaHCO₃, and 10 mM glucose (pH 7.3; 310 mOsm) until recording (recovery for 1–2 h). A single slice was transferred to the recording chamber mounted on an Olympus BX51W1 upright microscope. The slice was continually perfused with warm (35°C), oxygenated aCSF at 1.5 mL/min and allowed to acclimate for 15 minutes prior to patching. Targeted neurons were viewed with an Olympus x40 water-immersion lens. Uteri were removed, and wet weight was recorded to confirm the effect of oil or EB treatment. Mean uterine weights from oil-treated females were 0.041 g ± 0.0013, while EB-treated females were 0.15 g ± 0.0043 (p< .0001).

Electrophysiology Recordings

Electrophysiology was performed as described previously (33, 39, 40). In hypothalamic slices, standard whole-cell patch-clamp recording procedures and pharmacological testing were used. Whole-cell patch-clamp recordings were performed using pipettes made of borosilicate glass and pulled using a vertical PC-10 Narishige Glass Micropipette Puller (Narishige Group, Japan). An Axopatch 700B amplifier, Digidata 1322A Data Acquisition System, and pCLAMP software (version 10.2; Molecular Devices) were used for data acquisition and analysis.

Current-voltage (I-V) plots were constructed by voltage steps from −50 to −140 mV at 10-mV increments applied at 1-second intervals from a holding potential of −60 mV. The input resistance (Rin) was determined from the slope of each I-V plot in the range between −60 and −80 mV. Input resistance, series resistance, and membrane capacitance were monitored throughout the experiments. Only cells with stable series resistance (<30 MΩ; <20% change over the course of the recording) and input resistance (>500 MΩ) were used for analysis. Low-pass filtering of the currents was conducted at a frequency of 2 kHZ. The liquid junction potential was calculated to be −10 mV and corrected for during data analysis using pClamp software. All voltage-clamp recordings were performed from a holding potential of −60 mV.

To record M-currents, pipettes (3- to 5 MΩ resistance) were filled with an internal solution containing 10 mM NaCl, 128 mM K-gluconate, 1 mM MgCl, 10 mM HEPES, 1 mM ATP, 1.1 mM EGTA, and 0.25 mM GTP (pH 7.3; 300 mOsm). In voltage-clamp, a standard deactivation protocol (33, 39) was used to measure potassium currents elicited during 500-millisecond voltage steps from −30 to −75 mV in 5-mV increments after a 300-millisecond prepulse to −20 mV. The amplitude of the M-current relaxation or deactivation was measured as the difference between initial (<10 ms) and sustained current (>475 ms) of the current trace in the control conditions (1 μM TTX; 5 min). After baseline recording (~5 min), XE-991 (40 μM with 1 μM TTX) or ghrelin (100 nM with 1 μM TTX) was perfused for 10 minutes, and the protocol was repeated. Deactivation protocol was repeated twice for each perfusion condition and averaged for analysis.

The ghrelin dose response was recorded in current- and voltage-clamp, where continuous recording was monitored ~2 min before perfusion to establish a baseline. Ghrelin was perfused at a range of concentrations (0.1 to 1000 nM; 3 concentrations per neuron for 5 minutes each in the presence of 1 μM TTX) to determine if there is a change in sensitivity to ghrelin (cells depolarize at lower concentrations of ghrelin). Miniature Excitatory Post-Synaptic Current (mEPSC) frequency and amplitude were analyzed from voltage-clamp recordings using pCLAMP software (version 10.2; Molecular Devices). Briefly, using event detection, for each voltage-clamp trace a mEPSC template was chosen. Once a template was chosen, event detection was utilized to select mEPSCs and each individual mEPSC was manually accepted or rejected. At the end of the trace, the total number of mEPSCs were used to determine frequency. To determine amplitude, mEPSCs were averaged and the resulting trace was measured for amplitude. This process was repeated for each concentration of ghrelin.

To determine the second messenger pathway, the same M-current protocol was utilized where control conditions were 1 μM TTX + inhibitor for 5 minutes followed by 1 μM TTX + inhibitor + 100 nM ghrelin for 10 minutes. For this experiment, all Tac2-EGFP females were gonadally intact and in diestrus, determined by vaginal cytology the morning of the experiment. All uteri were dissected and weighed for confirmation.

