Mice expressing the genetically encoded calcium indicator GCaMP6f (47) in kisspeptin neurons were generated by crossing mice that express the Cre recombinase enzyme (Cre) in kisspeptin cells (Kiss1-Cre; Jackson laboratory stock #023426) (48) with mice that express Cre-dependent GCaMP6f at the ROSA26 locus (flox-STOP-GCaMP6f; Jackson laboratory stock #028865) (49). Male and female offspring heterozygous for the Kiss1-Cre and flox-STOP-GCaMP6f alleles (Kiss1-Cre::GCaMP6f) were used in calcium imaging experiments. Female mice expressing the humanized renilla green fluorescent protein (hrGFP) in kisspeptin neurons (Kiss1-hrGFP; Jackson laboratory stock #023426) (50) were used in electrophysiology experiments. All experimental mice were adults (2-6 months). Female estrous cycle stage was determined by vaginal lavage (5 µL H₂O) taken between zeitgeber time (ZT) 1 and 3 or post-mortem (ZT2.5-5.5). Aqueous vaginal smears were stained with methylene blue and cytology examined under light microscopy to assess estrous cycle stage (51). Mice were group-housed with littermates under controlled temperature (23 ± 2°C) and lighting (12h light/dark) conditions with ad libitum access to food and water. Mice were assigned to experiments based on their genotype, sex, and on their estrous cycle stage as needed. All experiments were approved by Kent State University’s Institutional Animal Care and Use Committee.

To visualize the distribution of GCaMP6f-expressing neurons in the RP3V of male and female mice, two male and two female (one diestrus and one proestrus) Kiss1-Cre::GCaMP6f mice were deeply anesthetized via intraperitoneal injection of 3 mg/mL pentobarbital before being perfused through the heart with 4% paraformaldehyde (PFA). The brains were then dissected out and placed in 4% PFA for 1 hour before being transferred to 20% sucrose in 0.1 M phosphate-buffered saline. The brains were then cut in 3 series of coronal sections (30 μm thick) using a freezing microtome. The tissue was mounted onto superfrost charged slices, air dried, and coverslipped using an aqueous mounting medium (ProLong™ Gold antifade mountant, Thermofisher). Epifluorescence images were taken in the RP3V region using a DMR microscope (Leica Microsystems) equipped with an ORCA-FLASH 4.0 V3 Digital CMOS (Hamamatsu, Japan), controlled by the MicroBrightField Neurolucida Software (MBF Bioscience).

Brain slices were prepared as previously described (52, 53). Briefly, mice were decapitated following isoflurane anesthesia, and their brains quickly removed. Coronal brain slices (200 μm thick) containing the RP3V were cut using a vibrating blade microtome (HM650V, Microm International GmbH) in an ice-cold solution containing (in mM): 87 NaCl, 2.5 KCl, 25 NaHCO₃, 1.25 NaH₂PO₄, 0.5 CaCl₂, 6 MgCl₂, 25 glucose and 75 sucrose. Brain slices were left to incubate at 30-34°C for at least 1 hour in artificial cerebrospinal fluid (aCSF) containing (in mM): 125 NaCl, 2.5 KCl, 26 NaHCO₃, 1.25 NaH₂PO₄, 2.5 CaCl₂, 1.2 MgCl₂ and 11 glucose. All solutions were equilibrated to pH 7.4 with a mixture of 95% O₂/5% CO₂. Female mice were killed between ZT2.5 and 5.5 and males between ZT2 and 4.5.

Individual brain slices obtained from Kiss1-Cre::GCaMP6f or Kiss1-hrGFP mice were placed under an upright epifluorescence microscope (either Scientifica, UK or Prior Scientific, UK) and constantly perfused (1.5 mL/min) with warm (32-34°C) aCSF.

