Previously, we have shown that an increase in the intracellular calcium concentration can potentiate TRPC5 channel current in POMC neurons (Qiu J. et al., 2010). Additionally, chelating intracellular calcium with BAPTA abolishes the slow excitatory postsynaptic potential (EPSP) and persistent firing in Kiss1^ARH neurons (Qiu Jian et al., 2021). Here, to investigate the contributions of voltage-activated calcium channels (VGCCs) to the increase in intracellular calcium, we measured the peak calcium current contributed by both the low and high voltage-activated calcium channels. To assess VGCC activity, we employed 150-ms test pulses starting from a holding potential of -80 mV with 10-mV increments, ultimately reaching a test potential of +40 mV in Kiss1^ARH neurons. The inward currents evoked by the voltage pulses were identified as calcium (Ca²⁺) currents, as they were blocked by the universal VGCC inhibitor Cd²⁺ (200 μM) (McNally et al., 2020) (Figures 2A-E). The maximum total inward and Cd²⁺-sensitive currents reached their peak amplitudes at -10 mV. To differentiate between various calcium channel subtypes present in Kiss1^ARH neurons, we applied selective antagonists individually, allowing us to isolate the drug-sensitive current for each cell. When we individually applied specific antagonists, we observed partial inhibition of the Ca²⁺ currents. Treatment with 10 μM nifedipine (Hiraizumi et al., 2008; Kato et al., 2003; Lee et al., 2002), an L-type Ca²⁺ channel inhibitor, resulted in a partial inhibition (26.0%) of the whole-cell calcium current. Additionally, application of 1 μM ω-conotoxin MVIIC (ConoMVIIC, targeting N/P/Q-type channels) (Nunemaker et al., 2003), 2 μM ω-conotoxin GVIA (conoGVIA, targeting N-type channels) (Lee et al., 2002), 200 nM ω-agatoxin IVA (AgaIVA, targeting P/Q-type channels) (Kato et al., 2003; McNally et al., 2020), 100 nM SNX-482 (targeting R-type channels) (Hiraizumi et al., 2008), or 1 μM TTA-P2 (TTAP2, targeting T-type channels) (McNally et al., 2020) also led to partial inhibition of the Ca²⁺ currents (Figure 2F). The observed reduction in current with each inhibitor indicates the presence of all the major subtypes of Ca²⁺ currents in Kiss1^ARH neurons. Among the inhibitors used, the largest components of the whole-cell calcium current was found to be sensitive to nifedipine (26.1%), conoGVIA (25.1%), and SNX-482 (31.1%) (Figures 2A, B, D and 2F). Subsequently, we documented contributions from TTA-P2-sensitive channels, accounting for approximately 6.7% of the total Ca²⁺ current, and AgaIVA-sensitive channels, which constituted approximately 3.9% (Figures 2E, C). These findings indicate that high voltage-activated L-, N-, and R-type channels constitute the largest components of the voltage-activated Ca²⁺ current in Kiss1^ARH neurons that not only increase the overall excitability but also greatly facilitate TRPC5 channel opening (Blair et al., 2009), which is the major downstream target of TACR3 activation by NKB (Qiu J. et al., 2016).

Our previous study utilizing optogenetics demonstrated that high-frequency photostimulation of Kiss1^ARH neurons releases NKB. This release of NKB induces slow excitatory postsynaptic potentials (EPSPs) and facilitates the recruitment of other Kiss1^ARH neurons, resulting in synchronous firing of the Kiss1^ARH neuronal population (Qiu J. et al., 2016). Additionally, chelating intracellular calcium with the fast chelator BAPTA abolishes the slow EPSP and persistent firing in Kiss1^ARH neurons, highlighting the role of calcium signaling in these processes (Qiu Jian et al., 2021). To access the involvement of HVA channels in the generation of the slow EPSP, we conducted experiments where we blocked the L-type Ca²⁺ channels with nifedipine (10 µM) and the N- and P/Q-type Ca²⁺ channels with ω-conotoxin MVIIC (1 µM). We then measured the slow EPSP. Indeed, both nifedipine and ω-conotoxin MVIIC significantly inhibited the slow EPSP by 42.9% and 60.4%, respectively (Figure 3). In addition, we used SNX (100 nM), which selectively blocks R-type Ca²⁺ channels, and the slow EPSP was reduced to 28.7% of its control value (Figure 3C). The selective T-channel blocker TTA-P2 (5 µM) inhibited the slow EPSP by 28.6% (data not shown). Therefore, it appears that all of the calcium channels contribute to maintaining the sustained depolarization underlying the slow EPSP.

The neuropeptides NKB (tachykinin2, Tac2) and kisspeptin (Kiss1), which are expressed in Kiss1^ARH neurons, are crucial for the pulsatile release of gonadotropin-releasing hormone (GnRH) and reproductive processes. E2 decreases the expression of Kiss1 and Tac2 mRNA in Kiss1^ARH neurons but enhances the excitability of Kiss1^ARH neurons by amplifying the expression of Cacna1g, Hcn1, and Hcn2 mRNA, as well as increasing T-type calcium currents and h-currents (Qiu, 2018, 20613}. Moreover, E2 drives Slc17a6 mRNA expression and enhances glutamatergic synaptic input to arcuate neurons and Kiss1^AVPV neurons (Qiu J. et al., 2018). As a result, the E2-driven increase in Kiss1^ARH neuronal excitability and glutamate neurotransmission may play a crucial role in triggering the surge of GnRH, ultimately leading to the LH surge.

To assess the impact of E2 on the modulation of voltage-activated calcium channels and Kiss1^ARH neuronal excitability, we employed real-time PCR (qPCR) to measure the relative expression levels of ion channel subtypes in Kiss1^ARH neurons. We compared the expression in E2-treated females to those treated with oil, using the specific primers listed in Table 1. The quantification was conducted on pools of 5 or 10 neurons, as indicated in the Methods. In both oil- and E2-treated females, we quantified the expression of Cav1.2 (L-type), Cav2.1 (P/Q-type), Cav2.2 (N-type) and Cav2.3 (R-type) mRNAs. Remarkably, all of these mRNA transcripts exhibited increased expression levels in response to E2 treatment (Figure 4A and B). Congruent with our previous findings (Qiu J. et al., 2018), we observed that the mRNA expression of Cav3.1, Hcn1, and Hcn2 was also upregulated in response to E2 treatment (Figure 4C). These results suggested that E2 has a regulatory effect on the expression and function of all of these ion channels in Kiss1^ARH neurons. To determine whether the increased mRNA expression translated into functional changes at the cellular level, we measured the whole-cell calcium current in Kiss1^ARH neurons obtained from ovariectomized mice treated with either vehicle or E2. We discovered that E2 treatment led to a significant increase in the peak calcium current density in Kiss1^ARH neurons, which was recapitulated as predicted by our computational modeling (Figure 5A-D). These findings indicate that the upregulation of the mRNA expression of calcium channels by E2 translated to an augmented peak calcium current in Kiss1^ARH neurons.

