Multiplexed Neurochemical Monitoring Reveals Glutamate Modulates Dopamine Neurotransmission in the Nucleus Accumbens

Abstract

Dopamine (DA) and glutamate (Glu) play vital roles in cognitive function, motor control, reward systems, and addiction. However, neurotransmitters are typically measured one at a time instead of simultaneously, which would allow a better picture of neurotransmitter interactions. In this study, we multiplexed dopamine detection using fast-scan cyclic voltammetry (FSCV) and glutamate detection using genetically encoded iGluSnFR fluorescent probes to simultaneously investigate dopamine and glutamate signaling in mouse nucleus accumbens slices. We perturbed glutamate transmission to identify how faster synaptic glutamate regulated slower dopamine volume transmission. After bathing brain slices in a buffer containing 100 μM glutamate, dopamine release decreased by nearly half. We then used drugs to inhibit different glutamatergic proteins to test their effects on dopamine release. Blocking the excitatory amino acid transporter (EAAT2) using DL-TBOA (20 μM) increased glutamate concentrations and decreased dopamine release. Blocking metabotropic glutamate receptors (mGluR2/3) with LY 341495 (40 μM) did not affect dopamine release and the mGluR2 agonist LY 379268 (10 μM) also did not affect dopamine release, so the effect was not mediated by metabotropic group 2/3 receptors. Blocking NMDAR with D-AP5 (50 μM) decreased glutamate but did not change dopamine release. Blocking AMPA/kaintate receptor function with NBQX (5 μM) did not change glutamate release but did increase dopamine release. This work demonstrates the importance of simultaneous detection of related neurotransmitters while showing that glutamate and dopamine release have an inverse relationship that is partially mediated by binding to AMPA and Kainate receptors.

Graphical Abstract

1). Introduction

Dopamine (DA) and glutamate (Glu) affect many neurological processes in the mammalian brain such as reward systems, motor action, pain management, and addiction.^1–6 Both dopamine and glutamate act as neuromodulators; dopamine acts as an excitatory or inhibitory neuromodulator⁷ while glutamate is primarily an excitatory neuromodulator.⁸ Recent research has shown that dopamine and glutamate are co-released and that their co-release is important to the governance of neurochemical function (Fig. 1A).^5,6 In mice, the nucleus accumbens (NAc) region is noted for its significant populations of both dopaminergic and glutamatergic neurons, which project from Ventral Tegmental Area (VTA); some of these neurons cotransmit dopamine and glutamate.^1,6 Dopamine and glutamate release are modulated by receptors, namely glutamate receptors and D1 and D2 dopamine receptors, on NAc presynaptic medium spiny neurons.^2,4,6,9–12 Metabotropic and ionotropic glutamate receptors are differentiated by their overall structure, with metabotropic receptors lacking the ligand-gated ion channel of ionotropic receptors.^13,14 Despite structural differences, ionotropic receptors have been shown to possess metabotropic actions that modulate neurotransmitter release.^13,15–18 Often via these metabotropic actions, presynaptic ionotropic receptors still play important roles in neuromodulation and are found in the NAc.^{10–13,16,18,19} Thus, the NAc is an ideal region to simultaneously detect dopamine and glutamate to determine how they modulate each other.

Figure 1.

(A) Cartoon depiction of the key components of a NAc synapse relevant to this study. The presynaptic neuron depicted here co-releases dopamine and glutamate for simplification, but the release could also come from separate neurons. (B) Experimental workflow for pharmacological experiments. The first 3 stimulations (20 minutes) are the recorded in the absence of drug, followed by 4 stimulations in the presence of drug (40 minutes). An electrical stimulation is applied every 10 minutes following the initial stimulation at 0 mins.

Detecting real-time release of dopamine and glutamate simultaneously is challenging because few methods exist to monitor them with the same technique. Dopamine is electroactive and therefore it has been monitored extensively at carbon-fiber microelectrodes (CFMEs) using fast-scan cyclic voltammetry (FSCV). FSCV uses rapid voltage scanning to oxidize and reduce electroactive neurotransmitters and produce a current-voltage response cyclic voltammogram that is a fingerprint for identification.^20,21 Dopamine is an electroactive, easily-oxidized catecholamine, but glutamate is difficult to detect in standard potential ranges as it does not undergo redox at physiological conditions.²² To overcome this limitation and detect glutamate electrochemically, biosensors have been developed with glutamate enzymes attached that create an electroactive product when glutamate is detected.^22–27 However, these sensors are difficult to make and have a limited shelf life. A newer technique to measure real-time glutamate signaling is a genetically-encoded fluorescent probe called Glutamate-Sensitive Fluorescent Reporters (iGluSnFRs).^23,27 These fluorescent reporter proteins leverage a membrane-inserted cpGFP attached to a ligand-binding domain with specificity for glutamate, and so they are fluorescent only when glutamate transiently binds. Our lab has demonstrated that fluorescent reporters can be used simultaneously with FSCV to measure neurotransmitters in mouse brain slices, allowing dopamine and serotonin to be monitored in the synapse (with fluorescent GRAB sensors) and in the extracellular space (with FSCV).^28,29 Recently, we combined FSCV and iGluSnFR to demonstrate an inverse relationship between dopamine and glutamate release in the caudate-putamen region.³⁰ Thus, to further explore this relationship, this study will combine FSCV and iGluSnFR to detect dopamine and glutamate to explore how glutamate influences dopamine signaling.

