Introduction

A fundamental feature of neural circuits is their functional and structural plasticity, which includes the formation, maintenance (stabilization), and elimination (pruning) of synapses [1,2,3]. The total number of synapses changes dynamically throughout the lifespan of the organism, and even small changes in synapse number and/or their turnover rate can significantly affect neural circuit connectivity [3, 4]. Aberrant neural circuit connectivity is associated with neuropsychiatric disorders [5, 6].

During the early postnatal period, spontaneous activity promotes the growth and differentiation of axonal projections and dendritic arbors, as well as the formation of connections [7,8,9]. Synaptic activity, which comes online slightly later, and persists throughout adult life, is critical for formation and maturation of neural circuits, as well as interaction between the individual and the environment. Synaptic activity can be regulated through Hebbian and homeostatic mechanisms [10,11,12,13,14]. While most researches on synaptic function and plasticity have focused on the classical neurotransmitters of glutamate and GABA [15,16,17], representing respectively the main forms of excitatory and inhibitory synaptic transmission in the brain, neuromodulators, including but not limited to, acetylcholine, noradrenaline, serotonin, and dopamine [18, 19] have received increasing attention.

Growth hormone (GH), also called somatotropin or human growth hormone, is a peptide hormone secreted by the anterior lobe of the pituitary gland. It is vital for normal physical growth in children, by stimulating the growth of all tissues of the body, including the bone, up until puberty [20]. In the nervous system, it stimulates brain growth, development and myelination, and has reported pro-cognitive and neuroprotective functions [21,22,23,24]. GH signals through GH receptor (GHR)-associated tyrosine kinases, expressed widely in the central nervous system (CNS) [25,26,27]. Clinical studies showed that patients with GH deficiency (GHD) exhibit cognitive impairment, and that GH administration significantly improves memory and cognitive capacity in children, or in adults with GH deficiency [28,29,30,31]. In animals, systemic administration of GH facilitates long-term memory and improves spatial memory in young, but not in aged rats [32,33,34]. On the other hand, spontaneous dwarf rats, which have a deficient version of the GH-gene, showed impaired memory performance in the Morris water maze [35]. GH knock-out mice have reduced brain size, loss of myelination and of parvalbumin-containing neurons, as well as cognitive impairments [36]. Together, these observations suggest that GH is critical for brain development and function, especially memory-related functions, although its detailed mechanism of action remains largely unknown.

The detection of GH in the cerebrospinal fluid (CSF) [37, 38] indicates that it can cross the blood–brain and/or blood–CSF barriers. Consistently, GHR is present at high levels in the choroid plexus, hypothalamus, hippocampus, pituitary and spinal cord [26, 27]. In cortex and hippocampus, GHR is most highly expressed in pyramidal neurons [25, 27]. Previous studies, mostly in adult rodents, showed that GH, signaling through GHR, can regulate excitatory synaptic transmission and/or the number/size of excitatory synapses, suggesting that it has neuromodulatory functions [39,40,41,42,43]. However, since GH is most often administered to young children for general growth promoting purposes, it is important to assay its effects on the developing brain.

Here, we systematically examine the neuromodulatory function of GH in young (pre-weaning) mice, measuring its effects on excitatory and inhibitory synaptic transmission, at two developmental time points and in two different brain regions. We found that subcutaneous (sc) injection of recombinant human growth hormone (rhGH) rapidly promoted excitatory synaptic transmission in S1BF L2/3 pyramidal neurons and in hippocampal CA1 pyramidal neurons, in P14 mice. At the later time point of P21, in addition to elevating excitatory synaptic transmission, rhGH also increased inhibitory synaptic transmission, an effect blocked by knockdown of GHR. Together, these results demonstrate that GH regulates synaptic transmission in the cerebral cortex and hippocampus in a developmental dynamic manner, affecting excitatory/inhibitory balance (E/I balance) differentially at different time points.

Materials and methods

Animals

All animal procedures complied with the animal care standards set forth by the US National Institutes of Health and were approved by the Institutional Animal Care and Use Committee at the Institute of Neuroscience (Chinese Academy of Sciences), Peking University, and Soochow University. Mice on C57BL/6 background were raised in a specific pathogen-free (SPF) environment and group housed under a 12 h–12 h light-dark cycle with food and water provided ad libitum from the cage lid. Health status was routinely checked, and only naïve mice (no previous procedures or tests) were used. Both male and female mice were used; no differences were observed, so data from both sexes were pooled. In all assays, control and experimental groups were littermates. The P14 group included mice aged P13-P15; the P21 group included mice aged P20-P22.

Acute brain slice preparation

Brain slices from deeply anesthetized mice (0.14 g/kg sodium pentobarbital) were essentially prepared as previously described [44, 45]. Mouse brains were rapidly removed and immersed in ice-cold dissection solution, choline-based dissection buffer contained (in mM): CholineCl 110, KCl 2.5, NaH2PO4 1.3, MgCl2 7, CaCl2 0.5, NaHCO3 25, glucose 20, bubbled with 95%O2/5% CO2, pH 7.4. Coronal slices were cut at 300–350 μm using a vibratome VT 1200 s (Leica, Germany) microslicer. Brain slices were allowed to recover in a submerged chamber containing artificial cerebrospinal fluid (aCSF) consisting of (in mM): NaCl 125, KCl 2.5, NaH2PO4 1.3, MgCl2 1.3, CaCl2 2, NaHCO3 25 and D-glucose 20, bubbled with 95%O2/5% CO2 for 30 min at 37 °C, and a further 60 min at 25–28 °C prior to recordings. All slices were used within 6 h of the first recording.

