GnRH pulse generator frequency is modulated by kisspeptin and GABA‐glutamate interactions in the posterodorsal medial amygdala in female mice

Abstract Kisspeptin neurons in the arcuate nucleus of the hypothalamus generate gonadotrophin‐releasing hormone (GnRH) pulses, and act as critical initiators of functional gonadotrophin secretion and reproductive competency. However, kisspeptin in other brain regions, most notably the posterodorsal subnucleus of the medial amygdala (MePD), plays a significant modulatory role over the hypothalamic kisspeptin population; our recent studies using optogenetics have shown that low‐frequency light stimulation of MePD kisspeptin results in increased luteinsing hormone pulse frequency. Nonetheless, the neurochemical pathways that underpin this regulatory function remain unknown. To study this, we have utilised an optofluid technology, precisely combining optogenetic stimulation with intra‐nuclear pharmacological receptor antagonism, to investigate the neurotransmission involved in this circuitry. We have shown experimentally and verified using a mathematical model that functional neurotransmission of both GABA and glutamate is a requirement for effective modulation of the GnRH pulse generator by amygdala kisspeptin neurons.

generator frequency. 1 This finding built upon our previous neuropharmacological approach using intra-MePD infusions of kisspeptin receptor (Kiss1r) agonists and antagonists, which respectively increased LH secretion or decreased LH pulse frequency. 2 However, the mechanisms underlying this neuronal population's stimulatory role over pulsatile LH secretion have not been studied. Glutamate and GABA are the major stimulatory and inhibitory neurotransmitters in the mammalian brain, and many neuronal networks rely on the balance between these two to regulate their activity. 3 Therefore, these two neurotransmitters are sensible candidates to be used by the amygdala neuronal networks underlying the upstream, extra-hypothalamic regulation of the GnRH pulse generator. Unsurprisingly, both GABA and glutamate neurons are found in the MePD, 4,5 and pharmacological antagonism of both has deleterious effects on several aspects of reproductive physiology. [6][7][8] In rats, blocking AMPA and NMDA glutamate receptors with CNQX and AP5, respectively, impedes activation of MePD neurons in response to vaginal-cervical stimulation, thereby preventing the pregnancy or pseudopregnancy response following intromission; AMPA antagonism also disrupts oestrous cycles. 6,7 Furthermore, MePD NMDA and GABA A receptor (GABA A R) antagonism causes a weight-independent advancement of puberty. 8 A reciprocal relationship between kisspeptin and GABA in the limbic system has also been shown, with i.v. kisspeptin administration in humans resulting in a reduced GABA signal in the anterior cingulate cortex. 9 Thus, GABA and glutamate play an important role in the MePD in reproductive physiology.
The interesting dichotomy between the well-known suppressive role of the MePD in reproductive physiology and the emerging activatory function of kisspeptin within this amygdaloid subnucleus has led to a hypothesis that MePD kisspeptinergic activity may stimulate GABAergic interneurons within the MePD that in turn synapse with, and inhibit, GABAergic projection efferents from the MePD, resulting in an overall disinhibition of the latter. Evidence to support this hypothesis stems from the knowledge that there is a significant population of GABAergic neurons that project from the MePD to reproductive neural centres such as those in the hypothalamus, 4,10 and inhibitory GABA interneurons, specifically, have also been detected in this subregion. 11,12 This is in line with the fact that the MePD is a pallidal subnucleus, as a result of its embryological origins in the caudoventral medial ganglionic eminence, indicating its similarity to other neural complexes which contain a classical GABA-GABA disinhibitory system. 13 Furthermore, other subnuclei of the amygdala, such as the basolateral amygdala and posteroventral medial amygdala, have been shown to exhibit functional glutamatergic signalling onto GABA interneurons, 12,14,15 supporting the hypothesis of an alternative glutamate-GABA-GABA pathway by which kisspeptin may activate the disinhibitory system.
It is therefore critical to investigate the GABAergic and glutamatergic signalling within the MePD with respect to kisspeptin and its action over GnRH pulse generator activity. To achieve this, a dual approach of simultaneously combining optogenetic activation and pharmacological antagonists was used via the implantation of an