The goal of this study is to understand the effect of glutamate on dopamine signaling by simultaneously detecting stimulated dopamine and glutamate release in the mouse nucleus accumbens (NAc). Previously, our laboratory identified an inverse relationship between glutamate and dopamine release; thus, we hypothesized that glutamate could be acting as a neuromodulator of dopamine. In this work, we test the effect of glutamate, glutamate transporter inhibitors, metabotropic receptor drugs, and ionotropic receptor antagonists to identify the effects on stimulated glutamate and dopamine release. Directly increasing glutamate by perfusing exogenous glutamate over the brain slice decreased dopamine, showing an inverse effect. Similarly, inhibiting excitatory amino acid transporters (EAATs) increased glutamate and decreased dopamine. Group 2 metabotropic glutamate receptors (mGluRs) did not independently affect dopamine release, so ionotropic receptors were explored instead. Ionotropic glutamate receptors did not independently affect collective glutamate and dopamine release, so they are not responsible for mediating the glutamatergic effects on dopamine. Thus, glutamate has an inhibitory effect on dopamine release, but that effect is not predominantly mediated through mGluR2, AMPA, or NMDA receptors.

2). Methods

2.1). Chemicals and Materials

Glutamate was purchased from Sigma-Aldrich (St. Louis, Missouri). Ingredients for standard αCSF buffer (125 mM NaCl, 2.5 mM KCl, 1.25 mM NaH₂PO₄, 25 mM NaHCO₃, 1 mM MgCl₂, 25 mM glucose, and 2 mM CaCl₂, pH 7.4) were also purchased from Sigma-Aldrich (St. Louis, Missouri) as well. D-AP5 ((2R)-amino-5-phosphonovaleric acid), DL-TBOA (DL-threo-β-Benzyloxyaspartate), NBQX (2,3-Dioxo-6-nitro-1,2,3,4-tetrahydrobenzo[f]quinoxaline-7-sulfonamide) Disodium Salt, LY 379268, and LY 341495 Disodium Salt were purchased from Tocris Bioscience (St. Louis, Missouri).

2.2). iGluSnFR Preparation and Expression

All procedures for animal experimentation were performed according to protocols approved by the Animal Care and Use Committee of the University of Virginia and in accordance with the US National Institutes of Health guidelines. Male C57BL/6J mice were bought from Jackson Labs. Animals were obtained at 4 weeks of age and held at the animal facility at the University of Virginia and family-housed in temperature-controlled animal rooms with 12-hr/12-hr light/dark cycles. Food and water were provided ad libitum. Mice were surgically injected with AAV-mediated iGluSnFR3 (see below) and then held at the vivarium for ~3-5 weeks before experimentation. Animals were roughly 8 weeks of age when experimental data was recorded.

Adeno-associated viral (AAV) vectors containing the genetically encoded glutamate-sensing fluorescent reporter iGluSnFR3 were purchased from Addgene (Addgene, Cat # 106174-AAV1). For expression of these constructs, animals were anesthetized via intraperitoneal injection of ketamine and xylazine at doses of 10 mg/kg and 2 mg/kg, respectively. Then, animals were placed within a stereotaxic frame, the skin retracted, and a hole was drilled above the desired injection site with a surgical drill. Viral solutions containing iGluSnFR3 were injected into the nucleus accumbens using a Hamilton 5 μL Microliter Neuros Syringe (Franklin, Massachusetts). Accurate perfusion of AAV iGluSnFR3 was achieved using a MICRO2T SMARTouch controller from World Precision Instruments (Sarasota, Florida). For imaging experiments, 150 nL of AAV iGluSnFR3 diluted with 150 nL phosphate buffer solution was injected into the nucleus accumbens (AP: 1.8 mm; ML: 1.00 mm; DV: −4.6) at a rate of 100 nL/min. The injection needle was left in the brain for ~10 minutes after the injection to ensure full virus diffusion. The syringe was then removed, and the skin was closed; the animals were given ketoprofen (10 mg/mL) as an analgesic. Animals were returned to the University of Virginia animal facility, monitored after surgery, and held for at least 3 weeks to allow for viral proliferation and optimal genetic expression.