Whole cell recordings

Whole-cell recordings were made with a MultiClamp 700B amplifier (Molecular Devices, Sunnyvale, CA, USA), as previously described [44, 45]. Signals were low-pass filtered at 2 kHz and digitized at 10 kHz using Digidata 1440 A (Molecular Devices). Cells were visualized with an Olympus microscope (Olympus BX51WI, Japan) and a 40× water immersion objective under infrared optics.

For mEPSC recordings, Cs+-based internal solution contained (in mM): CsMeSO4 100, CsCl2 5.5, HEPES 10, D-glucose 10, MgATP 2 and Na3GTP 0.3 (pH 7.4, 290 mOsm). For mIPSC recordings, high chloride Cs+-based internal solution contained (in mM): CsCl 110, NaCl 10, MgCl2 5, EGTA 0.5, MgATP 2, Na3GTP 0.3 and HEPES 40 (pH 7.4, 270–280 mOsm). For excitatory/inhibitory ratio (E/I ratio) recordings, Cs+-based low Cl– internal solution contained (in mM): CsMeSO3 130, MgCl2 1, CaCl2 1, HEPES 10, QX-314 2, EGTA 11, MgATP 2, Na3GTP 0.3 (pH 7.3, 295 mOsm).

Typically, recordings were made from 2-3 cells per brain slice, and 2–3 brain slices per mouse. Baseline recordings (Ctrl) were from 0–5 min before rhGH application, while rhGH recordings were taken 5–10 min after rhGH (1 μM) application.

For mEPSC recordings, cells were held at −70 mV in voltage-clamp, in the presence of TTX (0.5 μM) and gabazine (50 μM) to block Na+ channels and GABAA receptors, respectively. For mIPSC recordings, cells were held at −60 mV in the presence of TTX and NBQX (10 μM) to block Na+ channels and AMPA receptors respectively. For E/I ratio experiments, recordings were made in the presence of TTX, and cells were first clamped at −70 mV to record mEPSCs, followed by mIPSCs at 0 mV in the same cells.

For all mPSC recordings, a brief hyperpolarization (−10 mV, 100 ms) was given to monitor the series and input resistances every 10 s. Cells with changes in input resistance greater than 20% were excluded from analyses. All cells analyzed had a series resistance of <20 MΩ; liquid junction potential and series resistance were uncompensated. Data were analyzed in MiniAnalysis (Synaptosoft, Fort Lee, NJ) with detection thresholds of 5 pA and 6 pA, for mEPSC and mIPSC, respectively. Data were analyzed blinded to the experimental condition.

For assaying the contribution of NMDA receptors in mediating the effects of rhGH, recordings were as illustrated in Fig. 6g. Each experimental day, the effect of rhGH was first assayed in aCSF containing TTX and gabazine in 2-3 cells. Subsequently, the effect of rhGH was assayed in aCSF containing TTX, gabazine and D-AP5 (50 μM) in a further 3–4 cells. Recordings were made from one cell per brain slice. To assay the contribution of intracellular calcium, 5 mM BAPTA was added to the internal solution.

Evoked EPSCs were recorded from L2/3-L2/3 synapses. Briefly, a glass pipette (~4 MΩ) filled with aCSF was placed in L2/3, and stimulation (1 ms duration) was given through an Iso-Flex isolator (A.M.P.I., Israel). Multiple locations were tested until reliable monosynaptic currents were recorded. For AMPA/NMDA ratio experiments, cells were voltage-clamped at −70 mV and +40 mV in aCSF containing gabazine. The AMPA current was measured as the peak current at −70 mV, while the NMDA current was measured at +40 mV at 30 ms after stimulus onset.

The spike frequency versus injected current experiments were performed by measuring the average action potential firing rate during 500 ms depolarizing current injections at 20 pA increments, in current-clamp mode. The maximum current injected in each experiment was below the current that induced adaptation. Synaptic blockers (50 μM D-APV, 10 μM NBQX, and 50 μM gabazine) were included as indicated. The internal solution contained (in mM): K-gluconate 110, KCl 20, HEPES 20, MgCl2 5, EGTA 0.6, MgATP 2, Na3GTP 0.2 (pH 7.3, 290 mOsm).

Drug administration and in vivo stereotaxic viral injections

The mouse Ghr shRNA sequence (5′-GCTGCAAGAATTGCTCATGAA-3′) was subcloned into the hU6-MCS-CMV-EGFP RNAi lentiviral vector by GeneChem (Shanghai, China); it expressed shRNA under the U6 promoter and GFP under the ubiquitin promoter. Ctrl-RNAi lentiviral vector expressed the sequence (5′-TTCTCCGAACGGGTCACGG-3′).

For viral injections, P7-P10 mice were anesthetized with 30 mg/kg ketamine and 0.02 mg/kg dexmedetomidine, as previously described [46]. Lentiviruses (packaged by GeneChem, Shanghai, China; titers of 1 × 108 TU/ml or higher) were slowly injected into S1BF at a depth of 300 μm, a volume of 0.8 μl, and a speed of 0.06 μl/min, as previously described [47]. Micropipettes were withdrawn 5–10 min after completion of injection. Mice were then placed on a heating pad to recover, before being returned to their homecage.