The largest components of the calcium currents in Kiss1^ARH neurons from the E2-treated, ovariectomized females were found to be sensitive to nifedipine, conoGVIA, and SNX-482, accounting for approximately 24.9%, 24.6%, and 27.0% of the total current across cells, respectively, which is very similar to their contributions to the whole-cell current from vehicle-treated, ovariectomized females (Figure 3). In addition, contributions from TTA-P2-sensitive channels accounted for approximately 11.1% of the total Ca²⁺ current, while agaIVA-sensitive channels contributed to approximately 11.0% (Figure 5E). These results highlight the prevalence of L-, N-, and R-type calcium channels as the major contributors to the whole-cell calcium current in Kiss1^ARH neurons from E2-treated, ovariectomized females, but also the involvement of T-type and P/Q-type channels.

In order to determine whether the increase in peak voltage-activated calcium current density was the result of E2 regulating calcium channel kinetics, mRNA expression or both, we examined the voltage dependence of activation and inactivation. By measuring the voltage dependence of activation, we assessed how E2 affects the ability of calcium channels to open in response to membrane potential changes. Similarly, by examining the voltage dependence of inactivation, we determined how E2 influences the inactivation kinetics of calcium channels. Based on our results, there was no difference in the voltage dependence of activation between cells from the vehicle-treated control group (V_1/2 = -32.3 ± 2.1 mV; n = 13) and cells from estrogen-treated females (V_1/2 = -33.6 ± 2.5 mV; n = 11). Similarly, there was not a significant difference in the voltage dependence of inactivation between control cells (V_1/2 = -48.9 ± 4.8 mV; n = 6) and estrogen-treated cells (V_1/2 = -44.1 ± 1.9 mV; n = 5) (Figure 6). Therefore, although E2 increased the mRNA expression of HVA calcium channels, it did not affect the channel kinetics in Kiss1^ARH neurons. Furthermore, our previous studies established that there is no difference in the voltage dependence of activation and inactivation of T-type calcium channels in hypothalamic arcuate neurons between the vehicle-treated and E2-treated, ovariectomized females (Qiu J. et al., 2006). Also, E2 downregulated the expression of Kcnd2 mRNA encoding Kv4.2, which is expressed in Kiss1^ARH neurons (Mendonça et al., 2018) and has similar kinetics of activation as the T-type calcium channels (Oil-treated, ovariectomized females relative mRNA expression: 1.053, n = 5 animals versus E2-treated expression: 0.5643, n=5; t-test p = 0.0061). This opposing K⁺ current would dampen the inward calcium current. Therefore, it appears that E2 does not modulate calcium channel kinetics directly but rather alters the mRNA expression to increase the conductance.

In the brain, calcium plays a crucial role in sculpting neuronal firing by activating potassium channels, which subsequently influence neuronal behavior (Nicoll, 1988; Storm, 1990). Since HVA and LVA calcium channels were expressed in Kiss1^ARH neurons, all of which contribute to the elevation of intracellular calcium concentration ([Ca²⁺]_i) that facilitates TRPC5 channel opening (Blair et al., 2009), our next step involved measuring the changes in Ca²⁺-activated K⁺ channel conductances and assessing their mRNA expression. In various cell types increases in cytosolic calcium levels, whether resulting from extracellular influx or intracellular release, lead to the activation of plasma membrane calcium-dependent potassium channels (Sah Pankaj and Louise Faber, 2002). Similarly, in Kiss1^ARH neurons, these channels would be activated by calcium influx through all four types of high voltage-gated calcium channels, as well as the low voltage-activated calcium channel, which are all active during action potential firing. The activity of Ca²⁺-activated K⁺ channels play a crucial role in numerous physiological processes, including secretion and the regulation of neuronal firing properties. Two main families of Ca²⁺-activated K⁺ channel channels have been characterized, distinguished by their biophysical and pharmacological properties. These families are known as BK (Big Conductance K⁺) and SK (Small Conductance K⁺) channels in the CNS (Kshatri et al., 2018). BK channels are known for their high potassium selectivity and large single channel conductance, typically ranging from 100 to 300 pS. Activation of BK channels requires both calcium binding and membrane depolarization (Blatz and Magleby, 1987; Marty, 1989; Sah P., 1996; Storm, 1990). On the other hand, SK channels are simply activated by increases in cytosolic calcium levels, with their half-maximal activation at 0.3 µM (Bond et al., 1999).

To investigate K⁺ currents, the cells were maintained at a holding potential of -70 mV while being exposed to blockers CNQX, AP5, picrotoxin and TTX. Subsequently, the membrane potential was stepped by depolarizing voltages, ranging from -60 mV to +40 mV in 10 mV increments, for a duration of 500 ms (Brereton et al., 2013). This protocol was employed to activate K⁺ currents (Figure 7A). First, we examined SK currents. The mean current density was determined at the end of the voltage pulses. In vehicle-treated, OVX females the application of the SK channel blocker apamin (100 nM) (Spergel, 2007) led to a significant reduction in whole-cell currents in the +20 to +40 mV range (Figures 7A, B). The mean outward current density at +40 mV in the control group was 125.5 ± 13.1 pA/pF (n = 3) with the apamin-sensitive component contributing 56.2 ± 2.3 pA/pF (n = 3) (Figures 7A, C). In contrast, in the E2-treated females, the overall mean outward current density at +40 mV was 191.8 ± 17.4 pA/pF (n = 5), which was significantly greater than the vehicle control group (Figures 7D, E). However, there was no significant difference in the apamin-sensitive component between the vehicle-treated and E2-treated females, 56.2 ± 2.3 pA/pF versus 63.8 ± 5.5 pA/pF (n = 4), respectively (Figure 7F). Our computational model was calibrated so that SK channels contributed ∼50 pA/pF to the whole-cell outward K⁺ current in E2-treated females (Figure 7H).