2.3). Acute Brain Slice Preparation

To prepare acute nucleus accumbens slices, animals were anesthetized by isoflurane and decapitated. The brain was removed and immersed in cold (0-4 °C), oxygenated αCSF buffer as described above. 400 μm-thick coronal slices were cut from the brain using a Leica VT1000 S vibratome (Teaneck, New Jersey). Slices were equilibrated in oxygenated αCSF buffer at a temperature of 37.0±0.5 °C for ~0.5 hr prior to data acquisition. During data acquisition, the brain slices were submerged in an optically-accessible chamber, continuously perfused with oxygenated αCSF buffer with a solution exchange half-time of about 6 seconds; temperature of the bath was held at 34.0±0.5 °C. All drugs were bath-applied. Dopamine and glutamate release was induced using electrical stimulation. A bipolar stimulating electrode was implanted (Biphasic Stimulus Isolator, Microprobes) in the nucleus accumbens. Twelve biphasic stimulation pulses were applied, with currents of 300 μA and duration 4 ms, once every ten minutes (Fig. 1B).

2.4). Fluorescence Imaging using iGluSnFR Sensors

Brain slices expressing AAV-iGluSnFR3 (Addgene, Cat # 106174-AAV1) in the nucleus accumbens were imaged using wide-field epifluorescence imaging (Olympus) equipped with a Hamamatsu ORCA-Flash 4.0 camera. iGluSnFR3-expressing cells in the caudate were excited with a 488 nm wavelength light source (Lumencor Light Engine, Sola), and emission was recorded using a 530 nm filter. Acquired movies were analyzed using imageJ. Movies were background subtracted using imageJ by taking the average of three nonspecific regions where no significant change of iGluSnFR ΔF/F₀ was detected and then subtracting this average from the obtained iGluSnFR ΔF/F₀ signal from responsive cells.

2.5). Fast-scan Cyclic Voltammetry

Fast-scan cyclic voltammetry (FSCV) measurements were obtained using carbon-fiber microelectrodes (CFMEs) positioned in close proximity to iGluSnFR3 expressing neurons. T-650 Carbon fibers with a diameter of 7 μm (Cytec Engineering Materials Inc, Tempe, AZ), which were aspirated into a glass capillary (1.2 mm OD and 0.68 mm ID, A-M system, Sequim, WA) and pulled into electrodes with a PE-22 puller (Narishige International USA Inc, Amityville, NY). The carbon fiber was trimmed to approximately 100 μm in length from the pulled glass tip. To detect dopamine electrochemically, a triangular waveform was applied with a holding potential of −0.4 V and a switching potential of 1.3 V at 400 V/s using a WaveNeuro FSCV System (Pine Research Instrumentation, Durham, NC). For data collection and analysis, HDCV software (provided by R.M. Wightman, University of North Carolina) was used. A dopamine stock solution was prepared in 0.1 M HClO₄ and diluted to 500 nM with the abovementioned phosphate buffer for calibrations after the experiment.

2.6). Statistical Analysis

Statistical analyses were reported as mean ± s.e.m. Animals were randomly assigned into control or experimental groups. For glutamate experiments, each fluorescing cell analyzed was taken to be a single sample within a group and 3 cells were typically analyzed per slice. For dopamine experiments, each slice analyzed was a single sample within a group. Statistical significance was set to p<0.05. Data were analyzed using one-way and two-way ANOVAs as well as Wilcoxon Tests in GraphPad Prism software. The data for this study are available from the corresponding authors upon request.

3). Results

3.1). Glutamate decreases stimulated dopamine release

First, to understand the effect of glutamate on dopamine release, we measured stimulated dopamine release in the nucleus accumbens core after perfusing the slice with exogenous glutamate (100 μM). For all experiments, dopamine was stimulated 3 times every ten minutes before treatment as a control, and then glutamate (100 μM) was added to the perfusion buffer and 4 more stimulations (40 minutes) were performed with drug present (Fig. 1B).

Fig. 2A shows an example current vs time trace for an electrical stimulation before glutamate perfusion, and a trace 40 minutes after glutamate is introduced. Dopamine release is reduced after glutamate perfusion compared to pre-drug and the width of the peak is smaller. Fig. 2B shows the responses over time for control and glutamate trials; the exogenous glutamate was added after 3 pre-drug stimulations, which were averaged and then used to normalize the currents. Dopamine current significantly decreases while glutamate is perfused. Fig. 2C shows the normalized averages of dopamine current pre-drug (at 0, 10, and 20 minutes) and the average dopamine current decreased when glutamate was perfused (at 30, 40, 50, and 60 minutes). Therefore, glutamate inhibits dopamine release in the NAc. To more thoroughly understand this relationship, we then tested pharmacological agents while performing simultaneous detection of both dopamine and glutamate in the NAc.

Figure 2.

Dopamine current signals detected via FSCV in the presence of Glutamate. (A) Current vs. time graphs of representative observed currents before and after the bathing the slice in glutamate (100 μM). (B) Normalized dopamine current signals over time when exposed to glutamate. Signals were normalized to the pre-drug currents. After the addition of glutamate (shaded area) there is a significant decrease in normalized dopamine current (Two-way ANOVA, n=6, Time Factor: 0.0621, F (2.796, 27.96) = 2.794; Drug Factor: p=0.0161, F(1, 10)=8.354. The starred point represents significance between control and glutamate bath at time 60 (p= 0.0248) (C) Normalized average dopamine current recorded before and after the introduction of glutamate (100 μM) (Wilcoxon Test, n=6, p=0.0312).