Western blots and ELISA assay

Mouse brain sample preparation and Western blots were performed as previously described [45, 47]. Briefly, P14 mice were deeply anaesthetized using 0.14 g/kg sodium pentobarbital. Brains were quickly removed, and hippocampi were quickly dissected. Brain samples were homogenized with a motorized tissue grinder in HEPES buffer (0.32 M sucrose and 4 mM HEPES, pH 7.4) containing freshly added protease inhibitor cocktail tablets and phosphatase inhibitor cocktail tablets (Roche, Basel, Switzerland). The total homogenates were centrifuged at 1,000 × g for 10 min to remove the nuclear fraction. The supernatant was collected and centrifuged at 10,000 × g for 20 min to yield the crude membrane fraction. The following antibodies were used: NR2B (Millipore, Cat# 06–600, 1:1,000), GluR1 (Millipore, Cat# AB1504, 1:1,000), and GAPDH (ZSGB-BIO, Cat# TA-08, 1:1,000).

For assaying the efficiency of GHR RNAi, the Hepa 1–6 cell line (SCSP-512, Cell Bank, Shanghai Institute of Biochemistry and Cell Biology; originally CRL-1830 from ATCC) was transfected with lentivirus expressing GHR-RNAi-GFP or Ctrl-RNAi-GFP (80% transfection efficiency as visualized by GFP). Cell lysates were collected, and immunoblotting was performed using the following antibodies: GHR (1:1000, Sigma, Cat# G8919) and β-tubulin (1:2000, Sigma, Cat# T4026).

The following secondary antibodies were used: HRP-conjugated goat anti-mouse IgG (Dingguo Changsheng Biotech, Cat# IH-0031), HRP-conjugated horse anti-goat IgG (Dingguo Changsheng Biotech, Cat# IH-0051), HRP-conjugated goat anti-rabbit IgG (ZSGB-BIO, Cat# ZB-2301), and HRP-conjugated goat anti-mouse IgG (ZSGB-BIO, Cat# ZB-2305). Signals were visualized using ECL chemiluminescence substrate (Thermo Fisher Scientific) or ECL Plus kit (Solarbio), and collected using Amersham Imager 600 (GE Healthcare, Chicago, IL, USA) or Tanon 4600SF (Tanon, Shanghai, China). Blots were quantitated using Fiji/ImageJ.

GH concentration was measured using the human GH ELISA kit (R&D Systems, Cat# SGH00), which only detects human GH.

Golgi staining, image acquisition and data analysis

Golgi staining was carried out using the FD Rapid GolgiStainTM Kit (FD NeuroTechnologies, PK401), according to the manufacturer’s instructions. Images were acquired using a confocal microscope (A1R or TiE-A1 plus, Nikon, Tokyo, Japan), with a ×60 oil-immersion objective (N.A. = 1.4), at 2× zoom and a Z-interval of 0.5 μm. Basal and apical oblique dendrites of S1BF L2/3 or hippocampal pyramidal neurons were imaged. Only secondary dendrites were imaged and quantitated [48, 49]. Images from control and rhGH groups were imaged on the same day. Dendritic spines were visually identified, and dendrite length (typically >30 μm) was measured using Image-Pro Plus (Media-Cybernetics); typically, 5–10 neurons were analyzed per mouse. Images were blind coded. Example images were projected at minimal intensity, intensity values were inverted, and background was subtracted.

Quantification and statistical analysis

Statistical analysis was carried out using GraphPad Prism 7 (GraphPad Software, La Jolla, CA, USA). The Gaussian distribution of the data was assessed using the Shapiro–Wilk normality test, the KS normality test or the D’Agostino & Pearson normality test. If the data passed the Gaussian distribution test, parametric tests (paired or unpaired t test) were used. Otherwise, nonparametric tests were used: Wilcoxon matched-pairs signed rank test for paired groups, Mann–Whitney test for unpaired groups. For excitability experiments, two-way ANOVA followed by Bonferroni’s multiple comparisons test was used.

Data were analyzed blinded to the experimental conditions. All data are represented as the mean ± SEM, and “n” refers to the number of cells (for electrophysiological recordings and spine density analysis) or number of mice (for ELISA and Western blots). The number of mice used is indicated in figure legends; typically, three or more mice were used per experimental condition. All conditions significantly different from control are indicated. *P < 0.05; **P < 0.01; ***P < 0.001.

Results

rhGH promotes excitatory synaptic transmission in S1BF and hippocampal CA1 pyramidal neurons at P14

To examine the effect of GH on synaptic transmission in young mice, we subcutaneously (sc) administered P14 C57BL/6 mice with saline (Ctrl) or rhGH (2 mg/kg), and examined their rhGH protein level in cortical brain lysates at 2 h and 24 h post injection. rhGH level was significantly elevated in cortex 2 h following rhGH injection (Fig. 1a, b), suggesting efficient entry of peripherally administered rhGH into the brain; these results are consistent with previous reports [37, 50]. By 24 h, rhGH level essentially returned to baseline (Fig. 1b). We focused on the effect of GH on excitatory synaptic transmission at 2 h post-injection, when its brain level is high. We recorded AMPA-type glutamate receptor-mediated miniature excitatory postsynaptic currents (mEPSCs) from L2/3 pyramidal neurons in acute S1BF brain slices (Fig. 1a), following sc injection of different doses of rhGH (1, 2 or 4 mg/kg); concentrations were selected based on previous reports [51, 52]. The frequency of mEPSCs significantly increased in the 2 mg/kg (Ctrl, 2.13 ± 0.22 Hz; rhGH, 3.64 ± 0.30 Hz; P < 0.001) and 4 mg/kg (Ctrl, 4.01 ± 0.44 Hz; rhGH, 6.32 ± 0.43 Hz; P < 0.001) groups (Fig. 1e–h), but not in the 1 mg/kg group (Fig. 1c, d). In all experiments, Ctrl and rhGH groups were littermates. mEPSC amplitude was not significantly altered in any of the conditions (Fig. 1c–h). Thus, rhGH promotes excitatory synaptic transmission in S1BF L2/3 pyramidal neurons in a dose-dependent fashion at P14, with 2 mg/kg being an effective dose.