Statistical analysis

Comparisons of the I-V plots between control and subsequent drug conditions were performed at each voltage (−30 to −75 mV) using a repeated-measures two-way ANOVA with post hoc Bonferroni’s multiple comparison test. Maximum current at −35 mV was analyzed with paired Student’s t test. All statistical analysis was conducted using GraphPad Prism (GraphPad). Data were considered significant when P < .05. All data values were presented as mean ± SEM. Each n represents the number of cells examined.

Results

Rundown of the M-current

Because the KCNQ channel current (M-current) in KNDy neurons has not been previously studied, we initially established M-current activity using the selective KCNQ channel blocker XE-991, 40 μM a dose based on previous investigation (32, 40). To establish efficacy, we performed current-clamp recordings from Tac2-EGFP neurons and applied XE-991 to the bath solution. This resulted in an average depolarization of 3.5 mV ± 0.86 (n=4 cells), confirming that XE-991 was able to block KCNQ channels in these neurons (Figure 1A), the average RMP was −66.75 mV ± 1.31. To explore whether ghrelin would do the same, we repeated the current-clamp recordings and applied ghrelin (100 nM) to the bath solution. This resulted in an average depolarization of 10.33 mV ± 4.63 (n=3 cells), confirming that ghrelin does depolarize these neurons potentially through blocking KCNQ channels (Figure 1B), the average RMP was −72.33 mV ± 2.03. To evaluate the M-current in voltage-clamp, we measured the deactivation or relaxation of the whole-cell K⁺ currents elicited by an established protocol (41, 42). M-current was calculated by subtracting the current relaxation (the difference between the instantaneous and steady state; arrows) (Figure 1C) during continuous extracellular perfusion of TTX (1 μM). Because the M-current has been previously reported to show a decrease or “rundown” after cell dialysis in whole-cell recordings in some neuronal cell types, but not all (32, 43), the whole-cell K⁺ currents evoked by the deactivation protocol were monitored over a period of 30 minutes to determine the change in the relaxation currents over time. In Tac2 neurons from oil-treated, OVX females, the outward K⁺ currents evoked decreased by approximately 50% over a period of 20 minutes (p< 0.01; Figure 1D). Therefore, all subsequent recordings were restricted < 20 minutes to limit the rundown and ensure that any changes observed were due to drug application (32).

Figure 1.

(A) Representative current-clamp trace of the modest depolarization caused by 40 XE-991 in an Oil-treated OVX female. (B) Representative trace of the depolarization caused by 100 nM Ghrelin in an intact, diestrus female. (C) M-current deactivation protocol. (D) I-V plot from −75 to −30 mV after 1 μM TTX for 0, 10, 20 and 30 minutes. After a time period of 20–30 minutes, the M-current ran down by ~50%. All subsequent data was collected in under 20 minutes. Data were analyzed by two-way ANOVA and post hoc Bonferroni’s multiple comparison tests. (b=P<.01; c=P<.001; d=P<.0001). Dark blue letters indicate comparison between 10 and 30 minutes, light blue letters indicate a comparison between 10 and 20 minutes. n=7 cells.

Regulation of M-current activity by E2

To determine if estrogen replacement alters M-current activity in KNDy neurons, we examined the effects of XE-991 in Tac2 neurons from oil- and EB-treated OVX Tac2-GFP females. In the presence of TTX (1 μM) to block Na⁺-spike-dependent synaptic input, XE-991 (40 μM) suppressed evoked currents in both oil- and EB-treated OVX females (Figure 2A–B). XE-991 application resulted in an inhibition of the M-current in the range of −60 to −30 mV in both oil- and EB-treated females (p<0.05; Figure 2C–D). However, there was no difference between steroid conditions in the amount of XE-991-sensitive current (Figure 2E). The change in maximum peak current at −35 mV after XE-991 perfusion was 75.7 ± 13.8 pA (p<0.001, n=10 cells, Figure 2F) in oil-treated females and 68.0 ± 14.7 pA (p<0.001, n=11 cells, Figure 2F) in EB-treated females. In addition, steroid condition did not alter the mean or change in Resting Membrane Potential (RMP) or Rin due to XE-991 perfusion (Figure 2G–H). The RMP mean for oil-treated females before XE-991 was −56.4 mV ± 6.2, after XE-991 was −40.2 mV ± 7.41. The average change in RMP after XE-991 was 16.2 mV ± 5.9. The RMP mean for EB-treated females before XE-991 was −54.82 mV ± 5.04, after XE-991 was −46.82 mV ± 5.39. The average change in RMP after XE-991 was 8.0 mV ± 3.43. The Rin mean for oil-treated females before XE-991 was 0.76 GΩ ± 0.16, after XE-991 was 0.46 GΩ ± 0.08. The average change in Rin after XE-991 was a decrease of 0.31 GΩ ± 0.13. The Rin mean for EB-treated females before XE-991 was 0.93 GΩ ± 0.25, after XE-991 was 0.62 GΩ ± 0.16. The average change in Rin after XE-991 was a decrease of 0.37 GΩ ± 0.31. RMP and Rin values are summarized in Table 1.

