Glutamatergic Neurons in the Caudal Zona Incerta Regulate Parkinsonian Motor Symptoms in Mice

Abstract

Parkinson’s disease (PD) is the second most common and fastest-growing neurodegenerative disorder. In recent years, it has been recognized that neurotransmitters other than dopamine and neuronal systems outside the basal ganglia are also related to PD pathogenesis. However, little is known about whether and how the caudal zona incerta (ZIc) regulates parkinsonian motor symptoms. Here, we showed that specific glutamatergic but not GABAergic ZIc^VgluT2 neurons regulated these symptoms. ZIc^VgluT2 neuronal activation induced time-locked parkinsonian motor symptoms. In mouse models of PD, the ZIc^VgluT2 neurons were hyperactive and inhibition of their activity ameliorated the motor deficits. ZIc^VgluT2 neurons monosynaptically projected to the substantia nigra pars reticulata. Incerta-nigral circuit activation induced parkinsonian motor symptoms. Together, our findings provide a direct link between the ZIc, its glutamatergic neurons, and parkinsonian motor symptoms for the first time, help to better understand the mechanisms of PD, and supply a new important potential therapeutic target for PD.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12264-021-00775-9.

Introduction

Neurological disorders are now the leading source of disability globally. PD, a neurodegenerative disease with prominent death of dopaminergic neurons in the substantia nigra pars compacta (SNpc), is the fastest growing of these disorders [1, 2]. The resultant dopamine deficiency within the basal ganglia (BG) leads to a movement disorder characterized by the classical parkinsonian motor symptoms, bradykinesia/akinesia, muscular rigidity, rest tremor, and postural and gait impairment [3, 4]. Since the introduction of dopaminergic therapy, levodopa has provided the greatest symptomatic benefit, but its long-term use is associated with motor complications and reduced efficacy. Then, deep brain stimulation (DBS) became another important surgical option [3]. DBS is the therapeutic use of electrical stimulation of specific brain regions via an implanted electrode. Regions within the BG such as the subthalamic nucleus (STN) and the internal segment of the globus pallidus are common DBS target sites for treating the motor symptoms of PD [5]. Because these therapies are still limited and the underlying causes of PD are complicated, it is crystal clear today that so many advances have been insufficient and much remains to be better understood [4]. Therefore, further study of new regulatory circuits, nuclei, neurons, and neurotransmitters of parkinsonian motor symptoms will not only deepen the understanding of the pathogenesis of PD, but also provide support for the discovery of new therapeutic strategies and methods.

In recent years, it has been recognized that the pathogenesis of PD is not only related to the prominent death of dopaminergic neurons in the SNpc and the resulting lack of dopamine. Neurotransmitters other than dopamine and neuronal systems outside the BG, such as degeneration of cholinergic neurons in the pedunculopontine nucleus, dysfunction of the nucleus basalis of Meynert, noradrenaline dysfunction in the locus coeruleus, and changes in cerebellar circuits have also been associated with the pathogenesis of PD [4, 6]. The zona incerta (ZI) is a highly heterogeneous diencephalic region which has great cellular and neurochemical diversity and is involved in diverse functions [7]. The ZI can be divided into four major sectors on the basis of cytoarchitecture: rostral (ZIr), dorsal (ZId), ventral (ZIv), and caudal (ZIc) sectors [7]. Early pharmacological studies have suggested involvement of the ZI in various functions such as visceral activities, arousal, attention, and locomotion [7], yet the specific roles of different ZI subdomains and subpopulations need to be further clarified. Interestingly, more recent findings have reported that ZI subdomains, including ZIr, ZId and ZIv, via different GABAergic neuronal subpopulations, regulate sleep [8], fear [9, 10], pain [11], binge-like eating [12], defensive behaviors [13, 14], predatory hunting [15, 16], and neural development [17], but the role of the ZIc remains uncharacterized. We noted that the ZI is a largely inhibitory nucleus composed of heterogeneous groups of cells [7]; the ZIc has a number of histological and gene-expression characteristics that distinguish it from the other ZI subdivisions. The expression of the GABA gene GAD67 in the ZIc is very low while the remainder of the ZI is packed with GAD67-expressing neurons [18]. Furthermore, according to the Allen Mouse Brain Atlas (http://www.brain-map.org), the expression of Slc17a6 [the gene for vesicular glutamate transporter 2 (VgluT2), a marker for glutamatergic excitatory neurons] is specifically expressed in the ZIc, being almost absent in the rest of the ZI. Thus, we postulated that the glutamatergic neurons in the ZIc may have a certain unique function. In addition, we noted that Plaha et al. reported that high-frequency DBS of the ZIc improves the contralateral motor scores in patients with medically refractory PD [19], suggesting a possible involvement of the ZIc in parkinsonian motor symptoms. However, little is known about how the ZIc regulates these symptoms.

Therefore, we set out to determine whether ZIc and its glutamatergic neurons regulate parkinsonian motor symptoms by optogenetically activating and inhibiting these neurons. Functional and circuit level connections of ZIc glutamatergic neurons were also explored using a wide range of methods, including viral tracing, slice electrophysiology, and optogenetics combined with fiber photometry.

Materials and Methods

Animal Housing and Ethics Permits

VgluT2-ires-Cre and GAD2-ires-Cre transgenic mice were housed with food and water available ad libitum in a temperature-controlled room under a 12-h light/dark cycle (lights on at 07:00). All mice were group-housed until surgery, after which they were singly-housed. Polymerase chain reaction analyses were run to confirm the genotype of the VgluT2-ires-Cre- and GAD2-ires-Cre-positive mice from ear biopsies. All experiments were conducted in accordance with the guidelines for the care and use of laboratory animals of Zhejiang University (ZJU) and were approved by the Animal Advisory Committee at ZJU. All surgery was performed under i.p. 1% pentobarbital sodium or isoflurane anesthesia, and every effort was made to minimize suffering.

