Significant outcomes
-
Outcome 1: We find a neurodevelopmental increase in synaptic activity in hippocampal GABA transmission in both control rats and the LBW rat model of depression.
-
Outcome 2: Altered paired-pulse behaviour of evoked GABA release indicates lowered GABA vesicle release probability in the LBW model.
-
Outcome 3: Defective synaptic GABA release in the LBW model is accompanied by increases in mRNA encoding for distinct presynaptic proteins, including Snap-25 and Scamp2, all pointing to a GABAergic synaptopathy in the LBW rat model of depression.
Limitations
-
Exogenous stress hormone was administered in a rodent animal model, and tissues were analysed ex vivo.
Introduction
Major depressive disorder (MDD) is one of the most common neuropsychiatric diseases with the core symptoms of depressed mood and anhedonia. MDD also presents with a broad spectrum of other symptoms, including cognitive deficits and sleep disturbances, and has a high rate of comorbidity with anxiety (Fava et al., Reference Fava, Rankin, Wright, Alpert, Nierenberg, Pava and Rosenbaum2000; Mineur et al., Reference Mineur, Obayemi, Wigestrand, Fote, Calarco, Li and Picciotto2013). MDD is influenced by genetic background but is also associated with environmental factors, such as adverse life events. Stress represents the most significant vulnerability factor in the development of depressive disorders (Silva et al., Reference Silva, Maffioletti, Gennarelli, Baune and Minelli2021). Moreover, an increasing number of studies suggest a neurodevelopmental origin of the disease (Markham & Koenig, Reference Markham and Koenig2011; Nugent et al., Reference Nugent, Tyrka, Carpenter and Price2011; Van den Bergh et al., Reference Van den Bergh, van den Heuvel, Lahti, Braeken, de Rooij, Entringer, Hoyer, Roseboom, Räikkönen, King and Schwab2020). During foetal or early life, stress exposure can lead to adverse changes in the developing brain both structurally and functionally, which may increase the risk of MDD in adulthood (Matthews, Reference Matthews2000). The underlying pathophysiology of MDD is still poorly understood, but existing models and hypotheses suggest dysregulation in several neurotransmitter systems (Krishnan & Nestler, Reference Krishnan and Nestler2008), including the GABAergic inhibitory system (Luscher et al., Reference Luscher, Shen and Sahir2011; Zhang et al., Reference Zhang, Hu, Dong, Huang, Jiao, Hu, Dai, Yi, Gong, Li, Wang and Xu2021).
GABA is the major inhibitory neurotransmitter in the mammalian brain and is released by interneurons and acts on ionotropic GABAA and metabotropic GABAB receptors, thereby exerting inhibitory control of neuronal excitability. Synaptic GABAA receptors (GABAARs) mediate phasic inhibition, a rapid form of neurotransmission, while extrasynaptic GABAARs are coupled to a more persistent tonic inhibition (Farrant & Nusser, Reference Farrant and Nusser2005). It is hypothesised that GABAergic deficits are closely linked to mood disorders (Luscher et al., Reference Luscher, Shen and Sahir2011). Clinical studies showed reduced GABA levels in plasma (Petty & Schlesser, Reference Petty and Schlesser1981) and cerebrospinal fluid (Gerner & Hare, Reference Gerner and Hare1981) of depressed individuals. Altered GABAAR subunit mRNA expression was shown in different brain regions of humans (Merali et al., Reference Merali, Du, Hrdina, Palkovits, Faludi, Poulter and Anisman2004), suggesting both genetic and epigenetic impact in MDD. Moreover, experimental studies showed GABAAR alteration in rodents with anxiety-like behaviour induced by early life stress exposure (Caldji et al., Reference Caldji, Francis, Sharma, Plotsky and Meaney2000, Reference Caldji, Diorio and Meaney2003), suggesting GABAergic involvement in the neurodevelopmental origin of mental illnesses. Previously, our group demonstrated that a chronic mild stress (CMS) rat model of depression shows a functional deficit in GABA release, which could be reversed by antidepressant treatment (Holm et al., Reference Holm, Nieto-Gonzalez, Vardya, Henningsen, Jayatissa, Wiborg and Jensen2011; Nieto-Gonzalez et al., Reference Nieto-Gonzalez, Holm, Vardya, Christensen, Wiborg and Jensen2015).
Here, we examined GABAergic neurotransmission in the low birth weight (LBW) rat model of depression (Hougaard et al., Reference Hougaard, Andersen, Kjaer, Hansen, Werge and Lund2005). In this model, animals are exposed to the synthetic glucocorticoid dexamethasone (DEX) during foetal life, mimicking elevated maternal stress hormone levels during pregnancy (Conti et al., Reference Conti, Spulber, Raciti and Ceccatelli2017; Spulber et al., Reference Spulber, Conti, DuPont, Raciti, Bose, Onishchenko and Ceccatelli2015). Compared to control offspring, LBW rats show several deficits. Of relevance for depression, administration of DEX during the last week of gestation significantly increased immobility in the forced swim test and reduced sucrose preference (Abildgaard et al., Reference Abildgaard, Lund and Hougaard2014; Wu et al., Reference Wu, Huang, Gong, Xu, Lu, Sheng and Ni2019). Furthermore, prenatal exposure to DEX increased the susceptibility to CMS and hence the propensity to express depressive-like behaviour phenotypes (Oliveira et al., Reference Oliveira, Bessa, Mesquita, Tavares, Carvalho, Silva, Pêgo, Cerqueira, Palha, Almeida and Sousa2006). Other deficits include altered hypothalamic–pituitary–adrenal axis (HPA) function and expression of corticotropin-releasing hormone (CRH) and CRH receptor type 1 (CRHR1) in the hippocampus, reduced mobility, and altered startle behaviour (Hougaard et al., Reference Hougaard, Andersen, Kjaer, Hansen, Werge and Lund2005, Reference Hougaard, Mandrup, Kjaer, Bøgh, Rosenberg and Wegener2011; Kjaer et al., Reference Kjaer, Wegener, Rosenberg and Hougaard2010; Xu et al., Reference Xu, Sheng, Wu, Bao, Zheng, Zhang, Gong, Lu, You, Xia and Ni2018). Additionally, LBW rats show increased anxiety-like behaviours in adulthood (Oliveira et al., Reference Oliveira, Bessa, Mesquita, Tavares, Carvalho, Silva, Pêgo, Cerqueira, Palha, Almeida and Sousa2006; Nagano et al., Reference Nagano, Ozawa and Suzuki2008).