Figure 2.

Representative current traces of the M-current inhibition caused by 40 μM XE-991 in (A) Oil- and (B) EB-treated OVX female mice. I-V plot from −75 to −30 mV after XE-991 perfusion in (C) Oil- and (D) EB-treated OVX female mice. (E) I-V plot from −75 to −30 mV of the XE-991-sensitive M-current in Oil- and EB-treated OVX female mice.(F) XE-991 reduced the maximum peak current in Oil- and EB-treated mice. (G–H) The mean Resting Membrane Potential (RMP) and Input Resistance (Rin). (C–F) Data were analyzed by two-way ANOVA with Bonferroni’s multiple comparison tests. (G–H) Data were analyzed by unpaired t-test (a=P<.05; b=P<.01; c=P<.001; d=P<.0001). Sample sizes for C-H were Oil n=10 cells and EB n=11 cells.

Table 1.

Resting membrane potentials (RMP) and input resistance (Rin) from Figure 2 and 3 in both Control (before) and after either XE-991 (40 μM) or Ghrelin (100nM) perfusion conditions.

		Figure 2: Control	XE-991(40μM)	Figure 3: Control	Ghrelin (100nM)
Oil	RMP(mV)	−56.4 ± 6.2	−40.2 ± 7.41	−58.22 ± 4.7	−47.67 ± 5.69
	Rin (GΩ)	0.76 ± 0.16	0.46 ± 0.08	1.42 ± 0.39	1.70 ± 0.73
E2	RMP (mV)	−54.82 ± 5.04	−46.82 ± 5.39	−52 ± 4.11	−46.33 ± 4.72
	Rin (GΩ)	0.93 ± 0.25	0.37 ± 0.31	1.57 ± 0.67	0.82 ± 0.15

Suppression of the M-current by ghrelin

Our previous data in arcuate NPY neurons indicates that the M-current is a target for the peptide hormone ghrelin. To ascertain if ghrelin also suppresses the M-current in Tac2 (KNDy) neurons, ghrelin (100 nM) was perfused for 10 minutes after an initial (control) deactivation protocol. Ghrelin inhibited the M-current in EB-treated, OVX females with a more robust effect compared to oil-treated, OVX females (Figure 3A–D, EB: (F(1, 16)=9.690, P<0.0067)), although steroid condition did not change the magnitude of M-current inhibition by ghrelin (Figure 3E). The change in maximum peak current at −35 mV after ghrelin perfusion was 58.5 ± 15.8 pA (p<0.05, n=9 cells, Figure 3F) in oil-treated females and 59.2 ± 11.9 pA (p<0.01, n=9 cells, Figure 3F). In addition, steroid condition did not alter the ghrelin-induced change in RMP or Rin (Figure 3G–H). The RMP mean for oil-treated females before ghrelin was −58.22 mV ± 4.7, after ghrelin was −47.67 mV ± 5.69. The average change in RMP after ghrelin was 10.56 mV ± 3.17. The RMP mean for EB-treated females before ghrelin was −52 mV ± 4.11, after ghrelin was −46.33 mV ± 4.72. The average change in RMP after ghrelin was 5.67 mV ± 1.82. The Rin mean for oil-treated females before ghrelin was 1.42 GΩ ± 0.39, after ghrelin was 1.70 GΩ ± 0.73. The average change in Rin after ghrelin was an increase of 0.25 GΩ ± 0.77. The Rin mean for EB-treated females before ghrelin was 1.57 GΩ ± 0.67, after ghrelin was 0.82 GΩ ± 0.15. The average change in Rin after ghrelin was a decrease of 0.75 GΩ ± 0.7. RMP and Rin values are summarized in Table 1.

