Substance P excites GABAergic neurons in the mouse central amygdala through neurokinin 1 receptor activation

Abstract

Substance P (SP) is implicated in stress regulation and affective and anxiety-related behavior. Particularly high expression has been found in the main output region of the amygdala complex, the central amygdala (CE). Here we investigated the cellular mechanisms of SP in CE in vitro, taking advantage of glutamic acid decarboxylase-green fluorescent protein (GAD67-GFP) knockin mice that yield a reliable labeling of GABAergic neurons, which comprise 95% of the neuronal population in the lateral section of CE (CEl). In GFP-positive neurons within CEl, SP caused a membrane depolarization and increase in input resistance, associated with an increase in action potential firing frequency. Under voltage-clamp conditions, the SP-specific membrane current reversed at −101.5 ± 2.8 mV and displayed inwardly rectifying properties indicative of a membrane K⁺ conductance. Moreover, SP responses were blocked by the neurokinin type 1 receptor (NK1R) antagonist L-822429 and mimicked by the NK1R agonist [Sar⁹,Met(O₂)¹¹]-SP. Immunofluorescence staining confirmed localization of NK1R in GFP-positive neurons in CEl, predominantly in PKCδ-negative neurons (80%) and in few PKCδ-positive neurons (17%). Differences in SP responses were not observed between the major types of CEl neurons (late firing, regular spiking, low-threshold bursting). In addition, SP increased the frequency and amplitude of GABAergic synaptic events in CEl neurons depending on upstream spike activity. These data indicate a NK1R-mediated increase in excitability and GABAergic activity in CEl neurons, which seems to mostly involve the PKCδ-negative subpopulation. This influence can be assumed to increase reciprocal interactions between CElon and CEloff pathways, thereby boosting the medial CE (CEm) output pathway and contributing to the anxiogenic-like action of SP in the amygdala.

substance p (SP), a member of the tachykinin family of neuropeptides initially known as neurogenic inflammation-related substance (for review see Harrison and Geppetti 2001), has recently been proposed to be implicated in depression disorders based on the observation that SP concentration was increased in the serum (Bondy et al. 2003) and the cerebrospinal fluid (Geracioti et al. 2006) of depressed patients. Tachykinin levels are altered also in patients with anxiety disorders and schizophrenia and, together with their derivates, are considered promising therapeutic targets (for review see Ebner et al. 2009). SP and other neurokinins [neurokinin A (NKA) and neurokinin B (NKB)] comprise the mammalian tachykinin family (Pennefather et al. 2004). Biological action of tachykinins is mediated through neurokinin receptors (NKRs): SP predominantly acts through NK1R, whereas NKA and NKB are preferred ligands for NK2R and NK3R, respectively (for review see Maggi 1995).

SP and NK1R display a widespread distribution in the brain (Cuello and Kanazawa 1978; Mantyh et al. 1989), with high expression levels in the amygdala (Cassell and Gray 1989; Shigematsu et al. 2008; Singewald et al. 2008; Sreepathi and Ferraguti 2012). The amygdala is a key region of the brain involved in processing and propagating fear- and anxiety-related signals. In fact, in different experimental models it has been shown that SP is involved in stress-related and anxiety-like behavior, as well as in alcohol dependence (Ebner and Singewald 2006). For instance, SP modifies passive avoidance learning in rats after microinjection into globus pallidus and central amygdala (CE) (Kertes at al. 2009) and produces anxiogenic-like effect in the elevated plus maze test in CE but not in basolateral amygdala (BLA) (Carvalho et al. 2013). In contrast, infusion of the NK1R antagonist GR 82334 into the BLA or medial amygdala blocks fear-potentiated startle but does not significantly disrupt fear-potentiated startle when infused into the CE (Zhao et al. 2009). The CE is traditionally considered the main output of the amygdala complex, with projections to brain stem, hypothalamus, and basal forebrain (for review see Ehrlich et al. 2009; Pape and Paré 2010). The prevalent view on the CE as a mere output station has been gradually replaced by a model in which highly organized synaptic circuits define functional entities related to specific components of fear processing (for review see Duvarci and Paré 2014). Specifically, the CE is composed of the lateral (CEl) and medial (CEm) subnuclei, and subpopulations of CEl GABAergic neurons characterized by the absence or presence of protein kinase C δ (PKCδ) connect to CEm GABAergic neurons to gate fear expression and regulate fear generalization (Ciocchi et al. 2010; Haubensak et al. 2010; Li et al. 2013). While SP has been found to stimulate inhibitory transmission in BLA (Maubach et al. 2001), little is known about the cellular mechanisms of SP action in CE. Therefore, we thought it timely to investigate in more detail the cellular mechanisms of SP action in CE. We focused on the GABAergic neurons in the CEl, using electrophysiological techniques in acute slices in vitro and immunocytochemical methods in glutamic acid decarboxylase-green fluorescent protein (GAD67-GFP) knockin mice (Tamamaki et al. 2003) where GABAergic neurons are reliably visualized by GFP.

