Neurotensin speeds inhibition of dopamine neurons through temporal modulation of GABA_A and GABA_B receptor-mediated synaptic input

Abstract

Midbrain dopamine neurons play physiological roles in many processes including reward learning and motivated behavior, and are tonically inhibited by γ-aminobutyric acid (GABA)ergic input from multiple brain regions. Neurotensin (NT) is a neuropeptide which acutely modulates midbrain dopamine neuron excitability through multiple mechanisms, one of which is a decrease of GABA-mediated inhibition. However, the mechanisms through which NT depresses GABA signaling are not known. Here we used whole cell patch-clamp electrophysiology of dopamine neurons in mouse brain slices to show that NT acts both presynaptically to increase GABA_A and postsynaptically to decrease GABA_B receptor-mediated currents in the substantia nigra. The active peptide fragment NT_8–13 enhanced GABA_A signaling presynaptically by causing an increase in the size of the readily releasable pool of GABA via activation of the NT type-1 receptor and protein kinase A. Conversely, NT_8–13 depressed GABA_B signaling postsynaptically via the NT type-2 receptor in a process that was modulated by protein kinase C. Both forms of plasticity could be observed simultaneously in single dopamine neurons. Thus, as the kinetics of GABA_A signaling are significantly faster than those of GABA_B signaling, NT functionally speeds GABAergic input to midbrain dopamine neurons. This finding contributes to our understanding of how neuropeptide-induced plasticity can simultaneously differentiate and integrate signaling by a single neurotransmitter in a single cell and provides a basis for understanding how neuropeptides use temporal shifts in synaptic strength to encode information.

1. INTRODUCTION

Midbrain dopamine neurons encode both rewarding and aversive stimuli (Schultz, 1998; Ungless et al., 2004; Matsumoto and Hikosaka, 2007; Lammel et al., 2011; Zweifel et al., 2011). The firing patterns that relay this information are the result of synaptic signals that are conveyed via fast neurotransmitters and modulatory neuropeptides. The majority of fast neurotransmission in the midbrain is produced by γ -Aminobutyric acid (GABA)ergic input (Bolam and Smith, 1990) that arises from multiple brain regions and inhibits dopamine neuron activity. The midbrain also contains a dense plexus of inputs expressing the neuropeptide neurotensin (NT; Geisler and Zahm, 2006). These neurotensinergic inputs maintain tonic levels of NT in the midbrain (Frankel et al., 2011) and, like GABAergic inputs, arise from multiple brain regions.

NT modulates reward-associated behavior including feeding and drug self-administration (Kelley et al., 1989; Hanson et al., 2013) and is capable of maintaining self-administration directly into the midbrain (Glimcher et al., 1987). The mechanisms through which NT has these effects are complex and incompletely understood. NT excites dopamine neurons in the substantia nigra pars compacta (SNc) and ventral tegmental area (VTA) through direct depolarization at high concentrations (Jiang et al., 1994, Wu et al., 1995) and disinhibition via decreased dopamine D2 autoreceptor signaling at lower concentrations (Shi and Bunney, 1992; Nimitvilai et al., 2012; Piccart et al., 2015; Stuhrman and Roseberry, 2015). However, little is known about how neurotensin modulates synaptic GABA signaling in the midbrain.

Dopamine neurons express two classes of GABA receptors. Fast ionotropic GABA_A receptors hyperpolarize dopamine neurons by activating a chloride conductance while slower metabotropic GABA_B receptors couple to G-protein coupled inwardly rectifying potassium channels (GIRKs) to hyperpolarize dopamine neurons by activating a potassium conductance. Each receptor exerts distinct effects on dopamine neuron excitability, as blockade of GABA_A receptors increases burst firing while blockade of GABA_B receptors increases firing rate (Tepper and Lee, 2007). NT decreases GABA_B inhibitory postsynaptic currents in VTA dopamine neurons (Stuhrman and Roseberry, 2015), but it is not clear if this effect occurs at the level of the receptor or the presynaptic terminal, or if GABA_A signaling is affected.

Here we show that NT acts presynaptically to increase GABA_A currents and postsynaptically to decrease GABA_B currents in SNc dopamine neurons. This bidirectional modulation of GABAergic signaling can be observed simultaneously in single cells and proceeds through distinct NT receptors and intracellular signaling pathways. These findings indicate that NT exhibits temporal control over inhibitory input to midbrain dopamine neurons by increasing the strength of fast inhibitory signaling and decreasing the strength of slow inhibitory signaling. This neuropeptide-induced temporal shift is a product of a dual mechanism that has not previously been described in single neurons.

2. MATERIALS AND METHODS

2.1. Animals

Male DBA/2J mice were either purchased from The Jackson Laboratory or were the first generation offspring of previously purchased mice. Animals were group-housed under a reverse light/dark cycle (lights off from 9:00 A.M. to 9:00 P.M.). Humidity and temperature were controlled in the housing facility, and food and water were available ad libitum. Animal usage was reviewed and approved by Institutional Animal Care and Use Committees at the University of Texas Health Science Center at San Antonio and the Oklahoma Medical Research Foundation.

2.2. Brain slice electrophysiology

On the day of the experiment, mice (post-natal day range 42–120) were anesthetized with isoflurane and immediately decapitated. The brains were quickly extracted and placed in ice-cold carboxygenated (95% O₂ and 5% CO₂) artificial cerebral spinal fluid (aCSF) containing the following (in mM): 126 NaCl, 2.5 KCl, 1.2 MgCl₂, 2.5 CaCl₂, 1.4 NaH₂PO₄, 25 NaHCO₃, and 11 D-glucose. Kynurenic acid (1 mM) was added to the buffer for the slicing procedure. Horizontal midbrain slices (200 µm) containing the substantia nigra pars compacta were obtained using a vibrating microtome (Leica). Slices were incubated for at least 30 min at 34–36°C with carboxygenated aCSF that also contained the NMDA receptor antagonist MK-801 (10–20 µM).