Figure 3.

Representative current traces of the M-current inhibition caused by 100 nM ghrelin in (A) Oil- and (B) EB-treated OVX female mice. I-V plot from −75 to −30 mV after ghrelin perfusion in (C) Oil- and (D) EB-treated OVX female mice. (E) I-V plot from −75 to −30 mV of the ghrelin-sensitive M-current in Oil- and EB-treated OVX female mice.(F) Ghrelin reduced the maximum peak current in Oil- and EB-treated mice. (G–H) The mean Resting Membrane Potential (RMP) and Input Resistance (Rin). (C–F) Data were analyzed by two-way ANOVA with Bonferroni’s multiple comparison tests. (G–H) Data were analyzed by unpaired t-test (a=P<.05; b=P<.01; c=P<.001; d=P<.0001). Sample sizes for (C–H) were Oil n=9 cells and EB n=9 cells.

Ghrelin dose response and Tac2 neuronal activity

Because of our previous findings demonstrating that E2 increases GHSR expression in KNDy neurons (30), we hypothesize that E2 increases KNDy neuronal sensitivity to ghrelin. As we have established that ghrelin can inhibit the M-current more robustly with EB-treatment, we perfused ghrelin at increasing doses (0.01–1000 nM) and monitored cell excitability and depolarization in current- and voltage-clamp, respectively. During current-clamp recordings, Tac2 neurons from EB-treated, OVX females responded to ghrelin at much lower concentrations eliciting action potential firing at lower doses compared to Tac2 neurons from oil-treated females (Figure 4 A–B). The average RMP for all cells before drug application was −65mV ± 6.2mV. However, not all neurons responded to ghrelin in the same way. In voltage-clamp, cells exhibited an inward current, an outward current, or did not respond at all (Figure 4C). Cells were considered responsive if they showed more than an 8% change in current from the maximum current observed. On average the inward currents for 100 nM ghrelin, the only dose where a difference was observed, in oil-treated females was −4.5 pA ± 0.88, while EB-treated females were −11.43 pA ± 2.8 (p=0.0503). Because there was not an outward current for every dose of ghrelin, the average outward currents across all ghrelin doses from oil-treated females were 4.2 pA ± 1.62, while EB-treated females were 7.86 pA ± 2.19. In oil-treated females, 54 cells responded to ghrelin with an inward current, 5 responded with an outward current, and 14 had no response to ghrelin. In EB-treated females, 66 cells responded to ghrelin with an inward current, 7 responded with an outward current, and 10 had no response to ghrelin. Amongst the cells that responded with an inward current, ghrelin perfusion (100 nM) elicited a more robust current in EB-treated females compared to oil-treated females (p<0.05; Figure 4D). Interestingly, perfusion of 1000 nM ghrelin in EB-treated females did not elicit a stronger current, potentially due to desensitization of GHSR (44). During our voltage-clamp recordings, (Figure 4E–F) we observed a potential effect of steroid on miniature excitatory postsynaptic currents (mEPSC). Analysis of these traces revealed a reduction of mEPSC frequency, but not amplitude, in EB-treated females compared to oil-treated (p<0.05; Figure 4 G–H), indicating an interaction of E2 on presynaptic glutamate release.

Figure 4.

Representative current-clamp traces of an estrogen-dependent increase in neuronal sensitivity to increasing doses of ghrelin (0.01 nM to 100 nM) in (A) Oil- and (B) EB-treated OVX female mice. Average RMP for all cells before ghrelin application was −65mV ± 6.2mV (C) Table indicating the number of cells from voltage-clamp experiments that responded with either an inward current, outward current or no response. (D) Ghrelin dose response curve. Representative voltage-clamp traces of an estrogen-dependent decrease in mEPSC frequency in (E) Oil- and (F) EB-treated OVX female mice. mEPSC frequency (G) and amplitude (H) before (baseline) and after 100 nM ghrelin. Data were analyzed by two-way ANOVA with Holm-Sidak’s multiple comparison tests. (G-H) were analyzed by unpaired t-test (a=P<.05). Sample sizes are indicated in table (C).

