Abstract
Dysfunctions in growth hormone (GH) secretion increase the prevalence of anxiety and other neuropsychiatric diseases. GH receptor (GHR) signaling in the amygdala has been associated with fear memory, a key feature of posttraumatic stress disorder. However, it is currently unknown which neuronal population is targeted by GH action to influence the development of neuropsychiatric diseases. Here, we showed that approximately 60% of somatostatin (SST)-expressing neurons in the extended amygdala are directly responsive to GH. GHR ablation in SST-expressing cells (SSTΔGHR mice) caused no alterations in energy or glucose metabolism. Notably, SSTΔGHR male mice exhibited increased anxiety-like behavior in the light-dark box and elevated plus maze tests, whereas SSTΔGHR females showed no changes in anxiety. Using auditory Pavlovian fear conditioning, both male and female SSTΔGHR mice exhibited a significant reduction in fear memory. Conversely, GHR ablation in SST neurons did not affect memory in the novel object recognition test. Gene expression was analyzed in a micro punch comprising the central nucleus of the amygdala (CEA) and basolateral (BLA) complex. GHR ablation in SST neurons caused sex-dependent changes in the expression of factors involved in synaptic plasticity and function. In conclusion, GHR expression in SST neurons is necessary to regulate anxiety in males, but not female mice. GHR ablation in SST neurons also decreases fear memory and affects gene expression in the amygdala, although marked sex differences were observed. Our findings identified for the first time a neurochemically-defined neuronal population responsible for mediating the effects of GH on behavioral aspects associated with neuropsychiatric diseases.
SIGNIFICANCE STATEMENT Hormone action in the brain regulates different neurological aspects, affecting the predisposition to neuropsychiatric disorders, like depression, anxiety, and posttraumatic stress disorder. Growth hormone (GH) receptor is widely expressed in the brain, but the exact function of neuronal GH action is not fully understood. Here, we showed that mice lacking the GH receptor in a group of neurons that express the neuropeptide somatostatin exhibit increased anxiety. However, this effect is only observed in male mice. In contrast, the absence of the GH receptor in somatostatin-expressing neurons decreases fear memory, a key feature of posttraumatic stress disorder, in males and females. Thus, our study identified a specific group of neurons in which GH acts to affect the predisposition to neuropsychiatric diseases.
Introduction
Millions of people suffer from neuropsychiatric diseases, including depression, anxiety, and posttraumatic stress disorder, causing numerous impairments to these individuals. Growing evidence suggests that hormones may play a role in regulating different neurological aspects, altering the predisposition to these disorders. For example, changes in sex hormone levels, particularly estradiol, affect anxiety-like or depression-like behavior and fear memory in rodents and humans (Walf et al., 2009; Walf and Frye, 2010; Zeidan et al., 2011). Other studies have shown that glucocorticoids may be involved in the pathogenesis of neuropsychiatric diseases (Binder, 2009; Claes, 2009; Chiba et al., 2012).
There is evidence that children and adults with growth hormone (GH) deficiency present a higher prevalence of anxiety and depression (Karachaliou et al., 2021). However, it is still unknown how GH can affect the prevalence of neuropsychiatric diseases. In the last years, our group has shown that several brain structures associated with neuropsychiatric disorders contain a large number of neurons directly responsive to GH (Furigo et al., 2017; Wasinski et al., 2021a). These GH-responsive areas include the hypothalamus, hippocampus, amygdala, bed nucleus of the stria terminalis (BNST), and some brainstem nuclei. A previous study showed that approximately two-thirds of corticotropin-releasing hormone (CRH)-expressing neurons in the mouse paraventricular nucleus of the hypothalamus (PVH) exhibit phosphorylation of the signal transducer and activator of transcription 5 (pSTAT5) after a systemic GH injection, which indicates activation of GH receptor (GHR) and its downstream intracellular pathway (Quaresma et al., 2021). Since PVHCRH neurons are important regulators of the stress response, GH action on this neuronal population may be associated with neuropsychiatric disorders. However, GHR ablation in CRH-expressing cells causes no behavioral alteration in mice (Dos Santos et al., 2021).
GH-responsive cells are largely found in the extended amygdala, particularly in the central nucleus of the amygdala (CEA) and dorsolateral BNST (Furigo et al., 2017; Wasinski et al., 2021a). These nuclei contain several neurochemically-defined neuronal populations, including cells that express CRH, protein kinase Cδ, somatostatin (SST), and tachykinin 2 (Ye and Veinante, 2019). Inhibition of CEACRH neurons causes anxiolytic effects via projections to the dorsolateral BNST (Pomrenze et al., 2019). Although part of CEACRH and BNSTCRH neurons is responsive to GH, CRH-specific GHR knock-out mice exhibit normal anxiety and stress response (Dos Santos et al., 2021). In addition to pituitary-derived GH, previous studies have suggested that basolateral amygdala (BLA) may locally produce GH. Chronic stress not only increases fear memory but also the expression of GH in the BLA (Meyer et al., 2014). Virus-mediated GH overexpression in the BLA increases the number of cells activated by fear memory formation and dendritic spine density in the amygdala (Gisabella et al., 2016). Thus, amygdala neurons may be affected by either pituitary-derived or locally-derived GH. However, no information exists about the neuronal population in the amygdala that expresses the GHR to regulate fear responses.