Variations in intracellular calcium concentration ([Ca²⁺]_i) in RP3V^KISS1 neurons were estimated by measuring fluorescence changes in individual GCaMP6f-expressing RP3V neurons in slices from Kiss1-Cre::GCaMP6f mice. Slices were illuminated through a 40x immersion objective, using a light-emitting diode light source (pE300^ultra LED; CoolLED, UK) filtered for blue light excitation (460-487 nm; Semrock, USA). Epifluorescence (emission 500-546 nm; Semrock, USA) was collected using an ORCA-FLASH 4.0 LT+ CMOS camera (Hamamatsu). LED and camera were controlled and synchronized with the μ-manager 1.4 software (54). After a ≥15-minute stabilization period in the recording chamber, a focal plane including several fluorescent cell bodies was chosen and acquisitions (100 ms light exposure at 2 Hz for 10-15 minutes) started. Low intensity LED illumination was used (≈0.1 to 0.8 mW) to minimize GCaMP6f photobleaching.

RP3V^KISS1 neuron spontaneous action potential firing was recorded in brain slices from Kiss1-hrGFP mice. GFP-expressing RP3V neurons were visualized using brief LED illumination (excitation and emission as above) and subsequently approached with glass recording micropipettes using infrared differential interference contrast illumination. Recording micropipettes (tip resistance: 3-6 MΩ) were made with borosilicate glass (cat. #BF150-110-7.5, Sutter Instruments, USA), pulled using a Model P-1000 micropipette puller (Sutter Instruments). Action potential firing was recorded in voltage-clamp mode (no holding potential applied) in the minimally invasive cell-attached configuration (12-30 MΩ initial seal resistance). Glass micropipettes were filled with aCSF and the recording configuration was achieved by applying the lowest amount of suction required to detect spontaneous, fast, downward deflections in the current trace (spikes), which correspond to single action potentials (55). Electrical signals were recorded, filtered at 2 kHz, and digitized at 10 to 20 kHz using a double integrated patch amplifier (Sutter Instruments, USA), and acquired with the SutterPatch software (Sutter Instruments, USA).

All calcium imaging and electrophysiology experiments were performed between ZT3.5 and 10.

All drugs were dissolved to the appropriate stock concentration in water or in DMSO, aliquoted and stored at -20°C. Stock were diluted to working concentrations in aCSF prior to performing experiments. Final DMSO concentration never exceeded 0.1%. All drugs were bath-applied. Agonists were applied for one or two minutes after a ≥ three-minute baseline period in electrophysiology and in calcium imaging experiments. Antagonists and blockers were applied continuously to the slice for ≥ five minutes before a recording started and for the duration of the recording. In calcium imaging experiments, cell viability was routinely tested at the end of the experiments by applying the glutamate receptor agonists (S)-α-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA; 10-20 μM) or L-glutamate (100 μM), or by raising extracellular [K⁺] (+ 10 mM KCl).

AMPA (cat. #0254), the 5-HT₂ receptor antagonists 6-[2-[4-[Bis(4-fluorophenyl)methylene]-1-piperidinyl]ethyl]-7-methyl-5H-thiazolo[3,2-a]pyrimidin-5-one (ritanserin; cat. #1955) and 1,2,3,4,10,14b-Hexahydro-2-methyldibenzo[c,f]pyrazino[1,2-a]azepine hydrochloride (mianserin; cat. #0997), the 5-HT_2A receptor agonist (4-Bromo-3,6-dimethoxybenzocyclobuten-1-yl)methylamine hydrobromide (TCB-2; cat. #2592), the 5-HT_2C receptor agonist 8,9-Dichloro-2,3,4,4a-tetrahydro-1H-pyrazino[1,2-a]quinoxalin-5(6H)-one hydrochloride (WAY161503; cat. #1801), the 5-HT₃ receptor agonist 1-(6-Chloro-2-pyridinyl)-4-piperidinamine hydrochloride (SR57227; cat. #1205), the 5-HT₄ receptor agonist (±)-4-Amino-5-chloro-N-[1-[(3R*,4S*)-3-(4-fluorophenoxy)propyl]-3-methoxy-4-piperidinyl]-2-methoxybenzamide (cisapride; cat. #1695), the 5-HT₆ receptor agonist 3-[(-3-Fluorophenyl)sulfonyl]-N,N-dimethyl-1H-pyrrolo[2,3-b]pyridine-1-ethanamine dihydrochloride (WAY208466; cat. #3904), the 5-HT₇ receptor agonist (2S)-5-(1,3,5-Trimethylpyrazol-4-yl)-2-(dimethylamino)tetralin (AS19; cat. #1968) and 5-HT hydrochloride (cat. #3547) were purchased from Tocris (Bio-techne, USA). L-glutamate (cat. #HB0383) was purchased from HelloBio (USA) and tetrodotoxin citrate (cat. #T-550) from Alomone labs (Israel).