Ghrelin signals through a PLC-mediated pathway

Previous publications have established KCNQ channels are negatively modulated by Gq-coupled G-protein coupled receptors (GPCR), including GHSR (27). The binding of ghrelin to GHSR activates a phospholipase C (PLC) signaling cascade that hydrolyzes phosphatidylinositol 4,5-bisphosphate (PIP₂) for inositol triphosphate (IP₃) and diacylglycerol (DAG). IP3 is released into the cytosol and increases release of intracellular calcium from internal stores. DAG activates protein kinase C (PKC) which then either directly decreases KCNQ channel (M-current) conductance (27, 31) or activates PKA (40). To confirm the role of this pathway (PLC-PKC) in attenuating the M-current in Tac2 neurons, we applied a series of specific pharmacological interventions. First, to confirm the role of GHSR, we co-perfused a specific GHSR antagonist [D-Lys³]-GHRP-6 (50 μM) with ghrelin (100 nM) and TTX (1 μM) (Figure 5A). The GHSR antagonist eliminated the effect of ghrelin on the M-current, suggesting the M-current inhibition by ghrelin is indeed mediated by GHSR. Next, we inhibited PLC with U73122 (10 μM) and the PKC inhibitor, rottlerin (5 μM) in separate experiments (Figure 5 B–C). Both U73122 and rottlerin blocked the actions of ghrelin, confirming previous reports. GHSR signaling also may activate PKA as a second messenger (45), which blocks Gq-GPCR inhibition of the M-current in hypothalamic neurons (40). Using H89, the selective PKA inhibitor (10 μM), we observed a reduction of the ghrelin-induced suppression of the M-current (Figure 5D), indicating that PKA may play a role in Tac2 neurons.

Figure 5.

I-V plots from −75 to −30 mV before (black) and after (green) ghrelin [100 nM] perfusion in diestrus Tac2 female mice with GHSR antagonist, [D-Lys³]-GHRP−6 [50 μM] n=5 cells (A), PLC inhibitor, U73122 [10 μM] n=6 cells (B), PKC inhibitor, Rottlerin [5 μM] n=6 cells (C), and PKA inhibitor H89 [10 μM] n=7 cells (D). Data were analyzed by two-way ANOVA with Bonferroni’s multiple comparison tests.

Discussion

The results of the present study confirm previous findings that 17β-estradiol upregulates Ghsr expression (30), increases KNDy neuronal sensitivity to ghrelin, and subsequently enhances the ghrelin-induced inhibition of the M-current via a PLC-PKA pathway in ARC KNDy neurons. These conclusions are based on the following observations: 1) 100 nM Ghrelin was effective in inhibiting the M-current similar to XE-991, the selective KCNQ channel inhibitor; 2) ghrelin was able to elicit action potentials at a lower dose in EB-treated females; 3) there was a more robust inward current in response to ghrelin in EB-treated females compared to oil-treated controls; and 4) blocking the PLC-PKA pathway blocked the M-current inhibition by ghrelin.

The 17β-estradiol-induced increase in KNDy sensitivity to ghrelin is consistent with previous findings (30), and confirmed with our electrophysiological assessment. Additionally, we have confirmed and corroborated that ghrelin binds to GHSR, a Gq-coupled GPCR, activating PLC to hydrolyze PIP2 for IP3 and DAG. DAG in turn activates PKC which decreases KCNQ channel conductance (M-current inhibition). This results in a more depolarized and excitable cell (27, 31). Ghrelin inhibition of KCNQ channels in KNDy neurons is similar to our previous research on the M-current in other ARC neurons (32, 33, 46). This is further supported by the identification of the second messenger pathway (PLC-PKC) in which ghrelin has previously been described to inhibit the M-current (27).

The current study is also an initial characterization of the M-current in KNDy neurons as XE-991 substantially blocked the current and depolarized all KNDy neurons examined. Interestingly, we found that the M-current in ARC KNDy neurons rundown in less than 20 min, which has not been observed in other ARC neurons, like NPY or POMC (32). This could be due to cell dialysis during whole-cell patch clamp recordings over time (43) or differences in intracellular signaling that impinges on KCNQ activity (43, 47, 48). Because the M-current is a common target for Gq-coupled GPCR such as 5HT2c serotoninergic and M_1/5 muscarinic receptors in many types of hypothalamic neurons (33), robust M-current activity in KNDy neurons may be a key, although not the only, mechanism for modulation of KNDy activity.