SST-expressing neurons are involved in several functions, including the control of GH secretion (Steyn et al., 2016), food intake (Luo et al., 2018; Stengel and Taché, 2019; Kumar and Singh, 2020) and anxiety (Engin et al., 2008; Engin and Treit, 2009; Wohleb et al., 2016; Fuchs et al., 2017). While hypothalamic SST neurons regulate the somatotropic axis and metabolism, SST neurons in the extended amygdala are associated with changes in anxiety (Pantazopoulos et al., 2017; Ahrens et al., 2018; Sun et al., 2020; Bruzsik et al., 2021; Xiao et al., 2021) and fear memory (H. Li et al., 2013; Penzo et al., 2015; Besnard et al., 2019; Sun et al., 2020; Bruzsik et al., 2021). Whether SST neurons in the extended amygdala are responsive to GH, and SST-expressing cells mediate the effects of GH regulating anxiety and fear memory is still unknown. To understand how central GHR signaling can alter the predisposition to neuropsychiatric diseases, our study aims to determine whether GHR expression in SST neurons is necessary to regulate anxiety and fear memory in male and female mice.
Materials and Methods
Mice
To visualize SST-expressing cells, the SSTCre mouse model (The Jackson Laboratory; RRID: IMSR_JAX:018973) was crossed with mice expressing a Cre-dependent enhanced green fluorescent protein (eGFP; The Jackson Laboratory; RRID: IMSR_JAX:026175). GHR ablation specifically in SST-expressing cells was achieved by breeding the SSTCre mice with animals carrying loxP-flanked Ghr alleles (List et al., 2013). The experiments used only Ghrflox/flox mice and the control group was composed of littermates negative for the Cre allele, whereas SSTΔGHR mice carried the SSTCre gene in heterozygosity. Mice were in the C57BL/6J background and were produced and maintained in standard conditions of light (12/12 h light/dark cycle; lights on at 8 A.M.) with ad libitum access to regular rodent chow and filtered water. The experimental procedures were approved by the Ethics Committee on the Use of Animals of the Instituto de Ciencias Biomedicas at the Universidade de São Paulo.
Identification of GH-responsive and estrogen-responsive neurons
Adult SST-eGFP reporter male and female mice (n = 3–5/group) received an intraperitoneal injection of saline or GH (20 µg/g) extracted from porcine pituitary (National Hormone and Pituitary Program, Torrance, CA) and were perfused 90 min later. Mice were anesthetized with isoflurane and perfused transcardially with saline, followed by a 10% buffered formalin solution. Brains were collected and postfixed in the same fixative for 45 min and cryoprotected overnight at 4°C in 0.1 m PBS containing 20% sucrose. Brains were cut into 30-µm-thick sections using a freezing microtome. Brain slices were rinsed in 0.02 m potassium PBS, pH 7.4 (KPBS), followed by pretreatment in a water solution containing 1% hydrogen peroxide and 1% sodium hydroxide for 20 min. After rinsing in KPBS, sections were incubated in 0.3% glycine and 0.03% lauryl sulfate for 10 min each. Next, slices were blocked in 3% normal serum for 1 h, followed by incubation in a primary antibody cocktail containing anti-phosphoTyr694-STAT5 (1:1000; Cell Signaling Technology; catalog #9351; RRID: AB_2315225) and anti-eGFP (1:5000; Aves Labs; catalog #GFP-1020; RRID: AB_10000240) for 40 h. Subsequently, sections were rinsed in KPBS and incubated for 90 min in Alexa Fluor-conjugated secondary antibodies (1:500, Jackson ImmunoResearch Laboratories). After rinsing in KPBS, sections were mounted onto gelatin-coated slides and covered with Fluoromount G mounting medium (Electron Microscopic Sciences). The colocalization between estrogen receptor α (ERα) and eGFP (SST neurons) was performed as described above, except that we used the primary antibody anti-ERα (1:20,000, MilliporeSigma; catalog #04-1564; RRID: AB_10618636). A Zeiss Axiocam 512 color camera adapted to an Axioimager A1 microscope (Zeiss) was used to obtain the photomicrographs. Single-labeled and double-labeled cells were manually counted using the counting tool available in the Adobe Photoshop software.
Evaluation of energy and glucose homeostasis
Body weight and body composition were analyzed in approximately five-month-old male and female mice using time-domain nuclear magnetic resonance (LF50 body composition mice analyzer; Bruker). Subsequently, mice were single-housed for acclimation. Then, food intake was measured by recording the amount of daily food ingested for 5 consecutive days. O2 consumption (VO2), CO2 production, respiratory quotient (CO2 production/O2 consumption), and ambulatory activity (by infrared sensors) were analyzed by the Oxymax/Comprehensive Lab Animal Monitoring System (Columbus Instruments) for approximately 7 d. The data from the first 2 d of analysis were discarded because we considered the acclimatization period. The results used for each animal were the average of the analyzed days. In the experiments that evaluated glucose homeostasis, food was removed from the cage 4 h before each test, and after determining basal glucose levels (time 0), mice received intraperitoneal injections of 2 g/kg of glucose, 1 IU/kg of insulin or 0.5 g/kg of 2-deoxy-D-glucose (2DG; MilliporeSigma), followed by serial determinations of blood glucose levels using a glucose meter through samples collected from the tail tip. The area under the curve of blood glucose levels was calculated for each test per animal using the Prism software version 8.4.3 (GraphPad).