Slices were placed in a recording chamber attached to an upright microscope (Nikon Instruments) and maintained at 34–36°C with aCSF perfused at a rate of 1.5 ml/min. Dopamine neurons were visually identified using gradient contrast optics based on their location in relation to the midline, third cranial nerve, and the medial terminal nucleus of the accessory optic tract. Neurons were further identified physiologically by the presence of spontaneous pacemaker firing (1–5 Hz) with wide extracellular waveforms (> 1.1 ms) and a hyperpolarization-activated current (I_H) of > 100 pA. Recording pipettes (1.5–2.5 MΩ resistance) were constructed from thin wall capillaries (World Precision Instruments) with a PC-10 puller (Narishige International). Iontophoretic pipettes (100–150 MΩ) were made with a P-97 puller (Sutter Instruments). Whole-cell recordings for GABA_B experiments were obtained using an intracellular solution containing the following (in mM): 115 K-methylsulfate, 20 NaCl, 1.5 MgCl₂, 10 HEPES, 2 ATP, 0.4 GTP, and either 10 BAPTA or 0.025 EGTA, pH 7.35–7.40, 269–274 mOsm. For GABA_A experiments the intracellular solution differed only by a lower concentration of K-methylsulfate (57.5 mM) and the addition of KCl (57.5 mM). In our initial experiments, we observed no effect of calcium chelation on the effects of the active peptide fragment NT_8–13 and thus these data were pooled. Afterward, all experiments were performed using an intracellular solution containing 10 mM BAPTA.

GABA_B inhibitory postsynaptic currents (IPSCs) were evoked during voltage-clamp recordings (holding voltage, −55 mV) using electrical stimulation in the presence of the following receptor antagonists: picrotoxin (100 µM, GABA_A), sulpiride (100–200 nM, D2), DNQX (10 µM, AMPA), hexamethonium (100 µM, nAChR), MPEP hydrochloride (300 nM, mGluR5), JNJ 16259685 (500 nM, mGluR1) and AM 251 (500 nM, CB₁). GABA_A IPSC experiments used a holding voltage of −70 mV and were conducted in the presence of the GABA_B antagonist CGP 55845 (100 nM), sulpiride, DNQX, and hexamethonium. A bipolar stimulating electrode was inserted into the slice 100–200 µm caudal to the cell, and GABA_B IPSCs were evoked with a train of 3 stimulations (0.5 ms) applied at 100 Hz once every 30 seconds. GABA_A IPSCs were evoked with a pair of single pulses (0.5 ms) separated by 50 ms. Currents were also evoked by episodic iontophoresis of GABA (1 M in the pipette, pH 4.0). These experiments were carried out in the presence of picrotoxin (100 µM) to isolate GABA_B receptor-mediated currents or CGP 55845 (100 nM) to isolate GABA_A receptor-mediated currents. Leakage from the iontophoretic pipette was prevented by the application of a negative backing current, and GABA was ejected as a cation with a pulse of 200–500 nA for 10–300 ms with an ION-100 single-channel iontophoresis generator (Dagan Corporation). For antagonist and inhibition experiments, compounds were perfused for at least 10 minutes immediately before NT_8–13 application and continued throughout NT_8–13 perfusion, except for experiments using phosphatase inhibitor cocktails 2 or 3 in which slices were preincubated for at least 35 minutes before the recording.

2.3. Drugs

Kynurenic acid, MK-801, DNQX, GABA, sulpiride, hexamethonium, SR142948, H89, and phosphatase inhibitor cocktails 2 and 3 were purchased from Sigma-Aldrich. Staurosporine, SR48692, CGP55845, AM 251, JNJ 16259685, and MPEP hydrochloride were purchased form Tocris Bioscience. Calphostin C was purchased from Cayman Chemical. The active fragment NT_8–13 was purchased from American Peptide Company.

2.4. Experimental design and statistical analyses

Data were collected on a Dell computer running Windows 7 using Axograph version 1.5.4 and LabChart (AD instruments). Statistical analyses were performed using Prism Graphpad version 6. Sigmoidal curve fit analysis was used to determine EC₅₀ value for NT on GABA_B signaling. One-way or two-way ANOVAs or Student’s t-tests were used for between- and within-group comparisons. In most cases the effect of NT_8–13 was analyzed in the 5 sweeps immediately following perfusion (i.e., the beginning of washout) with two exceptions. The effect of 150 nM NT_8–13 on GABA_B signaling was analyzed both at the peak effect and at a later time point in order to compare the rate of recovery following a maximum response. Similarly, the peak effect of JMV-431 on GABA_B signaling was also analyzed as opposed to a single time point because recovery from depression was observed before the end of the actual perfusion. Dunnett’s post hoc tests were performed subsequent to significant ANOVAs. Data are presented as mean ± SEM. In all cases, α was set a priori at 0.05.