3.2). Simultaneous Measurement of Glutamate and Dopamine in the Mouse Nucleus Accumbens

After identifying that dopamine release is inhibited by increased glutamate, we performed simultaneous codetection experiments. Changes in dopamine were detected using FSCV by collecting background-subtracted cyclic voltammograms and changes in glutamate release were detected using iGluSnFR using ΔF/F₀.

Fig. 3A shows traces of stimulated glutamate from iGluSnFR ΔF/F₀ signals in a slice (n=4 cells) above a characteristic fluorescence image of fluorescent neuronal cells. Stimulation was applied after 20 seconds of recording fluorescence, resulting in a noticeable increase in ΔF/F₀. Fig. 3B shows stimulated dopamine current vs. time trace above the characteristic color plot below it that proves the signal is dopamine. Stimulation occurs at 5 seconds during recording, resulting in an increase in dopaminergic current. The background current for background subtraction was taken at 4 seconds which is denoted by a vertical green line.

Figure 3.

Simultaneous detection of dopamine and glutamate using FSCV and fluorescence microscopy. (A) Glutamate ΔF/F₀ traces (4 cells) displayed above a characteristic fluorescence microscope image of iGluSnFR-expressing cells. The two large black objects protruding in the top and bottom of the image are the arms of the stimulating electrode. The relatively thin black line protruding from the right of the image is the working electrode. (B) Dopamine current vs. time curve displayed above a characteristic FSCV color plot of dopamine. Electrical stimulation occurs at 5 seconds. A background (green line) was taken roughly 1 second prior to stimulation for background subtraction. (C) Normalized overlay of both dopamine current (blue) and glutamate ΔF/F₀ (green) on the same timescale. Current and ΔF/F₀ are normalized to set their respective maximum values to 100%. (D) Graph of average stimulated dopamine current (blue) and average glutamate ΔF/F₀ (green) over 60 minutes (Two-Way ANOVA, n=6, Time Factor: p=0.0091, F (2.309, 53.10) = 4.799 Group Factor: p=0.4560, F(1, 22)=0.5758).

Fig. 3C shows a normalized overlay of both dopamine and glutamate traces collected simultaneously. This overlay shows both signals normalized to their maximum respective values as 100%, demonstrating that glutamate-derived ΔF/F₀ signal (green) falls much faster than dopamine current signal (blue). Glutamate is a synaptic signal while dopamine is measured extrasynaptically, so the time course is not expected to be the same, and glutamate clearance is known to be fast. Fig. 3D shows the stimulated dopamine and glutamate signals over time. Both signals are stable for at least 60 minutes when stimulations are repeated every 10 minutes. By establishing that both sensors were stable for extended periods of time, these experiments are the control for pharmacological experiments.

3.3). Glutamate reuptake inhibitor, DL-TBOA, increases glutamate release and decreases dopamine release

After finding that increasing glutamate to supraphysiological levels dramatically decreased detected dopamine release, we tested if increasing glutamate concentration by perturbing endogenous glutamate pathways would have a similar effect. To increase glutamate levels, we administered DL-TBOA, a competitive inhibitor to block the EAAT responsible for presynaptic glutamate reuptake and transport to local astrocytes (Fig. 4). The time course of the experiment is shown in Fig. 1B. The first 20 minutes (3 stimulations) were used to establish a baseline for both dopamine and glutamate. After 21 minutes, DL-TBOA was perfused for the next 40 minutes (Fig. 1B). ^31,32

Figure 4.

Influence of glutamate transporter inhibitor DL-TBOA on dopamine and glutamate signals over time. (A) Comparison between ΔF/F₀ recorded under control and DL-TBOA (20 μM) conditions. Once DL-TBOA is applied, ΔF/F₀ increases significantly (Two-way ANOVA, n=22, Time Factor: p<0.0001, F(1.809, 70.53)=14.10; Drug Factor: p=0.0002, F(1, 38)=16.59). The starred points represent significance between control and DL-TBOA at time 30 (p=0.0252), 40 (p=0.0010), 50 (p=0.0026), and 60 (p=0.0314). (B) Comparison between peak dopamine currents recorded under control and DL-TBOA conditions (Two-way ANOVA, n=6, Time Factor: p=0.0474, F(3.025, 30.25)=2.963; Drug Factor: p=0.0177, F(1, 10)=8.039). The starred point represents significance between control and DL-TBOA at time 60 (0.0148). (C) Paired-comparison graphs depicting the change between the averages of all time points for pre-drug vs drug (Wilcoxon Test, n=22, p<0.0001). (D) Paired-comparison graphs depicting the change of the averages of dopamine current for pre-drug vs drug (Wilcoxon Test, n=6, p = 0.0312). (E) Cartoon image of EAAT blockage via the introduction of DL-TBOA at a glutamate synapse.