Fig. 1: rhGH increases excitatory synaptic transmission and spine density in S1BF L2/3 pyramidal neurons at P14 in vivo and in vitro.
figure 1

a Schematic of experimental timeline, showing sc injection of rhGH (or saline Ctrl), followed by whole-cell recordings in acute cortical slices of P14 mice. All recordings were from S1BF L2/3 pyramidal neurons. b Protein level of rhGH in the cortex of P14 mice 2 h and 24 h after sc injection of 2 mg/kg rhGH. Assay is specific for human GH, n = 7–9 mice per condition. Representative traces, average waveforms (c, e, g) and summary graphs (d, f, h) of the effects of sc injection of different doses of rhGH (c, d, 1 mg/kg; e, f, 2 mg/kg; g, h, 4 mg/kg) on mEPSC frequency and amplitude. n = 12–16 cells from 3 mice per condition. Representative traces, average waveforms (i) and summary graphs (j) of the effect of sc injection of rhGH (2 mg/kg) on mIPSC frequency and amplitude. n = 23 cells from 3 mice per condition. Representative traces (k, m) and summary graphs (l, n) of the effect of perfusing rhGH (1 μM) onto acute cortical slices on mEPSC (k, l) or mIPSC (m, n) frequency and amplitude. n = 12–13 cells from 3 mice per experiment. Representative images (o, p) and summary graphs (q, r) of basal and apical dendrite spine density in rhGH (2 mg/kg)-injected mice. n = 26–47 neurons from 6 to 7 mice per condition (Scale bar, 10 μm). Data are presented as mean ± SEM. Statistics are measured using unpaired t test (d, f, h, q, r), Mann–Whitney test (j), paired t test (l, n), or two-way ANOVA followed by Tukey’s multiple comparisons test (b). *P < 0.05, ***P < 0.001, n.s., no significance.

Full size image

The effect of rhGH was rapid, as its perfusion (1 µM) onto acute brain slices significantly increased the mEPSC frequency of S1BF L2/3 pyramidal neurons over the course of minutes (Fig. 1k, l) (Ctrl, 3.64 ± 0.36 Hz; rhGH, 4.18 ± 0.41 Hz; P < 0.05). To examine whether in addition to increasing excitatory synaptic transmission, rhGH also regulated spinogenesis, we performed Golgi staining 2 h following rhGH administration. The density of dendritic spines in both basal and apical oblique dendrites of L2/3 pyramidal neurons increased significantly, suggesting that rhGH promoted spine formation in vivo (Fig. 1o–r) (basal: Ctrl, 0.72 ± 0.02/μm; rhGH, 0.91 ± 0.02/μm; P < 0.001; apical: Ctrl, 0.78 ± 0.02/μm; rhGH, 0.87 ± 0.02/μm; P < 0.001). Inhibitory synaptic transmission, measured as the frequency and amplitude of miniature inhibitory postsynaptic currents (mIPSCs), was not significantly affected by rhGH administration or its perfusion onto brain slices, in P14 mice (Fig. 1i, j, m, n).

To examine if the effects of rhGH on regulating synaptic transmission were more general, we recorded from CA1 pyramidal neurons, and observed increased mEPSC frequency minutes following rhGH perfusion (Ctrl, 1.27 ± 0.32 Hz; rhGH, 1.79 ± 0.52 Hz; P < 0.05) (Fig. 2a, b), similar to its effects in S1BF. mEPSC frequency of hippocampal CA1 pyramidal neurons also increased 2 h following sc injection of rhGH (Ctrl, 1.61 ± 0.16 Hz; rhGH, 2.78 ± 0.32 Hz; P < 0.01), while its amplitude was not affected (Fig. 2e, g, h). The protein concentration of rhGH in the hippocampus, as well as in plasma, was also higher 2 h after in vivo rhGH injection, and returned to baseline at 24 h (Fig. 2f, k). Inhibitory synaptic transmission was not significantly affected either in vivo or in vitro (Fig. 2c, d, i, j). Consistent with the electrophysiological results, we observed significant increases in spine density in basal and apical dendrites, in both dorsal and ventral hippocampal CA1 pyramidal neurons (Fig. 2l–s).

Fig. 2: Pyramidal cells in hippocampal CA1 region show the same mEPSC and mIPSC changes as in S1BF.
figure 2

Representative traces, average waveforms (a, c) and summary graphs (b, d) of the effect of perfusing rhGH (1 μM) onto acute hippocampal CA1 slices on mEPSC (b) or mIPSC (d) frequency and amplitude. n = 11–14 cells from 3 mice per experiment. e Schematic of experimental timeline, showing sc injection of rhGH (or saline Ctrl), followed by whole-cell recordings in CA1 pyramidal neurons of P14 mice. f Protein level of rhGH in the hippocampus of P14 mice 2 h and 24 h after sc injection of 2 mg/kg rhGH. n = 7–9 mice per condition. Representative traces, average waveforms (g, i) and summary graphs (h, j) of the effects of sc injection of rhGH (2 mg/kg) on mEPSC and mIPSC frequency and amplitude. n = 15–16 cells from 3 mice per condition. k Protein level of rhGH in the plasma of P14 mice 2 h and 24 h after sc injection of 2 mg/kg rhGH. n = 5–6 mice per condition. Representative images (l–o) and summary graphs (p–s) of hippocampal basal and apical dendritic spine density in rhGH (2 mg/kg)-injected mice. n = 29–45 neurons from 3 to 5 mice per condition (Scale bar, 10 μm). Data are presented as mean ± SEM. Statistics are measured using paired t test (b, d), unpaired t test (h left, j, p–s), Mann–Whitney test (h right), or two-way ANOVA followed by Tukey’s multiple comparisons test (f, k). *P < 0.05, **P < 0.01, ***P < 0.001, n.s., no significance.