3. RESULTS

3.1. Neurotensin increases the readily releasable pool of GABA through NTSR1 and PKA

We performed whole-cell electrophysiological recordings of SNc dopamine neurons in horizontal brain slices from mice to determine the effect of NT receptor activation on GABA receptor-mediated neurotransmission. We obtained recordings in dopamine neurons voltage clamped at −70 mV of GABA_A inhibitory postsynaptic currents (IPSCs). GABA_A IPSCs were pharmacologically isolated using a cocktail of antagonists and blocked by the GABA_A receptor antagonist picrotoxin (Figure 1A). Bath perfusion of the active fragment of the peptide, NT_8–13 (50 nM), produced a gradual and sustained increase in the amplitude of GABA_A IPSCs (average of 5 sweeps before vs. immediately following perfusion, n = 19 cells from 10 mice, 46.41 ± 8.67% increase; t₁₈ = 5.31; p < 0.0001, Figure 1A,B). NT_8–13 (50 nM) also induced a modest inward shift in the holding current (n = 11 cells from 5 mice, 21.38 ± 3.44 pA; t₁₀ = 7.08, p < 0.0001, data not shown) and an increase in noise indicated by an increase in the variability of the holding current. A previous report indicated that intracellular calcium levels can modulate the effect of NT_8–13 on GABA_B signaling in VTA dopamine neurons (Stuhrman and Roseberry, 2015); however, we found that the NT_8–13-induced increase in the GABA_A IPSC in the SNc was not affected by chelating calcium in the neuron being recorded (chelator in intracellular solution; 0.025 mM EGTA: n = 7 cells from 5 mice, 51.25 ± 12.49% increase; 10 mM BAPTA: n = 12 cells from 5 mice, 43.59 ± 11.97% increase, p = 0.96, see Figure 2C). As such, data from experiments conducted using the two levels of intracellular calcium chelation were pooled (Figure 1B). To determine if NT_8–13 acts pre- or postsynaptically to enhance GABA_A IPSCs, we next elicited GABA_A receptor-mediated currents via iontophoresis of exogenous GABA (Beckstead et al., 2009). Bath perfusion of NT_8–13 had no effect on the amplitude of these currents (average of 5 sweeps before vs. immediately following perfusion; n = 10 cells from 5 mice, t₉ = 0.36, p = 0.73, Figure 1B). To further investigate presynaptic effects of NT_8–13 we next attempted to measure spontaneous GABA_A receptor-mediated IPSCs (sIPSCs). However, the increase in noise caused by NT_8–13 made measurements of sIPSC frequency unreliable and limited our ability to measure the amplitude of smaller sIPSCs. We therefore analyzed, in each cell, the amplitude of the 10 largest sIPSCs during a 150-second sampling taken both before and after drug application. Using this analysis, NT_8–13 had no effect on sIPSC amplitude (average of 10 largest sIPSCs; n = 6 cells from 4 mice, baseline: 54.18 ± 8.74 pA; NT_8–13: 51.00 ± 7.42 pA, t₅ = 0.51, p = 0.62, data not shown), consistent with a presynaptic locus of effect.

Figure 1. NT enhances GABA_A neurotransmission presynaptically via an increase in the size of the readily releasable pool.

Representative traces of the effect of the GABA_A receptor antagonist picrotoxin (Picro) or the active fragment of neurotensin (NT_8–13) on GABA_A IPSCs and GABA_A receptor-mediated currents elicited by iontophoresis of GABA (A). Summary data indicate that NT_8–13 enhances GABA_A receptor-mediated currents elicited via electrical stimulation (IPSC) but not GABA iontophoresis (B). Representative traces of trains of GABA_A IPSCs (40 stimuli at 20 Hz, artifacts digitally removed) before and after NT_8–13 perfusion (C). Representative graph from a single cell used for RRP_train analysis with inset zoomed on y-intercept of regression lines (D). NT_8–13 increased the size of RRP_train but did not affect ρ_train (E). Representative graph from a single cell used for RRP_EQ analysis (F). NT_8–13 increased the size of RRP_EQ and decreased ρ_EQ (G). *represents difference from baseline*p < 0.05; **p < 0.01; ***p < 0.001.

Figure 2. NT enhances GABA_A neurotransmission via NTSR1 and protein kinase A.

NT_8–13 (15–150 nM) bi-phasically modulated the amplitude of GABA_A IPSCs (A). NT_8–13 (50 nM) did not enhance GABA_A IPSCs in cells pre-incubated with SR48692 or from mice lacking NTSR1 (B). The NT_8–13-induced increase in GABA_A IPSCs is blocked either by pre-incubation with SR48692 or the protein kinase A (PKA) inhibitor H89, but not by chelating postsynaptic calcium (cells from the control and BAPTA groups were previously pooled to provide the time course in Fig. 1B) (C). A high concentration of NT_8–13 alone or in the presence of the NT type-1 receptor (NTSR1) antagonist SR48692 (SR48, 1 µM) had no effect on the amplitude of GABA_A IPSCs (D). *represents difference compared to control; *p < 0.05; **p < 0.01; ***p < 0.001. N.S. = not significant change from baseline.