Anxiety tests
Another group of mice was evaluated in the behavioral experiments. In the open field test, adult mice were placed in an arena [40 cm (w) × 40 cm (d) × 30 cm (h)] and their behavioral states were evaluated for 5 min. In this regard, total distance traveled, time spent and number of entries in the center of the arena, and duration and frequency of grooming, walking, rearing, sniffing, and stationary state were determined. The light-dark box test used a box divided into two compartments (light and dark), where each compartment measures 21.5 × 40.5 × 21 cm and with an open central access (5 × 5 cm) between them, positioned on a flat surface. The light compartment was entirely white, except for the translucent lid, with a luminosity of 400-lx considered aversive, and the dark compartment, entirely black, with a luminosity of 3 lx. All animals were placed individually at the end of the white compartment with their muzzle facing the central access, with complete freedom to explore the entire box for 5 min. The following parameters were analyzed: time spent in the light compartment, number of crossings between the compartments and distance traveled in the light zone. After each behavioral test, the boxes were cleaned with 5% alcohol to avoid olfactory cues. In the elevated plus maze test, we used an apparatus that is elevated 40 cm from the ground and it is composed of two opposite open arms measuring 35 × 5 cm, crossed by two closed arms of similar dimensions. Each animal was placed in the center of the apparatus and the number of entries, the distance traveled, and the time spent in the open arms were measured for 5 min. The animal’s position and movements were recorded and analyzed by the ANY-maze software (Stoelting Co.). After each test, the apparatuses were cleaned with 70% ethanol and air-dried before a new trial.
Auditory Pavlovian fear conditioning
This test initially used a conditioning box (28 × 26 × 23 cm) consisting of four white walls, a transparent acrylic lid, and a base formed by metallic bars that conduct electric current arranged in parallel (diameter of 0.4 cm and spacing between them of 1.05 cm) and connected to an electric shock generator (Insight Equipamentos). For the fear conditioning procedure, the mice were placed in the conditioning chamber for 2 min to habituate. After that, they were presented to three tone-shock pairings, consisting in tones of 3 kHz, 80 dB, 30 s each (conditioned stimulus, CS) paired with foot shocks (unconditioned stimulus, US: 0.6 mA, presented in the last 2 s of each sound). The intertrial interval was 60 s. Two minutes after the last CS-US pairing, the animals returned to their respective housing boxes. Between each session, the box was cleaned with a 79.5% water/19.5% alcohol/1% vanilla essence solution (Dr. Oetker). After 24 h, the fear of the tone was tested in a novel context consisting of a box measuring 22.5 × 19.5 × 27 cm, with a conical wall and floor covered with white self-adhesive paper. The session consisted of 2 min of habituation (Pre), followed by 40 CS presentations with a duration of 30 s each and 5-s intervals between them, in the absence of the US. The percentage of time in freezing was displayed in 200-s intervals. The mean percentage in freezing was also calculated for each animal.
Recognition memory
For the novel object recognition test, mice were acclimatized in the arena used in the open field for 5 min. On the next day, the training session consisted of a 5-min-long session in which the mice were placed at the bottom of the arena with their back to the two identical objects, and the time spent exploring each object was recorded. Approximately 24 h after the training session, animals were reinserted into the arena and one of the two objects used in the training session was replaced by a new one with a similar size. The time exploring the new and old objects was measured. During behavioral sessions, objects were fixed with tape to the floor so that the animals could not move them and none of the objects used evoked innate preference. The arena and stimulus objects were cleaned thoroughly between trials with 70% ethanol. The results are expressed as the percentage of time exploring each object during the training or test session in relation to total exploration time.
Quantitative real-time PCR
Control and SSTΔGHR male and female mice were anesthetized with isoflurane, euthanized by decapitation, and the entire brain was removed. A coronal brain section (600 μm) was cut with a vibratome. A bilateral micro punch comprising the CEA and BLA (bregma: –1.06 to –1.66 mm) were collected using an 18-gauge needle and stored at −80°C. Total RNA from the bilateral micro punch was extracted using the Arcturus PicoPure RNA Isolation kit (Thermo Fisher Scientific; catalog #KIT0204; RID: SCR_008817), following the manufacturer’s instructions. Assessment of RNA quantity and quality was performed with an Epoch Microplate Spectrophotometer (BioTek; RRID: SCR_017317). cDNA was synthesized by reverse transcription using 500 ng of RNA, SuperScript II Reverse Transcriptase (Invitrogen), and random primers p(dN)6 (MilliporeSigma). Quantitative real-time PCR was performed using the 7500TM Real-Time PCR System (Applied Biosystems), SYBR Green Gene Expression PCR Master Mix (Applied Biosystems), and specific primers for target genes (Table 1). Data were normalized to the geometric average of Actb and Ppia and reported as fold changes compared to values obtained from the control group (set at 1.0). Relative quantification of mRNA was calculated by 2-ΔΔCt.
Primer sequences
Statistical analysis
The differences between groups were analyzed by the unpaired two-tailed Student’s t test. Two-way ANOVA was used to analyze the histological data. Freezing behavior over time and the percentage of exploration of different objects were analyzed by repeated measures two-way ANOVA. Statistical analyses were performed using the Prism software. All results were expressed as mean ± SEM, and only p values <0.05 were considered to be statistically significant.
