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

Kisspeptin neurons in the arcuate nucleus of the hypothalamus generate 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 LH 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 pharmacological receptor antagonism, to investigate the neurotransmission involved in this circuitry. We have shown that functional neurotransmission of both GABAA and glutamate is a requirement for effective modulation of the GnRH pulse generator by amygdala kisspeptin neurons.


Introduction 2 9
Recent investigations have revealed a significant stimulatory role for kisspeptin in the 3 0 posterodorsal subnucleus of the medial amygdala (MePD) in GnRH pulse generator 3 1 modulation. Our previous optogenetic studies showed that sustained low-frequency blue 3 2 light stimulation activated these neurons to increase LH pulse frequency, a proxy for GnRH 3 3 pulse generator frequency (Lass et al., 2020). This finding built upon our previous 3 4 neuropharmacological approach using intra-MePD infusions of kisspeptin receptor (KISS1R) 3 5 agonists and antagonists, which respectively increased LH secretion or decreased LH pulse 3 6 frequency. (Comninos et al., 2016). However, the mechanisms underlying this neuronal 3 7 population's stimulatory role over pulsatile LH secretion have not been studied. Glutamate 3 8 and GABA are the major stimulatory and inhibitory neurotransmitters in the mammalian 3 9 brain, and many neuronal networks rely on the balance between these two to regulate their 4 0 activity (Petroff, 2002). Therefore, these two neurotransmitters are sensible candidates to 4 1 be used by the amygdala neuronal networks underlying the upstream, extra-hypothalamic 4 2 regulation of the GnRH pulse generator. Unsurprisingly, both GABA and glutamate neurons 4 3 are found in the MePD (Choi et al., 2005;Westberry and Meredith, 2016), and 4 4 pharmacological antagonism of both has deleterious effects on several aspects of 4 5 reproductive physiology (Oberlander et al., 2009;Polston et al., 2001, Li et al., 2015. In rats, 4 6 blocking AMPA and NMDA glutamate receptors with CNQX and AP5, respectively, impedes 4 7 activation of MePD neurons in response to vaginal-cervical stimulation, thereby preventing 4 8 the pregnancy or pseudopregnancy response following intromission; AMPA antagonism also 4 samples (4 µl) were collected every 5 mins for 2.5 h. The first hour of blood collection 1 3 3 consisted of no stimulation and in the subsequent 1.5 h, Kiss-Cre mice received optic 1 3 4 stimulation at 5 Hz (Lass et al., 2020). 1 3 5 For the neuropharmacological manipulation of GABA or glutamate receptor signalling with or 1 3 6 without simultaneous optogenetic stimulation, mice were connected to the laser as described 1 3 7 above, but additionally an injection cannula connected to extension tubing preloaded with 1 3 8 drug solution was inserted into the guide cannula of the optofluid implant immediately after 1 3 9 connection of the fibre optic cannula. Upon completion of experimental procedures, the mice were sacrificed with a lethal dose of 1 5 3 ketamine and transcardially perfused with heparinised saline for 5 mins, followed by 10 mins 1 5 4 of ice-cold 4 % paraformaldehyde in PBS (pH 7.4) for 15 mins using a pump (Minipuls, 1 5 5 Gilson, Villiers Le Bel, France). Their brains were rapidly collected and post-fixed 1 5 6 sequentially at 4 °C in 15% sucrose in 4% PFA and in 30% sucrose in 1 x PBS until they 1 5 7 sank. Brains were then snap frozen on dry ice and stored at -80 °C until processing. 1 5 8 Coronal brain slices (30 μ m thick) were sectioned using a cryostat (Bright Instrument, Luton, 1 5 9 UK). Every third section was collected between -1.34 mm to -2.79 mm from the bregma.