For calcium imaging, GCaMP6f fluorescence image time-series were processed in the FIJI software (56). Regions of interest (ROIs) were drawn around individual, in-focus fluorescent somata [12.1 ± 0.5 (ranging 4 to 24) and 9.2 ± 1.0 (ranging 7 to 14) ROIs per slice in females and males, respectively]. Mean fluorescence intensity within each ROI was measured in each frame. Fluorescence intensity data were analyzed using scripts written in R (http://www.r-project.org/). For each ROI, normalized fluorescence was calculated as F t F × 100 , where F is the baseline fluorescence intensity calculated as the mean fluorescence intensity over a one-minute period preceding agonist applications and F_t is the fluorescence measured at any time point. Time-dependent bleaching was corrected by subtracting the normalized fluorescence obtained in an ROI that did not contain a fluorescent cell within the RP3V (empty ROI). This approach could not be used for recordings of the effect of AMPA, L-glutamate or KCl as these treatments induced very large increases in fluorescence that contaminated the empty ROI, thereby causing unacceptable distortions of traces upon subtraction. ROIs in which normalized fluorescence traces reversibly increased from baseline for a period of at least 2 minutes, around the time of agonist application, were recorded as being excited. For illustration and statistical comparisons, agonist effects were calculated as the mean normalized fluorescence over 30 seconds at the peak of the effect minus that over 30 seconds of baseline.

For electrophysiology, spikes were detected using the threshold crossing method. Spike time stamps were organized into ten-second bins and the mean firing rate calculated for each bin. To determine if 5-HT affected the spontaneous firing of RP3V^KISS1 neurons, baseline firing was first measured as the mean firing rate during the two minutes preceding 5-HT application. Recordings in which the mean firing rate changed by greater than twice the standard deviation of baseline (2 × SD) during the five minutes after 5-HT first entered the bath were recorded as displaying a response to 5-HT. For statistical comparisons, firing rates were averaged over a one-minute period during baseline and at the peak of 5-HT effect.

In total, 800 GCaMP6f-expressing cells in 64 slices from 45 female mice and 55 GCaMP6f-expressing cells in 6 slices from 5 male mice were analyzed in calcium imaging experiments, while 17 GFP-expressing cells in 11 slices from 7 female mice were analyzed in electrophysiology experiments. These numbers are broken down by experiments in the results section.

Statistical analyses were performed using Prism 9.0 (GraphPad, USA). Data are reported in the text and tables as mean ± SEM, and in figures as mean ± SEM or ± 95% confidence intervals. Comparisons between two independent groups were made using the Mann-Whitney test and those between two paired groups using the paired t-test or the Wilcoxon signed rank test, as appropriate. Comparisons between multiple groups were made with the Kruskal-Wallis test with Dunn’s post-tests. Comparisons of proportions were made using Fisher’s exact tests. Differences were considered statistically significant for p< 0.05.

The effects of 5-HT on RP3V^KISS1 neuron activity were assessed using calcium imaging in brain slices from male and female Kiss1-Cre::GCaMP6f mice ( Figures 1A, B ). In slices from female mice, two-minute bath applications of 5-HT (10 μM) resulted in transient increases in GCaMP6f fluorescence in the majority of RP3V^KISS1 neurons ( Figures 1C, E ). In slices obtained from male Kiss1-Cre::GCaMP6f mice, 5-HT-induced increases in fluorescence could also be observed ( Figures 1D, F ), but these responses were seen much less frequently (see below).