Results
Distribution of SST-expressing neurons that are responsive to GH in the mouse brain
Several brain nuclei contain SST-expressing neurons, including the dorsolateral BNST, PVH, periventricular nucleus (PV), arcuate nucleus of the hypothalamus (ARH), tuberal nucleus (TU), and lateral CEA (Fig. 1). To identify specific neuronal populations that express functional GHR, we used the ability of a systemic GH injection to induce the activation of the GHR/STAT5 pathway, as previously shown (Furigo et al., 2017; Wasinski et al., 2021a). SST-eGFP reporter mice that received an intraperitoneal saline injection before the perfusion exhibited a small amount of pSTAT5 immunoreactive cells in the mouse brain and very few co-localizations with SST-expressing neurons (Fig. 1A–H). In contrast, an intraperitoneal GH injection induced a robust pSTAT5 expression in several brain nuclei (Fig. 1I–L). Importantly, numerous co-localizations between GH-induced pSTAT5 and SST were observed in the BNST (Fig. 1M), PVH (Fig. 1N), PV (Fig. 1N), and CEA (Fig. 1P). Several GH-responsive SST neurons were also observed in the ARH (Fig. 1O) but just a few co-localizations were observed in the TU (Fig. 1O).
SST neurons of different brain nuclei are responsive to GH. A–H, Epifluorescence photomicrographs showing pSTAT5 (red) and eGFP (green representing SST-expressing cells) immunoreactive neurons in the brain of an SST-eGFP reporter mouse that received an intraperitoneal injection of saline. I–P, Photomicrographs showing pSTAT5 and eGFP immunoreactive neurons in the brain of an SST-eGFP reporter mouse that received an intraperitoneal injection of GH (20 µg/g b.w.) and was perfused 90 min later. Double-labeled neurons may appear as yellow. 3V, third ventricle; ac, anterior commissure; ARH, arcuate nucleus of the hypothalamus; BLA, basolateral amygdala; BNST, bed nucleus of the stria terminalis; CEA, central nucleus of the amygdala; DMH, dorsomedial nucleus of the hypothalamus; LV, lateral ventricle; opt, optic tract; PV, periventricular nucleus, PVH, paraventricular nucleus of the hypothalamus; TU, tuberal nucleus; VMH, ventromedial nucleus. Scale bar = 500 μm.
A previous study has shown that approximately 60% of PVHSST and PVSST neurons express pSTAT5 after a systemic GH injection (Chaves et al., 2022). In the current study, we quantify this co-expression in other brain nuclei of control male and female mice. We observed that approximately 20% of ARHSST neurons contained GH-induced pSTAT5 (Fig. 2A). Notably, 62% and 60% of BNSTSST (Fig. 2B) and CEASST neurons (Fig. 2C), respectively, were responsive to GH using a SST-eGFP reporter mouse model. The percentage of SST neurons that were responsive to GH was similar between males and females (Fig. 2G–I). Furthermore, male and female mice exhibited similar numbers of ARHSST, BNSTSST, and CEASST neurons (Fig. 2J–L). To determine the physiological importance of GHR signaling in SST neurons, mice carrying inactivation of the Ghr gene exclusively in SST-expressing cells were produced (SSTΔGHR mice). In contrast to the results observed in control animals, SSTΔGHR mice exhibited few SST neurons containing GH-induced pSTAT5 in the ARH (Fig. 2D,G), BNST (Fig. 2E,H) and CEA (Fig. 2F,I). Moreover, pSTAT5 expression was intact in cells surrounding the SST neurons of SSTΔGHR mice. Thus, these findings validate the efficacy and specificity of the GHR ablation only in SST-expressing cells. Interestingly, SSTΔGHR male mice showed a reduced number of ARHSST neurons, compared to control males, whereas the number of ARHSST neurons was similar between control and SSTΔGHR females (Fig. 2J). GHR ablation in SST-expressing cells did not affect the number of SST neurons in the BNST and CEA (Fig. 2K,L).
Validation of mice carrying GHR ablation specifically in SST-expressing cells. A–C, Epifluorescence photomicrographs showing the co-localization between pSTAT5 (red) and eGFP (green representing SST-expressing cells) in SST-eGFP reporter control mice that received an intraperitoneal injection of GH before perfusion. The insets represent higher magnification photomicrographs of the selected area. Double-labeled neurons may appear as yellow. D–F, Photomicrographs showing the co-localization between pSTAT5 and eGFP in GH-injected SSTΔGHR mice. G–I, Quantification indicating the percentage of ARHSST (G), BNSTSST (H), and CEASST (I) neurons that are responsive to GH (express pSTAT5). J–L, Quantification indicating the number of ARHSST (J), BNSTSST (K), and CEASST (L) neurons. 3V, third ventricle; ARH, arcuate nucleus of the hypothalamus; BNST, bed nucleus of the stria terminalis; CEA, central nucleus of the amygdala. Scale bar = 100 μm. *p < 0.05, **p < 0.01, ****p < 0.0001 significant difference between groups (two-way ANOVA followed by Holm–Sidak’s multiple comparisons test).