Full size image

Together, these results demonstrated that rhGH promotes excitatory synaptic transmission and spinogenesis in P14 mice, in both S1BF L2/3 and CA1 pyramidal neurons, without significantly affecting GABAergic transmission.

rhGH promotes both excitatory and inhibitory synaptic transmission of hippocampal CA1 and S1BF L2/3 pyramidal neurons at P21

To determine if the effects of rhGH on synaptic transmission were constant or dynamic over the course of development, we performed electrophysical recordings at the later developmental stage of P21. We found that sc injection of rhGH at P21 increased the mEPSC frequency of pyramidal neurons in both L2/3 of S1BF (Ctrl, 6.92 ± 0.68 Hz; rhGH, 9.91 ± 0.84 Hz; P < 0.01) (Fig. 3a–c) and in hippocampal CA1 (Ctrl, 1.09 ± 0.18 Hz; rhGH, 1.98 ± 0.24 Hz; P < 0.05) (Fig. 4a, b). Consistently, perfusing brain slices from P21 mice with rhGH increased mEPSC frequency in both hippocampal CA1 (Ctrl, 0.75 ± 0.26 Hz; rhGH, 1.11 ± 0.36 Hz; P < 0.001) (Fig. 4e, f) and S1BF L2/3 (Ctrl, 8.19 ± 0.69 Hz; rhGH, 10.21 ± 1.04 Hz; P < 0.01) (Fig. 3f, g) pyramidal neurons. mEPSC amplitude was not affected by either rhGH injection in vivo or perfusion in vitro. Thus, the effect of rhGH in promoting excitatory synaptic transmission was similar at the developmental stages of P14 and P21.

Fig. 3: rhGH increases both excitatory and inhibitory synaptic transmission in S1BF at P21.
figure 3

a Schematic of experimental timeline, showing sc injection of rhGH (or saline Ctrl), followed by whole-cell recordings in acute cortical slices of P21 mice. Representative traces, average waveforms (b, d) and summary graphs (c, e) of the effect of sc injection of rhGH (2 mg/kg) on mEPSC or mIPSC frequency and amplitude. n = 18–19 cells from 3 mice per condition. Representative traces (f, h) and summary graphs (g, i) of the effect of perfusing rhGH (1 μM) onto S1BF L2/3 pyramidal neurons on mEPSC or mIPSC frequency and amplitude. n = 11–14 cells from 3 mice per experiment. Representative traces (j, l) and summary graphs (k, m) of firing responses to stepwise current injections with or without rhGH in S1BF L2/3 pyramidal neurons at P14 or P21. n = 20–22 cells from 3 mice per condition. Cells were held at −60 mV, in NBQX, D-AP5, and gabazine. Data are presented as mean ± SEM. Statistics are measured using paired t test (g, i), unpaired t test (c, e) or two-way ANOVA, followed by Bonferroni’s post hoc test (k, m). *P < 0.05, **P < 0.01, ***P < 0.001, n.s., no significance.

Full size image
Fig. 4
figure 4

rhGH increases both excitatory and inhibitory synaptic transmission in hippocampal CA1 neurons at P21. Representative traces, average waveforms (a, c) and summary graphs (b, d) of the effect of sc injection of rhGH (2 mg/kg) at P21 on mEPSC or mIPSC frequency and amplitude. n = 15–17 cells from 3 mice per condition. Representative traces, average waveforms (e, g) and summary graphs (f, h) of the effect of perfusing rhGH (1 μM) onto acute hippocampal CA1 slices on mEPSC or mIPSC frequency and amplitude. n = 11–12 cells from 3 mice per experiment. Representative traces (i, k) and summary graphs (j, l) of firing responses to stepwise current injections with or without rhGH in CA1 pyramidal neurons at P14 or P21. n = 25 cells from 3 mice per condition. Cells were holding at −60 mV, in NBQX, D-AP5, and gabazine. Data are presented as mean ± SEM. Statistics are measured using paired t test (f right, h), unpaired t test (b right, d left), Mann–Whitney test (b left, d right), Wilcoxon matched-pairs signed rank test (f left), or two-way ANOVA, followed by Bonferroni’s post hoc test (j, l). *P < 0.05, **P < 0.01, ***P < 0.001, n.s., no significance.

Full size image

Interestingly, rhGH significantly increased mIPSC frequency in both S1BF L2/3 (Ctrl, 8.22 ± 0.95 Hz; rhGH, 12.94 ± 0.77 Hz; P < 0.001) (Fig. 3d, e) and hippocampal CA1 (Ctrl, 7.09 ± 0.65 Hz; rhGH, 9.65 ± 0.57 Hz; P < 0.01) (Fig. 4c, d) pyramidal neurons following sc injection, in contrast to its lack of effects at P14. Brain slice perfusion of rhGH had similar effects as in vivo injection in both brain regions (Figs. 3h, i and 4g, h). Thus, at the later developmental stage of P21, rhGH promotes both excitatory and inhibitory synaptic transmission, in contrast to affecting only excitatory synaptic transmission at P14. In other words, it has developmentally dynamic effects in regulating neural circuit development and function.