Furthermore, to investigate the expression of the mRNAs encoding SK channel subunits in Kiss1^ARH neurons from vehicle-treated and E2-treated OVX females, qPCR experiments were performed on 10-cell Kiss1^ARH neuronal pools (Figure 7G). We focused on SK3 channels because these channels exhibit the highest expression in the hypothalamus and E2 regulates their expression (Bosch et al., 2002). E2 treatment had no effect on the mRNA expression of the SK3 subunit. These findings support our electrophysiology results.

Additionally, following the same protocol and in the presence of the same cocktail of blockers (CNQX, AP5, picrotoxin and TTX), we investigated the contribution of BK channels to Kiss1^ARH neuronal excitability. In the OVX females, the application of the BK channel blocker iberiotoxin (ibTx, 200 nM) (Niday and Bean, 2021) resulted in only a slight attenuation of the outward current (n = 5) (Figures 8A, B). The ibTx-sensitive current density measured at +40 mV was 31.1 ± 8.4 pA/pF (Figures 8A, C). However, in the E2-treated females, the application of ibTx significantly attenuated the whole-cell K⁺ current from +30 to +40 mV, (Figures 8D, E). Additionally, the ibTx-sensitive current was significantly larger in the +0 to +40 mV range in the E2-treated females compared to the OVX females (Figure 8F). These findings indicate that E2 treatment modulates the activity of ibTx-sensitive BK current in Kiss1^ARH neurons, resulting in increased current density (100.9 ± 11.7 pA/pF versus 31.1 ± 8.4 pA/pF at +40 mV). Hence our computational model was calibrated so that BK channels contributed ∼100 pA/pF to the whole cell outward K⁺ current in the E2-treated females (Figures 8H).

To investigate the expression of mRNA encoding BK channel subunits in Kiss1^ARH neurons from vehicle-treated and E2-treated OVX females, qPCR experiments were performed on 10-cell Kiss1^ARH neuronal pools (Figure 8G). E2 treatment significantly increased the mRNA expression of the BKα1 (Kcnma1) subunit. These findings support our electrophysiological findings that there is a significant increase in BK channel activity in Kiss1^ARH neurons with E2 treatment (Figure 8F). In addition, E2 increased the mRNA expression of Kcnb1 encoding Kv2.1 (E2-treated relative mRNA expression: 1.672, n = 5 versus oil-treated mRNA expression: 1.086, n = 5; t-test p-value = 0.0024). The combination of the up-regulation of the two of these K⁺ channels would facilitate rapid repolarization of Kiss1^ARH following an action potential.

Traditionally, the after hyperpolarization is divided into three distinct phases: fast (fAHP), medium (mAHP), and slow after hyperpolarization (sAHP) (Storm, 1990; Vogalis et al., 2003). The fast after hyperpolarization (fAHP) is primarily mediated by the BK family of potassium channels (Storm, 1987). The medium after hyperpolarization (mAHP) is predominantly mediated by apamin-sensitive SK2 channels (Bond et al., 2004; Peters et al., 2005). However, KCNQ family members contribute to both the mAHP and sAHP (Tzingounis A. V. et al., 2010; Tzingounis A. V. and Nicoll, 2008). Therefore, to investigate the contribution of KCNQ channels to Kiss1^ARH neuronal excitability, voltage clamp experiments were conducted in the presence of TTX, CNQX, AP5, and picrotoxin, and a standard M-current protocol was run using the M-channel blocker XE-991 to isolate the M-current (Greene et al., 2017; Roepke et al., 2011) (Figure 9 A, B). The application of XE-991 resulted in the inhibition of M-current within a physiologically relevant voltage range of -60 to -30 mV in E2-treated OVX females but exhibited minimal impact in OVX females (Figure 9 C, D). Although the XE991-sensitive current was relatively small compared to other voltage-activated K⁺ conductances, it demonstrated a significant increase in E2-treated, OVX females (Figures 9 E). The maximum peak current density sensitive to XE-991 at -30 mV was found to be four times higher in E2-treated OVX females when compared to OVX females. This would contribute to the repolarization following burst firing. Furthermore, E2 increased the mRNA expression of Kcnq2, (Figure 9F), which suggests that KCNQ channels play a key role in repolarizing Kiss1^ARH neurons following burst firing. Indeed, our modeling predicted that M-current contributed to the repolarization following burst firing (Figure 9G).

Based on our electrophysiological results, Kiss1^ARH neurons appear to transition from peptidergic to glutamatergic neurotransmission through E2-mediated changes in the expression of voltage-activated Ca²⁺ channels and K⁺ channels, and their respective conductances. Therefore, we asked the question is there a difference in peptide and glutamate mRNA expression mediating this transition? Therefore, we ran a comparison between Tac 2 (NKB) and Vglut2 (surrogate for glutamate) expression. The cycle threshold (CT) was compared between Tac2 and Vglut2 as well as Kiss1, TRPC5 and GIRK2 in Kiss neuronal cell pools from OVX oil-treated and OVX E2-treated animals (Figures 10A, B). It is worth noting that lower number of cycles illustrate a higher quantity of mRNA expression because the fluorescence is detected earlier, and one cycle difference represents a doubling in expression. As expected, the reference gene Gapdh did not change with E2-treatment. However, quantitative PCR results revealed that E2 treatment of OVX females significantly reduced Tac2 expression (Figure 10C), whereas Vglut2 mRNA was significantly increased in Kiss1^ARH neurons (Figure 10D). Moreover, both Trpc5 and Girk2 expression were significantly reduced in E2-treated, OVX females (Figures 10E, F).