| Stereotaxic injection of channelrhodopsin viral construct and implantation optofluid cannula
All surgical procedures were carried out under a combination of ketamine anaesthesia (Ketamidor, 100 mg kg À1 , i.p.; Chanelle Vet, Hungerford, UK) and xylazine (Rompun, 10 mg kg À1 , i.p.; Bayer, Leverkusen, Germany) under aseptic conditions. Mice were bilaterally ovariectomised (OVX) to mitigate negative feedback of endogenous oestrogen on LH secretion. Stereotaxic injection of the viral construct and implantation of the brain cannula was carried out concurrently with ovariectomy. Mice were placed in a Kopf motorised stereotaxic frame (Kopf, Tujinga, CA, USA) and procedures were carried out using a robot stereotaxic system (Neurostar, Tubingen, Germany). Following an incision of the scalp, a small hole was drilled into the skull at a location above the MePD. The stereotaxic injection coordinates used to target the MePD were obtained from the mouse brain atlas of

| Blood sampling procedure for LH measurement
Following a 1-week recovery period from surgery, the mice were handled daily to acclimatise them to the tail-tip blood sampling procedure for measurement of LH. 1 The blood samples were processed via an enzyme-linked immunosorbent assay as reported previously. 18 Mouse

| Validation of AAV injection site
Upon completion of experimental procedures, the mice were killed with a lethal dose of ketamine and transcardially perfused with heparinised saline for 5 min, followed by 10 min of ice-cold 4% paraformaldehyde in phosphate-buffered saline (pH 7.4) for 15 phosphate-buffered saline using a pump (Minipuls, Gilson, Villiers Le Bel, France). Their brains were rapidly collected and post-fixed sequentially at 4 C in 15% sucrose in 4% paraformaldehyde and in 30% sucrose in 1 Â phosphate-buffered saline until they sank. Brains were then snap frozen on dry ice and stored at À80 C until processing. Coronal brain slices (30 μm thick) were sectioned using a cryostat (Bright Instrument, Luton, UK). Every third section was collected between À1.34 mm and À2.79 mm from the bregma. Sections were mounted on glass microscope slides, air-dried and cover-slipped with Prolong Antifade mounting medium (Molecular Probes, Inc., Eugene, OR, USA). Only animals expressing enhanced yellow fluorescent protein (EYFP) in the MePD were analysed by using Axioskop 2 Plus microscope equipped with Axiovision, version 4.7 (Zeiss, Oberkochen, Germany).

| Statistical analysis
Appropriate sample sizes were predetermined by way of power analyses that were performed using the SigmaStat software (Systat Software Inc., San Jose, CA, USA) and expected variances and effect sizes that were based on our preliminary and published studies. 1 The Dynpeak algorithm was used to establish LH pulses. 19 The effect of optogenetic stimulation and neuropharmacology studies was established by comparing the mean LH interpulse interval (IPI), from the 1 h pre stimulation or drug administration control period to the 1.5 h experimental period. IPI are reported as change from the baseline condition for each mouse. On occasions where there were no LH pulses observed in the post treatment interval, the IPI was given a value of 90 min. LH pulse parameters were analysed by a two-way repeated measures analysis of variance (time and AAV condition and/or drug administration, respectively) and a subsequent Tukey's post-hoc test.
All statistics were performed using SigmaPlot, version 14 (Systat Software Inc.). Data are presented as the mean ± SEM. p < .05 was considered statistically significant.     Figure 2C). However, there were no statistic significant differences in the means between these two groups at the same time point (interaction between groups was not significant;     Only once the optic laser was switched on did the detection of LH pulses reliably cease. Therefore, it is reasonable to posit that, although this protocol indeed blocked GnRH pulse generator activity, this occurs via a mechanism of potential over-stimulation which sends the KNDy system into a state of inertia because it is unable to respond.

| In silico confirmation of the pulse generator responses to MePD regulation
The proposed hypothesis is that activation of MePD kisspeptin drives showed that administration of NMDA, a potent neuronal activator, evoked a multiunit electrical activity volley, the electrophysiological correlate of GnRH pulse generator activity 33 and corresponding LH pulse, followed by neuronal silence and cessation of LH pulses. Thus, the GnRH pulse generator is highly sensitive to incoming stimuli and may be prone to silencing by excessive activation. Importantly, glutamate antagonism alone did not result in a significant decrease in LH pulse frequency, and this is in line with the abovementioned theory of basal quiescence of the MePD kisspeptin system.
The present study also investigated the role of GABA B signalling in the activity of MePD kisspeptin and the GnRH pulse generator using CGP-35348, a GABA B R selective antagonist. By contrast to the significant reduction in LH pulse frequency observed with bicuculline and optic stimulation, the interference of GABA B R signalling in conjunction with optogenetics only went so far as to prevent the increase in LH pulsatility, indicating a present, yet smaller, influence. The reason for this difference remains unclear, but can be potentially explained by the pharmacological differences between GABA A and GABA B receptors, with the former accounting for fast inhibition, whereas the latter is responsible for slow inhibition. 34