Here, we examined postsynaptic GABAAR-mediated inhibitory signalling onto granule cells in the dentate gyrus using whole-cell patch-clamp recordings in brain slices from LBW rats. In addition, quantitative real-time polymerase chain reaction (real-time qPCR) was used to analyse the levels of presynaptic protein transcripts in the hippocampus. In comparison to control offspring, we found in LBW rats a dysfunction in the GABAergic synaptic plasticity and elevated expression levels of two genes, Snap-25 and Scamp2, of which the corresponding proteins are elements of the vesicle release machinery (Brand et al., Reference Brand, Laurie, Mixon and Castle1991; Söllner et al., Reference Söllner, Whiteheart, Brunner, Erdjument-Bromage, Geromanos, Tempst and Rothman1993). These results suggest that altered GABA release may be an essential feature of depressive disorders.
Material and methods
Ethics statement, animals, and dexamethasone exposure
Forty time-mated young adult rats (Wistars, HanTaC:WH, SPF) arrived at gestational day (GD) 3 at the Danish National Research Centre for the Working Environment, Copenhagen, Denmark. The rats were randomly distributed pairwise to white plastic cages (Eurostandard III, Scanbur, Denmark) with bedding of pine wood shavings and nesting material (Enviro-Dri, Brogaarden, Denmark). Environmental conditions were automatically controlled with a 12-h light–dark cycle with lights off at 06.00 am. Food (Altromin Standard Diet 1324) and tap water were provided ad libitum.
From GD 14 to 21, dams were injected subcutaneously in the nape of the neck with DEX (Sigma-Aldrich, Denmark) (150 µg/kg) once daily (LBW group; Fig. 1). Vehicle control animals were injected with 4% ethanol/isotonic saline. Injections were given during the dark phase of the light–dark cycle. For the study of 4- to 5- and 6- to 8-week-old rats, a maximum of 1 male per litter was selected randomly at weaning on postnatal day 21 (P21) and housed in groups of four with cage-mates of similar prenatal exposure until transfer to Translational Neuropsychiatry Unit, Department of Clinical Medicine, Aarhus University. Here, they were housed in clear Type 3 Makrolon cages and kept as described above. Offspring for the study of 2-week-old rats were also generated here, using similar procedures as for the older age groups.
DEX protocol in LBW rats. Female rats were exposed daily to dexamethasone (150 µg/kg) during the last week of gestation. Electrophysiological measurements were carried out on brain slices of male offspring at the age of 2, 4–5, and 6–8 weeks, respectively.
All efforts were made to minimise animal suffering and to reduce the number of animals used. The Danish National Committee for Ethics in Animal Experimentation, appointed by the Danish Ministry of Justice, granted ethical permission for the studies. The Laboratory Animal Welfare Committee at the National Research Centre for the Working Environment reviewed the LBW model of depression and found it in compliance with the ethics of the NRCWE regarding experiments of laboratory animals. All procedures were carried out in compliance with the EC Directive 86/609/EEC and Danish law regulating experiments on animals (permission 2007-561/1378 and 2007-561/1396). Two, 4- to 5-, and 6- to 8-week-old male offspring were used for electrophysiological studies.
Quantitative real-time polymerase chain reaction
Hippocampi were dissected (left and right) from 8-week-old LBW and control vehicle animals. Dissection, tissue homogenisation, RNA extraction, RNA characterisation, and cDNA synthesis were carried out as described previously (Elfving et al., Reference Elfving, Bonefeld, Rosenberg and Wegener2008). cDNA was stored undiluted at −80°C until use. The cDNA samples were diluted 1:30 with DEPC water before being used as a real-time qPCR template.
Real-time qPCR reactions were carried out in 96-well PCR plates using the Mx3000P (Stratagene, USA) and SYBR Green. The gene expression of 8 reference genes (18sRNA, ActB, CycA, Gapd, Hmbs, Hprt1, Rpl13A, Ywhaz), 14 genes encoding for synaptic proteins (Scamp2, Snap-25b (the protein encoded by the Snap-25b isoform will be referred to as Snap-25 hereafter), Snap-29, Snapin, Syntaxin 1A, Synapsin I-III, Synaptophysin, Synaptotagmin I-III, Vamp1, and Vamp2), and 7 other genes (Gr, Mr, Munc13, Munc18, Rims1, Scna, and Vgat) were investigated as previously described (Bonefeld et al., Reference Bonefeld, Elfving and Wegener2008; Elfving et al., Reference Elfving, Bonefeld, Rosenberg and Wegener2008). Briefly, each SYBR Green reaction (10 µl total volume) contained 1x SYBR Green master mix (BIORAD, CA, USA), 0.5 µM primer pairs, and 3 µl of diluted cDNA. All samples were run in duplicate. A standard curve, performed in duplicate, was generated on each plate. Essential gene-specific data about primer sequence and amplicon sizes are given in Tables 1 and 2. Primers were obtained from DNA Technology A/S and Sigma-Aldrich, Denmark.
Characteristics of gene-specific real-time qPCR primers – Reference genes
* GenBank accession number of cDNA and corresponding gene, available at http://www.ncbi.nlm.nih.gov/.
† Amplicon length in base pairs.