GHR ablation in SST neurons does not affect energy and glucose metabolism
Several studies revealed that SST neurons of different brain areas regulate food intake and other metabolic aspects (Campbell et al., 2017; Zhu et al., 2017; Luo et al., 2018; Zséli et al., 2018; Stengel and Taché, 2019; Kumar and Singh, 2020; Suresh Nair et al., 2022). Central GH action also modulates metabolism and stimulates feeding in different species, such as fish, rodents, and humans (Nyberg, 2000; Bohlooly-Y et al., 2005; Egecioglu et al., 2006; Deepak et al., 2008; Zhong et al., 2013; Kim et al., 2015; Furigo et al., 2019a, b; Quaresma et al., 2019; Teixeira et al., 2019; Donato et al., 2021; Gupta et al., 2021; Wasinski et al., 2021b; Stilgenbauer et al., 2023; Tavares et al., 2023). Thus, SST neurons may mediate GH’s metabolic effects. To test this possibility, the metabolism of control and SSTΔGHR mice were detailed assessed. No differences between control and SSTΔGHR mice were observed in the body weight, fat mass, lean body mass, food intake, VO2, respiratory quotient, and ambulatory activity, either in males (Fig. 3A–I) or females (Fig. 3J–R). Glucose tolerance (Fig. 4A,B), insulin sensitivity (Fig. 4C,D), and counter-regulatory response to 2DG (Fig. 4E,F) did not differ between control and SSTΔGHR mice. Thus, despite the numerous SST neurons that are responsive to GH, GHR ablation in these cells does not affect energy and glucose metabolism.
GHR ablation in SST neurons does not affect energy homeostasis. A–D, Body weight, fat mass, lean body mass, and daily food intake of control (n = 14–16) and SSTΔGHR (n = 9–11) male mice. E–I, Oxygen consumption (VO2), respiratory quotient (VCO2/VO2), and ambulatory activity of control (n = 5) and SSTΔGHR (n = 9) male mice. J–M, Body weight, fat mass, lean body mass, and daily food intake of control (n = 8–14) and SSTΔGHR (n = 9–10) female mice. N–R, VO2, respiratory quotient, and ambulatory activity of control (n = 7) and SSTΔGHR (n = 8) female mice. Possible differences between groups were analyzed by t test.
GHR ablation in SST neurons does not affect glucose homeostasis. A–F, Blood glucose levels and area under the curve of the glucose tolerance test, insulin tolerance test, and counter-regulatory response to 2-deoxy-D-glucose (2DG) in control males (n = 11–14), SSTΔGHR males (n = 7–9), control females (n = 6–7) and SSTΔGHR females (n = 5–8). Possible differences between groups in the areas under the curve were analyzed by t test.
GH action in SST neurons regulates anxiety in a sex-dependent manner
Possible changes in anxiety were determined in SSTΔGHR mice, compared to control animals. Using the open-field test, we observed a tendency (t(23) = 1.91, p = 0.06) of the SSTΔGHR male mice to spend less time in the center of the arena (Fig. 5A), suggesting increased anxiety-like behavior. However, no difference in the time spent in the center of the open field was observed between control and SSTΔGHR female mice (Fig. 5B). Number of entries in the center and total distance traveled as well as duration and frequency of grooming, walking, rearing, sniffing, and stationary state were also analyzed in the open field, but no significant differences were observed between the groups either in males or females (Table 2). Anxiety was further evaluated by measuring the number of transitions across the different compartments in the light-dark box test (Crawley and Goodwin, 1980). Confirming the previous results obtained in the open-field test, SSTΔGHR male mice showed a reduced number of transitions in the light-dark box test compared to control animals (Fig. 5C), whereas SSTΔGHR female mice presented similar anxiety-like behavior compared to control females (Fig. 5D). No differences between groups were observed in the time spent and the distance traveled in the light zone during the light-dark box test (Table 2). Next, the elevated plus maze test was also employed to measure anxiety. The total distance traveled during the elevated plus maze test was similar between control and SSTΔGHR mice, either in males (Fig. 5E) or females (Fig. 5I). However, SSTΔGHR male mice showed a decrease in the time spent in the open arms (Fig. 5F), in the time moving in the open arms (Fig. 5G), and in the distance traveled in the open arms (Fig. 5H), when compared to control animals. SSTΔGHR female mice exhibited normal behavior in the elevated plus maze test (Fig. 5J–L). Taken together, distinct behavioral tests indicate that GHR ablation in SST neurons increases anxiety-like behavior in male mice, but not in females.
Additional behavioral states
GH action in SST neurons regulates anxiety in a sex-dependent manner. A–D, Time spent in the center of the open field and number of transitions in the light-dark box test in control males (n = 14–15), SSTΔGHR males (n = 10–11), control females (n = 10) and SSTΔGHR females (n = 12–14). E–L, Total distance traveled in the elevated plus maze test, time spent in the open arms, time moving in the open arms and distance traveled in the open arms in control males (n = 15), SSTΔGHR males (n = 11), control females (n = 11) and SSTΔGHR females (n = 14). *p < 0.05, **p < 0.01 significant difference between groups (t test).