Fig. 4A shows the normalized ΔF/F₀ signal for glutamate for control and with perfusion of the EAAT blocker DL-TBOA. Glutamate release (pink) increases for twenty minutes after DL-TBOA and remains significantly increased above control values (black). The dopamine time course is shown in Fig. 4B. Dopamine changes more slowly than glutamate, as the current begins to significantly decrease after twenty minutes of drug perfusion. Fig. 4C shows the average iGluSnFR ΔF/F₀ pre and post drug, with lines connecting the paired points. These paired-comparison graphs demonstrate that iGluSnFR ΔF/F₀ increases in the drug interval (p<0.0001). Fig. 4D shows paired comparison data of the average dopamine current at the two experimental conditions: control and drug. Dopamine current decreases significantly with DL-TBOA (p=0.0267). Thus, increasing the endogenous glutamate release decreases dopamine release. Next, we tested whether glutamate was acting through ionotropic glutamatergic receptors to regulate dopamine release.

3.4). Glutamate Acts Through AMPA and Kainate Receptors to Modulate Dopamine

Following previous studies, we investigated glutamatergic receptors linked to dopamine release. Group 2 and 3 metabotropic glutamate receptors (mGluRs) are found on both pre- and postsynaptic neurons. To study group 2 mGluRs, we selected the agonist LY 379268 (10 μM) and antagonist LY 341495 (40 μM); concentrations were selected based on previous literature and IC₅₀ values, respectively, to effectively block all potential targets.³³ Using either drug, we found no significant effect on dopamine release. The experimental design followed the same workflow as the DL-TBOA experiments: a 60-minute experiment where the first 20-minute period has αCSF buffer without drug, followed by a 40-minute period of αCSF buffer with drug (Fig. 1B). Fig. S1A shows the normalized current over time for stimulated dopamine release under control and both mGluR2 agonist and antagonist drug conditions, and there is no effect of drug. Fig. S1B shows no change in the average dopamine current for agonist LY 379268s. Fig S1C shows no change in average dopamine current for the antagonist LY 341495. Thus, there is no statistically significant change for the average dopamine current with either mGluR drug and mGluR2/3 receptors do not appear to be mediating the effects of glutamate on dopamine.

To investigate roles of glutamatergic receptors to regulate dopamine release, we next targeted the ionotropic glutamate receptors. AMPA and Kainate (KAR) receptors were blocked with NBQX and NMDARs with D-AP5 (Fig 5G–H). NBQX is an effective blocker of AMPA receptors with an IC₅₀ of 0.15 μM for AMPARs and 4.8 μM for Kainate receptors, and 5 μM was selected to ensure full blockage of AMPARs with partial blockage of Kainate receptors.³⁴ D-AP5 is the more active isomer of AP5, a competitive inhibitor of NMDARs. D-AP5 inhibits NMDAR activity at a wide range of concentrations, but 50 μM is commonly used for rodent studies.^35–37 The experiments to test the effects of blocking the ionotropic receptors followed the same workflow as the DL-TBOA experiments: a 60-minute experiment where the first 20-minute period has αCSF buffer without drug, followed by a 40-minute period of αCSF buffer with drug (Fig. 1B).

Figure 5.

Influence of ionotropic receptor antagonists D-AP5 (NMDAR antagonist) and NBQX (AMPAR/KAR antagonist) on dopamine and glutamate. (A) Comparison of stimulated ΔF/F₀ for glutamate recorded under control, D-AP5 (50 μM, n=19) and NBQX (5 μM, n=23) (Two-way ANOVA, Time Factor: p=0.0007, F(3.120, 174.7)=5.857; Drug Factor: p=0.0024, F(3.120, 174.7)=5.857. The starred points represent significance between control and D-AP5 at time 30 (p<0.0001), 40 (p<0.0001), 50 (p=0.0005), and 60 (p<0.0001); no points were significantly different for NBQX. (B) Comparison of stimulated dopamine currents recorded under control, D-AP5 (50 μM, n=6) and NBQX (5 μM, n=7) conditions (Two-way ANOVA, Time Factor: p=0.0562, F(2.704, 43.26)=2.804; Drug Factor: p=0.0015, F(2, 16)=10.01). The starred points represent significance between control and NBQX at time 30 (p=0.0227), 40 (p=0.0331), and 50 (p=0.0397); no points were significantly different for D-AP5. (C-D) Paired-comparison graphs depicting the change between the averages of glutamate ΔF/F0 at different experimental conditions for (C) D-AP5 (Wilcoxon Test, n=19, p<0.0001) and (D) NBQX (Wilcoxon Test, n=23, p=0.0172). (E-F) Paired-comparison graphs depicting the change between the averages of dopamine current at different experimental time points for (E) D-AP5 (Wilcoxon Test, n=6, p=0.5625) and (F) NBQX (Wilcoxon Test, n=7, p=0.0156). (G) Cartoon image of presynaptic NMDAR blockage via the introduction of D-AP5. (H) Cartoon image of presynaptic AMPAR blockage via the introduction of NBQX.