Characteristics of gene-specific real-time qPCR primers – Target genes
* GenBank accession number of cDNA and corresponding gene, available at http://www.ncbi.nlm.nih.gov/.
† Amplicon length in base pairs.
For data normalisation, we first measured mRNA levels for the reference genes. Stability comparison of the expression of the eight reference genes was conducted with the Normfinder software (www.moma.dk/normfinder-software) (Andersen et al., Reference Andersen, Jensen and Ørntoft2004), and the best combination of two was selected. Values for each individual were normalised with the geometric mean of the reference genes CycA and ActB in the hippocampus.
Brain slice preparation
Rats were anaesthetised with isoflurane and decapitated. The brain was dissected and placed in ice-cold artificial cerebrospinal fluid (ACSF) composed of (in mM): 126 NaCl, 2.5 KCl, 2 CaCl2, 2 MgCl2, 1.25 NaH2PO4, 26 NaHCO3, 10 D-glucose (osmolality 305–315 mOsmol/kg), 2 kynurenic acid, pH 7.4 when bubbled with carbogen (5% CO2 – 95% O2). 350-μm-thick coronal slices were prepared with Leica Microtome VT1200S (Leica Biosystems, Germany) in ice-cold bubbled ACSF. Slices rested in a holding chamber for at least 1 h in bubbled ACSF at room temperature (22–25°C). 0.2 mM ascorbic acid and 0.2 mM pyruvic acid were added to the ACSF during slicing and storage to improve slice quality.
In vitro electrophysiology
For whole-cell patch-clamp recordings, slices were transferred into a recording chamber perfused with 33 ± 1°C bubbled ACSF at 2–3 ml/min. Neurons were visualised using a custom-built infrared microscope with a × 40 water-immersion objective (Olympus, Ballerup, Denmark) and a CCD100 camera (DAGE-MTI, Michigan City, IN, USA). Recordings were carried out using a MultiClamp 700B amplifier (Axon Instruments, Union City, CA, USA). Patch-pipettes were pulled from borosilicate glass (OD = 1.5 mm, ID = 0.8 mm; Garner Glass Company, Claremont, CA, USA) using a DMZ Universal Puller (Zeitz Instruments, Munich, Germany). After filling the pipettes with intracellular solution containing (in mM): 140 CsCl, 2 MgCl2, 0.05 EGTA, 10 HEPES, adjusted to pH 7.2 with CsOH (280–290 mOsmol/kg), their resistances were 3–5 MΩ. Giga seals (>1 GΩ) were obtained before break-in, and cells were held in voltage-clamp at a holding potential (Vhold) of -70 mV, while resistance was compensated by around 70% (lag 10 μs). During the experiments, whole-cell capacitance and series resistance were monitored. The observed range of the series resistance was 5.7 MΩ to 19.8 MΩ. The average series resistance change was 1.9 ± 0.19 MΩ, in percentage 19.5% ± 1.9% (n = 107). In some experiments, where the exact changes in resistances were not recorded, the increases were in the order of 2–3 MΩ. In all experiments, the neurons were discarded if resistance increased more than 50% or exceeded 20 MΩ.
Data acquisition and analysis
Currents were low-pass filtered (8-pole Bessel) at 3 kHz before being digitised at 20 kHz using a DA converter (BNC-2110), a PCI acquisition board (PCI-6014, National Instruments, Austin, TX), and a custom-written LabView 6.1 (National Instruments)-based software containing an acquisition interface and analysis module (EVAN v. 1.4, courtesy of Istvan Mody). This software was also used to detect and analyse spontaneous and miniature inhibitory postsynaptic currents (IPSCs) with a 6–8 pA amplitude detection threshold. All events were visually inspected before making an average of around 50 spontaneous inhibitory postsynaptic currents (sIPSCs). Event amplitude, 10–90% rise time, and frequency were measured. sIPSC weighted decay time constant (τw) was calculated using double-exponential fits:
$$I\left( t \right) = {A_1}*\exp \left( { - {t \over {{\tau _1}}}} \right) + {A_2}*\exp \left( { - {t \over {{\tau _2}}}} \right)$$
where I(t) is the current as a function of time (t), A 1 and A 2 are amplitude constants, and τ1 and τ2 are the two decay time constants. The goodness of fit was determined by visual inspection of the residuals.
IPSCs were also electrically evoked by paired-pulse stimulation every 10 s with an inter-event interval of 50 ms using a bipolar matrix microelectrode (FHC Inc, ME, USA) placed in the granule cell layer 200–300 µm from the recorded cell. The stimulation intensity was kept constant at 20–40% above the threshold for evoking single IPSCs. The paired-pulse ratio of average IPSC amplitudes was measured from averages of 10–20 sweeps.
Tonic GABAA receptor-mediated currents were assessed by a high concentration of the GABAA receptor antagonist 2-(3-carboxypropyl)-3-amino-6-methoxyphenyl-pyridazinium bromide (SR95531 > 100 μM), which produced an outward shift in the holding current. For quantification, 5 ms long samples were taken from the recording every 100 ms and plotted against time, omitting baseline points falling onto the decay of IPSCs (Drasbek & Jensen, Reference Drasbek and Jensen2006). Mean currents were calculated in 4 s long segments at three time points (denoted a, b and c): just before SR95531 application (b), 20 s before (a) and after (c). The tonic current was calculated as c – b. If the baseline variation (b – a), which defines the stability of the recording, was more than 6 pA, recordings were discarded.
Statistical analysis
For electrophysiological sIPSC and paired-pulse data, two-way ANOVA for group comparison was followed by post hoc Bonferroni tests to determine effects of age and treatment (control vs. LBW). For mIPSCs and tonic GABA currents consisting of one age group, unpaired Student’s t-tests were used. Data are presented as means ± SEM, with n indicating the number of neurons. Real-time qPCR data were analysed using unpaired Student’s t-test followed by Holm–Sidak correcting for multiple comparisons. Statistical analyses were performed with GraphPad Prism version 5.00/8.00 for Windows (GraphPad Software, San Diego, CA, USA).