MATERIALS AND METHODS

Animals.

All experiments were carried out in accordance with European Commission Council Directive 86/609/EEC. Protocols were approved by the Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen (AZ8.87-50.10.46.09.A170 and 87-51-04.2010.A322). Recordings were obtained from adult male GAD67-GFP mice (Tamamaki et al. 2003) and in addition from C57BL/6J mice (Charles River) for some of the pharmacological and immunocytochemical experiments.

Preparation of slices.

Male GAD67-GFP and C57BL/6J mice, aged from postnatal day(P)24 to P51, were deeply anaesthetized with Forene (isoflurane, 1-chloro-2,2,2-trifluoroethyl-difluoromethylether; 2.5% in O₂) and decapitated. Their brains were rapidly removed and placed in ice-cold oxygenated preparation solution containing (mM) 2.5 KCl, 1.25 NaH₂PO₄, 10 MgSO₄, 0.5 CaCl₂, 20 piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES), 10 glucose, and 200 sucrose (pH 7.35). Coronal slices (250 μm thick), containing the amygdala complex, were cut with a Leica VT-1200S microtome (Leica Microsystems, Wetzlar, Germany) at 4°C. Slices were incubated in the incubation solution containing (in mM) 125 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 24 NaHCO₃, 2 MgSO₄, 2 CaCl₂, and 15 glucose (pH 7.35) at 34°C for 20 min and then allowed to equilibrate for at least 1 h at room temperature before electrophysiological recording.

Electrophysiological recordings in vitro.

Recordings were obtained from visually identified neurons in the lateral section (CEl) of the CE as described previously (Kamprath et al. 2010; Sosulina et al. 2010). In brief, slices were superfused continuously (2.0 ml/min) with an oxygenated (95% O₂-5% CO₂) artificial cerebrospinal fluid (ACSF) solution at ∼30°C containing (in mM) 120 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 22 NaHCO₃, 2 MgSO₄, 2 CaCl₂, and 10 glucose (pH 7.35). CE neurons were identified by differential interference contrast-infrared videomicroscopy (S/W-camera CF8/1, Kappa, Gleichen, Germany). Whole cell recordings were acquired with a patch-clamp amplifier (EPC-7, List Medical Systems, Darmstadt, Germany or EPC10 amplifier, HEKA, Germany) at a sampling rate of 10 kHz and analyzed off-line with Clampfit 10.0 (Molecular Devices, Sunnyvale, CA) and Mini Analysis Program (Synaptosoft, Fort Lee, NJ). Recording electrodes were filled with (in mM) 95 K-gluconate, 20 K₃citrate, 10 NaCl, 10 HEPES, 1 MgCl₂, 0.1 CaCl₂, 1.1 EGTA, 3 MgATP, and 0.5 NaGTP (pH 7.2 with KOH). For some experiments a high-Cl⁻ intracellular solution was used containing (in mM) 110 CsCl, 30 K⁺-gluconate, 1.1 EGTA, 10 HEPES, 0.1 CaCl₂, 4 MgATP, and 0.3 NaGTP (Cl⁻ reversal potential = ∼0 mV). Correction for a liquid junction potential of 10 mV was applied. Series resistance was monitored throughout the experiments, and whole cell currents were analyzed only from recordings in which series resistance varied by <15%.

Resting membrane potential (RMP) was measured immediately after whole-cell configuration was obtained. Cells displaying an RMP positive to −55 mV were discarded. Neurons were held in current-clamp mode at −60 mV, and the following electrophysiological parameters were determined: membrane input resistance (IR), sag ratio, and medium afterhyperpolarization (mAHP). IR was calculated at steady state from a hyperpolarizing pulse (at −30 pA for 500 ms) according to Ohm's law. Sag ratio was calculated as the difference between the steady-state membrane potential of a 1-s hyperpolarizing step (−140 pA) and the most negative membrane potential at the beginning of the step divided by the most negative value of the membrane potential. The mAHP was calculated from the membrane hyperpolarization (baseline at −60 mV to the most negative potential) at 300 ms after termination of a depolarizing current step (250 pA, 1-s duration, at −60 mV) evoking spike firing (Schiess et al. 1993).

Effects of SP were determined under current- and voltage-clamp conditions. For each current-clamp experiment, the amplitude of the voltage deflection (V) evoked by the application of negative current injections (I) (−10 pA, 500 ms) at resting potential was measured, and IR was calculated based on Ohm's law (R = V/I). To determine excitability, action potentials were evoked by positive current injections with varying amplitudes and 500-ms duration. The amplitude of the depolarizing step was adjusted (range 5–150 pA) to evoke a minimum of two action potentials in an individual cell. Manual clamping was performed by injection of direct negative current through the recording pipette in order to hyperpolarize the neuron to the initial membrane potential (prior to SP application). Changes in membrane potential and firing frequencies were estimated.