Fig. 5A shows the normalized ΔF/F₀ signal from iGluSnFR in response to D-AP5 (50 μM), and NBQX (5 μM) as well as control. AP5 lowered glutamate ΔF/F₀ signals by roughly half. However, stimulated glutamate release increased slightly but did not differ significantly from control conditions in the presence of NBQX; it instead only significantly increases due to a time-related response (p<0.0001). Fig. 5B shows the normalized current for stimulated dopamine release under control, D-AP5, and NBQX conditions. D-AP5 did not statistically change dopamine currents when compared to control. Conversely, NBQX increased dopamine currents when compared to the control.

Fig. 5C shows the average iGluSnFR glutamate ΔF/F₀ at the two experimental conditions: pre-drug and drug for D-AP5 experiments. AP5 decreases average glutamate ΔF/F₀. Fig. 5D shows the average glutamate ΔF/F₀ for NBQX increases when compared to pre-drug values. Fig. 5E shows the average dopamine current for D-AP5 experiments. There is no statistically significant change for the average dopamine current under AP5 when compared to the pre-drug baseline. Fig 5F shows the effect of NBQX on dopamine release using paired trials, demonstrating that NBQX exposure increases dopamine release. Fig. 5G and Fig. 5H are simplified cartoon depictions of D-AP5 and NBQX acting on a presynaptic NMDAR and AMPAR/KAR, respectively. Specifically, while AP5 decreased glutamate, dopamine is largely unaffected. Conversely, NBQX increases glutamate slightly over time and increases dopamine release. Together, these experiments demonstrate that, while NMDAR, KAR, and AMPAR may affect glutamate slightly, and that AMPAR and KAR inhibition increases dopamine release, they do not independently replicate the inverse effects observed with glutamate and dopamine.

4). Discussion

Understanding neurotransmitter release and regulation requires measurements of multiple neurotransmitters simultaneously. In this work, we multiplexed two techniques: FSCV for dopamine detection and a genetically-encoded fluorescent sensor (iGluSnFR3) for glutamate detection. Both techniques allow real-time tracking of neurotransmitter release events and they can be multiplexed to target multiple neurotransmitters despite differences in chemical structure and electroactivity. Our main finding is that increasing glutamate is followed by a decrease in dopamine release. This decrease in dopamine was observed both when exogenous glutamate (μM) was perfused over the brain slice and when DL-TBOA (20 μM), a potent blocker of glutamate uptake from the synapse, was administered. By measuring dopamine and glutamate simultaneously, we note the time-dependent nature of DL-TBOA action where the dopamine decrease is not immediate and only occurs when glutamate increases significantly. We then tested glutamatergic receptors to see if they are responsible for mediating the effects on dopamine release. Neither an agonist nor antagonist to group 2 mGluR function had no significant effect on dopamine release. Blocking the ionotropic receptors resulted in occasional small changes of glutamate or dopamine, but neither blockage of NMDAR, or AMPAR/KAR led to the same, inversely proportional relationship between glutamate and dopamine as directly increasing glutamate concentration did. Future studies could look at other pathways, such as GABAergic proteins, to further understand the opposite nature of glutamate and dopamine. Together, these results motivate the simultaneous detection of dopamine, glutamate, and other neurotransmitters, to identify how they can modulate each other.

4.1). Dual Detection of Dopamine and Glutamate

Through the use of iGluSnFR3 and FSCV, we recorded the stimulated release of both glutamate and dopamine. The light source used for glutamate detection can change the background charging current of the electrode via the photoelectric effect, but turning the light on 20 s before FSCV detection allows the background to stabilize before dopamine is detected.^30,38 These techniques are complementary when considering that FSCV primarily detects extrasynaptic release^20,30 whereas iGluSnFR3 and other genetically-encoded fluorescent sensors detect synaptic release events.^28,29 The glutamate signals are faster than the dopamine signals, likely because of the extrasynaptic nature of the dopamine being measured is slower to be cleared than the more tightly-regulated uptake of synaptic glutamate.³⁰ The genetically-encoded fluorescent protein iGluSnFR3 is expressed on neuronal membranes at and around the synapse and allows for specific detection of glutamate at the synapse.^23,27 This location differs from recently-published enzymatic biosensors for glutamate multiplexed with CFMEs, as the biosensor-coupled CFMEs only detect extrasynaptic neurotransmitters.²² Thus, in this work we report how synaptic glutamate release modulates extrasynaptic release of dopamine.