Stereotaxic Viral Injection

Mice were deeply anesthetized with pentobarbital sodium (1%, wt/vol) and placed on a stereotaxic frame (RWD, 68030, 68025, Shenzhen, China). Body temperature was kept stable throughout the procedure with a heating pad. A scalp incision was made with eye scissors. The skull was exposed and perforated with a stereotaxic drill over the target region. Injections were performed with a 10 μL syringe (Hamilton, Nevada, USA) connected to a glass micropipette with a 10–15 μm diameter tip. A syringe pump (78–8130, KD Scientific, USA) was used to inject the viruses with speed and volume control. We injected 60–90 nL of virus into the ZIc (AP, –3.05 mm; ML, ±1.72 mm; DV, –3.96 mm relative to bregma) or substantia nigra pars reticulata (SNpr; AP, –3.05 mm; ML, ±1.72 mm; DV, –4.55 mm relative to bregma) at 30 nL/min. The micropipette was left in place for an additional 10 min to allow diffusion of the virus and then slowly withdrawn. After injection, mice were allowed 3–4 weeks for recovery.

For fiber photometry recordings, AAV2/9-DIO-GCaMP6m (titer: 4.8 × 10¹² genome copies/mL, from Taitool Bioscience, Shanghai, China) was injected stereotaxically into the ZIc of VgluT2-ires-Cre mice.

For simultaneous fiber photometry recordings and optogenetic activation, AAV2/9-hSyn-FLEX-ChrimsonR-mCherry-WPRE-pA (5.13 × 10¹² genome copies/mL, Taitool Bioscience) was injected stereotaxically into the ZIc of VgluT2-ires-Cre mice. AAV2/9-hSyn-GCaMP6m-WPRE-pA (7.23 × 10¹² genome copies/mL, Taitool Bioscience) was injected stereotaxically into the SNpr of VgluT2-ires-Cre mice.

For selective optical activation or inhibition of ZIc^VgluT2 neurons, AAV2/8-hSyn-DIO-hChR2(H134R)-mCherry-WPRE-pA (4.03 × 10¹² genome copies/mL), AAV2/8-hSyn-DIO-WPRE-pA (4.43 × 10¹² genome copies/mL), or AAV2/5-DIO-NpHR-mCherry (3.53 × 10¹² genome copies/mL) viruses (Taitool Bioscience) were bilaterally injected into the ZIc of VgluT2-Cre mice.

For optical activation of SNpr neurons, AAV2/9-hSyn-hChR2 (H134R)-mCherry-WPRE-pA (2.38 × 10¹² genome copies/mL) and AAV2/9-hSyn-mCherry-WPRE-pA (5.1 × 10¹² genome copies/ mL (Taitool Bioscience) were bilaterally injected into the SNpr of VgluT2-Cre mice.

For output tracing, AAV2/9-hSyn-DIO-mGFP (4.87 × 10¹² genome copies/mL, Taitool Bioscience) was unilaterally injected into the ZIc of VgluT2-ires-cre mice.

6-OHDA Model

To make the striatal 6-OHDA injections, mice were deeply anesthetized as previously described and 6-OHDA (5 mg/mL; Sigma-Aldrich, USA) or saline (as control) was unilaterally or bilaterally injected into the dorsomedial striatum (AP, +0.5 mm; ML, ±1.50 mm; DV, –3.00 mm relative to bregma; 1 μL per side) [20, 21]. Animals were allowed to recover for five days before post-lesion behavioral testing commenced.

Cannula Implantation and In Vivo Optogenetic Manipulation

For somatic stimulation of ZIc^VgluT2 neurons, after 3 weeks of ChR2, NpHR, or mCherry virus expression, optical fibers (200 μm O.D., 0.22 NA; Inper Ltd, Hangzhou, China) were bilaterally implanted into the ZIc (200–300 μm above the viral injection coordinates). To allow for optogenetic manipulation of ZIc–SNpr circuit axon terminals, optical fibers were placed 200–300 μm above the SNpr viral injection coordinates. All fibers were secured to the skull with bone screws and dental cement. Mice were allowed to recover for 5–7 days after implantation and were habituated for 15 min after connection to a laser source, after which behavioral tests were performed. Lasers at wavelengths of 473 nm (blue) or 589 nm (yellow) were applied and controlled with an intelligent optogenetic system (Aurora-200, Inper Ltd) at 5 mW for blue illumination (473 nm, 30 Hz, 5 ms pulse width) and 10 mW for yellow illumination (589 nm, continuous mode) in VgluT2-ires-cre mice expressing ChR2, NpHR, and mCherry.

In Vivo Fiber Photometric Calcium Recordings

A mono-fiber optic patch cord (MFO-1x2-F-W1.25-200-0.37-100, Inper Ltd) connected to the fiber photometry system (Inper Ltd) was attached to the implanted fiber optic cannula using a ceramic sleeve with black heat-shrink tubing. To record fluorescence signals from GCaMP6m, light from a 470-nm LED was bandpass filtered, collimated, reflected by dichroic mirrors, and focused by a 20× objective. LED light at 470 nm was delivered at a power of 40 μW at the tip of the fiber optic cannula. Emitted fluorescence from GCaMP6m was bandpass filtered and focused on the sensor of a CMOS camera. The end of the fiber was imaged at a frame rate of 60 fps, with Inper Studio, and the mean value of an ROI of the end-face of the fiber was calculated using Inper Analysis software. To serve as an isosbestic control channel, 405-nm LED light was delivered alternately with 470-nm LED light. Recorded Ca²⁺ signals were aligned with behavior videos using Inper Analysis software. The heatmap and averaged Ca²⁺ traces were plotted using a self-developed MatLab program.