Our computational modeling suggested that TRPC5 channels play a dominant role in regulating cell excitability. Therefore, as proof of principle, we utilized a CRISPR approach to mutate TRPC5 channels in Kiss1^ARH neurons similar to our previous studies (Hunker et al., 2020; Stincic et al., 2021). Hunker et al. developed a single viral vector for conditional expression of the smaller Staphylococcus aureus (SaCas9) and sgRNA that yields high-efficiency mutagenesis in specific cell types (Hunker et al., 2020). To selectively mutate Trpc5 in Kiss1^ARH neurons, we generated two guide RNA’s, one targeting exon 2, which is conserved across all splice variants, and the other targeting exon 7, the pore forming domain (Figures 11A, B). A cohort of Kiss1^ARH mice were given bilateral stereotaxic injections into the ARH of the two AAV1-FLEX-SaCas9-sgTrpc5’s or a control virus containing the Trpc5 guide with three base pairs in the seed region mutated (SaCas9-control) as described (Hunker et al., 2020). An additional Cre-dependent virus of the same serotype (AAV1) that drove expression of a fluorophore (YFP or mCherry) was co-administered in order to visualize injection quality and facilitate harvesting of cells (Figure 11C). After three weeks, mice underwent ovariectomy since OVX mice express the maximum slow EPSP amplitude (Qiu J. et al., 2016). Brain slices were prepared, and cells harvested as previously described (Qiu J. et al., 2018) and analyzed with qPCR. We found that the Trpc5 mutagenesis group displayed a reduction in relative expression of Trpc5 in Kiss1^ARH neurons compared to the control group (Figure 11D). Hence, the qPCR data verified that in the sgTrpc5-targeted mice we can selectively reduce Trpc5 gene expression in targeted cells.

As predicted, mutagenesis of Trpc5 in Kiss1^ARH neurons significantly attenuated the slow EPSP (Figures 12A, B, C) such that the postsynaptic excitation was reduced to a “trickle” of action potential firing. What we would not have predicted is that the double sgRNA mutagenesis of Trpc5 channels in Kiss1^ARH neurons significantly hyperpolarized the resting membrane potential by 7 mV (Figure 12D). Moreover, the rheobase (minimum current required to induce firing) significantly increased by ∼20% in females bearing the sgTrpc5 double mutagenesis (Figure 12E). The firing frequency versus injected current (F-I) curve for sgTrpc5 double mutagenesis Kiss1^ARH neurons was also significantly attenuated (Figure 12F). In agreement with these experimental findings, simulations of our mathematical model confirmed that TPRC5 channels should lower the rheobase and greatly enhance the firing activity of Kiss1^ARH neurons (Figures 12G, H). Finally, we employed our mathematical model to further investigate the transition from synchronous firing driven by NKB release and TRPC5 channel activation to burst firing generated by E2-mediated upregulation of endogenous conductances. Our simulations suggest that synchronous firing is indeed sculpted by the interplay between TRPC5 and GIRK channels, whereas burst firing is controlled by the E2-dependent increase of calcium and calcium-activated K⁺ conductances (Figure 13).

TRPC5 channels are highly expressed in Kiss1^ARH neurons (Figure 11), and TRPC5 channels are essentially ligand-activated calcium channels with a high permeability to calcium (P_Ca/P_Na = 9:1) (Venkatachalam and Montell, 2007). In general, mammalian TRPC channels are activated by both G protein-coupled receptors and receptor tyrosine kinases (Ambudkar and Ong, 2007; Clapham, 2003), and are one of the major downstream effectors activated by glutamate binding to group I metabotropic glutamate receptors (mGluR1 and mGluR5) in CNS neurons (Bengtson et al., 2004; Berg et al., 2007; Faber et al., 2006; Tozzi et al., 2003). In substantia nigra dopamine neurons mGluR1 agonists induce a current that exhibits the tell-tale double-rectifying current-voltage plot of TRPC channel activation (Tozzi et al., 2003), similar to what we see with the effects of the NKB agonist senktide in Kiss1^ARH neurons (Qiu Jian et al., 2021). Both mGluR1 and TacR3 are Gq-coupled to phospholipase C (PLC) activation which leads to hydrolysis of phosphatidylinositol 4,5-bisphosphate (PIP₂) to diacylglycerol (DAG) and inositol 1,4,5 triphosphate (IP₃), which is involved in channel activation (Birnbaumer, 2009). TacR3 (NKB) signaling has additional consequences since many K⁺ (e.g., GIRK, KCNQ) channels are dependent on PIP2 for channel opening, and depletion of PIP2 by PLC leads to channel closure (Brown and Passmore, 2009) (Zhang C. et al., 2013) (Whorton and MacKinnon, 2011) (Zheng et al., 2022). Therefore, depletion of PIP2 by NKB signaling would further facilitate the sustained firing of Kiss1^ARH neurons during synchronization.

A plateau potential has been characterized in hippocampal and cortical neurons (Arboit et al., 2020; Zhang Z. et al., 2011) as such these neurons express biophysical properties that allow them to continue to persistently fire even after a triggering synaptic event has subsided (Zylberberg and Strowbridge, 2017). The persistent firing activity of these neurons is linked to I_CAN (Zylberberg and Strowbridge, 2017), and TRPC5 channels appear to be responsible for the I_CAN (Zhang Z. et al., 2011). With TacR3 activation in Kiss1^ARH neurons there is an influx of Ca⁺² through TRPC5 channels leading to greater build-up of [Ca²⁺]_i that facilitates the opening of more TRPC5 channels in a self-sustaining (autocatalytic) manner (Qiu J. et al., 2016). Using the fast intracellular calcium chelator BAPTA, which has been shown to robustly inhibit TRPC5 channel activation in heterologous cells (Blair et al., 2009), we abolished the slow EPSP and persistent firing of Kiss1^ARH neurons following optogenetic stimulation (Qiu Jian et al., 2021). Moreover, HVA calcium channel blockers attenuated the generation of the slow EPSP (Figure 3) so it appears that they also contribute to the I_CAN since calcium influx via both LVA and HVA calcium channels can also facilitate TRPC5 channel opening in Kiss1^ARH neurons. In the ovariectomized female, treatment with E2 upregulated Cav1.2, Cav2.1, Cav2.2, and Cav2.3 mRNA by 1.5-to 2-fold and Cav3.1 mRNA expression by ∼3-fold. Hence, E2 significantly increased whole-cell calcium currents Kiss1^ARH neurons, which greatly enhanced the excitability and contributed to the burst firing of Kiss1^ARH neurons (present findings and (Qiu J. et al., 2018)). However, the amplitude of the slow EPSP with E2 treatment is only ∼25% of the amplitude in the ovariectomized state (Qiu J. et al., 2016). Therefore, there appears to be a physiologic transition of Kiss1^ARH neurons from the slow EPSP firing mode in the OVX state to the burst firing mode in the presence of E2, which has important physiologic ramifications as discussed below.