GHR ablation in SST-expressing neurons decreases fear response in male and female mice
In this experiment, mice were exposed to auditory Pavlovian fear conditioning. Control and SSTΔGHR mice exhibited similar fear acquisition in both sexes (Fig. 6A,D). Notably, when the fear of the tone was later tested in a novel context, SSTΔGHR mice exhibited a significant reduction in fear response, which was determined by the time the animal spends in freezing. Differently from what had been observed in the evaluation of anxiety, the fear response was reduced in both male (Fig. 6B,C) and female (Fig. 6E,F) SSTΔGHR mice. To determine whether SSTΔGHR mice have defects in any type of memory or specifically regarding fear memory, the novel object recognition test was used. Control and SSTΔGHR mice explored each object approximately 50% of the time during the training period, indicating no preference for any object (Fig. 7A,C). On the test day (24 h later), mice explored significantly more the new object, in comparison with the old one, without differences between male and female control and SSTΔGHR mice (Fig. 7B,D). Thus, GHR ablation in SST-expressing neurons decreases fear memory in both male and female mice but does not affect recognition memory.
GHR ablation in SST-expressing neurons decreases fear response in male and female mice. A, Percentage of time in freezing demonstrating the fear acquisition of male mice during the auditory Pavlovian fear conditioning. B, C, Fear response to sequential tones (40 conditioned stimulus presentations with a duration of 30 s each and 5-s intervals between them) evaluated 24 h after the fear conditioning in control (n = 13) and SSTΔGHR (n = 10) male mice. The percentage of time in freezing was displayed in 200-s intervals and the mean percentage in freezing was calculated for each animal. D, Percentage of time in freezing demonstrating the fear acquisition of female mice during the auditory Pavlovian fear conditioning. E, F, Fear response to sequential tones evaluated 24 h after the fear conditioning in control (n = 11) and SSTΔGHR (n = 14) female mice. #p < 0.05, ##p < 0.01 main effect of GHR ablation (repeated measures two-way ANOVA). *p < 0.05, **p < 0.01 significant difference between groups (t test).
GHR ablation in SST neurons does not affect recognition memory. A, Training session in the novel object recognition test showed that none of the objects used evoked innate preference in male mice. B, Evaluation of the recognition memory (24 h after the training session) by replacing a familiar object (old) with a new one in control (n = 9) and SSTΔGHR (n = 7–8) male mice. C, Training session in the novel object recognition test showing that none of the objects used evoked innate preference in female mice. D, Evaluation of the recognition memory (24 h after the training session) by replacing a familiar object (old) with a new one in control (n = 6) and SSTΔGHR (n = 9) female mice. *p < 0.05, ***p < 0.001 significant difference between the old and new objects (repeated measures two-way ANOVA).
Transcripts associated with synaptic plasticity and function are differently expressed in the amygdala of SSTΔGHR mice with marked differences between males and females
The amygdala plays a fundamental role in the regulation of neuropsychiatric diseases, such as depression, anxiety, and posttraumatic stress disorder (Shackman and Fox, 2016). Furthermore, previous studies have suggested that GH action in the amygdala regulates fear memory (Meyer et al., 2014; Gisabella et al., 2016). Here, we showed that a high percentage of CEASST neurons are responsive to GH (Fig. 2C). Thus, we decided to analyze the gene expression in the amygdala through a bilateral punch comprising the CEA and BLA. In male mice, a 20% reduction in the Ghr mRNA levels was observed in the amygdala of SSTΔGHR mice, compared to control animals (Fig. 8). This finding further confirms the Ghr inactivation in a subset of amygdala (SST-expressing) neurons. Sst mRNA levels were reduced by 50% in SSTΔGHR mice, whereas no changes in Crh, Gh, and Ghsr (ghrelin receptor) expression were observed between groups (Fig. 8). Different subunits of the glutamate receptors (Gria1, Grin1, Grin2a, and Grin2b), GABA receptors (Gabra1 and Gabbr1), potassium channels (Kcnj8 and Kcnj11), dopamine receptors (Dr2d), and genes related to memory and synaptic neurotransmission and plasticity (Slc32a1, Creb1, Dlg4, and Nrg1) were analyzed, but no differences between control and SSTΔGHR mice were found (Fig. 8). However, a significant increase in the mRNA expression of Gabbr2 (encodes the GABAB2 receptor), Syp (encodes the synaptophysin), and Bdnf (encodes the brain-derived neurotrophic factor) was observed in the amygdala of SSTΔGHR mice, compared to control animals (Fig. 8). Slc17a7 mRNA levels (encodes the vesicular glutamate transporter 1) tended to be increased (t(19) = 2.01, p = 0.059) in the amygdala of SSTΔGHR mice, when compared to control animals (Fig. 8). In female mice, Ghr and Sst mRNA levels were also reduced in the amygdala of SSTΔGHR mice, compared to control animals (Fig. 8). However, differently than the findings observed in males, SSTΔGHR female mice presented a reduction in Grin1 (encodes a subunit of the glutamate NMDA receptor), Slc17a7, and Dlg4 mRNA expression, whereas Gabra1 (encodes the GABAA receptor) and Gabbr1 (encodes the GABAB1 receptor) mRNA levels were increased, compared to control females (Fig. 8). Thus, GHR ablation in SST neurons caused sex-dependent changes in the amygdalar expression of factors involved in synaptic plasticity and function.
Gene expression analysis in a micro punch comprising the central nucleus of the amygdala and basolateral complex. mRNA levels (arbitrary units, a.u.) were determined in control (n = 13) and SSTΔGHR (n = 10) male mice, and in control (n = 9) and SSTΔGHR (n = 9) female mice. *p < 0.05 significant difference between groups (t test). #p < 0.06 (t test).