The fiber photometry coupled with the optogenetics system (ThinkerTech Nanjing Bioscience Inc.) was as follows. Briefly, the excitation light of the blue (470 nm) and yellow (593.5 nm) light sources were merged into one beam through a dichroic mirror. Then the beam was coupled into one multimode fiber by an object mirror and then transmitted to the specified brain section to excite the marker green fluorescent protein (such as GCaMP6) and corresponding channelrhodopsin (such as ChrimsonR). The emitted green light was collected by a multimode fiber and transformed into an electrical signal by weak signal detector, which reflected the activity information. Heatmaps and averaged Ca²⁺ traces were plotted using a self-developed MatLab program.

Behavioral Testing

To minimize stress, mice were handled extensively and placed in the test room to become habituated for at least 3 days before all behavioral tests. Room temperature and humidity remained stable for all experiments, which were carried out in daytime.

After ~4 weeks of recovery from viral injection, the mice were analyzed in validated behavioral paradigms. The duration and conditions of stimulation were specified for each test. At the end of the experiment, all animals were perfused. Only data from animals with correct optical fiber implantation sites and virus expression were included in the analysis.

Rotation and Rearing Tests

Mice were placed in a transparent resin cylinder (diameter 16 cm, height 25 cm) and video-monitored from above. Rearing times and net ipsilateral or contralateral rotations were scored blindly. Rotations were defined as each 360° rotation that contained no turn of >90° in the opposite direction. Rearing was defined as the body raised and both forelimbs off the ground. The test consisted of 3 consecutive 1-min epochs ‘Pre’, ‘Light’, and ‘Post’ periods (Pre–Light–Post) for each mouse. The ‘Pre’ period was 1 min immediately preceding laser illumination; the ‘Light’ period was 1 min with the laser on; and the ‘Post’ period was 1 min immediately after laser illumination. For mice unilaterally injected with 6-OHDA, the test consisted of 3 consecutive 2-min epochs in the Pre–Light–Post periods for each mouse. This test was based on open field observations showing that mice turned towards the side of the lesion when placed in a novel environment [22, 23].

Open Field Test (OFT)

The OFT was used to evaluate locomotor activity. Mice were placed in the center of a plastic arena (40 × 40 × 40 cm³) and allowed to move freely for 3 min while they were videotaped individually. The center was defined as the central 25% of the arena. The locomotor activity in the open field was video-recorded and analyzed with automatic behavioral tracking software (ANY-maze, Stoelting Co., USA). The open field chamber was cleaned with 70% alcohol between animals. The test time was divided into three consecutive 1-min epochs in the Pre–Light–Post periods. Total distance traveled (cm), movement speed (cm/s), immobility time (%), and time spent in the center area (%) were analyzed.

Pole Test

The pole test was applied as described [24] previously, with minor modifications. Each mouse was placed head-upward on the top of a vertical rough-surfaced pole (diameter 8 mm; height 50 cm) and the time until it descended to the floor (total landing time) was recorded with a maximum duration of 120 s. Even if the mouse descended part-way and fell the rest of the way, the behavior was scored when it reached the floor. When the mouse was not able to turn downward and instead dropped from the pole, total landing time was taken as 120 s (default value) because of the maximal severity. The test consisted of three consecutive epochs (Pre–Light–Post) for each mouse.

Rotarod Test

A rotarod assay was applied to investigate the motor coordination of mice using an accelerating rotating rod (LE8205, Panlab Harvard Apparatus, Barcelona, Spain). Mice were trained twice per day over two consecutive days separately at 11 rpm on day 1 rpm and 22 rpm on day 2. Each training session lasted for 300 s and the inter-training interval was >1 h. On the test day, the rod was programmed to accelerate from 4 rpm to 40 rpm over 120 s. Mice were individually placed in a neutral cage to recover from fiber connection for 15 min before starting the experiment. The time mice stayed on the rod after it started to accelerate was recorded. The test consisted of three consecutive epochs in the Pre–Light–Post periods for each mouse.

Histology and Imaging

Following behavioral tests, mice were deeply anesthetized and transcardially perfused with 0.9% saline followed by 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS, pH 7.4). Brains were carefully removed and post-fixed in 4% PFA for an additional 4–6 h. They were then transferred to 30% sucrose in PBS until they sank to the bottom of the container and were then cut into 40-μm coronal sections on a freezing microtome (CM3050 S, Leica, Germany). The sections were stored at −20°C in cryoprotectant solution containing 30% glycerol (v/v), 20% ethylene glycol (v/v), and PBS until staining.

For immunofluorescent staining, free-floating sections were washed three times with PBS (5 min each) and incubated with blocking buffer containing 5% goat serum and 3% bovine serum albumin (BSA) dissolved in 0.5% PBST (0.5% Triton X-100 in PBS) for 1 h at room temperature. The sections were then incubated with primary antibodies diluted in blocking buffer overnight at 4°C for 12 h (rabbit anti-tyrosine hydroxylase 1:100, Cell Signaling Technology, USA and rabbit anti-DsRed 1:500, Clontech, USA). After incubation, the sections were rinsed four times with PBS (15 min each) and incubated with a fluorescent dye-conjugated secondary antibody (1:1000 Alexa Fluor 488 goat anti-rabbit, Abcam, UK; 1:1000 Alexa Fluor 568 donkey anti-rabbit, Invitrogen, USA) for 2 h at room temperature. Following four washes with PBS (15 min each), the sections were incubated with DAPI (1:1000, Invitrogen, USA) for 5 min, then washed and mounted under coverslips with Fluoromount aqueous mounting medium (Sigma-Aldrich, USA).