Sections were mounted on glass microscope slides, air-dried and cover-slipped with Prolong 1 6 1 Antifade mounting medium (Molecular Probes, Inc, OR, USA). Only animals expressing 1 6 2 EYFP fluorescent protein in the MePD were analysed by using Axioskop 2 Plus microscope 1 6 3 equipped with Axiovision version 4.7 (Zeiss). 1 6 4 Statistical analysis 1 6 5 The Dynpeak algorithm was used to establish LH pulses (Vidal et al., 2012). The effect of 1 6 6 optogenetic stimulation and neuropharmacology studies was established by comparing the 1 6 7 mean LH inter-pulse interval (IPI), from the 1 h pre stimulation or drug administration control 1 6 8 period to the 1.5 h experimental period. On occasions where there were no LH pulses 1 6 9 observed in the post treatment interval the IPI was given a value of 90 mins. LH pulse 1 7 0 parameters were analysed by a two-way repeated measures ANOVA and subsequent Tukey 1 7 1 post-hoc test. All statistics were performed using SigmaPlot (version 14). The threshold 1 7 2 level for statistical significance was set at P < 0.05 with data presented as mean ± SEM. 1 7 3

7 4
Validation of AAV injection site and cannula position 1 7 5 The AAV-ChR2 virus used to infect the cells in Kiss-Cre mice was tagged with fluorescent 1 7 6 EYFP in order to be visualised under a microscope. The mean number of tdTomato-1 7 7 expressing kisspeptin cells in unilaterally-injected brain sections was 24.50 ± 5.20 1 7 8 (mean+SEM) per animal and 20.33 ± 4.50 (~83%) of tdTomato-expressing neurons were 1 7 9 EYFP fluorescence positive. Analysis of images acquired from coronal sectioning of the 1 8 0 mouse brains showed that 7 out of 9 animals had successful stereotaxic injection of AAV-1 8 1 ChR2 virus into the MePD, and all 7 also had successful cannula implantation into the 1 8 2 MePD. A representative example of a coronal brain section is shown in Figure 1 Effects of sustained optical stimulation at 5 Hz, with and without a control 1 8 6 administration of aCSF, on LH pulse frequency 1 8 7 After a 1 h control blood sampling period in the absence of optical stimulation, Kiss-Cre mice 1 8 8 were stimulated at 5 Hz for 90 mins with and without administration of aCSF. In both 1 8 9 experimental protocols, the stimulation resulted in a significant increase in LH pulse 1 9 0 frequency ( Fig. 2A and B). The mean inter-pulse interval (IPI) decreased from 25.00 ± 2.99 1 9 1 mins to 18.60 ± 2.07 mins (n = 7; p = 0.002) after 5 Hz stimulation only, and from 31.25 ± 1 9 2 1.25 mins to 22.73 ± 1.94 mins (n = 4; p = 0.002) after 5 Hz stimulation and infusion of aCSF 1 9 3 (Fig.2C). 1 9 4 Effects of bicuculline, a GABA A R antagonist, on LH pulse frequency in the presence 1 9 5 and absence of sustained 5 Hz optical stimulation 1 9 6 In the second part of the experiment, Kiss-Cre mice received a unilateral intra-MePD 1 9 7 infusion of bicuculline with and without sustained 5 Hz optical stimulation. After the dual 1 9 8 treatment of light and bicuculline, the average LH IPI significantly increased from 27.00 ± 1 9 9 2.55 mins to 42.42 ±. to 42.42 ± 7.39 mins (n = 5; p = 0.008), indicating reduced LH pulse 2 0 0 frequency ( Fig. 2D and F). There was no significant change in the IPIs before and after 2 0 1 bicuculline administration alone, with a pre-infusion IPI of 29.00 ± 3.78 mins and a post-2 0 2 infusion IPI of 29.75 ± 3.78 mins (n = 5; p = 0.771) ( Fig. 2E and F). In the third part of the experiment, Kiss-Cre mice received a unilateral infusion of CGP-2 0 6 35348, an antagonist selective for the GABA B R, both with and without continuous 2 0 7 optogenetic stimulation. There was no significant difference in LH IPI between the initial 60 2 0 8 mins control period and subsequent period of intervention in either of these experimental 2 0 9 protocols, however a trend towards increased LH pulse frequency during administration of 2 1 0 CGP-35348 alone in the absence of light was observed ( Fig. 2G-I). The pre-and post-2 1 1 intervention average IPIs for CGP-35348 with 5 Hz stimulation were 30.80 ± 4.67 mins and 2 1 2 30.53 ± 5.10 mins, respectively (n = 5; p = 0.954). For CGP-35348 administration alone, the 2 1 3 pre-and post-intervention average IPIs were 26.67 ± 2.28 mins and 21.25 ± 0.72 mins, 2 1 4 respectively (n = 5; p = 0.115).