Compounds
Kynurenic acid, ascorbic acid, SR95531, tetrodotoxin, and dexamethasone were obtained from Sigma-Aldrich (Denmark), while pyruvic acid was purchased from MP Biomedicals (Irvine CA, USA). SR95531 (5 mM) was dissolved in 50% DMSO (dimethyl sulfoxide) and 50% dH2O (distilled water), while dexamethasone was dissolved in 4% ethanol/isotonic saline. Compounds, except dexamethasone, were stored at – 20oC until use.
Results
Earlier studies suggested that hippocampal GABAA receptor-mediated signalling is disturbed in depressive phenotypes observed after exposure of rodents to CMS (Holm et al., Reference Holm, Nieto-Gonzalez, Vardya, Henningsen, Jayatissa, Wiborg and Jensen2011; Nieto-Gonzalez et al., Reference Nieto-Gonzalez, Holm, Vardya, Christensen, Wiborg and Jensen2015). To further investigate the GABAergic involvement in models of depressive disorders and gain insight into GABAergic malfunction during development, we studied LBW rats that develop a depressive phenotype due to prenatal stress exposure.
Control rats show developmental changes in phasic GABAA receptor-mediated signalling
Hippocampal GABAergic interneurons and their synapses show morphological development from embryonic day 13 (E13) (Amaral & Kurz, Reference Amaral and Kurz1985) until adolescence (P28–P55) (Corbin & Butt, Reference Corbin and Butt2011) in rats. The maturation of GABAergic neurons involves changes in their intrinsic properties and a shift from slow to fast postsynaptic signalling (Hollrigel & Soltesz, Reference Hollrigel and Soltesz1997; Banks et al., Reference Banks, Hardie and Pearce2002; Doischer et al., Reference Doischer, Hosp, Yanagawa, Obata, Jonas, Vida and Bartos2008). The development of electrophysiological properties of inhibitory neurons is not complete in early adolescence in the visual cortex, where it reaches a plateau around 8 weeks of age (Jang et al., Reference Jang, Cho, Park, Kim, Yoon and Rhie2010). Therefore, to investigate the postnatal development of GABAergic inhibition in the dentate gyrus, we studied three groups, aged 2 (developing), 4–5 (young adolescence) and 6–8 (later stage of adolescence – early adulthood) weeks, respectively. We recorded GABAA receptor-mediated IPSCs using whole-cell patch-clamp recordings from dentate gyrus granule cells. In the presence of kynurenic acid to block ionotropic glutamate receptors at a holding potential of –70 mV, IPSCs appeared as fast inward currents due to the CsCl-based pipette solution (ECl ∼ 0 mV). sIPSC parameters of untreated control and control vehicle groups showed no statistically significant differences (Table 3; n.s.). Therefore, these data were pooled and referred to as controls. In 2-week-old control offspring, the amplitude of sIPSCs was 39.4 ± 4.7 pA (n = 9). In 4- to 5-week-old control offspring, this had increased to 60.3 ± 5.6 pA (n = 26), while no further increase was observed at 6–8 weeks of age, yielding 59 ± 5.8 pA (n = 13; Fig. 2C, D). A two-way ANOVA confirmed that age did have a statistically significant effect on sIPSC amplitudes (F = 3.299; p = 0.0421). The frequency of sIPSCs did not change significantly with increasing age (Fig. 2A, B). These results suggest that phasic inhibition increases during postnatal development and, most prominently, between 2 and 4–5 weeks of age in our experiments.
Parameters of spontaneous GABAergic inhibitory postsynaptic currents (sIPSCs) in dentate granule cells of untreated control and control vehicle rats
No significant changes found.
Postnatal maturation of GABAergic activity in the dentate gyrus of control and LBW rats. (A) Representative traces showing whole-cell recordings from dentate gyrus granule cells in brain slices from control and LBW rats 4 weeks postnatally. sIPSCs occurred at 4.0 Hz in control and at 2.5 Hz in the LBW slice. (B) sIPSC frequency in control and LBW animals in three age groups (2, 4–5, and 6–8 weeks old). Average sIPSC frequencies were not significantly different, as examined by two-way ANOVA indicated to the right (n.s.). In LBW rats, sIPSC frequencies showed similar values throughout development. (C) Averaged sIPSCs from the three age groups are superimposed, showing the increase in the amplitude during development in control and LBW rats. (D) sIPSC amplitudes in control and LBW rats in different age groups. Amplitudes increased significantly from 39.4 ± 4.7 (2 weeks; n = 9) to 60.3 ± 5.6 (4–5 weeks; n = 26) and 59.0 ± 5.8 (6–7 weeks; n = 13) in control rats. A similar pattern was seen in LBW rats through development, with no significant differences compared to control.
Maturation of sIPSC kinetics in control rats
To study the kinetics of GABAA receptor-mediated synaptic currents, we analysed their waveform, including rise time (10–90%) and decay of sIPSCs. Rise times decreased during postnatal development in control rats. The rise time of 708 ± 168 µs (n = 9) in 2-week-old rats was shortened in 4- to 5- and 6- to 8-week-old rats to 318 ± 17 µs (n = 26) and 318 ± 42 µs, respectively (n = 13; F = 21.51; p < 0.0001; Fig. 3A, B). The decay of averaged sIPSCs showed a significant acceleration from 2 to 4–5 and 6–8 weeks of age, while weighted decay time constants were 8.0 ± 0.6 ms (n = 9), 6.2 ± 0.2 ms (n = 26), and 6.1 ± 0.6 ms (n = 13), respectively (F = 20.11; p < 0.0001; Fig. 3C, D). A decrease in both sIPSC rise time and decay suggests an acceleration of the receptor kinetics during postnatal development in dentate granule cells, and the developmental changes in kinetics seemed to occur particularly during the first month of age.