The expression levels of virus-infected cell bodies and axons were outlined on corresponding sections in the Mouse Brain in Stereotaxic Coordinates. Whole slides were imaged with a 10× objective on an Olympus VS120 (Japan) virtual slide microscope system or with a 20× objective on a NikonA1R (Japan) laser scanning confocal microscope.

Electrophysiology

Acute slices were prepared as follows. Briefly, mice were deeply anesthetized with isoflurane and quickly decapitated. The brains were harvested and submerged in ice-cold oxygenated (95% O₂ and 5% CO₂) cutting artificial cerebrospinal fluid (aCSF) containing (in mmol/L): 92 N-methyl-d-glucamine (NMDG), 2.5 KCl, 1.25 NaH₂PO₄, 30 NaHCO₃, 20 HEPES, 25 glucose, 2 thiourea, 5 Na-ascorbate, 3 Na-pyruvate, 0.5 CaCl₂·2H₂O, and 10 MgSO₄·7H₂O for 1–2 min. Then, 300-μm coronal slices containing the ZIc and SNpr were prepared with a vibratome (Leica, VT1200s, Germany) in NMDG-aCSF. The slices were allowed to recover for 13 min at 33°C in the oxygenated cutting solution. After that, the slices were transferred into the incubating aCSF (in mmol/L): 92 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 30 NaHCO₃, 20 HEPES, 25 glucose, 2 thiourea, 5 Na-ascorbate, 3 Na-pyruvate, 2 CaCl₂·2H₂O, and 2 MgSO₄·7H₂O and maintained at room temperature for at least 1 h.

For recordings, slices were transferred to a recording chamber and the temperature was controlled by a heat controller (Warner Instruments, Hamden, CT) and 26°C oxygenated artificial CSF (in mmol/L: 124 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 24 NaHCO₃, 12.5 glucose, 5 HEPES, 2 CaCl₂·2H₂O, and 2 MgSO₄·7H₂O) was continuously perfused onto the slice. Whole-cell patch-clamp recordings were made with a MultiClamp 700B amplifier (2 kHz low-pass filter, 10 kHz digitization; Molecular Devices, USA) and a 1440A interface (Molecular Devices) with pClamp 11.1 software (Molecular Devices). Fluorescent cells were visualized under an Olympus BX51WI microscope (Japan) equipped with a 40× water-immersion lens and illuminated with a mercury lamp. Whole-cell patch-clamp recordings were collected 3 min after obtaining a stable whole-cell configuration.

To test the effectiveness of the ChR2 and NpHR viruses, current-clamp mode was used to record the membrane potential of ZIc^VgluT2 virus-expressing neurons. The internal solution consisted of (in mmol/L): 130 K-gluconate, 4 KCl, 10 HEPES, 0.3 EGTA, 10 phosphocreatine-Na2, 4 Mg-ATP, 0.3 Na2-GTP (pH 7.3). Voltage-clamp mode (holding potential, –70 mV) was used to record the photocurrent of ZIc^VgluT2 virus-expressing neurons.

To evoke synaptic transmission using ChR2, photostimulation (473 nm, ~6 mW, 5-ms pulses) was delivered by a LED light source, and postsynaptic recordings in voltage-clamp mode (holding potential, −70 mV) were made in the SNpr areas of ChR2 terminal infection using electrodes (3−5 MΩ) filled with a cesium-based internal solution (135 cesium methanesulfonate, 5 TEA-Cl, 5 CsCl, 20 HEPES, 0.4 EGTA, 2.5 Mg-ATP, 0.25 Na2-GTP, and 1 QX-314; pH 7.25; 290 mOsm). Then, tetrodotoxin (TTX; Tocris, UK) was bath-applied to block Na⁺ channels. Both TTX (1 μmol/L) and 4-aminopyridine (4AP, 100 μmol/L; Sigma-Aldrich, USA) were added into the external solution to verify the monosynaptic responses, and 6,7-dinitroquinoxaline-2,3-dione (DNQX, 10 μmol/L; Sigma-Aldrich, USA) and AP-V (50 μmol/L; Sigma-Aldrich, USA) were added to block glutamatergic currents.

Statistical Analysis

All data are presented as the mean ± SEM. GraphPad Prism v. 6.0 was used for all statistical analyses (GraphPad Software). All statistical data can be found in the figure legends with corresponding sample sizes. The following statistical tests were used for the behavioral data analyses: two-tailed paired or unpaired t-tests were used as indicated in the figure legends. Normal distribution was determined by the D’Agostino-Pearson, Shapiro-Wilk, and Kolmogorov-Smirnov normality tests. Statistical significance was accepted at the level of P < 0.05 (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Results