1 5
Effects of glutamate receptor antagonism on LH pulse frequency with and without 2 1 6 continuous 5 Hz optic stimulation 2 1 7 The final protocol of the experiment involved unilateral infusion of a drug cocktail consisting 2 1 8 of both AP5 and CNQX, antagonists for AMPA and NMDA receptors, respectively, in the 2 1 9 presence and absence of light in Kiss-Cre mice ( Fig. 2J-L). After the 60 mins control period, 2 2 0 sustained 5 Hz optogenetic stimulation together with infusion of the antagonist resulted in a 2 2 1 significantly decreased LH pulse frequency; indeed, in a number of cases LH pulsatility 2 2 2 ceased altogether and the average IPI increased from 27.33 ± 1.89 mins to 69.50 ± 10.26 2 2 3 mins (n = 6; p = 0.007). For AP5 and CNQX alone, there was a trend of increased IPI before 2 2 4 and after treatment from 23.04 ± 3.24 mins to 38.59 ± 15.73 mins, however this was not 2 2 5 significant (n = 4; p = 0.253).

6
Discussion 2 2 7 The present study highlights for the first time a potential mechanism by which kisspeptin 2 2 8 activity in the MePD is stimulatory over the hypothalamic GnRH pulse generator. Building 2 2 9 upon previous findings that low-frequency (5 Hz) optogenetic stimulation of MePD Kiss1 2 3 0 neurons increases the frequency of pulsatile LH secretion (Lass et al., 2020), the results 2 3 1 presented in this study support the hypothesis that this is reliant upon the activity of both 2 3 2 GABA and glutamate in the MePD. 2 3 3 It has been shown that the MePD, with its overwhelmingly GABAergic neuronal outputs, is a 2 3 4 significant inhibitor of gonadotrophic hormone secretion and wider facets of reproductive 2 3 5 physiology; its stimulation and lesioning delays and advances pubertal onset, respectively, 2 3 6 and the MePD's activation during stress is deleterious on reproductive function and 2 3 7 behaviour (Elwers and Critchlow, 1960;Bar-Sela and Critchlow, 1966;Lin et al, 2011).
However, the recent findings that MePD Kiss1 neurons stimulate the GnRH pulse generator 2 3 9 raises the hypothesis for intranuclear GABA-GABA disinhibitory interactions, typical of limbic 2 4 0 pallidal structures such as the MePD. The current study tested this utilising specific 2 4 1 optogenetic activation of MePD kisspeptin neurons in conjunction with bicuculline and CGP-2 4 2 35348 -intra-MePD antagonists for GABA A R and GABA B R, respectively. Indeed, either of 2 4 3 these drugs together with optic stimulation prevented the increase in LH pulse frequency 2 4 4 seen with 5 Hz stimulation alone. The result of intra-MePD infusion of bicuculline together 2 4 5 with 5 Hz optogenetic stimulation is a surprising one and poses an interesting question: how 2 4 6 can intra-MePD GABA A R antagonism not only prevent the stimulatory effects of optogenetic 2 4 7 stimulation, but in fact cause the opposite result of significantly reducing the activity of the 2 4 8 GnRH pulse generator, while bicuculline alone had no effect on LH pulse frequency? We 2 4 9 suggest a neuronal circuit, involving glutamatergic synaptic mechanisms, that may underly 2 5 0 this phenomenon ( Figure 3). Indeed, a stimulatory effect of the MePD kisspeptinergic 2 5 1 system over the GnRH pulse generator has been linked to glutamatergic activation; the 2 5 2 pubertal transition is tightly correlated with a developmental increase in the expression of almost-complete ablation of pulsatile LH secretion following MePD kisspeptin activation 2 5 9 combined with infusion of intranuclear glutamate antagonists provides further support to 2 6 0 MePD kisspeptin effect's dependence on glutamate. Therefore, the following mechanism is 2 6 1 proposed: optical stimulation of MePD kisspeptin in the presence of GABA A R antagonists 2 6 2 decreases LH pulse frequency by glutamatergically activating the hypothetical GABAergic 2 6 3 projections from the MePD to KNDy neurons; whether this occurs via glutamate secretion 2 6 4 from MePD kisspeptin cells themselves is unknown. In other words, cancelling the ability of 2 6 5 GABA interneurons to take part in the disinhibition during optical stimulation shifts the 2 6 6 balance from a stimulatory to inhibitory output from the MePD. The fact that GABA A R antagonism alone failed to cause any change in LH pulse frequency 2 6 8 is an important factor in our model: it is possible that under basal conditions MePD 2 6 9 kisspeptin neurons are relatively quiescent, and therefore solely pharmacologically blocking 2 7 0 the inputs to the GABAergic MePD projections without a corresponding increase in 2 7 1 glutamatergic activity would make little change to the net influence of the MePD over the 2 7 2 KNDy system. Although silent kisspeptin signalling under basal conditions supports the 2 7 3 current hypothesised model, it contradicts neuropharmacology studies that show 2 7 4 antagonising endogenous kisspeptin within the MePD causes a robust decrease in LH pulse 2 7 5 frequency (Comninos et al., 2016). However, this latter study used OVX rats that were 2 7 6 supplemented with 17 β -estradiol to mimic the hormonal profile in the diestrus phase of the 2 7 7 estrous cycle. It is important to note, the present study used OVX mice which were not 2 7 8 supplemented with 17 β -estradiol. Kiss1 expression within the MePD varies in relation to the 2 7 9 estrous cycle with low expression observed in OVX mice; estradiol treatment, however, 2 8 0 amplified Kiss1 expression in this brain region (Kim et al., 2011). This may explain why 2 8 1 under basal conditions the MePD kisspeptin system appears reduced in our OVX mouse 2 8 2 model.