Since rhGH upregulated both excitatory and inhibitory synaptic transmission 2 h post-injection at P21, we also tested its effects at the later time point of 24 h post-injection, to assay if rhGH had long-lasting effects. In mice receiving sc injection of rhGH at P20 and assayed 24 h later (Fig. s1a), mEPSC and mIPSC frequency and amplitudes were not significantly affected, suggesting that a single injection of rhGH did not have lasting effects on synaptic transmission (Fig. s1b–e). The lack of effects at 24 h consists with plasma, cortical and hippocampal rhGH levels returning essentially to baseline level 24 h post-injection (Figs. 1b, 2f, k; please also see Discussion).

Intrinsic excitability, measured in the presence of blockers of both excitatory and inhibitory synapses, was not significantly affected at P14 (Figs. 3j, k and 4i, j), and was reduced following GH application at P21 (Figs. 3l, m and 4k, l), suggesting that in addition to regulating excitatory and inhibitory synaptic transmission, rhGH also regulates ion channels that contribute to action potential firing, in a developmental dynamic fashion.

GHR mediates the effects of rhGH on synaptic transmission at P21

Once released into the circulation, GH binds and activates the cell-surface receptor GHR [53], which is highly expressed in cortex and hippocampus [27, 54]. To examine whether the effect of rhGH is mediated through GHR, we used RNA interference to specifically reduce the expression of GHR in S1BF L2/3 pyramidal neurons (GHR-RNAi-GFP+). The effectiveness of the RNAi was verified in vitro (Fig. 5a). Lentivirus expressing GHR RNAi or Ctrl was injected into S1BF, and recordings were made at P21 (Fig. 5b). In pyramidal neurons with GHR expression knocked down, rhGH-induced increase in mIPSC frequency was effectively blocked (Ctrl, 9.22 ± 1.91 Hz; rhGH, 9.55 ± 1.96 Hz, P = 0.57) (Fig. 5c, d). As controls, rhGH-induced increase in mIPSC frequency was unaffected in neighboring uninfected neurons (GHR-RNAi-GFP-) (Fig. 5e, f) or in neurons expressing a control lentivirus (Ctrl-RNAi-GFP+ and Ctrl-RNAi-GFP-) (Fig. S2). The rhGH-induced increase in mEPSC frequency was also effectively blocked in neurons expressing GHR RNAi (Ctrl, 4.55 ± 0.45 Hz; rhGH, 4.27 ± 0.54 Hz, P = 0.25) (Fig. 5g, h), while neighboring uninfected neurons were unaffected (Ctrl, 4.20 ± 0.73 Hz; rhGH, 5.13 ± 0.84 Hz, P < 0.05) (Fig. 5i, j). These results demonstrate that rhGH mediates its effects on excitatory and inhibitory synaptic transmission through GHR.

Fig. 5: GHR is required for regulation of synaptic transmission by rhGH.
figure 5

a Representative immunoblots and quantitation of GHR knockdown using lentiviral-mediated RNAi. b Schematics showing time line of lentiviral injection and whole-cell recordings. Representative trace and average waveforms from Lenti-GHR-RNAi-infected (c, g) S1BF L2/3 pyramidal neurons or neighboring uninfected neurons (e, i), before and after rhGH (1 μM) perfusion; corresponding summary graphs of mIPSC amplitude and frequency (d, f) or mEPSC amplitude and frequency (h, j). n = 8–9 cells from 3 to 5 mice per experiment. Data are presented as mean ± SEM. Statistics are measured using Wilcoxon matched-pairs signed rank test (d, f) or paired t test (h, j). *P < 0.05, **P < 0.01, n.s., no significance.

Full size image

The contributions of AMPA and NMDA receptors to GH-dependent increase in excitatory synaptic transmission

To further examine whether rhGH application increased the total strength of synchronized synaptic release, we measured the amplitude of evoked AMPA receptor-mediated and NMDA receptor-mediated currents at P14 in S1BF L2/3 pyramidal neurons. We found that rhGH application increased the amplitude of evoked NMDA receptor-mediated current (Ctrl, 206.70 ± 26.51 pA; rhGH, 342.63 ± 66.51 pA, P < 0.01) (Fig. 6a, b), without significantly affecting the amplitude of AMPA receptor-mediated current (Ctrl, 242.19 ± 40.36 pA; rhGH, 306.72 ± 51.71 pA, P = 0.06) or AMPA/NMDA ratio (Ctrl, 1.25 ± 0.16 pA; rhGH, 1.03 ± 0.15 pA, P = 0.19) (Fig. 6c, d). Consistent with these electrophysiological results, sc rhGH injection led to increased membrane-associate level of NMDA receptor subunit NR2B (Fig. 6e), without significantly affecting the level of AMPA receptor subunit GluR1 (Fig. 6f).

Fig. 6: Regulation of synaptic transmission by rhGH is dependent on NMDA receptors.
figure 6

a–d Example evoked NMDA- (top) and AMPA-receptor (bottom) mediated currents with (right) or without (left) rhGH. Filled circles indicate the time point at which measurements were taken (a). Summary graphs of the amplitude of evoked NMDA- or AMPA-receptor mediated current (b, c), and of AMPA/NMDA ratio (d). n = 16 cells from 4 mice. e, f Representative immunoblots and quantitation of NR2B and GluR1 levels from hippocampal membrane fractions of Ctrl and rhGH-treated mice. 6–9 mice per condition. g Schematic of experimental timeline, showing rhGH application, in the presence or absence of the NMDA receptor antagonist D-AP5. Representative traces, average waveforms (h, j) and summary graphs (i, k) of the effect of perfusing rhGH (1 μM) onto acute cortical slices on mEPSC or mIPSC frequency and amplitude at P14 or P21. n = 10–15 cells from 3 mice per experiment. Data are presented as mean ± SEM. Statistics are measured using Wilcoxon matched-pairs signed rank test (b, c), paired t test (d, i, k), unpaired t test (e, f). **P < 0.01, n.s., no significance.