Presynaptic enhancement of IPSCs may proceed by either an increase in the probability of release (ρ) or in the size of the readily releasable pool (RRP). To determine whether an increase in ρ contributes to NT-induced increases in the GABA_A IPSC we analyzed the effects of NT_8–13 on the ratio of the response to paired stimulation. Paired-pulse ratios were not affected by NT_8–13 (analysis performed on IPSC recordings from Figure 1B; average of 10 sweeps; n = 17 cells from 8 mice, baseline: 0.692 ± 0.064; NT_8–13: 0.723 ± 0.040; t₁₆ = 0.773, p = 0.451, data not shown, 2 cells excluded for having less than 10 sweeps to analyze), suggesting that NT_8–13 does not enhance GABA_A IPSCs via a change in ρ. We next investigated the size of the RRP and ρ via two different forms of analysis of high frequency (20 Hz, 4 s) trains of IPSCs. The first (Schneggenburger et al., 1999) consists of calculating the cumulative IPSC amplitude and fitting a line to the last 10 points. Using this technique, the Y-intercept of that line provides an estimate of the size of the RRP_train, while the ratio of the amplitude of the first IPSC to RRP_train provides an estimate of ρ_train. The second (Elmqvist and Quastel, 1965) consists of plotting cumulative IPSC amplitudes against individual IPSC amplitudes of the train. Fitting a line to the earliest IPSCs in the train provides an x-intercept which gives an estimate of size of the RRP_EQ, while the ratio of the amplitude of the first IPSC to RRP_EQ provides an estimate of ρ_EQ. Together, RRP_train and RRP_EQ are respectively thought to provide lower and upper bounds of the true RRP (Chu et al., 2012). As in earlier experiments using pairs of stimulations, NT_8–13 (50 nM) increased the amplitude of GABA_A IPSCs elicited via a train of stimuli by 54.7% (average of sweeps 10–20; n = 5 cells from 4 mice, baseline, 435.9 ± 89.15 pA, post-NT_8–13, 647.2 ± 62.36 pA; t₄ = 3.39, p = 0.028, Figure 1C,E,G). This increase in the GABA_A IPSC is attributed to an increase in the RRP as analyzed by both of the methods described above (RRP_train, 40.29 ± 11.70% increase; t₄ = 4.04, p = 0.016; RRP_EQ, 94.12 ± 15.12% increase; t₄ = 6.22, p = 0.0034, figure 1E, 1G). These two methods of analysis differed in the estimation of ρ, with no change in ρ_train (t₄ = 1.12, p = 0.33, Figure 1E) but a slight decrease in ρ_EQ (23.23 ± 5.90% decrease; t₄ = 3.69, p = 0.021, Figure 1G). The discrepancy between these two analyses in estimation of ρ may be attributed to the fact that the calculation of ρ_EQ fails to account for replenishment of the RRP while ρ_train incorporates replenishment into the analysis. Overall these analyses suggest that NT enhances GABA_A signaling presynaptically, primarily by increasing the size of the RRP.

There are three distinct NT receptors expressed in the midbrain, the NT type 1 (NTSR1) and type 2 (NTSR2) are G-protein coupled receptors while the NT type 3 (NTSR3, also known as sortilin) is a member of the vacuolar protein sorting 10 protein (Vps10p) receptor family (Petersen et al., 1997; Vincent et al., 1999; Sarret et al., 2003; Xia et al., 2013). We next investigated the NT receptor necessary for NT_8–13-induced increase in GABA_A IPSC amplitude. The effect of NT_8–13 on GABA_A IPSCs was biphasic and nonlinear (bell-shaped; 15 nM: n = 6 cells from 2 mice; 50 nM: n = 19 cells from 10 mice; 150 nM: n = 12 cells from 5 mice, Figure 2A). NT_8–13 (50 nM) failed to enhance GABA_A IPSCs in mice lacking the neurotensin type-1 receptor (NTSR1KO) or when slices were pre-incubated with SR48692 (1 µM, F_4,35 = 6.835, p = 0.0004). Dunnett’s post hoc analysis revealed a significant contribution of NTSR1 on the enhancement of the GABA_A IPSC (NTSR1KO, n = 8 cells from 3 mice, compared to control, p = 0.0156; SR48692, n = 5 cells from 2 mice, compared to control, p = 0.0026, Figure 2B,C). Interestingly, a higher concentration of NT_8–13 (150 nM) did not significantly enhance GABA_A IPSCs (average of 5 sweeps preceding vs. immediately following perfusion; NT_8–13 alone: n = 12 cells from 5 mice, 10.76 ± 7.91% increase, t₁₁ = 1.06, p = 0.31, Figure 2D). Furthermore, NT_8–13 (50 nM) induced a moderate and transient depression of GABA_A IPSCs in the presence of SR48692, which may be indicative of an NTSR2 effect. Since NTSR1 has higher affinity for NT than NTSR2 or NTSR3 (Vincent et al., 1999), it was possible that the biphasic concentration-response curve could be explained by higher concentrations of NT activating NTSR2 or NTSR3 which in turn blunted GABA release. However, NT_8–13 (150 nM) did not significantly decrease GABA_A IPSCs in the presence of NTSR1 antagonist SR48692 (average of 5 sweeps preceding vs. immediately following perfusion; n = 5 cells from 3 mice, 3.070 ± 11.33% decrease from baseline; t₄ = 0.738, p = 0.50, Figure 2D) suggesting that NTSR1 is responsible for the entire concentration-response curve. We next investigated the intracellular signaling mechanisms through which NTSR1 activation enhances GABA release. It has previously been shown that neurotensin modulates midbrain dopamine neuron activity in a protein kinase A (PKA)-dependent manner (Shi and Bunney, 1992). Consistent with this finding, when slices were pre-incubated with the PKA inhibitor H89 (1 µM), 50 nM NT_8–13 failed to enhance GABA_A IPSCs (Dunnett’s, n = 8 cells from 4 mice, p = 0.0082, Figure 2C). Together, these experiments indicate that 50 nM NT_8–13 enhances GABA_A signaling through NTSR1 and PKA.