ERα is expressed in a small group of ARHSST neurons
ERα is expressed in different hypothalamic areas and subnuclei of the amygdala (Simerly et al., 1990). Thus, to investigate whether SST neurons are sensitive to changes in estrogen levels, which could help to explain the sex differences observed in the present study, the percentages of ARHSST, BNSTSST and CEASST neurons that express ERα were determined in male and female mice. ERα expression was abundantly found in the ARH (Fig. 9A,D) and other areas, such as the ventrolateral part of the ventromedial nucleus of the hypothalamus and medial nucleus of the amygdala (data not shown). However, only approximately 7% of ARHSST neurons expressed ERα in males and females (Fig. 9G,H). Few ERα positive cells were found in the BNST (Fig. 9B,E) and CEA (Fig. 9C,F) and virtually no BNSTSST (dorsolateral subdivision) and CEASST neuron expressed ERα (Fig. 9G,H).
ERα is expressed only in a small group of ARHSST neurons. A–F, Epifluorescence photomicrographs showing the co-localization between ERα (red) and eGFP (green representing SST-expressing cells) in SST-eGFP reporter male and female mice. The insets represent higher magnification photomicrographs. G, H, Quantification indicating the percentage of ARHSST, BNSTSST and CEASST neurons that express ERα. Scale bar = 100 μm. Two-way ANOVA was used for the comparison between male and female mice in the different areas analyzed.
Discussion
In the current study, we observed that GHR ablation in SST-expressing cells led to increased anxiety-like behavior in male mice and reduced fear memory in both sexes. These findings are in accordance with studies indicating that GH deficiency increases anxiety (Karachaliou et al., 2021) and the blockade of GH action in the amygdala reduces fear memory (Meyer et al., 2014). The absence of memory deficit in the object recognition test indicates that SSTΔGHR mice exhibit specific alterations in fear memory and not in other types of memory (e.g., recognition or novelty). Of note, mice carrying a neuronal-specific STAT5 ablation also exhibit a reduction in contextual fear memory (Furigo et al., 2018). Thus, GHR signaling likely relies on the transcription factor STAT5 to modulate fear memory. Therefore, disruption of GHR signaling in SST-expressing neurons affects behavioral aspects associated with neuropsychiatric disorders.
A previous study already investigated the consequences on the somatotropic axis of GHR ablation in SST neurons. Unexpectedly, both male and female SSTΔGHR mice show no significant alterations in pulsatile GH secretion, serum insulin-like growth factor-1 (IGF-1) levels, hepatic IGF-1 expression or body growth (Chaves et al., 2022). Thus, early-life GHR ablation in SST cells is insufficient to disturb the somatotropic axis and body growth, probably via compensatory mechanisms (e.g., negative feedback loop mediated by GH and IGF-1 in other hypothalamic neuronal populations or somatotrophs). In the present study, we confirmed that SSTΔGHR mice exhibited normal body weight and composition, which is in accordance with the lack of changes in GH secretion.
It is known that central GH action regulates metabolism (Furigo et al., 2019b; Donato et al., 2021; Tavares et al., 2023), and SST-expressing cells control different metabolic aspects (Campbell et al., 2017; Zhu et al., 2017; Luo et al., 2018; Zséli et al., 2018; Stengel and Taché, 2019; Kumar and Singh, 2020; Suresh Nair et al., 2022). Our findings indicate that GHR signaling in SST-expressing cells is not involved in the regulation of energy and glucose homeostasis. Since metabolic alterations represent confounding factors that can affect behavior and neurological function, the absence of such changes in SSTΔGHR mice further supports the direct role of GH action on SST neurons in controlling the predisposition to neuropsychiatric disorders. However, we only studied ad libitum fed mice consuming regular chow. So, we cannot rule out possible metabolic alterations of SSTΔGHR mice under different metabolic conditions, such as calorie restriction or high-fat diet intake.
SSTΔGHR male mice exhibited increased anxiety-like behavior in the light-dark box and elevated plus maze tests, and a tendency to expend less time in the center of the open field. The lack of statistical significance in the open field may be explained by the reduced sensitivity of this test to evaluate unconditioned anxiety in certain conditions (Ennaceur, 2014; Sturman et al., 2018). Nonetheless, we obtained robust evidence that GHR ablation in SST neurons increases anxiety in male mice. Noteworthy, the open field generally produces reliable results evaluating locomotor activity and exploratory behaviors. In this regard, we found no evidence that SSTΔGHR mice present changes in locomotor activity during both the behavioral tests and the evaluation of metabolism.