Full size image

To examine the contribution of NMDA receptors to rhGH-induced change in synaptic transmission, we measure the effect of rhGH application in the presence of the NMDA receptor antagonist D-AP5 (50 µM) (Fig. 6g). For this set of experiments, each day, the effect of rhGH was assayed in the absence of D-AP5, as positive control (see Materials and Methods). D-AP5 application effectively blocked rhGH-induced increase in mEPSC frequency (Ctrl, 2.10 ± 0.28 Hz; rhGH, 1.84 ± 0.23 Hz, p = 0.06) (Fig. 6h, i) at P14.

Accumulating evidence suggests that extra-synaptic NMDA receptors can also modulate GABAergic synaptic transmission [55]. We thus measured whether the upregulation of inhibitory synaptic transmission by rhGH at P21 was also NMDA receptor-dependent. Indeed, in the presence of D-AP5, rhGH-induced increased in mIPSC was blocked (Ctrl, 14.15 ± 1.25 Hz; rhGH, 14.22 ± 1.21 Hz, P = 0.90) (Fig. 6j, k). Together, the above results suggest that NMDA receptors play important roles in mediating or modulating the effects of rhGH on both excitatory and inhibitory synaptic transmission.

Since activation of NMDA receptors leads to increased calcium influx, we loaded neurons with the calcium chelator BAPTA (5 mM) and measured synaptic transmission. The effect of rhGH application on increasing mEPSC frequency at P14 (Ctrl, 3.62 ± 0.35 Hz; rhGH, 3.45 ± 0.37 Hz, P = 0.18) (Fig. s3a, b) and mIPSC frequency at P21 (Ctrl, 8.19 ± 0.81 Hz; rhGH, 7.89 ± 0.88 Hz, P = 0.50) (Fig. s3c, d) were completely blocked in BAPTA-loaded cells. Taking together, the above results suggest that rhGH regulates excitatory and inhibitory synaptic transmission through postsynaptic NMDA receptors and their downstream calcium-dependent signaling.

Growth hormone differentially affects excitatory/inhibitory balance at P14 and P21

Since rhGH differentially regulated excitatory and inhibitory synaptic transmission at P14 and P21, we asked whether it affected excitatory/inhibitory balance differentially at these different ages. Recording synaptic transmission from the same set of neurons before and after rhGH perfusion (Fig. 7a), we found that rhGH significantly increased mEPSC frequency (Ctrl, 4.44 ± 0.51 Hz; rhGH, 5.37 ± 0.52 Hz, P < 0.01), but did not affect mIPSCs at P14 (Fig. 7b–e), consistent with the results of recording these two parameters separately in different neurons (Fig. 1k–n). Importantly, these changes led to a significant increase in the excitatory/inhibitory (E/I) ratio at P14 (Ctrl, 0.67 ± 0.12 Hz; rhGH, 2.86 ± 0.49 Hz, P < 0.01) (Fig. 7f). mIPSC (Ctrl, 5.86 ± 0.85 Hz; rhGH, 6.94 ± 1.07 Hz, P < 0.05) and mEPSC (Ctrl, 4.08 ± 0.58 Hz; rhGH, 5.67 ± 0.82 Hz, P < 0.05) frequency of P21 mice also increased following rhGH application (Fig. 7h–k), consistent with our previous results (Fig. 3f–i). In contrast to P14, The E/I ratio of mEPSCs and mIPSCs was unaffected by rhGH perfusion (Ctrl, 0.76 ± 0.11 Hz; rhGH, 0.95 ± 0.16 Hz, P = 0.09) (Fig. 7l) at P21. The amplitude of mIPSCs and mEPSCs, as well as their ratios, was not affect by rhGH application at either age (Fig. 7e, g, k, m). These results demonstrate that the effects of rhGH on E/I balance is dynamic over the course of development.

Fig. 7: Regulation of E/I ratio by rhGH in S1 L2/3 pyramidal neurons at P14 and P21.
figure 7

a Schematic of experimental timeline showing rhGH application and E/I ratio measurements. mEPSC and mIPSC pairs were recorded from each cell, in the following order: Ctrl mEPSC, Ctrl mIPSC, rhGH mIPSC, rhGH mEPSC. Representative traces, average waveforms (b, c) and summary graph (d–g) of the effect of perfusing rhGH (1 μM) on mEPSC or mIPSC frequency and amplitude, as well as E/I ratio in S1BF L2/3 pyramidal neurons of P14 mice. n = 10 cells. Representative traces, average waveforms (h, i) and summary graph (j–m) of the effect of rhGH on mEPSC or mIPSC frequency and amplitude, as well as E/I ratio of P21 mice. n = 10 cells. Data are presented as mean ± SEM. Statistics are measured using paired t test. *P < 0.05, **P < 0.01, n.s., no significance.

Full size image

Discussion

In this study, we found that rhGH dynamically regulates excitatory and inhibitory synaptic transmission in hippocampal CA1 and S1BF L2/3 pyramidal neurons, both in vivo and in vitro, in a GHR and NMDA receptor-dependent manner. These results demonstrate that the effect of rhGH on E/I balance alters over the course of development.