Recently, Tian et al. demonstrated that TRPC4, a close homolog of TRPC5 sharing ∼64 percent homology, is a “coincidence detector” of neurotransmission by both Gq/11 and Gi/o-coupled receptors in lateral septal (LS) neurons (Tian et al., 2022). In whole-cell recordings of LS neurons, TRPC 4 channels mediate a strong depolarizing plateau potential that in contrast to TRPC5 channel activation in Kiss1^ARH neurons, abrogates action potential firing as a result of a depolarization block. In many instances the plateau potential in LS neurons is followed by an AHP, which is dependent on the activation of Gi/o-coupled receptors. In contrast, we have not observed an AHP in Kiss1^ARH neurons following the slow EPSP (Qiu J. et al., 2016). Tian and colleagues showed that the depolarizing plateau in LS neurons is codependent on activation of both Gq/11-coupled mGluR1 glutamate receptors and Gi/o-coupled γ-aminobutyric acid type B receptors, the latter activating GIRK channels. Moreover, the firing patterns in LS neurons encodes information about the relative strengths of these contrasting inputs (i.e., Gq/11 versus Gi/o) such that only mGluR1 produces weak depolarization accompanied by increased firing of LS neurons, whereas pure GABA_B receptor activation hyperpolarizes the cells and abrogates firing activity. Coincident input of both mGluR1 and GABA_B receptors results in a brief burst of action potentials followed by a pause in firing, and both the pause duration and firing recovery patterns reflect the relative strengths of Gq/11 versus Gi/o inputs. Importantly, Tian and colleagues computationally simulated these various scenarios with computational modeling, and similar to our modeling, utilized only TRPC4 and GIRK channels. A notable difference between the Kiss1^ARH neurons and the LS neuronal circuitry is that the GIRK channel activity is predominately at the nerve terminal of Kiss1^ARH neurons, and GIRK channels are opened via dynorphin binding to kappa-opioid receptors, which does not translate into membrane hyperpolarization of the soma membrane from which we are recording (Qiu J. et al., 2016). Therefore, although the high frequency firing activity of both LS and Kiss1^ARH neurons can be modeled around TRPC and GIRK channels, the generated firing patterns are dramatically different based on the timing (co-incident activation of TRPC4 and GIRK channels in LS) and localization of the GIRK channels in the axon terminal of Kiss1^ARH neurons. Moreover, E2-treated, ovariectomized females show a significant down-regulation of both Trpc5 and Girk2 mRNA expression in Kiss1^ARH neurons (Figure 10), which is important for the physiological transitioning as described below.

Interestingly, the hypothalamic A12 (ARH) dopamine neurons show a rhythmic “oscillatory” firing behavior that transitions to a tonic firing mode with synaptic input from Thyrotropin-releasing hormone (TRH) neurons (Lyons D.J. et al., 2010) or feedback by circulating prolactin, released by pituitary lactotrophs (Lyons D. J. et al., 2012). A more recent paper has revealed that TRPC5 channels mediate the plateau potential and tonic firing in A12 dopamine neurons in response to prolactin (Blum et al., 2019). Similar to our findings with CRISPR deletion of Trpc5 in Kiss1^ARH neurons (Figures 11 and 12), conditional knockout of Trpc5 in dopamine neurons abrogated the prolactin-induced plateau potential and tonic firing. Although Blum and colleagues did not model the oscillatory firing or tonic firing of the A12 dopamine neurons, their findings are consistent with our results showing that the activation of TRPC5 channels underlies the slow EPSP (plateau potential) and sustained firing.

Beyond the ligand-gated (e.g., baclofen) GIRK channels, there are endogenous K⁺ channels that help sculpt the firing activity of kisspeptin neurons. We focused on the calcium-activated K⁺ channel family: the large-conductance, calcium-activated potassium (BK, also called BK_Ca, K_Ca1.1, MaxiK, Slo), small conductance, calcium-activated K⁺ (SK1, SK2, SK3) (Bond et al., 2005), and the K⁺ channels underlying the M-current (KCNQ, Kv7.1-7.5) (Brown and Passmore, 2009), which mediate the fast afterhyperpolarization (AHP), the intermediate AHP/slow AHP, respectively (Andrade et al., 2012)

BK channels are gated by both voltage and cytoplasmic calcium and sculpt action potential firing in CNS neurons (Blatz and Magleby, 1987; Marty, 1989; Sah P., 1996; Storm, 1990). Indeed, BK channels have been shown to mediate rapid spike repolarization—i.e., the fast AHP in hippocampal CA1 pyramidal neurons (Lancaster and Nicoll, 1987; Storm, 1987). Blockade of BK channels in CA1 neurons attenuates the initial discharge frequency in response to current injection, which is attributable to suppression of the BK channel-dependent rapid spike repolarization (Lancaster and Nicoll, 1987; Storm, 1987). Blockade of BK channels is thought to increase inactivation of the spike-generating transient Na⁺ current and activate more of the slower K⁺ currents, thereby enhancing refractoriness and reducing excitability during the immediate aftermath of the first action potential (Shao et al., 1999). Thus, BK channels facilitate high-frequency burst firing of CA1 neurons. Furthermore, extracellular field recordings confirmed that BK channels contribute to high-frequency burst firing in response to excitatory synaptic input to distal dendrites in CA1 neurons (Gu et al., 2007). Therefore, BK channels appear to play an important role for early high-frequency, rapidly adapting firing in hippocampal CA1 pyramidal neurons, thus promoting the type of bursting that is characteristic of these cells in vivo during behavior. Based on our in vitro electrophysiological recordings and computational modeling we see a similar physiological phenomenon in Kiss1^ARH neurons (Figure 8). Not only does E2 increase the mRNA expression of Kcnma1 (Figure 8G), but also the maximum BK (IbTx-sensitive) current by 4-fold. In addition, Kcnb1 mRNA was also up-regulated, and the combination of these two K⁺ conductances would facilitate rapid repolarization during burst firing and promote glutamate release similar to hippocampal CA1 neurons.

In contrast to BK channel expression, E2 did not affect the mRNA expression of SK3 channel mRNA. SK channels underlie the apamin-sensitive component of the medium duration AHP and are responsible for repolarization following a burst of action potentials (Andrade et al., 2012; Bond et al., 2005). The activation of SK channels is voltage-independent, but SK channels have a higher affinity for Ca²⁺ than BK channels (Andrade et al., 2012). SK channels are tightly coupled (within 100 nm) to L-type Ca²⁺ channels, and BK channels (within 30 nm) are tightly coupled to N-type Ca²⁺ channels in hippocampal CA1 pyramidal neurons (Marrion and Tavalin, 1998). The determination of the proximity of SK and BK channels to HVA calcium channels in kisspeptin neurons will require cell-attached patch recordings. However, in Kiss1^ARH neurons the SK channels may come into play during a short burst of action potentials but would become overwhelmed with the higher frequency synchronized, sustained firing as a result of NKB stimulation (Qiu J. et al., 2016). As discussed above what limits the synchronized firing of Kiss1^ARH neurons is the activation of GIRK channels.

