NKB Signaling in the Medial Amygdala Stimulates Gonadotropin Release in a Kisspeptin-Independent Manner in Female Mice

Neurokinin B (NKB) signaling is critical for reproduction in all studied species. The existing consensus is that NKB induces GnRH release via kisspeptin (Kiss1) stimulation in the arcuate nucleus. However, the stimulatory action of NKB is dependent on circulating estrogen (E2) levels, without which, NKB inhibits LH release. Importantly, the evidence supporting the kisspeptin-dependent role of NKB, derives from models of persistent hypogonadal state [e.g. Kiss1r knockout (KO) mice], with reduced E2 levels. Here, we demonstrate that in the presence of E2, NKB signaling induces LH release in a kisspeptin-independent manner. Moreover, senktide (NKB receptor agonist) delivery to the medial amygdala (MeA) increases LH in E2-treated Kiss1 KO females (but not males or sham-treated females) similar to controls, and thus, this increase is independent of Kiss1 neurons. These results document a novel kisspeptin-independent regulatory pathway of reproductive function in females mediated by NKB-responsive neurons in the MeA.

showed dense intermingling and multiple foci of close apposition with NK3R containing cells and 149 fibers, but no co-expression within GnRH cell bodies (>100 cells analyzed from a total of 16 mice; 150 Figure 3C). Interestingly, we observed no fibers containing NK3R immunoreactivity in the internal 151 or external zone of the median eminence (ME; Figure 3A, 3B). 152 To identify the brain area in which NK3R receptive neurons that mediate the kisspeptin-153 independent GnRH release reside, we stereotaxically administered senktide specifically into the 154 ARC, the POA (at the level of the MS) or the MeA. These areas are prime candidates to play a 155 role in LH stimulation because (a) they contain NK3R, the immunoreactivity of which is regulated 156 by E2, (b) there is an anatomical overlap of NK3R and GnRH protein, at least in the MS and ARC 157 and (c) they contain GnRH and/or Kiss1 cell bodies and fibers and are known to play an important 158 role in reproductive function (Smith et al., 2006, Kim et al., 2011. Senktide administration into the 159 ARC or POA of WTOVX+E2 mice stimulated LH secretion within 15 min from drug infusion ( Figure  160 3D, 3E; P<0.0001 for both) compared to Kiss1 KO animals. However, when senktide was 161 administered into the MeA of WTOVX+E2 and Kiss1 KO females supplemented with E2 a robust 162 increase in LH was observed within 15 min after senktide infusion that was similar in both 163 genotypes ( Figure 4D). Conversely, in the absence of E2, Kiss1 KO females did not show any 164 alteration in LH release ( Figure 4D), mimicking the LH responses we obtained after an ICV 165 injection of senktide in these animals ( Figure 1B). 166

Chemogenetic activation of the MeA Kiss1 neuron stimulates LH release in WT but not 167
Kiss1 KO female mice. 168 Similar to the experiments described above, we delivered Cre-dependent AAV5-DIO-169 hM3Dq:mCherry to the MeA of Kiss1 Cre/+ or Kiss1 Cre/Cre mice. HM3Dq:mCherry expression was 170 present in the MeA ( Figure 4F) and was limited to sections ranging from (from -1.6 mm to -2.0 171 mm from bregma; paxinos atlas). Two out of eight mice (one from each group) had primarily 172 unilateral spread of the DREADD, but they did not differ significantly from their respective groups 173 in LH concentrations, and were therefore, included in further analyses. Within the MeA, mCherry 174 cell bodies were co-expressed in ~89 % of GFP-immunoreactive cells (i.e. Kiss1 cells), and was 175 not observed in non-GFP cells. 176 Kiss1 Cre/+ mice expressing hM3Dq:mCherry in the MeA and treated with CNO to activate 177 the Kiss1 MeA neurons, had an increase in LH within 30 min after the injection (P=0.0107) compared 178 to animals receiving saline treatment, which was sustained for another 30 min before returning to 179 basal levels ( Figure 4G). No alteration in LH was observed in E2-supplemented Kiss1 KO animals 180 treated with either saline or CNO ( Figure 4H). 181