During voltage-clamp recordings the holding potential (V_h) was set at −70 mV. Current vs. voltage relationships were constructed from voltage ramps applied from −70 to −140 mV (duration 150 ms) at 0.05 Hz. The SP-specific membrane current was calculated by subtracting the ramp current during baseline recordings from those recorded at the maximum effect of the treatment (∼10 min of application) and plotted against the respective membrane potential. Spontaneous inhibitory postsynaptic currents (sIPSCs) mediated by GABA_A receptors (GABA_ARs) were isolated in the presence of the glutamate receptor blockers 6,7-dinitroquinoxaline-2,3-dione (DNQX, 10 μM) and dl-2-amino-5-phosphonovalerate (AP-5, 25 μM) and the GABA_B receptor antagonist (2S)-3-[(1S)-1-(3,4-dichlorophenyl)ethyl]amino-2-hydroxypropyl(phenylmethyl) phosphinic acid (CGP55845A, 5 μM). Miniature inhibitory postsynaptic currents (mIPSCs) were isolated in the presence of NMDA/non-NMDA and GABA_B receptor antagonist plus the addition of tetrodotoxin (TTX, 1 μM). Effects of SP on sIPSC and mIPSC frequency were analyzed and visually confirmed with a semiautomated threshold-based mini detection software (Mini Analysis, Synaptosoft). Means of sIPSC or mIPSC characteristics were determined at baseline and during experimental drug conditions containing a minimum of 60 events in a time period of 2 min under steady-state conditions. The tonic current was calculated and distinguished from sIPSCs with a routine written in MATLAB in which the mean holding current was obtained by a Gaussian fit to an all-points histogram from 2 min under steady-state conditions. Responses were quantified as the difference in holding current between baseline and experimental conditions.

All drugs were bath applied. SP was obtained from Bachem (Heidelberg, Germany), TTX from Biotrend (Cologne, Germany), and the NK1R agonist [Sar⁹,Met(O₂)¹¹]-SP from Tocris (Wiesbaden-Nordenstadt, Germany), and the NK1R antagonist L-822429 was synthesized in house as described previously (Singewald et al. 2008). Drug effects were determined by measuring peak amplitudes of near-maximal responses under current-clamp or voltage-clamp conditions. Because the SP effect displayed strong desensitization when application was repeated, only responses to the first application in a given slice were included in the analysis. In experiments with NK1R antagonist or TTX, SP or the NK1R agonist was added to the bathing solution 10–20 min after application of the pharmacologically active substance.

Immunofluorescence.

P30 GAD67-GFP and adult C57BL/6J (Charles River) mice were deeply anesthetized with pentobarbital (50 mg/kg body wt) and transcardially perfused with PBS, followed by an ice-cold solution containing 4% PFA with or without 0.3% glutaraldehyde in PBS for 35–40 min. Brains were removed, postfixed for 2–4 h in the same fixative, and cryoprotected with 25% sucrose if necessary. Coronal sections (40 μm) were cut at the level of the CEl and collected in Tris-buffered saline (TBS). After several washings with TBS, sections were blocked with 10% normal horse serum (NHS), 2% BSA, 0.3% Triton X-100 in TBS for 2 h or with 1% BSA and 0.1% Triton X-100 in TBS for 30 min to minimize nonspecific binding before incubation of slices with primary antibodies in either 2% NHS, 2% BSA, 0.3% Triton X-100 in TBS or 1% BSA and 0.1% Triton X-100 in TBS at 4°C for 16–18 h. The rabbit polyclonal anti-NK1R antibody [Chemicon, catalog no. AB5060, lot no. 0512016863, 1:10,000 (data Fig. 3), lot no. NG1809683, 1:500 (data Fig. 4)] was raised against the COOH-terminal 23 amino acids (residues 385–407) of rat NK1R. The mouse monoclonal anti-PKCδ antibody (BD Biosciences, catalog no. 610398, lot no. 13866, 1:1,000) was raised against the COOH terminus (amino acid residues 114–289) of human PKCδ protein containing the catalytic domain of the enzyme.

Fig. 3.

Double-immunofluorescence staining for NK1R and PKCδ in the CEl. Low (A; scale bars, 50 μm) and high (B; scale bars, 10 μm; magnified region as indicated) magnification of images showing the localization of NK1R (in red, left)- and PKCδ (in green, center)-like immunoreactivity in the CEl by confocal fluorescence microscopy. Their superimposed images (merged) are presented on right.

Fig. 4.

NK1R expression in the CEl. A: GAD67-GFP fluorescence (in green, left) of CEl neurons, NK1R immunostaining (in red, center), and superimposition of the 2 images (right). Scale bars, 50 μm. CEm, medial section of central amygdala; BLA, basolateral amygdala. B: enlarged individual neurons shown in confocal microscopic images, with GAD67 (left), NK1R (center), and superimposed (right) immunostaining. Scale bars, 10 μm.