3.2. Neurotensin depresses GABA_B receptor signaling postsynaptically via NTSR2

Our finding that NT_8–13 increases GABA release from terminals was surprising given a previous report of a NT_8–13-induced decrease in the amplitude of GABA_B receptor-mediated IPSCs in the VTA (Stuhrman and Roseberry, 2015). The discrepancy between these two findings could be attributed to a postsynaptic site of action for NT_8–13 on GABA_B signaling. We next obtained recordings in dopamine neurons voltage clamped at −55 mV of GABA_B inhibitory postsynaptic currents (IPSCs) that were pharmacologically isolated using a cocktail of antagonists and blocked by the GABA_B receptor antagonist CGP 55845 (Figure 3A). Consistent with a postsynaptic mechanism of action, we observed that NT_8–13 (50 nM) depressed GABA_B receptor-mediated currents elicited by both electrical stimulation (n = 8 cells from 5 mice; baseline: 55.63 ± 4.70 pA; 5 sweeps following NT_8–13 perfusion: 38.3 ± 6.29 pA; t₇ = 5.78, p = 0.0007) and iontophoresis of GABA (baseline: 89.77 ± 6.57 pA; last 5 sweeps of NT_8–13 perfusion: 61.22 ± 6.89; t₈ = 5.37, p = 0.0007; Figure 3A,B,C). Furthermore, these effects occurred to the same extent regardless of the source of GABA (IPSCs, 32.22 ± 6.40% decrease from baseline; iontophoresis, 31.50 ± 5.99% decrease from baseline; t₁₅ = 0.082, p = 0.936). As was observed with GABA_A IPSCs, NT_8–13-induced depression of GABA_B IPSCs was not dependent on intracellular calcium in the cell being recorded (5 sweeps following NT_8–13 perfusion; chelator in intracellular solution; 0.025 mM EGTA: n = 4 cells from 3 mice, 25.82 ± 11.4% decrease; 10 mM BAPTA: n = 4 cells from 2 mice, 38.61 ± 8.40% decrease; t₆ = 1.00, p = 0.356, data not shown). As such, data from experiments conducted using the two levels of intracellular calcium chelation were pooled (Figure 3C). Concentration-response analysis indicated an EC₅₀ value of 56.7 nM for the NT_8–13-induced depression of GABA_B signaling (5 nM: 4 cells from 3 mice; 15 nM: 6 cells from 2 mice; 50 nM: 9 cells from 4 mice; 150 nM: 6 cells from 3 mice; 500 nM: 2 cells from 2 mice, Figure 3D). We next sought to determine which NT receptor(s) are necessary for the effects of NT_8–13 on GABA_B currents. It is possible that both NTSR1 and NTSR2 activation or multiple intracellular signaling pathways contribute to NT-induced depression of GABA_B currents, but these effects would be difficult to dissect by analyzing only the moderate depression caused by lower concentrations of NT_8–13. As such, antagonist and inhibitor experiments were carried out using a higher concentration of NT_8–13 (150 nM). Slices were pre-incubated either with the NTSR1 antagonist SR48692 (1 µM) or the non-selective NTSR1/2 antagonist SR142948A (1 µM). NT_8–13-induced depression of GABA_B signaling was blocked by the NTSR1/2 antagonist SR142948A (n = 4 cells from 3 mice, 8.67 ± 3.54% decrease; t₃ = 2.17, p = 0.12) but not by the NTSR1 specific antagonist SR48692 (n = 4 cells from 2 mice, 36.35 ± 7.28% decrease; t₃ = 4.15, p = 0.025, Figure 3E), suggesting that NTSR2 activation is required to depress GABA_B signaling.

Figure 3. NT depresses GABA_B signaling via NTSR2 activation and requires the presence (but not activation) of NTSR1.

Representative traces of the effect of the GABA_B receptor antagonist CGP 55845 (CGP) or NT_8–13 on GABA_B receptor-mediated currents elicited by electrical stimulation (A) or iontophoresis of GABA (B). Unlike GABA_A signaling, NT_8–13 depresses GABA_B receptor-mediated currents elicited via both electrical stimulation (IPSC) and iontophoresis to a similar magnitude of effect (C). The relationship between the concentration of NT_8–13 and maximum degree of depression of GABA_B receptor-mediated currents elicited via iontophoresis of GABA indicates an EC₅₀ of 56.7 nM (D). NT_8–13 depresses GABA_B receptor-mediated currents elicited via iontophoresis of GABA in the presence of the NTSR1 antagonist SR48692 (SR48) but not the dual NT type-1/NT type-2 receptor antagonist SR142948A (SR14, E). NT_8–13-induced depression of GABA_B receptor-mediated currents does not occur in mice lacking NTSR1 (F). * represents difference compared to baseline; *p < 0.05; ***p < .001; N.S. = not significant.

Surprisingly, NT_8–13 (50 nM) had a significantly reduced effect on GABA_B signaling in dopamine neurons from NTSR1KO mice compared to wild-type mice (NTSR1KO, n = 5 cells from 3 mice, 9.73 ± 9.72% decrease from baseline; t₄ = 1.18, p = 0.30, Figure 3F). However, we obtained additional support for the role of NTSR2 in depression of GABA_B signaling by using the NTSR2-selective compound JMV-431 (1–10 µM) which also decreased GABA_B signaling (peak effect of JMV-431; 1 µM: n = 3 cells from 1 mouse, −16.06 ± 2.23% change from baseline, t₂ = 30.92, p = 0.001; 10 µM: n = 6 cells from 2 mice, −33.19 ± 7.732% change from baseline, t₅ = 3.264, p = 0.0223, data not shown). These results suggest that NT_8–13 depresses GABA_B signaling postsynaptically in a manner that is dependent on the activation of NTSR2, and the presence but not the activation of NTSR1.