4.2). Increasing Extracellular Glutamate Decreases Dopamine Release

To understand potential neuromodulatory effects glutamate has on dopamine, we directly perfused glutamate over a mouse brain slice. Perfusing glutamate decreased dopamine release by nearly half. While glutamate is traditionally an excitatory neurotransmitter, our findings show that increasing glutamate decreases dopamine release, a surprising finding. Other studies have reported indirect and direct effects of glutamate on dopaminergic signaling.^4,6 For example, aberrations in glutaminergic signaling found in diseases such as schizophrenia lead to indirect changes in dopaminergic signaling.³⁹ As FSCV primarily detects extrasynaptic (volume) transmission, our results show that glutamate inhibits the extrasynaptic release of dopamine.

4.3). Increasing Glutamate Concentration Using a EAAT Blocker Decreases Dopamine Release

Next, we increased glutamate in the NAc by blocking the EAAT (EAAT2) responsible for glutamate reuptake. EAAT2 is located on presynaptic neurons and neighboring astrocytes, and it takes up glutamate and then recycles it via the glutamate-glutamine cycle.^40,41 Our hypothesis was that blocking EAAT2 would lead to glutamate accumulation around the synaptic cleft, increasing the signal at the genetically-encoded iGluSnFR sensors. As predicted, glutamate signals from iGluSnFR sensors increased dramatically in as little as 10 minutes of DL-TBOA perfusion. Stimulated glutamate release reached a maximum after 20 minutes of perfusion and then decreased slightly. This increase and then fall does not appear to be saturation of the sensors, themselves, considering that maximum at t = 40 mins is considerably larger than the stable values acquired for the rest of the experiment’s duration. DL-TBOA primarily inhibits EAAT2, but other EAATs, such as EAAT1, are also inhibited to a lesser extent. EAATs transport glutamate into the intracellular space at a relatively slow rate of ~30 glutamate/s,^41–43 so it is possible that this stabilization occurs after the recruitment of additional EAAT1 proteins to normalize glutamate reuptake. This notion is supported by recent studies that have attributed neuroplasticity to EAAT2 function and NMDAR signaling.^41,44 There is also the possibility that the slight decrease is due to less available glutamate release due to less glutamate being recycled through astrocytic EAAT2 uptake.

Dopamine release was measured simultaneously with glutamate, to provide a picture of how multiple neurotransmitters change concurrently. Although glutamate significantly changed in as little as 10 minutes from the onset of DL-TBOA, dopamine signals only significantly changed after glutamate reached its peak release, 30-40 minutes after the drug was perfused. Dopamine then decreased over time, and never reached a stable level. Thus, while dopamine changes in the opposite direction as glutamate with DL-TBOA, the time is different and the effect on dopamine is slower, suggesting that the effects of glutamate may be downstream signaling cascades and not direct effects of rapid glutamate receptor signaling. For example, glutamate may increase downstream GABA levels which might inhibit dopamine release in future stimulations.^45,46 Ionotropic and metabotropic GABA receptors (GABA_A and GABA_B, respectively) have both been shown to contribute to dopamine regulation, specifically its inhibition, in the rodent NAc.^47,48 The DL-TBOA data corroborate the data with glutamate perfusion which showed high levels of glutamate lead to decreased dopamine release.

The results of both experiments using glutamate and DL-TBOA demonstrate that increasing glutamate to supraphysiological levels decreases stimulated dopamine release. This finding is shown via two separate perfused drugs showing that the result is consistent, regardless of the method used to increase glutamate levels. Differences in the time course with DL-TBOA results highlight the importance of simultaneous measurement of both dopamine and glutamate. Our results demonstrate that DL-TBOA indirectly inhibits dopamine release via a stepwise process where first glutamate is increased and then dopamine is inhibited, which shows the advantage of simultaneous measurements.

4.4). Glutamate Receptors Are Not Independently Responsible for Glutamate’s Inhibition of Dopamine

To elucidate the role of glutamate receptors on dopamine release, we tested group 2 metabotropic receptors. Group 2 mGluRs have previously been shown to decrease dopamine release via agonist drug studies performed in vivo and in other brain regions.^12,33 Using the agonist LY 379268 and the antagonist LY 341495 in a pilot study, dopamine release was unaffected by either drug (Fig. S1). While some studies in intact animals have found a decrease by the agonist, mGluRs may have different impacts on dopamine release dependent on the brain region of interest³³ or when complete circuits are not present, as is true in brain slice experiments here.

The next question we addressed was whether ionotropic glutamate receptors mediate that regulation of dopamine release. Past studies have shown that glutamatergic ionotropic receptors are localized on dopaminergic presynaptic neurons, leading to neuromodulatory effects on dopamine release.^19,49 These effects occur across various brain regions such as the prefrontal cortex,^9,50 hippocampus,^49,49 and NAc.^9,51 Additionally, there is evidence that glutamatergic ionotropic receptors can modulate neurotransmitter release via metabotropic mechanisms^15,16,52,53 or by working collectively with dopaminergic D1 receptors.⁵⁴ Our primary targets were NMDARs and AMPARs, which were blocked using the competitive antagonists D-AP5 and NBQX, respectively. KARs were also blocked, to a lesser extent, by NBQX.