ZIc Glutamatergic Neuronal Activation Induces Parkinsonian Motor Symptoms

To test the role of ZIc neurons in motor activity, we injected AAV-hSyn-DIO-ChR2-mCherry and control virus into the ZIc of VgluT2-ires-Cre and GAD2-ires-Cre mice. We set the stimulation frequency to 30 Hz based on a clinical report that the mean firing rate of ZI neurons is 29.5 Hz in PD patients [25]. The neurons in the ZIc expressing VgluT2⁺ ChR2 readily followed 5-ms light pulses delivered at 30 Hz (Fig. 1A–C). Three weeks after viral injection, we bilaterally implanted optic fibers in the ZIc of these mice (Fig. 1D, E). After allowing recovery from implantation, we performed behavioral tests 7 days later. Unexpectedly, unilateral photostimulation (473 nm, 30 Hz, 5 ms) of ZIc^VgluT2 neurons elicited robust light-locked ipsiversive rotations and significantly impaired rearing (Fig. 1F, G). Thus, unilateral activation of ZIc glutamatergic neurons mimicked the rotational behavior induced by unilateral dopamine depletion [26]. Bilateral photostimulation of ZIc^VgluT2 neurons induced bradykinesia (significant decreases in movement speed and total distance moved) (Fig. 1H, I) and akinesia (a sharp increase in immobility) (Fig. 1J). ZIc^VgluT2 neuronal activation did not affect center zone exploration time (Fig. 1K), indicating that anxiety-like behavior was unaffected. To further analyze the role of ZIc neurons in bradykinesia and motor coordination, we used the pole and rotarod tests. Bilateral photostimulation of ZIc^VgluT2 neurons induced a significant increase in pole descending time (Fig. 1L) and a marked decrease in time staying on the accelerating rotarod (Fig. 1M). No rotational bias or motor impairment was detected with stimulation of ZIc GABAergic neurons in GAD2-ires-Cre mice (Fig. S1). These data establish a causal role of ZIc glutamatergic neurons in inducing bradykinesia and akinesia, and disrupting motor coordination, mimicking the motor deficits induced by bilateral dopamine depletion [27]. Together, these results indicate that ZIc glutamatergic neuronal activation induces parkinsonian motor symptoms.

Fig. 1.

Optogenetic activation of ZIc^VgluT2 neurons induces parkinsonian motor symptoms. A Schematic of light stimulation of and patch-clamp recordings from glutamatergic ZIc neurons transfected with AAV-DIO-ChR2-mCherry in brain slices from VgluT2-ires-Cre mice. B Example patch-clamp recordings showing light stimulation induces time-locked action potential firing in ChR2-expressing VgluT2 neurons in the ZIc [blue bars, application of light stimuli (473 nm, 5 ms, ~6 mW)] at the indicated frequencies. C Mean evoked spike fidelity at different stimulation frequencies (n = 6 ZIc^VgluT2 neurons from 3 VgluT2-ires-Cre mice; spike fidelity of 100% means that each blue light stimulus generates an action potential). D Schematic showing viral injection and placement of the optical fiber. E Image showing optical fiber position and ChR2 expression in ZIc^VgluT2 neurons of a VgluT2-ires-Cre mouse (scale bar, 200 μm). F, G Number of net ipsilateral rotations (net ipsilateral rotations count = number of ipsilateral rotations – number of contralateral rotations) (F) and rearings (G) in ChR2 and control mice during Pre–Light–Post trials. H–K Performance of ChR2 and control mice in the OFT: movement speed (H), distance moved (I), time immobile (J), and time spent exploring the center zone (K) during Pre–Light–Post trials. L Total time spent descending from top of rod facing upward in ChR2 and control mice during Pre–Light–Post trials. M Latency to fall from the accelerating rotating rod in ChR2 and control mice during Pre–Light–Post trials. For all behavioral tests, n = 11 mice for the ChR2 group and n = 9 for the control group. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, n.s., no statistically-significant difference, paired t-test.

ZIc Glutamatergic Neurons are Hyperactive in Dopamine-Depleted (DD) Mice

We speculated that ZIc glutamatergic neuronal activity may be activated in the DD state. To measure the activity of ZIc^VgluT2 neurons in a 6-OHDA-induced PD mouse model, we infused a Cre-dependent adeno-associated virus (AAV) vector expressing a Ca²⁺-sensor protein (GCaMP6m) in the ZIc^VgluT2 neurons and measured the Ca²⁺ responses when mice were immobile 7 days after unilateral 6-OHDA injection into the dorsomedial striatum (Fig. 2A–C). Increased ZIc^VgluT2 neuronal activity (as indicated by significantly enhanced spontaneous Ca²⁺ responses: increased number of Ca²⁺ responses and the area under △F/F curves) was recorded by in vivo fiber photometry on the lesioned side compared with the non-lesioned side (Fig. 2D–I). These results indicated that ZIc glutamatergic neurons are activated in 6-OHDA PD model mice. To measure the activity of ZIc^VgluT2 neurons in a more severe PD model, we repeated the above procedures, but with bilateral 6-OHDA injection into the dorsomedial striatum (Fig. 2J). Increased ZIc^VgluT2 neuronal activity (as indicated by significantly enhanced spontaneous Ca²⁺ responses - increased number of Ca²⁺ responses and the area under △F/F curves) was recorded by in vivo fiber photometry in DD mice compared to control mice (Fig. 2K–N). These results further corroborate that ZIc glutamatergic neurons are activated in 6-OHDA PD model mice and hyperactivity of these neurons could be responsible for the parkinsonian motor deficits.

Fig. 2.

ZIc^VgluT2 neurons are hyperactive in DD mice. A Schematic of recording system for Ca²⁺ signals in ZIc^VgluT2 neurons with fiber photometry in VgluT2-ires-Cre mice. B Image showing optical fiber position and GCaMP6m expression in the ZIc of a VgluT2-ires-Cre mouse (scale bar, 200 μm). C Visualization of striatal dopaminergic afferents by tyrosine hydroxylase (TH) staining in coronal slices (scale bar, 1 mm). D, E Example ΔF/F of GCaMP6m signals in ZIc^VgluT2 neurons on the intact (D) and lesioned (E) side in unilaterally DD mice (green, 470 nm signal; purple, 405 nm control signal). F, G Heatmaps reporting ΔF/F of the 470 nm signal from recordings on the intact (F) and lesioned (G) side in individual mice. H Quantification of the area under the curve per second (AUC) in ZIc^VgluT2 neurons on the lesioned and intact side in unilaterally DD mice (n = 4 mice; *P <0.05, paired t test). I Quantification of the ΔF/F count (3% as threshold) in ZIc^VgluT2 neurons in the lesioned and intact side in unilaterally DD mice (n = 4; ***P < 0.001, paired t test). J Visualization of striatal dopaminergic afferents by TH staining in coronal slices (scale bar, 1 mm). K, L Examples ΔF/F of GCaMP6m signals in ZIc^VgluT2 neurons from control (K) and lesioned (L) mice (green, 470 nm signal; purple, 405-nm control signal). M, N Heatmaps reporting ΔF/F of the 470 nm signal of the recordings from control (M) and lesioned (N) mice. O Quantification of the AUC in ZIc^VgluT2 neurons in control and bilaterally DD mice (n = 4 control + 4 lesioned mice; ****P < 0.0001, unpaired t test). P Quantification of the ΔF/F count (3% as threshold) in ZIc^VgluT2 neurons in control and bilaterally DD mice (n = 4 control mice + 4 lesioned mice; ****P < 0.0001, unpaired t-test).