Postnatal maturation of sIPSC kinetics in dentate gyrus of control and LBW rats. (A, C) Representative sIPSCs from different age groups superimposed after normalisation to the same peak amplitude to illustrate the developmental changes in rise time and decay in control and LBW rats. (B) Histograms summarising the sIPSC rise time in control and LBW animals in the same age groups. Average rise time of sIPSCs decreased significantly from 708 ± 168 (2 weeks; n = 9) to 318 ± 17 (4–5 weeks; n = 26) and 318 ± 42 µs (6–8 weeks; n = 13) in control. LBW rats showed a similar developmental acceleration of the rise time. Rise times in LBW rats were not different in any age group compared to the age-matched controls. (D) sIPSC decay in control and LBW rats in the same age groups. Average decay time constant of sIPSCs significantly decreased from 8.0 ± 0.6 (2 weeks; n = 9) to 6.2 ± 0.2 (4–5 weeks; n = 26) and to 6.1 ± 0.6 ms (6- to 8-week-old group; n = 13) in control. LBW rats show a similar developmental change in the decay phase of sIPSCs. sIPSC decays in LBW rats were not different in any age group compared to control. n.s.: not significant, ***p < 0.001.
Development of phasic inhibition in LBW rats
In LBW rats, sIPSC amplitudes also increased with age (42 ± 5.4 pA, n = 6; 56 ± 8.4 pA, n = 13; 67.9 ± 12.8 pA, n = 17 in 2, 4- to 5- and 6- to 8-week-old rats, Fig. 2C, D). Comparison of sIPSC amplitudes of control and LBW rats in each age group showed no significant differences (Fig. 2D). In comparison to the developmental increase in sIPSC frequencies in controls, LBW rats showed similar frequencies of 2.49 ± 0.53 Hz (n = 6), 2.18 ± 0.45 Hz (n = 13), and 2.59 ± 0.32 Hz (n = 17) in the three age groups (Fig. 2A, B). In addition, there was a tendency of sIPSC frequency of 4- to 5-week-old LBW rats to be lower than that of controls; however, this result was not statistically significant (p = 0.08, Fig. 2A, B).
In LBW rats, sIPSC rise times (10–90%) followed the same developmental pattern as in controls, yielding 832 ± 138 µs (n = 6), 322 ± 27 µs (n = 13), 416 ± 50 µs (n = 17; Fig. 3A, B) with increasing age. Thus, comparison of rise time in the control and LBW animals in each age group showed no significant differences (Fig. 3B). The decay of sIPSCs in LBW rats also showed a similar developmental pattern with increasing age to that of controls (9.0 ± 0.4 ms, n = 6; 6.1 ± 0.4 ms, n = 13, and 5.5 ± 0.4 ms, n = 17, respectively; Fig. 3D). The observed developmental patterns of both rise time and decay in LBW rats suggest a normal development of synaptic GABAA receptor kinetics in LBW rats.
Release probability of GABAergic terminals is reduced in LBW rats
We next studied the short-term plasticity of GABAergic synapses, which correlates with release probability at the presynaptic terminals. We recorded electrically evoked IPSCs by stimulating within the granule cell layer using a paired-pulse protocol in three age groups. At an inter-pulse interval of 50 ms, IPSCs in control animals showed paired-pulse depression in all three age groups, yielding 0.86 ± 0.07 (n = 6), 0.85 ± 0.04 (n = 9), and 0.62 ± 0.08 (n = 5), respectively (Fig. 4A, B). These results indicate a developmentally regulated short-term plasticity in control rats. Interestingly, in 2-week-old LBW rats the paired-pulse ratio was similar to that of controls (0.94 ± 0.11, n = 6); however, at older ages the paired-pulse ratios were increased (at 4–5 weeks: 1.17 ± 0.11, n = 13, p < 0.05; at 6–8 weeks: 1.01 ± 0.07, n = 7, p = 0.10, post hoc Bonferroni test). Thus, we observed a switch from paired-pulse depression to paired-pulse facilitation in the LBW rats, which can be interpreted as a loss of the normal developmental increase in the release probability in LBW animals.
Reduction in GABAergic probability of release in the LBW rats in adolescence and adulthood. (A) Traces showing electrically evoked GABAA receptor-mediated IPSCs in a control and a LBW rat slice. In control, the paired-pulse ratio of the IPSCs was 0.66, whereas it was 1.04 in LBW. The ratios were calculated as the second IPSC amplitude normalised to the first. These exemplar recordings were carried out 8 weeks postnatally, and the responses indicate a reduction of GABA release probability in the LBW rat exposed to dexamethasone in utero. (B) Histogram showing the paired-pulse ratio of the evoked IPSCs in three age groups (2, 4–5, and 6–8 weeks old). Paired-pulse depression was increasing during development in control rats from 0.86 ± 0.07 (2; n = 6) and 0.85 ± 0.04 (4–5; n = 9) to 0.62 ± 0.08 (6–8 weeks old; n = 5). In LBW rats, the paired-pulse depression seen in the 2-week-old group turned into paired-pulse facilitation at 4–5 and 6–8 weeks. Between control and LBW rats, no changes were observed at 2 and 6–8 weeks, but at 4–5 weeks postnatally, paired-pulse ratios were significantly increased in LBW rats. These data suggest a presynaptic defect in evoked GABA vesicle release in LBW animals. n.s.: not significant, **p < 0.01.