We next investigated the intracellular mechanism through which NT_8–13 depresses GABA_B signaling. Inhibition of PKA by pre-incubation with H89 (1 µM) failed to block NT_8–13-induced depression of GABA_B currents (data not shown). Furthermore, the peak effect of NT_8–13 (150 nM) was not significantly changed by inhibiting: protein kinase C (PKC) with calphostin C (1 µM, n = 8 cells in 3 mice, Figure 4A,B) in the recording pipette, a large array of protein kinases by pre-incubation with staurosporine (1 µM, n = 4 cells in 2 mice, Figure 4A,B), and a large array of phosphatases with two different phosphatase inhibitor cocktails purchased from Sigma-Aldrich (PIC2: n = 7 cells from 3 mice; PIC3: n = 5 cells from 3 mice, 1% by volume pre-incubation, measured as the average of 5 sweeps at the peak effect of NT_8–13; one-way ANOVA; F_5,30 = 0.85, p = 0.52, data not shown). We next analyzed recovery of GABA_B signaling after NT_8–13 perfusion. A two-way ANOVA revealed a significant interaction between treatment and time (F_2,15 = 3.99, p = 0.41). Dunnett’s posthoc analysis revealed that inhibition of protein kinases by staurosporine (1 µM; n = 2 cells from 1 mouse, p = 0.054; Figure 4A,B) or PKC by calphostin C (1 µM; n = 8 cells from 3 mice, p = 0.012; Figure 4A,B) impeded recovery of GABA_B currents towards baseline, though this effect was only statistically significant in the latter (average of 5 sweeps at end of recording, change from baseline; NT_8–13 alone: −21.74 ± 7.0%; Staurosporine: −52.0 ± 6.2%; Calphostin C: −46.6 ± 6.8%, Figure 4B). Together these data indicate that NT-induced depression of GABA_B input proceeds through a distinct intracellular signaling pathway from the enhancement of GABA_A signaling and exhibits a recovery that is PKC-dependent.

Figure 4. NT_8–13-induced depression of GABA_B receptor-mediated currents recovers in a protein kinase C-dependent manner.

Time course of NT_8–13 (150 nM)-induced depression of GABA_B receptor-mediated currents elicited via iontophoresis of GABA, under control conditions or when the slice was pre-incubated with either the non-specific protein kinase inhibitor staurosporine (Staur) or the protein kinase C inhibitor calphostin C (Cal C, A). NT_8–13-induced depression of GABA_B receptor-mediated currents elicited via iontophoresis of GABA persisted when the slice preparation was pre-incubated with calphostin C (B). * represents difference from NT_8–13 alone; *p < 0.05.

3.3. Neurotensin shifts strength of inhibitory inputs toward fast GABAergic signaling

The experiments thus far indicating that NT_8–13 enhances GABA_A signaling were performed in the presence of the GABA_B receptor antagonist CGP55845, and likewise the GABA_B experiments were performed in the presence of the GABA_A receptor antagonist picrotoxin. GABA_B receptors are located both postsynaptically on dopamine neurons and presynaptically on terminals of GABAergic inputs to dopamine neurons (Paladini and Tepper, 1999; Boyes and Bolam, 2003). Our experiments to this point seemingly eliminated the possibility of a physiological contribution from the GABA_B receptors located on GABA terminals in the effects of NT_8–13 on GABA release. The question remained whether these seemingly disparate effects of a single neuropeptide on signaling through a single fast neurotransmitter can occur simultaneously within a single cell in the absence of receptor blockers. Therefore, to determine the net physiological effect of NT receptor activation on GABA transmission we took advantage of the fact that the ion concentrations in our aCSF and intracellular solutions result in GABA_A receptor-mediated Cl⁻ currents that are inward and GABA_B receptor-mediated K⁺ currents that are outward (when cells are voltage-clamped at −55 mV). Delivery of a train of three stimuli (0.5 ms, 100 Hz) produced a fast inward current characteristic of GABA_A IPSCs and a slow outward current characteristic of GABA_B IPSCs (Figure 5A). NT_8–13 significantly enhanced putative GABA_A IPSCs (measured as peak inward current occurring 0–10 ms after the last stimulation in the train) and depressed putative GABA_B IPSCs (measured as peak outward current 10–500 ms after the last stimulation in the train, Figure 5B). We also found that blockade of GABA_B receptors with CGP55845 (data from Figure 1B) did not alter NT_8–13-induced enhancement of GABA_A IPSCs (average of 5 sweeps before vs. immediately following perfusion; control: n = 7 cells from 3 mice, 48.93 ± 18.86% increase from baseline; CGP55845: n = 14 cells from 8 mice, 50.83 ± 10.91% increase from baseline; t₁₉ = 0.108, p = 0.915, Figure 5C). Similarly, blockade of GABA_A receptors with picrotoxin (data from figure 3C) failed to alter NT_8–13-induced depression of GABA_B IPSCs (average of 5 sweeps before vs. 5 sweeps following perfusion; control: n = 6 cells from 3 mice, 39.17 ± 5.30% decrease from baseline; picrotoxin: n = 8 cells from 4 mice, 32.22 ± 6.40% decrease from baseline; t₁₂ = 0.80, p = 0.44, Figure 5C). These data suggest that under physiological conditions (i.e., in the absence of GABA receptor antagonists) NT receptor activation can simultaneously enhance fast GABA_A signaling and depress slow GABA_B signaling in a single cell, producing a net acceleration of GABA signaling.

Figure 5. NT-induced plasticity temporally modulates GABAergic input to midbrain dopamine neurons.

Representative traces of the simultaneous effects of NT_8–13 on GABA_A and GABA_B receptor-mediated signaling in a single neuron, with the GABA_A portion of the dual IPSC being blocked by picrotoxin (Picro, A). NT_8–13 simultaneously increased rapid inward currents with peak amplitude between 0–10 ms after stimulation (putatively GABA_A) and decreased slow outward currents with peak amplitude between 10–500 ms after stimulation (putatively GABA_B, B). There is no effect of the GABA_B antagonist CGP55845 (CGP, data from figure 1B) on the NT_8–13-induced enhancement of GABA_A currents, nor is there an effect of the GABA_A antagonist picrotoxin (data from figure 3C) on the NT_8–13-induced reduction in GABA_B currents (C).