Because 5-HT signaling may be involved in the generation of the LH surge, we examined RP3V^KISS1 neuron responses to 5-HT across the estrous cycle. As illustrated in Figure 2A , 5-HT-induced increases in RP3V^KISS1 neuron fluorescence could be seen in slices from diestrous, proestrous and estrous mice and were larger than those seen in slices from male mice. The proportions of RP3V^KISS1 neurons excited by 5-HT were similar at all cycle stages (diestrus: 85.3%, proestrus: 87.5%, estrus: 82.0%; p > 0.18, Fisher’s exact tests) and were significantly higher than in males (29.1%; p < 0.001, Fisher’s exact tests; Figure 2B and Table 1 ). The magnitude of RP3V^KISS1 neuron responses to 5-HT significantly varied across these groups, with male responses significantly smaller than those in females at any estrous cycle stage, but no statistical differences across the estrous cycle ( Table 1 ; Figure 2C ). Smaller responses to 5-HT in males versus females could not be explained by a reduced ability of male RP3V^KISS1 neurons to mount changes in [Ca²⁺]_i, as responses to the glutamate receptor agonist AMPA were, in fact, larger in males than in females in those slices that were tested in this manner ( Table 2 ).

Together, these observations indicate that RP3V^KISS1 neuron responses to exogenous 5-HT are – to some extent – sex-dependent, being more prominent in females. Female responses, however, are not affected by the estrous cycle. The remainder of the study was conducted in brain slices from females.

We then investigated if the effect of 5-HT was associated with changes in electrical activity in RP3V^KISS1 neurons. In slices from female Kiss1-hrGFP mice, we recorded individual RP3V GFP-expressing neurons in the cell-attached configuration. Bath applications of 5-HT (one minute, 50 μM) increased action potential firing in just over half (52.9%) of RP3V^KISS1 neurons (9 out of 17 neurons in 8 slices from 5 mice [1 di- and 4 estrus]; Figures 3A, C ), decreased it in a few cells (2 cells [11.8% of total] in 2 slices from 2 mice [1 di- and 1 estrus]; Figures 3B, C ), or did not affect firing (6 neurons [35.3% of total] in 5 slices from 4 mice [1 pro-, 1 di- and 2 estrus]). In those RP3V^KISS1 neurons that were stimulated by 5-HT, action potential firing increased from 2.20 ± 0.42 to 6.33 ± 1.40 Hz (p < 0.01, t = 3.66, df = 8, paired t-test; Figure 3D ). In the two cells inhibited by 5-HT, spontaneous firing (0.73 and 0.12 Hz) was transiently silenced in both cases.

These observations indicate that 5-HT primarily stimulates RP3V^KISS1 neuron action potential firing. While 5-HT could also suppress firing in RP3V^KISS1 neurons, this was seen much less frequently.

We next tested whether the effect of 5-HT is direct or mediated through the release of an intermediary factor by cells within the slice. To do this, we used a protocol in which slices from Kiss1-Cre::GCaMP6f females were exposed to 5-HT (two minutes, 10 μM) twice at a 15- to 20-minute interval, the second 5-HT application being carried out in the presence of the voltage-gated sodium channel inhibitor tetrodotoxin (TTX; 0.5 μM) to block electrical activity in the slice. This concentration of TTX is sufficient to inhibit action potential generation in ex vivo preparations (57–59). As illustrated in Figure 4A , 5-HT-induced increases in RP3V^KISS1 neuron fluorescence persisted in the presence of TTX. In control conditions, 57 out of 68 cells (83.8%) were stimulated by 5-HT. In the presence of TTX, 53 of these neurons (77.9% of the total; p = 0.51, Fisher’s exact test; Figure 4B ) displayed 5-HT-induced increases in normalized fluorescence. On average, 5-HT increased RP3V^KISS1 neuron normalized fluorescence to a similar extent whether in the absence (2.72 ± 0.37%) or in the presence of TTX (2.34 ± 0.37%; n = 57 from 7 slices in 4 mice [2 pro-, 1 di- and 1 estrus]; p = 0.24, Wilcoxon test; Figure 4C ).