ZIc Glutamatergic Neuronal Inhibition Ameliorates Motor Deficits in DD Mice

As we had shown that activation of ZIc^VgluT2 neurons induced parkinsonian motor symptoms and their activity was abnormally high in the DD mice, we speculated that inhibition of this abnormal hyperactivity could rescue the parkinsonian motor deficits. We first injected AAV-DIO-NpHR-mCherry and control virus into the ZIc and demonstrated that the activity of ZIc^VgluT2 NpHR-expressing neurons was readily blocked following 589-nm light stimulation (Fig. S2). Three weeks after viral injection we rendered these mice semi-parkinsonian or parkinsonian by infusing 6-OHDA unilaterally or bilaterally into the dorsomedial striatum [20, 23] and then implanted optic fibers above the ZIc (Fig. 3A, B), resulting in a near-total loss of dopaminergic innervation of this region after 1 week (Fig. 3C, D). The unilateral DD mice were accompanied by parkinsonian motor deficits that included spontaneous ipsiversive rotations, decreased rearing (Fig. 3E–F), bradykinesia, increased immobility, and impaired motor coordination (Fig. 3G–J). We attempted to rescue these symptoms by unilateral photo-inhibition of ZIc^VgluT2 neurons. Unexpectedly, ZIc^VgluT2 neuronal inhibition significantly ameliorated the motor deficits. Unilateral inhibition rescued the spontaneous rotational bias (Fig. 3E) and ameliorated the bradykinesia (decreased pole descending time) (Fig. 3K), akinesia (decreased immobility in the OFT) (Fig. 3I), and motor coordination impairment (increased time on the rotarod) (Fig. 3L). Interestingly, in healthy naïve mice, unilateral photo-inhibition of ZIc^VgluT2 neurons did not affect motor behavior (Fig. S3A–D). Taken together, these findings demonstrate that ZIc^VgluT2 neuronal inhibition ameliorates the motor deficits in unilaterally DD mice. Consequently, we speculated that inhibition of this abnormal activity may rescue the motor deficits in the more severe bilaterally DD mice which were accompanied by parkinsonian motor deficits that included bradykinesia, increased immobility, and impaired motor coordination (Fig. 3M–R). We attempted to rescue these symptoms by bilateral photoinhibition of ZIc^VgluT2 neurons. Unexpectedly, ZIc^VgluT2 neuronal inhibition significantly ameliorated the motor deficits. Bilateral inhibition ameliorated the bradykinesia (increased speed and distance moved in the OFT, and decreased pole descending time) (Fig. 3M, N, Q), akinesia (decreased immobility) (Fig. 3O), and motor coordination impairment (decreased pole descending time and increased rotarod performance) (Fig. 3Q, R). ZIc^VgluT2 neuronal inhibition did not affect the center zone exploration time, indicating anxiety-like behavior was unaffected (Fig. 3P). Moreover, in healthy naïve mice, bilateral photo-inhibition of ZIc^VgluT2 neurons did not affect motor behavior (Fig. S3E–J). Taken together, these findings demonstrate that ZIc^VgluT2 neuronal inhibition ameliorates a constellation of motor deficits in a mouse model of PD.

Fig. 3.

Optogenetic inhibition of ZIc^VgluT2 neurons ameliorates dopamine depletion-induced parkinsonian motor deficits. A Schematic showing the viral injection, 6-OHDA injection, and placement of the optical fiber. B Image showing optic fiber position and NpHR expression in the ZIc of a VgluT2-ires-Cre mouse (scale bar, 200 μm). C Visualization of striatal dopaminergic afferents by TH staining in coronal slices (scale bar, 1 mm). D Graphical representation of the rotation and rearing tests, with the mouse receiving unilateral optical stimulation during light trials. E, F Number of net contralateral rotations (net contralateral rotation count = number of contralateral rotations – number of ipsilateral rotations) (E) and rearings (F) in NpHR and control DD mice during Pre–Light–Post trials. G–J Performance of NpHR and control DD mice in the OFT: movement speed (G), distance moved (H), time immobile (I), and time spent exploring center zone (J) during Pre–Light–Post trials. K Total time spent descending from the top of the pole facing upward in NpHR and control DD mice during Pre–Light–Post trials. L Latency to fall from the accelerating rotating rod in NpHR and control mice during Pre–Light–Post trials. In the unilateral behavioral tests, n = 10 mice for the NpHR group and n = 9 for the control group; *P <0.05, ***P <0.001, n.s., no statistically-significant difference, unpaired t-test. M–P Performance of NpHR and control DD mice in the OFT: movement speed (M), distance moved (N), time immobile (O), and time spent exploring the center zone (P) during Pre–Light–Post trials. Q Total time spent descending from the top of the pole facing upward in NpHR and control mice during Pre–Light–Post trials. R Latency to fall from the accelerating rotating rod in NpHR and control DD mice during Pre–Light–Post trials. For all behavioral tests, n = 11 mice for the NpHR group and n = 10 for the control group. *P < 0.05, **P < 0.01, ****P < 0.0001, n.s., no statistically-significant difference, unpaired t-test.