4. DISCUSSION

The present work shows that NT bi-directionally modulates GABAergic input to midbrain dopamine neurons through two distinct NT receptors and intracellular signaling cascades. NT increases GABA_A signaling to dopamine neurons presynaptically via NTSR1 and PKA. NT does not appear to increase GABA release at GABA_B synapses, but rather depresses GABA_B receptor signaling postsynaptically via NTSR2 and recovers through a mechanism that involves PKC. This dual plasticity results in a shift toward fast inhibitory signaling at dopamine neurons. These results inform our understanding of how neuropeptides can contribute to the encoding of diverse synaptic inputs to a single neuron.

4.1. Peptide modulation of synaptic transmission

Neuropeptides affect neuronal excitability and behavior via diverse mechanisms, and a growing number of reports are identifying neuropeptides as modulators of synaptic transmission. However, while neuropeptides have been shown to act pre- or postsynaptically and sometimes both at a single cell, these effects typically drive a single endpoint (e.g., excitation). Some neuropeptides specifically decrease excitatory neurotransmitter release (Tallent and Siggins, 1997) while others tend to increase it (Acuna-Goycolea and van den Pol, 2004). Neuropeptide Y and corticotropin releasing factor can act pre- and postsynaptically to synergistically depress activity of orexin neurons and dopamine neurons, respectively (Fu et al., 2004; Beckstead et al., 2009; Williams et al., 2014). Here we show that NT engages distinct receptors and intracellular mechanisms, with the net effect of speeding inhibitory input to dopamine neurons. These findings add an additional layer of complexity to the established mechanisms through which neuropeptides can affect synaptic transmission, neuronal excitability, and the coding of information.

Specifically in midbrain dopamine neurons, NT modulates both inhibitory and excitatory signaling in a way that does not provide a clear directional response in overall excitability. Intrinsically, NT directly depolarizes dopamine neurons through activation of a non-selective cation conductance (Jiang et al., 1994). Synaptically, NT affects input to dopamine neurons produced by at least three distinct neurotransmitters and five distinct receptors. NT produces biphasic effects on glutamate signaling in VTA dopamine neurons that are both NTSR1-dependent, with low concentrations enhancing and high concentrations depressing currents (Kempadoo et al., 2013). NT depresses dopamine D2 autoreceptor-mediated synaptic currents in the SNc and VTA (Piccart et al., 2015), as well as GABA_B receptor-mediated currents in the VTA (Stuhrman and Roseberry, 2015). Our work expands on these findings by demonstrating that the effects on GABA_B receptor signaling occur in the SNc, are postsynaptic, depend on NTSR2, and require PKC activation to terminate. Furthermore, NT also enhances GABA release preferentially at GABA_A receptor synapses in a NTSR1- and PKA-dependent manner. These findings reinforce the notion that the effects of NT on dopamine neurons are complex and involve multiple neurotransmitter systems, both excitatory and inhibitory.

NT also affects synaptic transmission in other brain areas. NT enhances serotonin signaling in the prefrontal cortex, likely via direct activation of serotonergic neurons through NTSR1 (Jolas and Aghajanian, 1997; Petkova-Kirova et al., 2008). Neurotensin can also decrease neurotransmission by engaging endocannabinoid systems, both in the dorsal striatum by decreasing glutamate release (Yin et al., 2008) and in the periaqueductal gray by decreasing GABA release (Mitchell et al., 2009). Similar to our findings, NT presynaptically enhances GABA signaling in the oval bed nucleus of the stria terminalis, (Krawczyk et al., 2013). However, that effect is attributed to increased probability of release rather than size of the readily releasable pool, occurs at higher concentrations of NT perfusion than used presently, and is not dependent on PKA (Krawczyk et al., 2013). The current findings are also congruent with previous work showing that PKA increases GABA signaling presynaptically in hippocampal neurons and in VTA dopamine neurons (Capogna et al., 1995; Ingram et al., 1998; Melis et al., 2002).

4.2. GABAergic input to dopamine neurons

Direct activation of either rapid, ionotropic GABA_A receptors or slower, metabotropic GABA_B receptors results in hyperpolarization of dopamine neurons and inhibition of firing (Grace and Bunney, 1979). However, the endogenous sources of GABA which activate each respective receptor are distinct and the ultimate effects on dopamine neuron excitability are complex. Tonic activation of GABA_A receptors decreases firing rate and shifts dopamine neurons away from a burst firing mode (Paladini and Tepper, 1999; Brazhnik et al., 2008). In contrast, there is no tonic activation of GABA_B receptors (Paladini and Tepper, 1999), possibly due to their extrasynaptic location requiring spillover of GABA out of the synapse for activation (Isaacson et al., 1993; Boyes and Bolam, 2003). When specific GABAergic inputs are stimulated however, dopamine neuron firing is reduced by either GABA_A or GABA_B receptor activation (Edwards et al., 2017). Afferents from the striatum or nucleus accumbens preferentially activate GABA_B receptors, while those from the substantia nigra pars reticulata, globus pallidus, and the VTA preferentially activate GABA_A receptors (Brazhnik et al., 2008; Edwards et al., 2017). In both cases, stimulation of these afferents and the resulting activation of GABA receptors results in powerful inhibition of dopamine neuron firing (Brazhnik et al., 2008; Edwards et al., 2017). Remarkably, in the SNc inputs which preferentially activate GABA_B receptors (from the dorsal striatum and nucleus accumbens) express NTSR1, while the inputs which preferentially activate GABA_A receptors (substantia nigra pars reticulata and globus pallidus) do not (Boudin et al., 1996; Fassio et al., 2000; Pickel et al., 2001). However, it is likely that SNc GABA neurons express NTSR1, as 45% of NTSR1 expressing neurons in the substantia nigra pars compacta are non-dopaminergic (Opland et al., 2013). Our results suggest that in vivo, neurotensin release might preferentially weaken striatal inputs to dopamine neurons and strengthen input from local GABA neurons, providing a mechanism through which both temporal and spatial modulation of GABAergic input could occur.