182
Our results provide evidence that the MeA is a component of the gonadotropin axis. 183 Specifically, we have identified two independent pathways within the MeA that can lead to the 184 stimulation of GnRH/LH release. The first involves Kiss1 MeA neurons, the activation of which, 185 stimulates LH release into the peripheral circulation. Furthermore, this is achieved by the release 186 of kisspeptin and not by any other signaling molecules produced within the Kiss1 MeA neuron, since 187 LH was increased only in animals with an intact kisspeptin signaling system. A second pathway, 188 involving NKB/NK3R signaling, was also identified, when senktide (NK3R agonist) administration 189 into the MeA induced LH release in Kiss1 KO mice (Padilla et al., 2018). Thus, kisspeptin is not 190 a required mediator between NK3R activation in the MeA and LH secretion. Interestingly, this 191 pathway is female-specific and estrogen-dependent as responses were absent in males and 192 hypogonadal females. 193 From a mechanistic point of view, the most likely kisspeptin-independent pathway for LH 194 stimulation by NKB would involve the direct regulation of GnRH release (Krajewski et al., 2005). 195 Despite there being an anatomical overlap of GnRH and NK3R protein, specifically in the ARC 196 and POA (at the level of the MS), we observed no instances of colocalization between NK3R 9 and GnRH cell bodies irrespective of the presence or absence of sex steroids. This reveals 198 certain anatomical differences to what has been previously demonstrated in the rat, where 199 ~16% of GnRH cell bodies were found to contain NK3R protein (Krajewski et al., 2005). slices, that senktide induces GnRH release from the ME and this effect is, in part, present in Kiss1 208 KO mice (Gaskins et al., 2013). We did not observe any NK3R immunoreactive fibers in the 209 internal or external zone of the median eminence, indicating a potential lack of direct NKB (or 210 senktide) regulation of the GnRH terminals in that area. Nonetheless, other signaling molecules 211 such as glutamate (Nestor et al., 2016) or galanin, γ-aminobutiric acid (Skrapits et al., 2015) can 212 potentially stimulate LH secretion and must also be considered. However, activation of the 213 Kiss1 ARC (KNDy) neuron stimulated LH release only in mice with an intact kisspeptin signaling 214 system and was completely absent in Kiss1 KO mice. This provides evidence that kisspeptin, but 215 no other signaling molecule produced by Kiss1 ARC (KNDy) neuron can stimulate LH release in 216

vivo. 217
The distribution of NK3R has been described in the human, rat, and ewe (Mileusnic et al., the mouse brain. Interestingly, in certain areas the immunoreactivity of NK3R-containing cell 220 bodies was highly dependent on sex steroid levels. Specifically, estrogen downregulated NK3R 221 expression in the ARC whereas the opposite was true for the MeA, with more NK3R containing 222 cell bodies evident when animals were supplemented with E2. High sensitivity of NK3R expression 223 to E2, has also been reported for the ARC with in situ hybridization studies (Navarro et al., 2009). 224 Interestingly, this regulation of NK3R expression is reminiscent of the regulation of Kiss1 by E2 in 225 these areas (Kim et al., 2011, Smith et al., 2006. indicating that this kisspeptin-independent NKB/NK3R signaling mechanism in the MeA becomes 248 activated only when E2 is present. This notion is further supported by our finding that the number 249 of NK3R cells increases with E2, and this upregulation is specific to the MeA. NK3R MeA expressing 250 cells do not co-localize with NKB, but are surrounded by a plethora of NKB fibers (Supplemental 251 it is compelling to hypothesize that the kisspeptin-independent, NK3R-dependent pathway is 275 employed for the generation and/or enhancement of the LH surge and/or female sexual behavior, 276 e.g. lordosis, given that this mechanism was absent in male mice, and is exclusively activated in 277 the presence of estrogen, similar to what is observed in the female AVPV/PeN (Smith et al., 2006). 278 In accordance, recent evidence demonstrated the enhancement of the LH surge in rats exposed 279 Rodent Diet 8664) and were given ad libitum access to tap water. For all studies, C57Bl/6 WT or 301 Kiss1 Cre/+ (heterozygous state) males or females between age 8 and 20 weeks were used and 302 studied in parallel to Kiss1 Cre/Cre (Kiss 1 knock-out state) littermates. In order to test the specificity 303 of the NK3R antibody, NK3RKO mice were used as described below. 304

Experiment 1: Effect of central (ICV) administration of senktide on LH release in male and 305 female WT and Kiss1 KO mice with or without the presence of sex steroids. 306
In this experiment we aimed to assess whether central activation of ΝΚ3R signaling with 307 senktide (an NK3R specific agonist), can stimulate LH release in Kiss1 KO mice (i.e., in a 308 kisspeptin-independent manner) in the presence or absence of sex steroids. Adult WT male and 309 female mice were GND and studied in parallel to hypogonadal (with low sex steroid levels) Kiss1 310 KO littermates (n=10/group). ICV injections (see below) of senktide (Tocris Biosience, Cat. No. 311 1068; 600 pmol diluted in 5µl 0.9% NaCl) were performed and blood samples were collected 312 before (basal) and 25 min after ICV injection for LH measurements as has been previously 313 described (Navarro et al., 2015). Next, animals were implanted with sex steroids (n=10/group) 314 and the ICV experiment was repeated a week later. The dose of senktide used, and the time of 315 blood collection were selected based on our previous studies (Navarro et al., 2015). 316

Experiment 2: Effect of ARC KNDy neuron chemogenetic activation on LH release in WT 317 and Kiss1 KO female mice in the presence of estradiol. 318
In order to determine whether the release of other components, besides kisspeptin, within 319 the Kiss1 ARC (KNDy) neuron can stimulate LH release, we used a chemogenetic approach to 320 specifically activate Kiss1 ARC neurons of Kiss1 Cre/+ or Kiss1 Cre/Cre mice treated with E2 (n=5- control, and to confirm that animals were appropriately treated and primed. 332

Experiment 3: Effect of senktide administration in to the ARC, POA or MeA on LH release 333 in female WT and Kiss1 KO mice with the presence of estrogen. 334
In this set of experiments, we aimed to locate the brain area which senktide is acting to 335 stimulate LH release. To this end, we first conducted neuroanatomical studies to confirm NK3R 336 protein expression in the mouse hypothalamus, as well as to investigate the potential anatomical