ZIc^VgluT2 Neurons Have a Monosynaptic Glutamatergic Projection to the SNpr

To unravel the underlying neural circuit mechanism of glutamatergic ZIc^VgluT2 neurons in the regulation of parkinsonian motor symptoms, we anatomically mapped their downstream projections. We observed efferent fields prominently referred to the SNpr (Fig. 4A). The SNpr is the major output of the BG in rodents [28]. As VgluT2⁺ neurons are glutamatergic, increased glutamate transmission from the ZIc would lead to over-activity of the SNpr. To determine whether this is indeed the case, we injected AAV-hSyn-FLEX-ChrimsonR-mCherry-WPRE-pA and control virus into the ZIc and AAV-hSyn-GCaMP6m-WPRE-pA into the SNpr of VgluT2-ires-Cre mice. Three weeks after viral injection, we implanted optical fibers in the SNpr of these mice (Fig. 4B, C), waited 7 days for recovery and performed simultaneous optogenetics and fiber photometry tests (Fig. 4D). Upon photo-activation, we found significant light-locked increases in Ca²⁺ activity in the SNpr (Fig. 4E–G). To further confirm and investigate the anatomical and functional connectivity results, we used ex vivo slice electrophysiology (Fig. 4H). The ChR2 virus was injected into the ZIc of VgluT2-ires-Cre mice, and light-evoked excitatory postsynaptic currents (eEPSCs) were recorded in cell bodies surrounded by ChR2-labeled terminals in the SNpr from brain slices. Recorded SNpr neurons with ChR2-infected terminals exhibited monosynaptic excitatory responses to the optogenetic activation of the axonal terminals, as evidenced by the finding that eEPSCs were blocked by TTX, rescued by 4AP, and abolished in the presence of the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor antagonist DNQX and the N-methyl-d-aspartate receptor antagonist DL-2-amino-5-phosphonovaleric acid (Fig. 4I, J). These results confirmed the connectivity of the monosynaptic glutamatergic ZIc–SNpr circuit.

Fig. 4.

ZIc has a monosynaptic glutamatergic connection to the SNpr. A Visualization of the AAV-DIO-mGFP injection site in ZIc^VgluT2 neurons and afferents in the SNpr (scale bar, 200 μm). B Schematic showing viral injection and placement of the optical fiber. C Image showing optical fiber position and ChrimsonR expression in ZIc ^VgluT2 neurons and GCaMP6m expression in SNpr neurons of a VgluT2-ires-Cre mouse (scale bar, 200 μm). D Schematic of the simultaneous recording and stimulation system for Ca²⁺ signals in the SNpr and photo-activation in ChrimsonR-expressing ZIc^VgluT2 neurons with fiber photometry in VgluT2-ires-Cre mice. E Upper panel, heatmaps reporting ΔF/F of the 470 nm signal of the recordings in individual mice (n = 4 mice injected with ChrimsonR); lower panel, averaged ΔF/F of the GCaMP6m signal (red trace) coupled with optogenetic activation (blue line) in the SNpr. F Upper panel, heatmaps reporting ΔF/F of the 470-nm signal of the recordings in individual mice (n = 4 mice injected with control virus); lower panel, averaged ΔF/F of the GCaMP6m signal (red trace) coupled with optogenetic activation (blue line) in the SNpr. G Peak fluorescence values during light stimulation in mice injected with ChrimsonR or mCherry (n = 4+4; ****P <0.0001, unpaired t-test). H Area under the curve per second (AUC) during light stimulation of mice injected with ChrimsonR or mCherry (n = 4+4; ****P <0.0001, unpaired t-test). I Schematic of light stimulation of glutamatergic ZIc neurons transfected with AAV-DIO-ChR2-mCherry and patch-clamp recording from SNpr neurons transfected with AAV-hSyn-GCaMP6m in brain slices from VgluT2-ires-Cre mice. J Examples of light-evoked EPSCs recorded from ChR2 terminal-infected SNpr neurons at a holding potential of –70 mV. EPSCs are blocked by 1 μmol/L TTX, rescued by 100 μmol/L 4-AP, and blocked again by 10 μmol/L DNQX and 50 mmol/L AP-V (right panel). K Amplitudes of light-evoked EPSCs as in I (n = 5 neurons from 3 mice).

ZIc–SNpr Circuit Activation Induces Parkinsonian Motor Symptoms

It has been proposed that increased glutamate transmission in the BG output regions leads to the hypokinetic motor symptoms of PD. For example, abnormally elevated firing of the STN in PD may contribute to motor symptom generation via increased glutamate release into the BG output regions [29–31]. To test whether activation of the ZIc–SNpr circuit leads to parkinsonian motor symptoms, we injected AAV-hSyn-DIO-ChR2-mCherry and control virus into the ZIc of VgluT2-ires-Cre mice. Three weeks after viral injection, we bilaterally implanted optical fibers into the SNpr of these mice (Fig. 5A, B). The mice were allowed 7 days for recovery from implantation, then we performed behavioral tests. Unilateral photo-activation (473 nm, 30 Hz, 5 ms) of ZIc^VgluT2 neuronal terminals in the SNpr induced robust and light-locked ipsiversive rotations and significantly decreased rearing (Fig. 5C, D). Bilateral photo-activation of ZIc^VgluT2 neuronal terminals in the SNpr induced bradykinesia (decreased speed and distance moved) (Fig. 5E, F) and akinesia (increased immobile time) (Fig. 5G) while leaving center zone exploration time unaffected in the OFT (Fig. 5H). Further, bilateral photo-activation of ZIc^VgluT2 neuronal terminals in the SNpr significantly impaired motor coordination (increased pole descending time and decreased time on the rotarod) (Fig. 5I, J). No effect on motor behavior was detected in the mice injected with control virus (Fig. 5). These data indicate that the ZIc^VgluT2 neurons regulate parkinsonian motor symptoms by affecting the basal ganglia output nucleus SNpr.