After washing (3 × 10 min with TBS), sections were exposed to Cy3-conjugated donkey anti-rabbit IgG (1:300, Dianova, Germany or 1:500, Jackson ImmunoResearch Laboratories) or Alexa Fluor 647-conjugated donkey anti-mouse IgG (1:500, Jackson ImmunoResearch Laboratories) for 1.5–2 h, washed again, and coverslipped with Immumount or Invitrogen ProLong Gold Antifade Reagent. For negative controls, occlusion of the primary antibody from the staining procedure was routinely performed with no positive immunological signal detected. Confocal images were performed with a confocal laser scanning microscope (Nikon eC1 plus). For quantitative analysis of the cell numbers, PKCδ immunopositivity was used for visualization of CEl boundaries and delineation from CEm (Haubensack et al. 2010), and the CEl in a given section was subdivided into three or four squares of 7,000–10,000 μm². The number of NK1R-positive cells in each square was set to 100%, and the percentage of double-labeled PKCδ/NK1R-positive cells was determined. Border areas of CEl were excluded from calculation. In GAD67-GFP mice, the percentage of GFP-positive cells was calculated in relation to NK1R-positive cells.

Data analysis.

Statistical analysis (1-way ANOVA, repeated-measures ANOVA, χ²-test, Mann-Whitney test, Wilcoxon test, 1-sample t-test) was done with GraphPad Prism 5.0 (GraphPad Software, San Diego, CA) or PASW Statistics 17 for Windows software (SPSS, Chicago, IL). The significance level for all measures was set at P < 0.05, and all data are presented as means ± SE.

RESULTS

Electrophysiological properties of GFP-positive neurons in the CEl.

Recordings were obtained from a total of 100 GFP-positive neurons in the CEl from 38 GAD67-GFP mice. Neurons were classified as regular-spiking, late-firing, and low-threshold bursting types according to a previously proposed scheme (Dumont et al. 2002). The large majority of neurons recorded (n = 87) fell into these three categories (Fig. 1), and only a minority of GFP-positive neurons (n = 13) could not be classified. Examples are illustrated in Fig. 1, and statistical data are presented in Table 1. Of note, regular-spiking neurons (n = 31) displayed small-amplitude mAHPs compared with those of late-firing (n = 30) and low-threshold bursting (n = 26) neurons. Furthermore, the sag of hyperpolarizing membrane responses was more pronounced in low-threshold bursting than in regular-spiking neurons, and afterdepolarization following spike activity was observed in a subpopulation of late-firing neurons (n = 14/30). These electrophysiological properties are in accordance with those previously reported for CE neurons (Dumont et al. 2002; Lopez de Armentia and Sah 2004; Martina et al. 1999; Tsetsenis et al. 2007).

Fig. 1.

Electrophysiological properties of green fluorescent protein (GFP)-positive neurons in the lateral section of central amygdala (CEl). Examples of 3 types of neurons with different firing behavior: late firing, regular firing, and low-threshold bursting in GFP-positive neurons; original traces obtained in response to a hyperpolarizing current step (−30 pA) and depolarizing steps (+50 pA, +90 pA, and +110 pA) from resting membrane potential, as indicated for each neuron. Action potentials are cut to visually emphasize the near-threshold firing behavior of the neurons. Top inset: example of an afterdepolarizing potential (ADP) following a depolarizing current-induced action potential (+90 pA). Bottom inset: burstlike firing behavior at the beginning of the depolarizing step.

Table 1.

Physiological properties of mouse GFP-positive CEl neurons

	n	Incidence, %	RMP, mV	IR, MΩ	Sag	mAHP, mV
Late firing (1)	30	30	−73.07 ± 1.31	303.3 ± 17.41	0.098 ± 0.009	−9.02 ± 0.92* (2)
Regular spiking (2)	31	31	−74.69 ± 1.59	322.2 ± 26.51* (4)	0.074 ± 0.012* (3)	−5.62 ± 0.91* (1, 3)
Low-threshold bursting (3)	26	26	−71.05 ± 0.97	353.5 ± 22.77* (4)	0.114 ± 0.009* (2)	−11.76 ± 0.68* (2)
Nonclassified (4)	13	13	−75.95 ± 1.35	211.1 ± 18.77* (2, 3)	0.072 ± 0.010	−8.83 ± 1.24

Values are means ± SE. Statistical differences between the 4 types of green fluorescent protein (GFP)-positive neurons with different firing patterns, estimated based on post hoc Bonferroni test, are shown under the corresponding parameter (P < 0.05).

n, No. of cells in a group; incidence, occurrence of this type of neuron within the recorded lateral central amygdala (CEl) neuronal population; RMP, resting membrane potential; IR, input resistance; mAHP, medium afterhyperpolarization.

SP-evoked membrane depolarization through decrease in K⁺ conductance.