The mechanism of action by which NTSR2 activation depresses GABA_B receptor-mediated currents is not clear. GABA_B receptor signaling can be reduced by many different mechanisms depending on the cells in which they are expressed. GABA_B receptors on dopamine neurons are sensitive to intracellular levels of calcium, with increased calcium leading to decreased GABA_B receptor-mediated currents (Beckstead and Williams, 2007; Sharpe et al, 2014). A previous report indicates that intracellular calcium levels can modulate the effect of NT_8–13 on GABA_B signaling in VTA dopamine neurons (Stuhrman and Roseberry, 2015); however, our findings suggest that this is not the case in SNc dopamine neurons. This discrepancy may be due to differences in calcium signaling in the SNc and VTA, which have differential expression of calcium activated ion channels and calcium-binding proteins (Wolfart et al., 2001; Neuhoff et al., 2002). GABA_B receptors on VTA GABA neurons slowly desensitize and are internalized in a protein phosphatase 2A dependent manner (Padgett et al., 2012). Further, GABA_B receptors on hippocampal neurons desensitize due to phosphorylation by PKC or Ca²⁺/calmodulin-dependent protein kinase II (Pontier et al., 2006; Guetg et al., 2010). Here we show that none of these pathways are responsible for the NT-induced depression of GABA_B receptor-mediated signaling in SNc dopamine neurons, although PKC does play a role in returning currents to baseline levels. Surprisingly, NT-induced depression of GABA_B signaling required the presence, but not the activation, of NTSR1. One possibility is that the constitutive NTSR1KO mice undergo developmental neuroadaptations that result in decreased NTSR2 expression and/or function. However, NTSR1KO mice do not exhibit a loss of NTSR2 expression or NT-induced analgesia (Remaury et al., 2001; Maeno et al., 2004). Furthermore, constitutive ablation of NTSR1 expressing cells in the VTA does not change expression of NTSR2 (Woodworth et al., 2017). Another possible explanation could be the formation of NTSR1/NTSR2 heterodimers, which affect trafficking of the receptors to the membrane in a cell expression system (Hwang et al., 2010). This suggests that perhaps there are not enough NTSR2 receptors at the cell surface in the NTSR1 knockout to produce an effect. Furthermore, we do not favor promiscuity of NT receptor antagonists as a pharmacological explanation, as 1 µM SR48692 has been previously reported to successfully block NTSR1 in brain slice preparations (Bose et al., 2015; Zhang et al., 2015; Zhang et al., 2016), and the NTSR2 selective compound JMV431 was here sufficient to depress GABA_B signaling.

4.3. Behavioral implications

GABA_A and GABA_B receptor-mediated signals in the midbrain play distinct roles in reward and substance use disorders. Intra-VTA infusion of a GABA_B receptor agonist, but not a GABA_A receptor agonist, results in a shift in responding for intracranial self-stimulation of the medial forebrain bundle (Willick and Kokkinidis, 1995). Similarly, intra-VTA delivery of a GABA_B antagonist, but not a GABA_A antagonist, prevents stress-induced reinstatement of cocaine seeking (Blacktop et al., 2016). Knockdown of GABA_A signaling in dopamine neurons increases locomotor activity elicited by morphine (Parker et al., 2011) but deletion of GABA_B receptors does not (Edwards et al., 2017). In contrast, GABA signaling at both receptors has been implicated in feeding behavior; either a GABA_A or a GABA_B receptor antagonist injected into the VTA blocks mu-opioid receptor-activated feeding by DAMGO or intra-accumbens GABA_B receptor-activated feeding by baclofen (Echo et al., 2002; Miner et al., 2010).

Our findings suggest that NT can temporally and spatially shift priority of GABA inputs to dopamine neurons. Given the established role for neurotensin in reinforcement, this shift of priority likely modulates the behavioral effects of natural rewards and abused drugs. Interestingly, NT analogs block cocaine conditioned place preference and reinstatement of cocaine seeking, which are dependent on interactions between external cues and rewards (Torregrossa and Kalivas, 2008; Boules et al., 2016). Conversely, NT receptor antagonists blunt amphetamine-induced hyperlocomotion, which is not dependent on external cues (Rompre, 1997; Rompre and Perron, 2000; Panayi et al., 2005). Given the complexity of NT-induced plasticity it is not surprising that conflicting evidence exists regarding how NT modulates the behavioral effects of abused drugs (recently reviewed by Ferraro et al., 2016). The diverse mechanisms of action through which NT modulates dopamine neuron excitability present both an abundance of targets and an ongoing challenge for the design of pharmacotherapeutic agents to treat substance use disorders.

Highlights.

• NT enhances GABA_A receptor-mediated signaling at midbrain dopamine neurons

• NT depresses GABA_B receptor-mediated signaling at dopamine neurons postsynaptically

• NT plasticity mechanism differs for GABA_A and GABA_B receptor-mediated signaling

• NT shifts the strength of inputs to dopamine neurons towards fast GABA signaling