Rapid effects of rhGH in promoting synaptic transmission and spinogenesis

A key finding of our study is that rhGH can rapidly upregulate excitatory synaptic transmission and increase dendritic spine density, within 2 h of sc injection in vivo and within minutes of perfusion onto brain slices (Figs. 1–4). Most previous studies examined the effects of multiple days of rhGH administration [56, 57]. We only found two previous studies which measured the acute effects of rhGH. One study found that rhGH enhanced long-term potentiation (LTP), while another found that it increased evoked AMPA and NMDA receptor-mediated currents, following 30–150 min of perfusion [40, 41]. Both studies used adult or aged rats. To our knowledge, our study was the first to examine the effects of rhGH on developing (pre-weaning) rodents, and the first to show that rhGH can rapidly increase mEPSC frequency, as well as the density of dendritic spines.

Field recordings provide evidence of how a group of cells change. By recording mEPSCs and measuring spine density, we provide evidence for how rhGH regulates excitatory synapses at the level of individual neurons and individual synapses. Our results showing increased spine density and increased mEPSC frequency suggest that rhGH increases the total number of excitatory synapses of hippocampal CA1 and S1BF L2/3 pyramidal neurons. Mechanistically, we showed that rhGH increased evoked NMDA receptor-mediated currents (Fig. 6a–d), suggesting important contribution of postsynaptic NMDA receptors in mediating its effects. Consistently, rhGH increased the membrane-associated level of the NMDA receptor subunit NR2B (Fig. 6e), while blocking NMDA receptor function or postsynaptic calcium-dependent signaling inhibited the effects of rhGH (Fig. 6h-k, S3). These results add to the previously reported NMDA receptors in mediating the effects of rhGH [32, 33, 40, 41, 58].

The effects of rhGH in promoting excitatory synaptic transmission was observed at both P14 and P21, in both hippocampal CA1 and S1BF L2/3 pyramidal neurons. Previous studies mostly focused on the effects of rhGH in hippocampal CA1 neurons in adult and aging rodents, and also showed increased excitatory synaptic transmission [40, 41]. Put together, these results suggest that rhGH likely has general effects in promoting excitatory synaptic transmission, across ages and across brain regions. More future studies in different brain regions and at different ages are needed to fully determine the generality of the effects of rhGH on excitatory synaptic transmission.

Regulation of inhibitory synaptic transmission and excitatory/inhibitory balance by rhGH

The most exciting and interesting result of our study is identification of a development-dependent effect of rhGH on inhibitory synaptic transmission. Specifically, we found that rhGH enhanced inhibitory synaptic transmission at P21, but not at the younger age of P14 (Figs. 1–4). To our knowledge, this is also the first report of an acute effect of rhGH on inhibitory synaptic transmission. Mechanistically, we showed that rhGH regulated inhibitory synaptic transmission through its receptor GHR (Fig. 5) and required NMDA receptor-dependent signaling (Fig. 6). Consistent with rhGH increasing excitatory synaptic transmission throughout development, but only promoting inhibitory synaptic transmission only at P21, rhGH perfusion altered E/I balance at P14, but not at P21 (Fig. 7).

The dynamic effects of rhGH on synaptic transmission and E/I balance draws attention to the importance of timing of rhGH administration, regarding its effects on brain development and plasticity. Much more extensive studies in different brain regions and at different developmental stages would be required to draw general conclusions. In most circumstances, rhGH is administrated to promote overall growth, in individuals with growth deficits. Thus, an administration protocol that maintains E/I balance would be optimal, as an increase in excitation without corresponding increase in inhibition may elevate the risk of seizures. Comparing rodent and human ages is complex. Whether rhGH also has developmentally dynamic effects on inhibitory synaptic transmission in the developing human brain remains to be determined. Our results draws attention to the importance of developmental age in rhGH administration.

Potential implications for rhGH therapy

While it is generally recognized that GH deficiency (GHD) is associated with cognitive impairment, and that GH administration, in addition to promoting overall growth, also significantly improves memory and cognitive capacity, its optimal mode of administration and safety remains under debate [28,29,30,31]. Our results demonstrating dynamic effects of rhGH administration on inhibitory synaptic transmission over the course of development add developmental age as an additional factor for consideration. By showing that in addition to regulating synaptic transmission in hippocampal CA1 pyramidal neurons, rhGH can also regulate synaptic transmission in S1BF L2/3 pyramidal neurons, we extend the target brain regions of rhGH. These results suggest that many more factors need to be taken into account when assessing the in vivo consequences of rhGH administration.

The rapid effects of rhGH on increasing excitatory and inhibitory synaptic transmission may draw concern from some, worried about accumulation of such effects following multiple administrations. We note that although plasma and brain tissue levels of rhGH are elevated 2 h following sc injection Figs. 1b, 2f, k, they return to baseline levels at 24 h. Consistently, 24 h after sc injection of rhGH, excitatory and inhibitory synaptic transmission are not significantly altered (Fig. S1). This result suggests that rhGH administration only has transient effects on synaptic transmission. However, it is not known whether multiple administrations of rhGH may have cumulative effects. Previous studies have shown that the pattern of rhGH administration, in addition to the dose, affected outcome [56, 57]. Based on existing evidence, we suggest that rhGH administration in P21 or older mice, or equivalent age in humans, would be preferable, as E/I balance is maintained.

In summary, by systematically characterizing the short-term effects of rhGH administration on excitatory and inhibitory synaptic transmission in multiple brain regions and multiple developmental time points, our results reiterate the complexity by which neuromodulators dynamically regulate neural circuit wiring and plasticity. Our results draw particular attention to dynamic regulation of E/I balance over the course of development. More extensive characterization of the short- and long-term effects of rhGH in animal models and humans are needed to better guide safe administration and optimal treatment effects in humans.