Responsiveness to SP was investigated in a total of 60 GFP-positive neurons in the CEl from 26 GAD67-GFP mice comprised of regular-spiking (n = 24), late-firing (n = 16), low-threshold bursting types (n = 13) and nonclassified (n = 7) neurons. Of these, SP responses were observed in regular-spiking (n = 17; 71%), late-firing (n = 10; 63%), low-threshold bursting (n = 11; 84%), and nonclassified (n = 4; 57%) cells. The percentage of SP responders was not different between classes of neurons (χ²-test, P = 0.51). In current-clamp mode at RMP, application of SP (300 nM) depolarized the GFP-positive neurons by 2.7 ± 0.6 mV (n = 19 of 23 in 8 mice, P < 0.01, Wilcoxon test; Fig. 2Aa). This response was associated with an increase in apparent IR from 397.5 ± 50.9 MΩ to 495.4 ± 61.4 MΩ at the maximal effect of the SP response and to 464 ± 55 MΩ at manually clamped resting potential [F(2,36) = 24.07; P < 0.001]. The mean number of spikes elicited by direct positive current injections increased from 2.2 ± 0.2 under control conditions to 4 ± 0.5 during steady-state responses to SP and to 2.4 ± 0.5 at manually clamped resting potential (n = 19), as estimated by repeated-measures ANOVA and Tukey's multiple comparison test [F(2,36) = 24.07; P < 0.001] (Fig. 2Ab). Upon washout of SP, the membrane potential typically returned to resting value, and the number of spikes and IR amounted to 3.1 ± 0.6 and 436.9 ± 85.1 MΩ, respectively, not significantly differing from preapplication levels (P = 0.14 and P = 0.48, Wilcoxon test).

Fig. 2.

Direct postsynaptic responses of GFP-positive neurons in CEl to application of substance P (SP). Aa: current-clamp recording showing the depolarizing effect of SP (300 nM) applied at resting membrane potential of −75 mV. Up- and downward deflections represent responses to repetitive injection of de- and hyperpolarizing current steps (+5 pA, 500 ms; −10 pA, 500 ms), respectively. Action potentials are cut for clarity. Ab: enlarged responses to depolarizing current steps of +50 pA, corresponding to the control condition (1), maximal SP effect (2), and manual clamp (3). B: difference currents obtained under voltage-clamp conditions by ramping the membrane from −70 mV to −140 mV within 150 ms, during near-maximal action of SP and control conditions (Ba) and during action of the neurokinin receptor type 1 (NK1R) agonist [Sar⁹,Met(O₂)¹¹]-SP and control conditions (Bc). I, current; V, voltage. Note inward rectification and a reversal potential (E_rev) of the current close to K⁺ equilibrium potential (E_K). The SP-specific current persists during action of TTX (Bb). Bd: averaged membrane currents recorded at holding potential (V_h) = −70 mV in response to SP (300 nM), during SP with added TTX (1 μM), during SP after application of the NK1R antagonist L-822429 (1 μM), and during application of [Sar⁹,Met(O₂)¹¹]-SP (300 nM). Data are means ± SE. *Significant differences between groups: SP vs. L-822429+SP, SP in TTX vs. L-822429+SP, and [Sar⁹,Met(O₂)¹¹]-SP vs. L-822429+SP (P < 0.05, Dunn's multiple comparison test).

To investigate the possible mechanisms underlying responses to SP, further experiments were performed under voltage-clamp conditions in a separate group of CEl neurons (from 9 GAD67-GFP mice). Application of voltage ramps (0.5 mV/ms, 0.05 Hz, −70 to −140 mV), and subtraction of the membrane current during and before SP action, demonstrated an SP-evoked inward current at around −70 mV that displayed moderate inward rectification and a reversal potential at −101.5 ± 2.8 mV (n = 10 of 18 cells) (Fig. 2, Ba and Bd). The reversal potential of the SP-evoked current was near the presumed K⁺ equilibrium potential (E_K) = −108.2 mV, as obtained by the Nernst equation for an external K⁺ concentration of 2.5 mM and an internal concentration of 155 mM at 30°C. These data indicated that the response to SP was mediated by a decrease in membrane K⁺ conductance. Bath application of TTX (1 μM) did not alter the SP response (n = 13 in 9 mice; Fig. 2, Bb and Bd), providing evidence for direct SP action and postsynaptic localization of the mediating SP receptors.

Involvement of NK1R.

In various regions of the brain, NK1Rs are involved in mediating neuronal response to SP, in particular in the amygdala (Maubach et al. 2001) and striatum (Govindaiah et al. 2010). In GFP-positive CEl neurons, application of the NK1R antagonist L-822429 (1 μM) prevented the SP (300 nM) response in all tested cells (n = 10 in 4 mice; Fig. 2Bd). Inward currents recorded during SP action under control conditions and in the presence of L-822429 were significantly different (P < 0.01, Kruskal-Wallis test; Fig. 2Bd). Furthermore, application of the NK1R agonist [Sar⁹,Met(O₂)¹¹]-SP (300 nM) evoked an inward current at around −70 mV displaying inward rectification and a reversal potential at −101.5 ± 2.3 mV (n = 6 of 11 tested cells in 8 mice), thereby resembling the SP-specific current (Fig. 2, Bc and Bd). Overall these data indicate that NK1Rs are involved in mediating the suppression of an inwardly rectifying K⁺ current in CEl neurons.