The involvement of SST-expressing neurons in the regulation of anxiety is well-established (Engin et al., 2008; Engin and Treit, 2009; Wohleb et al., 2016; Fuchs et al., 2017; Pantazopoulos et al., 2017; Ahrens et al., 2018; Sun et al., 2020; Bruzsik et al., 2021; Xiao et al., 2021). For example, optogenetic activation of the CEASST projections to the central sublenticular extended amygdala causes an anxiogenic response without affecting locomotor activity (Sun et al., 2020). CEASST neurons are mostly GABAergic (Raver et al., 2020; Hogri et al., 2022) and another study showed that mice lacking the enzyme glutamate decarboxylase 67 in SST-expressing cells exhibit anxiety-like behavior, but no differences in locomotor activity (Miyata et al., 2019). The fact that GHR ablation in SST neurons led to alterations in anxiety only in male mice is not surprising because the pattern of GH secretion is sexually dimorphic (Steyn et al., 2016) and numerous studies have demonstrated marked sex differences in the role played by the extended amygdala, and in particular of SST neurons, in the control of anxiety and other behaviors (Dao et al., 2020; Diab et al., 2020; Jefferson et al., 2020; Englund et al., 2021; Kirson et al., 2021; Urien et al., 2021; Suresh Nair et al., 2022; Urien and Bauer, 2022). Accordingly, alterations in the activity of CEASST neurons by early life stress induce anxiety-like behavior in males but not female rodents (Englund et al., 2021). Female rats on proestrus, which is a phase of the estrous cycle characterized by high levels of estrogen, exhibit greater anxiolytic effects of diazepam (Ravenelle et al., 2018). Thus, sex differences in anxiety-like behavior are likely modulated by sex hormone receptors (Borrow and Handa, 2017). However, since BNSTSST and CEASST neurons do not express ERα, estrogens act in other neuronal populations to modulate mood. Of note, the estrous cycle of the females was not determined in our study, which could have increased the variability in the evaluated responses.
Although the pituitary gland is responsible for secreting GH into the systemic circulation (Steyn et al., 2016), chronic stress enhances GH expression in the BLA (Meyer et al., 2014). Virus-mediated GH overexpression in the BLA enhances long-term fear memory, whereas GHR signaling blockade in the amygdala prevents ghrelin-induced increases in fear response (Meyer et al., 2014). Whether amygdala-derived GH presents a local or endocrine function is currently unknown. However, the present and previous studies (Furigo et al., 2017; Wasinski et al., 2021a) indicate that a systemic or central GH stimulus induces pSTAT5 immunoreactive cells only in the CEA, whereas BLA exhibits a small number of GH-responsive cells. Thus, a paracrine communication possibly exists between BLA and CEA. So, GH produced in BLA cells may act on GHR-expressing neurons of the CEA. Although it was not our scope to precisely identify the subgroup of SST neurons involved in this response, previous studies provided robust evidence that GHR signaling in the amygdala regulates fear memory (Meyer et al., 2014; Gisabella et al., 2016). Thus, GH’s effects on fear memory are likely mediated by CEASST neurons, although the participation of other neuronal populations, such as BNSTSST and ARHSST, is also possible. Thus, a limitation of our study was our incapacity in determining the specific nucleus involved in the behavioral phenotype of SSTΔGHR mice. On the other hand, our study highlighted the great importance of GHRSST neurons in controlling anxiety and fear memory.
CEA and BLA were collected to analyze possible changes in gene expression. The 20% reduction in the amygdalar Ghr expression is in accordance with the genetic ablation only in SST-expressing and the maintenance of this receptor in other cell types present in the CEA. SSTCre mouse model exhibits reduced SST expression, explaining our gene expression result (Viollet et al., 2017). However, even in homozygosity when SST expression is much lower compared to heterozygous animals (as in our study), SSTCre mice do not exhibit significant behavioral alterations, including tests that evaluated anxiety-like and depression-like symptoms (Viollet et al., 2017). GH administration regulates synaptic plasticity, and excitatory and inhibitory synaptic transmission (Olivares-Hernández et al., 2018; G.Y. Li et al., 2023), including in the amygdala (Gisabella et al., 2016). GHR ablation in SST neurons led to changes in the mRNA expression of GABA and glutamate receptors, neurotropic factors and proteins associated with synaptic transmission in the amygdala. Although it is unclear how the changes in the expression of these transcripts could explain the behavioral phenotype of SSTΔGHR mice, these findings suggest that alterations in GHR signaling is likely sufficient to cause a significant impact on synaptic plasticity and function in the amygdala. However, additional studies are necessary to establish a causal relationship between the alterations in the expression of these genes and the behavioral changes observed in SSTΔGHR mice. Furthermore, the striking differences observed between male and female SSTΔGHR mice on changes in gene expression in the amygdala further supports a sex-dependent effect of GH action regulating behavior.
In conclusion, the present study identified the SST neurons as a key cell population responsible for mediating the effects of GH on neurological aspects associated with neuropsychiatric diseases. Thus, behavioral and mood problems caused by endocrine dysfunctions, such as GH deficiency or acromegaly, can be better understood. It is worth mentioning that aging is accompanied by a well-documented decrease in GH secretion. Therefore, future studies should investigate the potential association between GH, SST neurons, and neuropsychiatric disorders during aging. Finally, our study reiterates the importance of GH action in the brain to regulate important physiological functions.
Footnotes
This work was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo FAPESP)/Brazil Grants 2017/21854-9 (to F.M.C.), 2019/21707‐1 (to R.F.), and 2020/01318-8 (to J.D.); the National Institutes of Health National Institute on Aging Grant R01AG059779 (to J.J.K. and E.O.L.); the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)/Brazil Finance Code 001 (to W.O.d.S. and H.R.V.); and the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)/Brazil Grant 303363/2019-3 (to J.D.). We thank Ana M. P. Campos for her technical assistance.
The authors declare no competing financial interests.
- Correspondence should be addressed to Jose Donato Jr at jdonato{at}icb.usp.br