Finally, the calcium-activated slow AHP probably plays a critical role in the repolarization of Kiss1^ARH after burst firing. The molecular identification of the channels mediating the slow AHP has long been an area of intense investigation (Andrade et al., 2012; Vogalis et al., 2003). A critical feature of the slow AHP is that it activates very slowly (hundreds of milliseconds) long after the rise in cytoplasmic Ca²⁺ (Sah P. and Clements, 1999) so an intermediate Ca²⁺ signaling molecule has long been thought to be involved. Indeed, Tzingounis and colleagues (Tzingounis A.V. et al., 2007) have provided compelling evidence that the diffusible calcium sensor hippocalcin is the critical intermediate molecule involved in Ca²⁺ sensing. The slow AHP is abrogated in hippocalcin KO mice, and transfection of hippocalcin into cultured hippocampal neurons generates a pronounced slow AHP in response to a depolarizing stimulus (Tzingounis A.V. et al., 2007). Importantly, the slow AHP is activated by Ca²⁺ with an EC₅₀ ≈ 300 nM, which is well within the operational range of hippocalcin but well below that of calmodulin (Andrade et al., 2012). Finally, two seminal papers from Tzingounis and colleagues demonstrate that KCNQ 2, 3 channels are responsible for the slow AHP in hippocampal dentate neurons (Tzingounis A. V. and Nicoll, 2008), and KCNQ 5 channels are responsible for the slow AHP in CA3 neurons (Tzingounis A. V. et al., 2010). Moreover, the KCNQ channel blocker XE991 attenuates the slow AHP in CA3 neurons (Tzingounis A. V. and Nicoll, 2008). Based on these seminal findings we investigated the role of the KCNQ channels, which “classically” underlie the M-current in Kiss1^ARH. The M-current was first identified in Kiss1^ARH neurons by Conde and Roepke (Conde and Roepke, 2019), and presently we found that Kcnq2 mRNA is expressed in Kiss1^ARH neurons and up-regulated by E2, which translated into a greater M-current in Kiss1^ARH neurons (Figure 9). Incorporating the M-current into our computational model indeed supports our hypothesis that this is a critical K⁺ conductance, along with SK and BK, for membrane repolarization after burst firing (Figure 9G). Importantly the slow AHP, as opposed to the fast AHP (BK) and medium AHP (SK) is highly regulated by multiple neurotransmitters (Andrade et al., 2012), which sets the stage for further modulation of the slow EPSP in Kiss1^ARH neurons.

Since the expression of the peptide neurotransmitters in Kiss1^ARH neurons are down-regulated by E2, the Kiss1^ARH neurons are believed to be under “inhibitory” control by E2 and are important for “negative-feedback” regulation of GnRH release (Lehman et al., 2013; Navarro V. M. et al., 2009; Smith et al., 2005) (Rance Naomi E. and Young, 1991) (Rance N.E., 2009). However, our past (Gottsch et al., 2011) and current findings document that these Kiss1^ARH neurons express HVA and LVA calcium and HCN (pacemaker) channels and are excited by co-released glutamate from neighboring Kiss1^ARH neurons, which indicates that these neurons have pacemaker electrophysiological properties similar to other CNS neurons (Bal and McCormick, 1993; Lüthi and McCormick, 1998). Additionally, in contrast to the neuropeptides, E2 increases Slc17a6 (Vglut2) mRNA expression, in addition to mRNA for the voltage-activated calcium and HCN channels, and increases glutamate release onto Kiss1^AVPV/PeN neurons (Qiu J. et al., 2018). Interestingly, Slc17a6 mRNA expression in Kiss1^ARH neurons and the probability of glutamate release are decreased along with the neuropeptides in intact versus castrated males (Nestor et al., 2016), which indicates a profound sex difference in the glutamate signaling by Kiss1^ARH neurons (Nestor et al., 2016; Qiu J. et al., 2018). Obviously, in the male there is no preovulatory LH surge so there is no need for excitatory glutamatergic input to the few Kiss1^AVPV neurons in the male.

In females, conditional knockout of Slc17a6 in Kiss1 neurons abrogates glutamate release from Kiss1^ARH neurons (Qiu J. et al., 2018). Kiss1^AVPV/PeN neurons do not express Slc17a6 and do not release glutamate. Within the Kiss1^ARH neurocircuitry the lack of glutamate transmission does not diminish the slow EPSP in ovariectomized females (Qiu J. et al., 2018). Indeed, a recent publication from the Herbison lab demonstrates that glutamate generates small “synchronizing” events that are dependent on the ionotropic receptors (Han et al., 2023), but the fast neurotransmitter is unable to support the sustained firing (i.e., slow EPSP) that is necessary for peptide release and synchronization of the KNDy network. Rather, we believe that glutamate neurotransmission is more important for excitation of Kiss1^AVPV/PeN neurons and facilitating the GnRH (LH) surge with high circulating levels of E2, when peptide neurotransmitters are at a nadir, but glutamate levels are high in female Kiss1^ARH neurons. Indeed, low frequency (5 Hz) optogenetic stimulation of Kiss1^ARH neurons, which only releases glutamate in E2-treated, ovariectomized females (Qiu J. et al., 2016), generates a surge-like increase in LH release during periods of optical stimulation (Lin et al., 2021; Voliotis et al., 2021). Therefore, there appears to be a clear role for glutamatergic transmission from the Kiss1^ARH to Kiss1^AVPV/PeN neurons in amplifying the LH surge in the female mouse. Finally it is important to keep in mind that even in the presence of high physiological levels of E2, the mRNA expression of Tac2 is many-fold higher than Kiss1 (Figure 10), which is essential for NKB maintaining synchronous firing of Kiss1^ARH neurons, albeit at a lower frequency, across all physiological states (Qiu J. et al., 2016). Indeed, there is a progressive change from a strictly pulsatile pattern of GnRH in the hypophyseal portal circulation to one containing both pulsatile and non-pulsatile components during the development of the GnRH surge in the ewe (Evans, Dahl, Mauger, & Karsch, 1995; Evans, Dahl, Mauger, Padmanabhan, et al., 1995), and a pulsatile mode of LH secretion during the preovulatory LH surge is also evident in other species including humans (Rossmanith et al., 1990). Therefore, we believe that our cellular molecular and electrophysiological findings in combination with our computational modelling provide a foundation for understanding the complex role of Kiss1^ARH neurons in controlling fertility in the mammal. Finally, our model provides the first comprehensive biophysical description of the conductances underlying the neuronal activity of Kiss^ARH neurons, which can serve as a basis for future computational modelling of the Kiss^ARH neuronal network and its interactions with other brain regions involved in the complex regulation of mammalian female reproduction.