To further elucidate the possible localization of NK1R on GABAergic neurons in CEl, immunohistochemical stainings were performed. Since the CEl neuronal population is comprised of PKCδ-positive and -negative cells (Haubensak et al. 2010), we examined the existence of NK1R in these two classes of neurons. Immunohistochemical double stainings were performed with NK1R-specific and PKCδ-specific antibodies in C57BL/6J mice (n = 7). PKCδ-immunopositive reactions indicated the boundaries of the CEl. Staining for NK1Rs was observed in the plasma membrane and cytoplasm of cell bodies by confocal microscopy. Although the reaction product appeared somewhat punctate, it clearly rimmed the nucleus, forming a round to oval-shaped ring, and thus the cell was considered to be NK1R-positive (Fig. 3). NK1R was expressed predominantly in PKCδ-negative cells (Fig. 3). Only 17.5% (n = 42/240 cells) of NK1R-positive neurons were PKCδ-positive, whereas 80.4% (n = 172/214) of PKCδ-stained neurons did not display NK1R labeling. In the next series of experiments, NK1R-specific antibodies were used in GAD67-GFP mice (n = 2). Results revealed that NK1R are expressed somatically and in the neuropil predominantly in GFP-positive cells (Fig. 4). In fact, 95.7% (n = 88/92 cells) of NK1R-positive neurons were GFP-positive, while only 4.3% (n = 4/92) of NK1R-stained neurons did not display GFP labeling (Fig. 4).

Effects of SP on inhibitory synaptic activity in CEl neurons.

In a separate population of neurons, we investigated the effects of SP on GABA_AR-mediated IPSCs in CEl neurons in order to clarify the possibility that SP influences excitability through GABAergic synaptic network activity. Since we cannot rule out possible influences of the GAD67 mutation on GABAergic synaptic activity, we performed this set of experiments in CEl of C57BL/6J mice, containing >90% GABAergic neurons (Nitecka and Ben-Ari 1987). A high-intracellular Cl⁻ solution was used, enabling recording of inwardly directed IPSCs at membrane potentials near rest. Under voltage-clamp conditions, sIPSCs appeared as fast inward currents at a V_h of −70 mV in the presence of NMDA receptor, non-NMDA glutamate receptor, and GABA_B receptor antagonists (Fig. 5A). Application of SP (300 nM) resulted in a significant increase in frequency (from 8.06 ± 1.30 Hz to 14.43 ± 2.54 Hz; paired t-test, P = 0.0068) and amplitude (from −56.81 ± 7.19 pA to −79.35 ± 10.27 pA; paired t-test, P = 0.0498) of sIPSCs (n = 13 in 12 mice; Fig. 5, Aa and B). Application of the NK1R agonist [Sar⁹,Met(O₂)¹¹]-SP (300 nM) resulted in a similar increase in frequency (from 7.89 ± 1.22 Hz to 10.87 ± 1.63 Hz; paired t-test, P = 0.0263) and amplitude (from −71.87 ± 6.08 pA to −85.55 ± 6.62 pA; paired t-test, P = 0.0341) of sIPSCs (n = 9 in 6 mice; Fig. 5, Ab and B). Properties of sIPSCs were not different between the two groups of CEl neurons under baseline conditions (unpaired t-test; frequency P = 0.9284, amplitude P = 0.2343) and during action of SP or [Sar⁹,Met(O₂)¹¹]-SP (unpaired t-test; frequency P = 0.3217, amplitude P = 0.6985). Of note, SP and [Sar⁹,Met(O₂)¹¹]-SP also induced a slight inward shift in holding current (−5.48 ± 2.98 pA, 1-sample t-test, P = 0.1026; −6.28 ± 1.6 pA, 1-sample t-test, P = 0.006). The presence of the NK1R antagonist L-822429 (1 μM) abolished responses to SP, in that frequency (baseline: 9.02 ± 2.21 Hz; L-822429 + SP: 10.31 ± 2.28 Hz; paired t-test, P = 0.1618) and amplitude (baseline: −86.53 ± 22.77 pA; L-822429 + SP: −78.64 ± 18.93 pA; paired t-test, P = 0.5039) of sIPSCs were not affected (n = 4 in 3 mice; data not shown).

Fig. 5.

SP effects on GABAergic synaptic activity. A: representative traces of spontaneous inhibitory postsynaptic currents (sIPSCs) before (top) and after (bottom) application of SP (Aa) and the NK1R agonist [Sar⁹,Met(O₂)¹¹]-SP (Ab). Note the increase in sIPSC frequency and amplitude during action of SP and the NK1R agonist, in the same respective CEl neuron. Ba: averaged instant frequency of IPSCs for each pharmacological condition. Numbers indicate the GABAergic neurons recorded in CEl for each pharmacological condition. Bb: averaged sIPSC amplitudes for each pharmacological condition. *P < 0.05, **P ≤ 0.01. n.s., Not significant.