Effect of action potential blockade on GABAergic activity
To examine the fraction of GABAergic activity that is not dependent on action potential firing, we added the Na+ channel blocker tetrodotoxin (TTX) to the slices from 4- to 5-week-old rats. We recorded action potential-independent miniature IPSCs (mIPSC). In slices from controls, TTX (1 µM) reduced the IPSC frequency to 1.13 ± 0.29 Hz (n = 7), which was similar to the resulting frequency in LBW slices (1.08 ± 0.24 Hz, n.s., n = 10, Fig. 5A), while the fractional decrease in GABAergic frequency was more pronounced in the control slices (by 69%) than in LBW slices (by 51%; Fig. 5B). Thus, action potential-driven GABA release appeared to be reduced in LBW slices, while TTX resistant miniature release, reflecting the number of active synaptic GABAergic boutons, was similar. The amplitudes and kinetics of averaged mIPSCs showed no significant differences in LBW rats compared to untreated control rats (data not shown).
Extrasynaptic and action potential-independent synaptic GABAA receptor activities are unaltered in adolescent LBW rats. (A) Whole-cell recordings of miniature IPSCs (mIPSCs) showed no change in frequency and amplitude of GABAergic single vesicle responses between groups. Experiments were performed in the presence of TTX, and averages of 50 mIPSCs from each group are shown. Experiments were carried out 4 weeks after birth. (B) The action potential blocker TTX reduced the sIPSC frequency fractionally more in LWB rats compared with controls in 4- to 5-week-old rats. The findings indicate that the action potential-dependent, TTX-sensitive component of GABA release is smaller in LBW rats than in controls. Action potential-independent GABA release was similar in both groups. (C) Tonic extrasynaptic GABAA receptor-mediated currents were revealed by the GABAA receptor antagonist SR95531 (∼100 µM). All-points histograms (left) represent the mean currents and noise levels before and after SR95531. Tonic GABAA-mediated currents were similar in this representative control (23.6 pA) and LBW rat (22.9 pA) experiment. The experiments were performed 4–8 weeks after birth. (D) No significant difference was found between control and LBW rats.
Tonic GABAergic activity is unchanged in LBW rats
Part of the postsynaptic response to GABA is mediated by peri- and extrasynaptic GABAA receptors, which respond to the low ambient GABA concentration in the extracellular space, giving rise to a persistent, tonic GABA current. To examine the tonic current in adolescent dentate granule cells, we recorded the GABA current in whole-cell configuration from 4- to 8-week-old rats. We measured the outward shift upon blockade of GABAA receptors using the competitive antagonist SR95531 (∼100 µM). The average SR95531-sensitive current density of control rats was 2.2 ± 0.5 pA/pF (n = 9), and in LBW rats, it was 1.7 ± 0.5 pA/pF (n = 9, n.s.) (Fig. 5C, D). Granule cell capacitances were similar (9.7 ± 0.7 vs. 12 ± 1.4 pF), suggesting an unaltered tonic inhibition in the dentate gyrus.
Synaptic vesicle proteins SNAP-25 and SCAMP2 are upregulated in LBW rats
As we found a failure of the normal developmental increase in the release probability at GABAergic terminals, we tested whether this could be due to impaired synaptic vesicle protein expression. The mRNA levels of 8 reference genes and 14 genes of synaptic vesicle proteins were investigated in the hippocampus from 8-week-old LBW and control rats (Fig. 6A). The mRNA levels of Snap-25 (percentage of control: 120% ± 4.7%, t = 4.301, df = 12, p = 0.001029) and Scamp2 (151% ± 6.8%, t = 9.051, df = 10, p = 0.000004) were significantly upregulated, while mRNA levels of Snap-29, Snapin, Synapsin I, II, III, Synaptophysin, Synaptotagmin I, II, III, Syntaxin 1A, Vamp1, and Vamp2 were unchanged. Additionally, 7 other genes were investigated (Gr, Mr, Munc13, Munc18, Rims1, Scna, and Vgat), which can be associated with GABAergic neurotransmission, but their mRNA levels were unchanged (Fig. 6B). The upregulation of Snap-25 and Scamp2 suggests a differential gene expression of synaptic vesicle proteins in LBW rats, which might be linked to the dysfunction in the GABAergic synaptic neurotransmission, and in particular, the apparent change in presynaptic vesicle release probability. When running multiple corrections, both Snap-25 and Scamp2 stayed significant (p = 0.020386 and p = 0.000083, respectively).
Increase in the mRNA expression levels of presynaptic proteins, Snap-25 and Scamp2 in 8-week-old LBW rats compared to vehicle controls. (A) Real-time qPCR was used to quantify mRNA expression levels of synaptic vesicle proteins in the hippocampus of the LBW rats. The normalised values are plotted as mean group values ± SEM and expressed as per cent of respective control. Control: n = 10, LBW: n = 7. Unpaired t-test, **p < 0.01, ***p < 0.001. After Holm–Sidak multiple corrections, values were p = 0.020386 for Snap-25 and p = 0.000083 for Scamp2, respectively. (B) Real-time qPCR was also used to quantify mRNA expression levels of seven selected genes in the hippocampus in the LBW rat model. No significant changes were observed. Control: n = 10, LBW: n = 7.