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 in accordance with the United Kingdom Animals (Scientific Procedures) Act 1986 and were approved by King’s College London (KCL) Ethical Review Committee.

Kiss1^Cre transgenic female mice version 2 (Padilla et al., 2018) were selectively bred at OHSU and KCL. They also were crossed with heterozygous Ai32 mice (RRID:IMSR_JAX:024109, C57BL/6 background) which carry ChR2 (H134R)–EYFP gene in their Gt(ROSA)26Sor locus (Madisen et al., 2012). All animals were maintained under controlled temperature and photoperiod (lights on at 0600h and off at 1800h at OHSU or 0700h and off at 1900h at KCL) and given free access to food (Lab Diets 5L0D) and water. Where specified, Kiss1^Cre mice received viral injections to express channelrhodopsin 2 (ChR2) in Kiss1^ARH neurons, fourteen to twenty-one days prior to each experiment as described (Qiu J. et al., 2016). Some of the females were ovariectomized seven days prior to an experiment. Each animal was injected on day 5 following OVX with 0.25 μg E2 or vehicle, followed on day 6 with 1.50 μg E2 or vehicle and used for experiments on day 7 (Bosch et al., 2013).

Fourteen to twenty-one days prior to each experiment, the Kiss1^Cre mice (>60 d old) received bilateral ARH injections of a Cre-dependent adeno-associated viral (AAV; serotype 1) vector encoding mCherry (AAV1-Ef1α-DIO-mCherry) or AAV1 vectors designed to encode SaCas9 and single-guide RNAs (sgRNAs) (See the SaCas9 section for specifics on the sgRNA design). Using aseptic techniques, anesthetized female mice (1.5% isoflurane/O₂) received a medial skin incision to expose the surface of the skull. The glass pipette with a beveled tip (diameter = 45 μm) was filled with mineral oil, loaded with an aliquot of AAV using a Nanoject II (Drummond Scientific). ARH injection coordinates were anteroposterior (AP): −1.20 mm, mediolateral (ML): ± 0.30 mm, dorsoventral (DV): −5.80 mm (surface of brain z = 0.0 mm); 500 nl of the AAV (2.0×10¹² particles/ ml) was injected (100 nl/min) into each position, and the pipette left in place for 10 min post-injection, then slowly retracted from the brain. The skin incision was closed using Vetbond (3M) and each mouse received analgesia (Rimadyl, 4-5 mg/kg, s.c.).

The generation of AAV1-FLEX-SaCas9-U6-sgTrpc5 viruses were done at the University of Washington using published methods (Gore et al., 2013; Hunker et al., 2020). The constructs of sgRNAs for Trpc5 (AAV1-FLEX-SaCas9-U6sgTrpc5) were designed to target exon2 and exon7, respectively (Figure 11A and B). To achieve Trpc5 mutagenesis in Kiss1^ARH neurons, Kiss1^Cre mice were co-injected with AAV1-DIO-mCherry, AAV1-FLEX-SaCas9-U6-sgTrpc5-exon2, and AAV1-FLEX-SaCas9-U6-sgTrpc5-exon7 at a ratio of 10%, 45%, and 45%, respectively. Control animals were co-injected with AAV1-FLEX-SaCas9-U6-sgRosa26 and AAV1-DIO-mCherry at a ratio of 90% and 10%, respectively. AAV1-DIO-mCherry was co-injected with Cas9 vectors to confirm the targeting of injections and visualize the infected Kiss1^ARH neurons.

Electrophysiological and optogenetic studies were made in coronal brain slices (250 μm) containing the ARH from AAV1-EF1α-DIO-mCherry injected Kiss1Cre:GFP or Kiss1-Cre:EYFP::AI32 mice, which were vehicle-treated OVX, and E2-treated OVX females 10 weeks and older as previously described (Qiu J. et al., 2016; Qiu J. et al., 2018). Whole-cell patch recordings were performed in voltage-clamp and current-clamp as previously described (Qiu J. et al., 2018) using an Olympus BX51 W1 fixed stage scope out-fitted 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 3.5–4 MΩ. In whole-cell configuration, access resistance was less than 30 MΩ; the access resistance 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 (Qiu J. et al., 2003; Qiu J. et al., 2014). Electrophysiological signals were digitized with a Digidata 1322A (Axon Instruments) and the data were analyzed using p-Clamp software (Molecular Devices, Foster City, CA). The liquid junction potential was corrected for all data analysis.

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, Newton, NJ) with the light path directly delivered through an Olympus 40x water-immersion lens. For high-frequency (20 Hz) stimulation the length of stimulation was 10 seconds (Qiu J. et al., 2016).

For studying the activation/inactivation characteristics of the Ca²⁺ current, the electrodes were filled with an internal solution as described (Qiu J. et al., 2003) consisting of the following (in mM): 100 Cs⁺ gluconate, 20 TEA-Cl, 10 NaCl, 1 MgCl₂, 10 HEPES, 11 EGTA, 1 ATP, 0.25 GTP, the pH was adjusted to 7.3 with CsOH at 300 mOsm. The bath solution as described (Zhang X. B. and Spergel, 2012) consisted of (in mM) 117.5 NaCl, 25 NaHCO₃, 1.25 NaH₂PO₄, 10 TEA-Cl, 2 CaCl₂, 1 MgCl₂, 20 sucrose and 5 glucose, gassed with 95% O₂ /5% CO₂ (pH 7.4, 309 mOsm) and supplemented with 1 µM TTX, 10 µM CNQX, 50 µM AP5 and 100 µM picrotoxin. The effects of estrogen treatment on the peak current, peak current density, and activation/inactivation characteristics of calcium current were measured. Activation curves were fitted by the Boltzmann equation: I/I_max=1/{1+exp[V_1/2-Vs)/k]}, where I is the peak current at the step potential Vs, Imax is the peak current amplitude, V_1/2 is the step potential yielding half-maximum current, and k is the slope factor. Inactivation curves were fit with the Boltzmann equation: I/I_max = 1 - 1/{1 + exp [(V_H - V_1/2)/k]}, where I is the peak current at the step potential V_H, I_max is the peak current amplitude, V_1/2 is the step potential at which half the current is inactivated, and k is the slope factor.