8 3
Therefore, the proposed model does well to explain why GABA A R antagonism in the 2 8 4 presence of MePD Kiss1 optical stimulation reduces LH pulse frequency. However, a 2 8 5 possible explanation as to why optogenetically stimulating these neurons in the presence of 2 8 6 intra-MePD glutamate antagonists essentially stops all GnRH pulse generator activity is 2 8 7 more complex. Complying with the model would suggest that blocking glutamate activity, 2 8 8 while activating the GABA-GABA disinhibitory system, would in fact increase LH pulse 2 8 9 frequency rather than prohibiting it altogether. Nevertheless, we provide a potential 2 9 0 explanation for this phenomenon. By examining all of the individual pulse profiles more 2 9 1 closely, a subtle, but potentially crucial aspect is identified. In over 80% (5 out of 6) of tests 2 9 2 in which glutamate antagonists were infused in conjunction with optical stimulation, a pulse 2 9 3 of LH was detected precisely 10 mins after the bolus infusion, and immediately before the 2 9 4 onset of light stimulation. Only once the optic laser was switched on did the detection of LH 2 9 5 pulses reliably cease. Therefore, it is reasonable to posit that while this protocol indeed 2 9 6 blocked GnRH pulse generator activity, this occurs via a mechanism of potential over-2 9 7 stimulation which sends the KNDy system into a state of inertia as it is unable to respond. 2 9 8 The proposed hypothesis is that activation of MePD kisspeptin drives i) the disinhibitory 2 9 9 GABA-GABA pathway from the MePD, ii) glutamatergic interneurons that in turn project to 3 0 0 the GABA-GABA pathway, and iii) glutamatergic projections from the MePD onto the ARC (a 3 0 1 schematic diagram describing this is shown in Figure 3A). Thus, optogenetic stimulation of 3 0 2 the MePD Kiss1 neurons combined with antagonism of MePD glutamate results in the net 3 0 3 effect of heightened activation of the GnRH pulse generator and resultant inertia ( Figure 3B). 3 0 4 The proposal of excessive GnRH pulse generator neuronal activity resulting in depolarisation 3 0 5 silencing is in line with recent findings from our research group. Using mathematical models 3 0 6 confirmed with in vivo optogenetics, it is now known that the ultradian oscillation of the 3 0 7 hypothalamic KNDy network works on a bifurcation system that is eventually terminated as Knobil, 1993) and corresponding LH pulse, followed by neuronal silence and cessation of LH 3 1 6 pulses. Thus, the GnRH pulse generator is highly sensitive to incoming stimuli and may be 3 1 7 prone to silencing by excessive activation. Importantly, glutamate antagonism alone did not 3 1 8 result in a significant decrease in LH pulse frequency, and this is in line with the 3 1 9 abovementioned theory of basal quiescence of the MePD kisspeptin system. 3 2 0 The present study also investigated the role of GABA B signalling in the activity of MePD 3 2 1 kisspeptin and the GnRH pulse generator using CGP-35348, a GABA B R selective 3 2 2 antagonist. In contrast to the significant reduction in LH pulse frequency observed with 3 2 3 bicuculline and optic stimulation, the interference of GABA B R signalling in conjunction with 3 2 4 optogenetics only went so far as to prevent the increase in LH pulsatility, indicating a 3 2 5 present, yet smaller, influence. The reason for this