During blockade of action potentials by TTX, mIPSCs occurred at an average frequency of 3.46 ± 0.4 Hz and amplitude of −49.37 ± 4.91 pA (n = 16 in 11 mice). Application of SP did not affect frequency or amplitude of the mIPSCs (3.62 ± 0.33 Hz; paired t-test, P = 0.6752; −45.01 ± 2.71 pA; P = 0.0824; Fig. 5B) but resulted in a slight inward shift of the holding current by −5.23 ± 1.36 pA (1-sample t-test, P = 0.0063).

DISCUSSION

The major results of the present study show that SP evokes a direct excitatory response in GAD67-GFP-positive, GABAergic neurons of the murine CEl, which is mediated by a decrease in membrane conductance reversing at the E_K. Pharmacological and immunocytochemical evidence indicates mediation via NK1R stimulation, predominantly (80%) localized in the PKCδ-negative subpopulation of GABAergic neurons in the CEl. In addition, SP increased the frequency and amplitude of GABA_AR-mediated sIPSCs but not mIPSCs, indicating a presynaptic spike-dependent facilitation of inhibitory transmission.

Nature of SP receptors and effector mechanisms of CEl neurons.

Previous immunocytochemical studies reported on expression of NK1R in the amygdala, with particularly high levels in central amygdaloid regions (Gadd et al. 2003; Singewald et al. 2008; Sreepathi and Ferraguti 2012; Truitt et al. 2009). The present study adds to these findings the notion that NK1Rs are expressed in GFP-positive GABAergic neurons in CEl (Fig. 4). In keeping with this, SP-evoked responses in these neurons persisted during TTX, were imitated by the NK1R agonist [Sar⁹,Met(O₂)¹¹]-SP, and were blocked by the NK1R-selective antagonist L-822429, indicating a direct postsynaptic response mediated via NK1R stimulation. These responses were associated with a decrease in membrane conductance and respective membrane current reversing close to the calculated E_K, thereby pointing to a decrease in K⁺ channel activity. Although we did not attempt to characterize the nature of involved K⁺ channels, we noted inwardly rectifying properties of the SP-evoked current. In fact, SP stimulation has been found to inhibit G protein-coupled inwardly rectifying K⁺ channels (GIRK) in neurons of the periaqueductal gray (Drew et al. 2005) and midbrain ventral tegmental area (Xia et al. 2010). Responsiveness to SP will thus require tonic activation of GIRK-type channels. An intriguing possibility is the involvement of a constitutively active inward rectifier K⁺ channel, termed “KirNB” (inward rectifier K⁺ channels in the nucleus basalis), which is deactivated through SP, resulting in neuronal excitation (Nakajima and Nakajima 2010). It is important to note that no differences were observed with respect to SP responsiveness between electrophysiologically classified types of CEl neurons, suggesting a ubiquitous mechanism of regulating excitability in GABAergic neurons in CEl.

Furthermore, SP increased the frequency and amplitude of sIPSCs in CEl neurons, most likely via stimulation of NK1Rs. The absence of SP effects on mIPSCs recorded in the presence of TTX indicates that this influence depends on TTX-sensitive spike activity, suggesting location of the NK1Rs upstream to the recorded CEl neuron. Presynaptic SP effects have indeed been observed in BLA projection neurons through NK1Rs localized in presynaptic GABAergic cells (Maubach et al. 2002). The increase in frequency of sIPSCs observed in CEl neurons of the present study is indeed suggestive of such a presynaptic effect. Interestingly, all tested neurons displayed this presynaptic response to SP, indicating a high degree of synaptic connectivity to neurons that are directly excited by SP. In line with this, CEl neurons are mutually interconnected through GABAergic synapses (Ciocchi et al. 2010), suggesting that the SP-induced increase in inhibitory synaptic activity reflects the increase in excitability and spike firing of presynaptic CEl neurons. Besides increases in sIPSC frequency, NK1R stimulation was also found to increase the amplitude of sIPSCs and to induce an inward shift in holding current. This effect is difficult to reconcile with a mere presynaptic site of action. The inward current occurred during recording with Cs⁺-containing internal solution and thus blockade of K⁺-mediated postsynaptic components of the NK1R response. Furthermore, the NK1R-induced inward current persisted during TTX-induced blockade of spike activity, suggesting a direct postsynaptic effect of the recorded CEl neuron. One intriguing possibility is NK1R-mediated stimulation of GABA_ARs at extrasynaptic sites in CEl neurons, which are activated by ambient GABA (Botta et al. 2015; Romo-Parra et al. 2015). Increasing activation of extrasynaptic GABA_ARs would lead to an increase in tonic inhibitory tone in the CEl synaptic circuitry, thereby adding a slow time course of SP modulation complementing the regulation of fast inhibitory synaptic GABAergic events.

Possible synaptic network correlates and functional significance of SP-mediated effects in CE.