To record M-currents, pipettes were filled with an internal solution consisting of 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; 290 mOsm). During voltage-clamp, we employed a standard deactivation protocol (Conde and Roepke, 2019; Roepke et al., 2011) to measure potassium currents. This involved 500-ms voltage steps ranging from –30 to –75 mV in 5-mV increments, following a 300-millisecond prepulse to –20 mV. The amplitude of the M-current relaxation or deactivation was quantified as the difference between the initial (<10 ms) and sustained current (>475 ms) of the current trace.

The bath solution for whole-cell recording of BK, SK and M currents was aCSF supplemented with 1 µM TTX, 10 µM CNQX, 50 µM AP5 and 100 µM picrotoxin.

A standard artificial cerebrospinal fluid (aCSF) was used (Qiu J. et al., 2003; Qiu J. et al., 2010). All drugs were purchased from Tocris Bioscience unless otherwise specified. 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, 10 mM Nifedipine (Sigma), 2 mM ω-conotoxin GVIA (ConoGVIA; Alomone Labs), 0.5 mM ω-conotoxin MVIIC (ConoMVIIC; Alomone Labs, 50 µM ω-agatoxin IVA (AgaIVA; Alomone Labs), 50 µM SNX-482 (Alomone Labs), 10 mM TTA-P2 (TTAP2; Alomone Labs), 200 mM CdCl₂ (Cd²⁺; Sigma), 40 mM XE991 (Alomone Labs), 100 µM Iberiotoxin (Alomone Labs), 500 µM Apamin (Alomone Labs) and 100 mM NiCl₂ (Ni²⁺; Sigma). Stocks (1000×) were prepared in dimethylsulfoxide DMSO (picrotoxin, TTAP2) or water (TTX, AP5, CNQX, ConoGVIA, ConoMVIIC, AgaIVA, SNX-482, Cd²⁺, Ni²⁺) and stored at −20°C. Aliquots of the stock solutions were stored as appropriate until needed.

Cell harvesting and qPCR was conducted as previously described (Bosch et al., 2013). The ARH was microdissected from basal hypothalamic coronal slices obtained from female Kiss1^Cre version 2 mice (Padilla et al., 2018) (n = 5-7 animals/group). The dispersed cells were visualized, patched, and then harvested (5 or 10 cells/tube) as described previously (Bosch et al., 2013). Briefly, ARH tissue was incubated in papain (7mg/ml in oxygenated aCSF) for 50 min at 37° C then washed 4 times in low Ca²⁺ aCSF and two 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, Navato, CA) 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. cDNA synthesis was performed as previously described (Bosch et al., 2013).

Primers for the genes that encode for low and high voltage-gated calcium channels, TRPC5, Vglut2, large conductance calcium-activated K⁺ (BKα) channels, small conductance calcium-activated K⁺ (SK3) channels and 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 (Bosch et al., 2013). We have already published primer sequences for the low voltage-activated calcium channels, Vglut2 and GAPDH (see Table 1) (Qiu J. et al., 2018). 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 utilized to determine the real-time PCR efficiency (E = 10^(−1/m) – 1) (Biosystems, 2006; Pfaffl, 2001). Only primers resulting in efficiencies of 90-100% were used for analysis. Primer sequences, qPCR parameters and efficiency calculations are provided in Table 1.

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 (Bosch et al., 2013). The comparative ΔΔC_T method (Livak and Schmittgen, 2001; Pfaffl, 2001; Schmittgen and Livak, 2008) 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 1x Power SYBR Green PCR Master Mix and 0.5 µM forward and reverse primers. Three to four 5-cell or 10-cell pools per animal were analyzed from 5-7 animals per group. The relative linear quantity was determined using the 2^-ΔΔCT equation (Livak and Schmittgen, 2001; Pfaffl, 2001; Schmittgen and Livak, 2008). Relative mRNA expression level of target genes in Kiss1^Cre neurons was obtained by comparing OVX Oil-treated controls to OVX E2-treated animals. The mean Δ CT for the target genes from the OVX Oil-treated 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.

For the visualized, whole-cell patch recording experiments, only one cell was recorded per slice. Two to three slices were analyzed from each Kiss1^Cre mouse, with at least 3-5 mice contributing to each group. For cell harvesting of dispersed Kiss1^Cre-YFP neurons and qPCR measurements, 10 cells per pool and 3-6 pools from each animal were used, unless otherwise specified. Statistical comparisons between two groups were performed using an unpaired two-tailed Student’s t-test. Comparisons between more than two groups were performed using the repeated measures, multifactorial 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 ± standard error of the mean (SEM). Differences were considered statistically significant if the probability of error was less than 5%.

A mathematical model of the Kiss1^ARH neuron was developed and calibrated based on our physiological findings. Here, we employed the Hodgkin-Huxley modelling approach (Hodgkin and Huxley, 1952). Accordingly, the equation describing the membrane potential V_m of the Kiss1^ARH neuron is given by where C_m is the membrane capacitance and I is the sum of 12 ionic currents:

I_NaT and I_NaP are the transient and persistent sodium currents, respectively; I_A represents the A current; I_M represents the M current; I_SK and I_BK are the potassium currents through the SK and BK channels respectively; I_h the HCN current; I_T the T-type calcium current; I_Ca represents other calcium currents (L-, N-, P/Q-, and R-type); I_TRPC5 represents Calcium current throught the TRPC5 channel; I_GIRK potassium current through the GIRK channels. Finally, I_leak represents the contribution of leak currents. We use the Hodgkin-Huxley formalism to model current dynamics and their dependence on the membrane potential. A complete specification of all currents along with the complete table of model parameters can be found in the supplementary material. Model simulations were conducted in Matlab using the built-in numerical solver (ode45; based on an explicit Runge-Kutta (4, 5) formula). The Matlab code is available at https://git.exeter.ac.uk/mv286/kiss1-arcuate-neuron-model.