We have also observed that with EB treatment, lower doses of ghrelin induced action potentials (AP) and at higher doses of ghrelin exhibited reduced APs, possibly due to over-excitation or internalization of the receptor, a common mechanism for Gq-coupled GPCRs (44). The EB-induced effects on ghrelin sensitivity is clearly indicated at the 100 nM dose of ghrelin where an increase in the percent maximum inward current is augmented in EB-treated females. Novaria et al., (2014) found that not all Tac2 neurons are active and some are quiescent (49). Furthermore, not all Tac2 neurons express detectable levels of GHSR (30). Thus, a significant subpopulation of KNDy neurons respond to ghrelin and in sufficient numbers to elicit known physiological control of the HPG axis (36, 50).

Lastly, we observed that EB-treatment reduces mEPSC frequency, but not the amplitude, indicating that there is a decrease in the probability of release from presynaptic neurons. This could be due to a reduction in glutamate signaling from neurons that project to KNDy neurons, including other KNDy neurons. Previous studies have demonstrated that ARC KNDy neurons exhibited increased glutamatergic transmission when ERα was knocked down in these neurons (51), which supports our study with proestrous levels of E2. Furthermore, it is critical to consider that high E2 inhibits neuropeptide expression in these neurons (30), and EB treatment can increase negative feedback via dynorphin in these neurons, ultimately reducing excitability. It is important to note that we only saw a reduction in mEPSC with the 100 nM dose. As mEPSC amplitude did not change, we expect that there is no change in the expression of NMDA or AMPA receptors in KNDy neurons due to E2 (52).

While the effects of E2 and ghrelin in ARC KNDy neurons have become more defined, it is critical to keep in mind that KNDy neurons are one node of a large circuit within the ARC. KNDy neurons not only project locally to POMC and NPY neurons, but also to preoptic GnRH and kisspeptin neurons of the AVPV, playing an important role in puberty and ovarian functions (37, 49). KNDy neurons activate POMC neurons via kisspeptin and glutamate release and inhibit NPY neurons directly via glutamate and indirectly through kisspeptin (16, 37). As all three neurons modulate the excitability of the other two neuronal populations and respond differently to peripheral peptide hormones, any modulation of one neuronal population by a hormone will have disparate effects on the other two populations. Therefore, we hypothesize that ghrelin’s excitation of KNDy neurons and it’s potentiation by E2 would reduce ghrelin-induced food intake via NPY neurons in female mice due to the excitation of POMC and the inhibition of NPY. Our hypothesis is supported by previous data demonstrating that E2 suppresses ghrelin-induced food intake in female rats (53). This pathway may be sensitive to energy states, both positive (obesity) and negative (fasting/caloric restriction), as it is known that obesity can cause ghrelin resistance in NPY/AgRP neurons (54) and a high fat diet can reduce the amount of plasma ghrelin in females (26). We are currently exploring the effects of diet-induced obesity on ghrelin’s actions in KNDy neurons.

Furthermore, as ghrelin administration suppresses LH pulsatility through β-endorphin produced by POMC neurons in fed female rats (29, 55), ghrelin activation of KNDy neurons may be involved, as GHSR is not highly expressed in POMC neurons (23). Indeed, KNDy neurons can excite POMC neurons via glutamate during states of elevated E2 (37), which may lead to greater β-endorphin release onto GnRH terminals, inhibiting LH pulsatility. Hence, during states of elevated ghrelin (fasting, food scarcity, etc.), the HPG axis would be suppressed diverting motivation towards food intake and away from reproduction. Because of this, the role of ghrelin signaling in KNDy neurons may prove to be relevant in the control of reproduction, energy homeostasis, and other physiological processes controlled by KNDy neurons. Future experiments will explore if androgens or E2 in males also regulate Ghsr expression in KNDy neurons and if ghrelin activation of KNDy neurons alters the downstream modulation of POMC or NPY neurons utilizing additional transgenic mouse models.

In conclusion, during the normal reproductive cycle, E2 fluctuates to control the growth of ovarian follicles via negative feedback regulated by ARC KNDy neurons. We have shown that when E2 is high (proestrus or OVX + E2), Ghsr expression is increased in KNDy neurons (30) to augment their sensitivity to states of elevated ghrelin (starvation/caloric restriction) to suppress the HPG axis. Ghrelin, in turn, suppressed the M-current with greater potency leading to depolarization and neurotransmitter (glutamate) release to potentially suppress the HPG axis during periods of food scarcity in females.