Discussion
A shift from slow to fast GABAergic synaptic signalling of GABAergic neurons has previously been documented from birth until adolescence in different brain regions in rodents (DG: (Hollrigel & Soltesz, Reference Hollrigel and Soltesz1997); CA1: (Cohen et al., Reference Cohen, Lin and Coulter2000); cortex: (Doischer et al., Reference Doischer, Hosp, Yanagawa, Obata, Jonas, Vida and Bartos2008); superior colliculus: (Jüttner et al., Reference Jüttner, Meier and Grantyn2001)). Consistent with these studies, we found that the phasic synaptic inhibitory signalling arriving at granule cells in the dentate gyrus shows an acceleration in the rising and decaying phases of sIPSCs until postnatal weeks 4–5. The maturation seemed to stabilise at around 4–5 weeks as no further change was observed at 6–8 weeks of age. The acceleration of fast synaptic GABAA currents could arise due to several developmental events during postnatal maturation. The morphology of dendritic and axonal arbours evolves until the second and third weeks of age (Seress & Ribak, Reference Seress and Ribak1990). It contributes to an increase in membrane capacitance and a decrease in input resistance, increasing the speed of postsynaptic events (Doischer et al., Reference Doischer, Hosp, Yanagawa, Obata, Jonas, Vida and Bartos2008). GABAA receptors with different subunit compositions show distinct kinetics (Mohler et al., Reference Mohler, Benke, Mertens and Fritschy1992), and GABAAR subunit expression is known to change during juvenile and adolescence in a region-dependent manner (McKernan et al., Reference McKernan, Cox, Gillard and Whiting1991; Fritschy et al., Reference Fritschy, Paysan, Enna and Mohler1994). The GABAAR α2-subunit mediates 10x slower kinetics over α1-containing GABAARs (Dixon et al., Reference Dixon, Sah, Lynch and Keramidas2014). Thus, developmental increases in α1 and γ2 and decreases in α2 subunit expression in the hippocampus (Killisch et al., Reference Killisch, Dotti, Laurie, Lüddens and Seeburg1991; Poulter et al., Reference Poulter, Barker, O’Carroll, Lolait and Mahan1992; Lopez-Tellez et al., Reference Lopez-Tellez, Vela, del Rio, Ramos, Baglietto-Vargas, Santa-Maria, Ruano, Gutierrez and Vitorica2004) may likely cause the observed acceleration in dentate inhibitory signalling (Okaty et al., Reference Okaty, Miller, Sugino, Hempel and Nelson2009). However, several other factors might contribute, such as changes in the intrinsic properties of the neurons (Doischer et al., Reference Doischer, Hosp, Yanagawa, Obata, Jonas, Vida and Bartos2008), postnatal maturation of the vesicle release machinery (Kirischuk & Grantyn, Reference Kirischuk and Grantyn2003), or changes in the firing rate of different inhibitory cell populations.
Our observation of a developmental increase in the amplitude of GABAAR-mediated sIPSCs can result from the increase in number and size of GABAergic synapses seen during postnatal development (Seress & Ribak, Reference Seress and Ribak1990). Our results in control rats also suggest a slight increase in the sIPSC frequency between 2 and 4–5 weeks of age, although these changes were not statistically significant. sIPSC frequency represents a summated picture of spontaneous activity of the synapses and the activity rate of the circuitry. Both the synapses and the connectivity rate of the network undergo quantitative and qualitative maturation during the postnatal period (Kilb, Reference Kilb2012). However, the number of interneurons shows a 40% decrease due to programmed developmental cell death during the same period (Southwell et al., Reference Southwell, Paredes, Galvao, Jones, Froemke, Sebe, Alfaro-Cervello, Tang, Garcia-Verdugo, Rubenstein, Baraban and Alvarez-Buylla2012). These opposing effects might be counterbalanced, resulting in a somewhat constant spontaneous GABAergic activity in the developing dentate gyrus after postnatal week 2. Our results also point to an increase in the release probability from GABAergic terminals between postnatal weeks 2 and 4–5, consistent with earlier studies (Jüttner et al., Reference Jüttner, Meier and Grantyn2001; Kirischuk et al., Reference Kirischuk, Jüttner and Grantyn2005), which correlates with an increase in readily releasable pool size after postnatal week 2 in hippocampal interneurons (Mozhayeva et al., Reference Mozhayeva, Sara, Liu and Kavalali2002).
Prenatal stress has been found to cause several changes in brain structures during embryonic development (Weinstock, Reference Weinstock2011; Franke et al., Reference Franke, Van den Bergh, de Rooij, Kroegel, Nathanielsz, Rakers, Roseboom, Witte and Schwab2020), changes that must be long-lasting to cause neuropsychiatric disorders later in life. This raises the question, how stress in utero affects the postnatal maturation of hippocampal GABAergic signalling. In 2-week-old (juvenile) LBW rats, GABAergic signalling was comparable to that of control animals. This suggests that GABAergic malfunction due to prenatal stress will manifest later in life, as we observed pronounced differences in 4- to 5-week-old animals. However, abnormal development of GABAergic neurons has been observed already at birth in a mouse model of prenatal stress (Stevens et al., Reference Stevens, Su, Yanagawa and Vaccarino2013). In the offspring of mice subjected to acute bright light stress, a delay in the interneuron progenitor migration was demonstrated (Stevens et al., Reference Stevens, Su, Yanagawa and Vaccarino2013). This is indicative of a delay in developmental processes.
The probability of GABA release decreased during postnatal development in LBW rats. This is consistent with earlier findings in our laboratory, showing decreased release probability of dentate gyrus GABAergic terminals in the CMS model of depression in adulthood (Holm et al., Reference Holm, Nieto-Gonzalez, Vardya, Henningsen, Jayatissa, Wiborg and Jensen2011). These similarities across different depression models suggest that malfunction in GABA release might be central to the underlying pathomechanisms of depressive disorders.
Our data showed a tendency of the frequency of spontaneous IPSCs to be downregulated (although this was not statistically significant), whereas the action potential-independent miniature IPSCs showed no difference compared to controls. The tendency of a decreased frequency of sIPSCs could reflect alterations in the action potential-driven GABA release from the presynaptic terminals of interneurons. This could occur due to a decrease in the number of presynaptic interneurons following the delay in interneuron progenitor migration reported by (Stevens et al., Reference Stevens, Su, Yanagawa and Vaccarino2013). Our finding of unchanged mIPSC frequency does, however, not support this notion. Alternatively, a decreased frequency in sIPSCs could occur due to a decrease in the probability of GABA release from nerve terminals. Our findings of changes in the paired-pulse ratio of evoked synaptic responses further support this idea of decreased probability of GABA release (Dobrunz & Stevens, Reference Dobrunz and Stevens1997).