We first tested NMDAR’s potential effects on dopamine release and found that there was no effect on dopamine when blocking NMDARs using D-AP5, a structural homolog of glutamate. D-AP5 is a competitive inhibitor of NMDARs, by binding the ligand-binding domain glutamate typically occupies on NMDAR.⁵⁵ Conveniently, D-AP5 is selective and does not have binding targets beyond NMDAR, so its effects should be limited to the inhibition of NMDAR.^56,57

Glutamate signals decreased as a result of D-AP5 perfusion. While postsynaptic NMDARs are not known to directly regulate glutamate release, novel research on presynaptic NMDARs has suggested that presynaptic NMDARs could regulate evoked release of glutamate.^10,11,58 Thus, blocking presynaptic NMDARs with D-AP5 could limit the evoked release of glutamate, which is demonstrated by our results. Furthermore, D-AP5 has, however, been shown to decrease overall excitatory neuronal function, which may have some effect on stimulated glutamate levels in the synapse.^59–61

Interestingly, dopamine was unaffected when D-AP5 was introduced, instead of being the opposite of glutamate as it was in previous experiment. The dopamine results provide an interesting contrast to other works that observed increased dopamine release in the NAc when selectively blocking NMDARs in the PFC.^9,62 There are also studies that have explicitly shown that NMDARs are not responsible for glutamatergic control of dopamine release within the nucleus accumbens.^51,63 Thus, our results follows previous literature demonstrating that NMDARs in the nucleus accumbens are not responsible for regulation of dopamine release.^51,58,63 This result shows that the change in magnitude of dopaminergic transmission is not always inversely proportional to glutamate transmission.

In the last two decades, research has shifted towards the hypothesis that AMPARs can regulate neurotransmitters beyond glutamate and GABA,⁶³ including the catecholamines noradrenaline and dopamine.^49,51,63,64 Furthermore, AMPAR effects on dopamine and glutamate have been implicated in drug-seeking and reward behaviors mediated by the NAc.^1,65 Our results are congruous with these findings as dopamine peak current is increased under NBQX compared to control (Fig. 5B); the peak current also increased between paired pre-drug and drug points (Fig. 5F). Additionally, glutamate increases in the paired comparison (Fig. 5D), but the increase is not significantly different when compared to control data (Fig. 5A). Thus, there is little change in glutamate that would be having a direct effect on dopamine signaling. Previously, AMPARs have been connected to D1 and D2 receptor activity,^54,66 which may be altered due to the blockage of AMPARs in the NAc, resulting in a minimally-increased dopamine current without seeing a significant change in glutamate. As dopamine transmission is increased by D2 antagonists,⁶⁷ and AMPARs and D2 receptors interact,^54,66 it follows that disrupting AMPARs would increase dopamine transmission, as our data demonstrates. AMPARs have also been shown to act through metabotropic mechanisms, activating G proteins and regulating neurotransmission.^52,53 Glutamate kainate receptors (KARs) are also blocked by NBQX, albeit to a lesser extent than AMPARs. Blocking KARs is protective for dopaminergic neurons and KARs may also play a role in the slight increased dopamine current observed when NBQX was perfused.⁶⁸ Furthermore, like AMPARs, KARs have been shown to act via metabotropic mechanisms that influence neurotransmitter release.^15–17 While effect of NBQX may be due to both AMPARs and KARs, the main finding is that there is little effect of NBQX on glutamate or dopamine release and it is not responsible for mediating the inverse effects reported with glutamate perfusion.⁶³

5). Conclusions

This study multiplexed FSCV using a CFME sensor and fluorescence microscopy using a GluSnFR sensor to simultaneously detect dopamine and glutamate in mouse NAc brain slices. The simultaneous detection of dopamine and glutamate allowed for the analysis of extracellular glutamate’s effects on dopamine and the effects of glutamate receptors on dopamine in the mouse NAc. This study demonstrated that an increase in glutamate decreased dopamine current, and that increasing glutamate by inhibiting glutamate transporters with DL-TBOA also decreased dopamine. NMDARs are not a major regulator of dopamine release. Blocking presynaptic NMDARs when using D-AP5 decreased dopamine but did not change dopamine. AMPARs have a small effect increasing glutamate and a small increase in DA, likely due to dopaminergic receptors interacting with AMPARs. Thus, dopamine and glutamate are not always inversely proportional. These results reinforce previous studies exploring reward seeking behaviors mediated by the nucleus accumbens through AMPARs and other receptors.^4,65,69,70 Taken together, these results demonstrate that increasing glutamate in the mouse nucleus accumbens core inhibits dopamine release, and the mode of action is partially controlled through Kainate and AMPA receptors.

Highlights.

• Simultaneous dual detection of glutamate and dopamine in the nucleus accumbens

• Glutamatergic inhibitory neuromodulation of dopamine

• Synaptic neuromodulation of extrasynaptic dopamine release