difference remains unclear, but can be 3 2 6 potentially explained by the pharmacological differences between GABA A and GABA B 3 2 7 receptors, with the former accounting for fast inhibition while the latter is responsible for slow 3 2 8 inhibition (Nicoll et al., 1990). Moreover, it has been shown that in the case of LH release, 3 2 9 only ICV activation of the GABA A R, and not GABA B R, both reduced LH release from the 3 3 0 pituitary and GnRH levels in the POA (Leonhardt et al., 1995), suggesting a differential role 3 3 1 for the two receptor subtypes in reproductive neuroendocrinology. Moreover, while it has 3 3 2 been shown that knockout of the GABA B R subtype in adult female mice results in subfertile 3 3 3 phenotypes such as decreased hypothalamic levels of GnRH and GnRH mRNA, it has no 3 3 4 diminishing effects on LH or FSH levels, or Kiss1 expression in the hypothalamus (Catalano These data have demonstrated, for the first time, the possible neuronal mechanisms by 3 4 0 which increased kisspeptinergic activity within the amygdala increases GnRH pulse 3 4 1 frequency, which could also provide the basis for the sexual development of puberty. It 3 4 2 would be of interest to further investigate the roles of GABA and glutamate in this network, 3 4 3 including in relation to perturbations of reproductive physiology associated with the 3 4 4 amygdala such as stress and abnormal food intake. (www.kcl.ac.uk/library/researchsupport/research-data-management/preserve/deposit-your-4 5 0 data-with-kings3  G  A  B  A  a  n  d  g  l  u  t  a  m  a  t  e  r  e  c  e  p  t  o  r  a  n  t  a  g  o  n  i  s  m  w  i  t  h  a  n  d  w  i  t  h  o  u  t  s  u  s  t  a  i  n  e  d  o  p  t  o  g  e  n  e  t  i  c  s  t  i  m  u  l  a  t  i  o  n  o  f  M  e  P  D  K  i  s  s  1  n  e  u  r  o  n  e  s  o  n  L  H  p  u  l  s  e  f  r  e  q  u  e  n  c  y  .   R  e  p  r  e  s  e  n  t  a  t  i  v  e  e  x  a  m  p  l  e  s  s  h  o  w  i  n  g  t  h  e  e  f  f  e  c  t  s  o  f  5  H  z  s  t  i  m  u  l  a  t  i  o  n  (  A  )  ,  5  H  z  +  a  C  S  F  (  B  )  ,  b  i  c  u  c  u  l  l  i  n  e  +  5  H  z  (  D  )  ,  C  G  P  -3  5  3  4  8  +  5  H  z  (  G  )  ,  A  P  5  +  C  N  Q  X  +  5  H  z  (  J  )  o  n  p  u  l  s  a  t  i  l  e  L  H  s  e  c  r  e  t  i  o  n  i  n  o  v  a  r  i  e  c  t  o  m  i  s  e  d  m  i  c  e  .  R  e  p  r  e  s  e  n  t  a  t  i  v  e  e  x  a  m  p  l  e  s  f  o  r  t  h  e  e  f  f  e  c  t  s  o  f  d  r  u  g  s  a  l  o  n  e  b  i  c  u  c  u  l  l  i  n  e  (  E  ) ,  F  I  G  U  R  E  3  .  M  o  d  e  l  s  h  o  w  i  n  g  t  h  e  p  r  o  p  o  s  e  d  i  n  t  e  r  a  c  t  i  o  n  s  a  n  d  p  a  t  h  w  a  y  s  i  n  v  o  l  v  e  d  i  n  t  h  e  M  e  P  D  r  e  g  u  l  a  t  i  o  n  o  v  e  r  t  h  e  h  y  p  o  t  h  a  l  a  m  i  c  G  n  R  H  p  u  l  s  e  g  e  n  e  r  a  t  o  r  (  A  R  C   K  N  D  y ) . (  A  )  A  c  c  o  r  d  i  n  g  t  o  t  h  e   h  y  p  o  t  h  e  s  i  s  m  o  d  e  l  ,  k  i  s  s  p  e  p  t  i  n  i  n  t  h  e  M  e  P  D  r  e  g  u  l  a  t  e  s  t  h  e  G  n  R  H  p  u  l  s  e  g  e  n  e  r  a  t  o  r  b  y  a  c  t  i  v  a  t  i  n  g   t  w  o  p  a  t  h  w  a  y  s  :  i  )  a  G  A  B  A  -G  A  B  A  d  i  s  i  n  h  i  b  i  t  o  r  y  p  a  t  h  w  a  y  a  n  d  i  i  )