Decreased GABA concentration is a common feature in brain tissues of depressed human patients (Sanacora et al., Reference Sanacora, Mason, Rothman, Behar, Hyder, Petroff, Berman, Charney and Krystal1999; Bhagwagar et al., Reference Bhagwagar, Wylezinska, Jezzard, Evans, Ashworth, Sule, Matthews and Cowen2007), which might influence the basal tonic GABA signalling in the hippocampus. However, when testing the extrasynaptic GABAA receptor-mediated tonic current without exogenous agonists in the dentate gyrus of the LBW rats, we found no abnormality compared to control rats, consistent with earlier findings in the CMS model of depression (Holm et al., Reference Holm, Nieto-Gonzalez, Vardya, Henningsen, Jayatissa, Wiborg and Jensen2011).
Testing across several presynaptic proteins, we showed that Snap-25 mRNA is upregulated in the LBW rats. This corresponds well with the increased expression in SNAP-25 protein levels in the hippocampus and prefrontal cortex in rats whose mothers underwent restraint three times daily during the last week of gestation (Cao et al., Reference Cao, Wang, Zheng, Cheng and Zhang2018).
Mechanistically, SNAP-25 regulates synaptic strength (Bark et al., Reference Bark, Bellinger, Kaushal, Mathews, Partridge and Wilson2004; Scullin et al., Reference Scullin, Tafoya, Wilson and Partridge2012) via the downregulation of voltage-gated Ca2+ channels (Condliffe et al., Reference Condliffe, Corradini, Pozzi, Verderio and Matteoli2010). Antonucci and colleagues showed a stronger paired-pulse depression at GABAergic synapses in SNAP-25 heterozygous (SNAP-25+/-) cell cultures, where the expression level of SNAP-25 is expected to be half of that in control cells (Antonucci et al., Reference Antonucci, Corradini, Morini, Fossati, Menna, Pozzi, Pacioni, Verderio, Bacci and Matteoli2013). This effect might be due to the decreased inhibitory effect of SNAP-25 on Ca2+ channels (Condliffe et al., Reference Condliffe, Corradini, Pozzi, Verderio and Matteoli2010), leading to an increased vesicle release probability at GABAergic terminals (Antonucci et al., Reference Antonucci, Corradini, Morini, Fossati, Menna, Pozzi, Pacioni, Verderio, Bacci and Matteoli2013; Kochlamazashvili & Haucke, Reference Kochlamazashvili and Haucke2013). However, it is essential to note that SNAP-25 is also involved in glutamate release (Antonucci et al., Reference Antonucci, Corradini, Morini, Fossati, Menna, Pozzi, Pacioni, Verderio, Bacci and Matteoli2013). We showed an increased expression level of Snap-25 mRNA in the hippocampus of adult LBW rats, which may cause an opposite effect in the GABAergic synapses, leading to a decrease in the vesicle release probability through downregulation of Ca2+ channel activity. Thus, if an increased Snap-25 mRNA is translated into protein, this presents one possible explanation of the observed decrease in release probability of GABAergic terminals in adult LBW rats. Since the latter was already prominent in adolescence, this raises the idea that upregulation of SNAP-25 might be a characteristic feature already in developing LBW rats, which may be a topic for future investigations.
Finally, Scamp2 was also upregulated in adult LBW rats. SCAMP2 is involved in both exocytic and endocytic secretory pathways (Brand & Castle, Reference Brand and Castle1993), and overexpression of SCAMP2 in neuroendocrine cells inhibits both exo- and endocytosis of secretory granules (Liu et al., Reference Liu, Guo, Tieu, Castle and Castle2002), while knockdown of SCAMP2 also leads to decreased exocytosis (Liao et al., Reference Liao, Zhang, Shestopal, Szabo, Castle and Castle2008). SCAMP2 brain expression is involved in the regulation of different monoamine transporters (Müller et al., Reference Müller, Wiborg and Haase2006; Fjorback et al., Reference Fjorback, Müller, Haase, Raarup and Wiborg2011). However, its precise role in the brain has yet to be elucidated. Furthermore, in humans, a correlation was found between single-nucleotide polymorphism of SCAMP2 and neuroticism (Luciano et al., Reference Luciano, Huffman, Arias-Vásquez, Vinkhuyzen, Middeldorp, Giegling, Payton, Davies, Zgaga, Janzing, Ke, Galesloot, Hartmann, Ollier, Tenesa, Hayward, Verhagen, Montgomery, Hottenga, Konte, Starr, Vitart, Vos, Madden, Willemsen, Konnerth, Horan, Porteous, Campbell, Vermeulen, Heath, Wright, Polasek, Kovacevic, Hastie, Franke, Boomsma, Martin, Rujescu, Wilson, Buitelaar, Pendleton, Rudan and Deary2012). Thus, the increased expression level of SCAMP2 found in the hippocampus of LBW rats may alter neurotransmission in the monoaminergic neurotransmitter systems of the brain, although its connections to GABA release in this animal model are not clear at the moment.
In conclusion, we demonstrated that prenatal stress exposure leads to decreased GABAergic signalling in the dentate gyrus in the LBW rat model of depression. Although this will require further investigations, we propose that increased expression of Snap-25 in the hippocampus is linked to the reduction in both spontaneous and evoked GABA release onto dentate granule cells. Our results further support the GABAergic hypothesis of depressive disorders, and thus, the multifaceted mechanisms of the GABAergic system might be a putative target for clinical treatment.
Acknowledgements
The authors thank Lone Overgaard and Lene Wind Steffensen for their excellent technical assistance.
Author contributions
ZD and JLNG performed electrophysiological experiments and analyses. BE carried out the molecular biology experiments, and KSH and GW generated the animals. BE, MMH, KJ, and GW designed and supervised the research. All authors contributed to the writing of the manuscript and the preparation of figures.
Conflict of interest
GW is the Editor-in-Chief of Acta Neuropsychiatrica, but actively withdrew and was not involved during the review or decision process of this manuscript.