This finding indicates that the effect of 5-HT on RP3V^KISS1 neuron activity is likely direct.

Because female RP3V^KISS1 neurons express genes for multiple stimulatory 5-HT receptor subtypes, including htr2a, htr2c, htr3a, htr4, htr6 and htr7 (60), we next examined if activating these 5-HT receptors could increase [Ca²⁺]_i in female RP3V^KISS1 neurons. Increases in RP3V^KISS1 neuron activity could be seen in response to bath-applications of the agonists TCB-2 (1 μM; 5-HT_2A receptors; 2 di-, 1 pro- and 1 estrous mice), WAY161503 (10 μM; 5-HT_2C receptors; 3 diestrous mice), SR57227 (1 μM; 5-HT₃ receptors; 4 diestrous mice), cisapride (1 μM; 5-HT₄ receptors; 3 diestrous mice), WAY208466 (10 μM; 5-HT₆ receptors; 3 diestrous mice) and AS19 (10 μM; 5-HT₇ receptors; 1 di-, 3 pro- and 1 estrous mice) ( Figures 5A, B and Table 3 ). However, activating 5-HT_2A and 5-HT_2C receptors resulted in greater proportions of stimulated RP3V^KISS1 neurons (≈55-70%) than activating other receptor subtypes ( Figure 5C and Table 3 ). In addition, activating 5-HT_2C receptors resulted in significantly larger responses than activating other 5-HT receptors, whereas responses to 5-HT_2A receptor activation were rather moderate in magnitude ( Figure 5D and Table 3 ).

Taken together, these observations indicate that the female RP3V^KISS1 neuron population expresses multiple functional excitatory 5-HT receptor subtypes. However, of these, activation of 5-HT₂ receptors, in particular 5-HT_2C, most closely mimics the effect of 5-HT on female RP3V^KISS1 neurons.

Based on these findings, we then sought to determine the role of 5-HT₂ receptors in mediating the effect of 5-HT on RP3V^KISS1 neuron activity. Using a dual 5-HT application protocol like that described above, we tested the impact of 5-HT₂ receptor antagonists on RP3V^KISS1 neuron responses to 5-HT. As illustrated in Figure 6A , the antagonist ritanserin (5 μM), which prevents the LH surge in rats (34), almost completely blocked the effect of 5-HT in RP3V^KISS1 neurons. 69 out of 85 cells (81.2%) were stimulated by 5-HT in control conditions. In the presence of ritanserin, 31 of these (36.5% of the total; p < 0.001, Fisher’s exact test) had 5-HT-induced increases in normalized fluorescence ( Figure 6B ). On average, 5-HT-induced changes in RP3V^KISS1 neuron normalized fluorescence were substantially reduced by ritanserin (0.29 ± 0.10% versus 3.37 ± 0.32% in the absence of antagonist; n = 69 in 5 slices from 5 mice [2 pro-, 2 di- and 1 estrus]; p < 0.001, sum of signed ranks (W) = -2387, Wilcoxon test; Figure 6C ). The same held true when only those cells that were excited by 5-HT in the presence of ritanserin were considered ( Table 4 ). Similar results were obtained with the antagonist mianserin (10 μM). In its presence, only 8 out of 75 cells (10.7% versus 80.0% [60 out of 75] in control; p < 0.001, Fisher’s exact test; Figure 6D ) were excited by 5-HT. Moreover, 5-HT-induced increases in normalized fluorescence were significantly suppressed by mianserin (-0.53 ± 0.22% versus 2.61 ± 0.42% in the absence of antagonist; n = 60 in 6 slices from 5 mice [1 pro-, 2 di- and 2 estrus]; p < 0.001, sum of signed ranks (W) = -1718, Wilcoxon test; Figure 6E ), even when only cells stimulated by 5-HT in the presence of the antagonist were considered ( Table 4 ).

Together with the results obtained using 5-HT receptor agonists, these data indicate that the effect of 5-HT on RP3V^KISS1 neurons is – for the most part – mediated by 5-HT₂ receptors, likely through the combined effect of 5-HT_2A and 5-HT_2C receptor activation.