The SP system is activated during stress in limbic structures, involved in regulation of mood, anxiety behaviors, and emotion processing (for review see Ebner at al. 2009). SP is highly expressed in the amygdala (Nakaya et al. 1994), in particular in the CE, and SP immunoreactivity is colocalized with GAD staining (Shigematsu et al. 2008). Moreover, tissue-specific expression of SP in the CE is supported by highly conserved long-range enhancer and MEIS1 transcription factor (Davidson et al. 2006). Oral application of NK1R antagonist can be used clinically for the treatment of depression (Keller et al. 2006; Kramer et al. 2004; Ratti et al. 2011). Clinical studies have also suggested that phobic reactions affect the SP-NK1R system in the right amygdala in female patients (Michelgård et al. 2007). More recently, SP-NK1R dysregulation has been proposed as a pathophysiological mechanism in posttraumatic stress disorder (PTSD; Mathew et al. 2011) and social anxiety disorder, where increased NK1R availability in the amygdala was observed (Frick et al. 2015). Significant elevations in SP concentrations in cerebrospinal fluid (CSF) were found in male combat veterans with PTSD, as well as phasic increases in SP following symptom provocation (Geracioti et al. 2006). While the NK1R antagonist GR205171 does not seem to be effective in reducing symptoms associated with chronic PTSD (Mathew et al. 2011) it shows anxiolytic effects in healthy humans (Pringle et al. 2011) and alleviates anxiety symptoms and anxiety-induced cerebral blood flow in the amygdala of social anxiety patients (Furmark et al. 2005).

Investigations in animal models have helped in looking precisely into NK1R mechanisms: mice with functional ablation of the NK1R gene (De Felipe et al. 1998) are hyperactive (Herpfer et al. 2005) and represent abnormalities similar to those seen in patients with attention deficit hyperactivity disorder (for review see Yan et al. 2009). Furthermore, in healthy young men infusion of SP caused an anxiogenic and memory-disturbing effect (Herpfer et al. 2007). On the other hand, mice genetically deficient in NK1R displayed suppressed alcohol consumption (George et al. 2008; Thorsell et al. 2010) and disrupted behaviors associated with morphine reward (Gadd et al. 2003). Furthermore, systemic application of an NK1R antagonist into mice with high anxiety-related behavior as well as microinjection of an NK1R antagonist into the amygdala of gerbils caused anxiolytic-like effects (Heldt et al. 2009; Sartori et al. 2011). In the medial amygdala, NK1R antagonists buffered stress-induced aggravation of fear (Ebner et al. 2004).

Our findings provide a possible explanation of these SP actions on the cellular level in the CEl by showing an NK1R-mediated increase in activity of GABAergic neurons at both the direct postsynaptic and presynaptic network levels. One important observation was that predominantly PKCδ-negative CEl neurons express NK1R, suggesting that these types of neurons are the dominant target of SP action in CEl. The CEl contains two functionally distinct subpopulations of neurons, which are excited (CElon) and inhibited (CEloff) upon presentation of fear-conditioned stimuli and which are largely overlapping with the genetically defined PKCδ-positive and -negative neuronal subtypes (Ciocchi et al. 2010; Haubensak et al. 2010). These findings suggest that SP via stimulation of NK1R mediates direct excitation mostly of the CElon pathway. It is important to add that NK1R immunoreactivity was also observed in PKCδ-positive neurons in CEl, although in a minority of cells (17%) compared with high expression in the population of PKCδ-negative neurons (80%). The CElon and CEloff neurons are mutually interconnected, both types connect to output neurons in CEm, and conditioned fear responses are driven by stimulus-induced disinhibition of these CEm output neurons (Ciocchi et al. 2010). The NK1R-mediated increase in excitability in the mutually interconnected CElon-CEloff pathway can be assumed to boost the disinhibition of CEm output neurons, and thereby to make an important contribution to the anxiogenic-like action of SP in the amygdala.

GRANTS

This work was supported by IMF (Innovative Medical Research, University Münster; SO220608 to L. Sosulina, ST 611004 to C. Strippel, and KA220804 to T. Kanyshkova), the German research foundation (DFG; SFB-TRR58, TPA03 to H.-C. Pape), and Austrian Science Funds FWF P25375 to N. Singewald.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: L.S. and H.-C.P. conception and design of research; L.S., C.S., H.R.-P., A.L.W., T.K., S.B.S., and M.D.L. performed experiments; L.S., C.S., H.R.-P., A.L.W., T.K., S.B.S., and M.D.L. analyzed data; L.S., H.R.-P., S.B.S., N.S., and H.-C.P. interpreted results of experiments; L.S., H.R.-P., A.L.W., and S.B.S. prepared figures; L.S., C.S., H.R.-P., S.B.S., and H.-C.P. drafted manuscript; L.S., H.R.-P., and H.-C.P. edited and revised manuscript; L.S., C.S., H.R.-P., A.L.W., S.B.S., M.D.L., N.S., and H.-C.P. approved final version of manuscript.