Fig. 5.

Optogenetic activation of the ZIc-SNr circuit induces parkinsonian motor symptoms. A Schematic showing viral injection and placement of the optical fiber. B Images showing optic fiber position and ChR2 expression in ZIc^VgluT2 neurons in a VgluT2-ires-Cre mouse (left panel scale bar, 200 μm) with magnified terminal region in the SNpr (right panel scale bar, 100 μm). C, D Number of net ipsilateral rotations (net ipsilateral rotations count = number of ipsilateral rotations – number of contralateral rotations) (C) and rearings (D) in ChR2 and control mice during Pre–Light–Post trials. E–H Performance of ChR2 and control mice in the OFT: movement speed (E), distance moved (F), time immobile (G), and time spent exploring the center zone (H) during Pre–Light–Post trials. I Total time spent descending from the top of the pole facing upward in ChR2 and control mice during Pre–Light–Post trials. J Latency to fall from the accelerating rotating rod in ChR2 and control mice during Pre–Light–Post trials. For all behavioral tests, n = 10 mice in the ChR2 group and n = 9 in the control group. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, n.s., no statistically-significant difference, paired t-test.

The SNpr is the main BG output nucleus in rodents. According to the classical model, dopamine depletion leads to excessive activity of neurons in these nuclei [32–34]. To test whether direct optogenetic activation of SNpr neurons leads to parkinsonian motor symptoms, we injected AAV-hSyn-ChR2-mCherry and control virus into the SNpr of VgluT2-ires-Cre mice. Three weeks after viral injection, we implanted optical fibers into the SNpr of these mice (Fig. S4A, B). Mice were given 7 days to recover from implantation before we performed behavioral tests. Unilateral photo-activation (473 nm, 30 Hz) of SNpr neurons induced robust and light-locked ipsiversive rotation and significantly decreased rearing (Fig. S4C, D). Bilateral photo-activation of SNpr neurons induced bradykinesia (decreased speed and total distance moved) (Fig. S4E, F) and akinesia (increased immobile time) (Fig. S4G), while center zone exploration time was unaffected in the OFT (Fig. S4H). Further, bilateral photo-activation of SNpr neurons induced significant impairment of motor coordination (increased pole descending time and decreased time on the rotarod) (Fig. S4I, J). No effect on motor behavior was detected in the mice injected with control virus (Fig. S4). These data are consistent with the classical model that excessive activity of neurons in the BG output nuclei leads to the hypokinetic motor symptoms in PD. Taken together, the present study demonstrates that hyperactivity of ZIc^VgluT2 neurons via monosynaptic glutamatergic transmission to the SNpr is responsible for parkinsonian motor symptoms. Inhibition of this neuronal hyperactivity ameliorated parkinsonian motor deficits (Fig. S4K).

Discussion

In the present study, we identified a critical role of ZIc^VgluT2 neurons and their glutamatergic transmission in the regulation of parkinsonian motor symptoms. We found that activation of VgluT2⁺ neurons in the ZIc induced time-locked parkinsonian motor symptoms and inhibition of this subset of neurons ameliorated the motor deficits in DD mice. ZIc glutamatergic neurons were hyperactive in both unilaterally and bilaterally DD mice. Interestingly, the peak counts of the spontaneous Ca²⁺ response in ZIc^VgluT2 neurons on the non-lesioned side of unilaterally DD mice was slightly higher than that of the control mice bilaterally injected with saline, possibly due to compensatory mechanisms and the difference in the general health of the mice. ZIc^VgluT2 neurons had a monosynaptic glutamatergic projection to the major BG output SNpr and activation of this circuit induced time-locked parkinsonian motor symptoms. Together, our findings not only provide a direct link between the ZIc and parkinsonian motor symptoms, but also define the role of ZIc^VgluT2 neurons and their glutamatergic transmission in the regulation of parkinsonian motor symptoms, which helps to better understand the mechanisms of PD and support the ZIc and its glutamatergic transmission as an important potential therapeutic target in PD treatment.

PD is the second most common and fastest-growing neurodegenerative disorder [1, 35]. The classical motor features include bradykinesia/akinesia, muscular rigidity, rest tremor, and postural instability. Other than the prominent death of SNpc dopaminergic neurons and resultant dopamine deficiency, PD also involves neurotransmitters besides dopamine and regions of the nervous system outside the BG [3, 4]. Although early studies suggested the possible involvement of the ZI in motor behavior and PD, the results were contradictory and lacked specificity [36]. Early pharmacological studies showed that microinjections into the ZI–hypothalamus region of bicuculline (a GABA_A receptor antagonist), picrotoxin (a GABA_A receptor channel blocker) or agonists of glutamate receptors, increase locomotor activity in rats [37]. Moreover, catalepsy induced by haloperidol is antagonized by blockade of GABA_A receptors in the ZI, while stimulation of these receptors induces catalepsy [38]. However, another study reported that increased rather than decreased mean firing of some (undefined) neurons in the central ZI (ZId and ZIv) occurs after 6-OHDA lesioning [39]. The ZI is a highly heterogeneous nucleus with a rich tapestry of neurochemically distinct cells and widespread connections to almost all regions of the brain [7]. With the latest advances in the technology of dissecting the functions of specific neuronal subgroups and neural circuits, it has become feasible to specify the roles of the subdomains, subpopulations, and circuit connections of the ZI. More recently, elegant studies have begun dissecting the various functions of the ZI based on its different subdomains, neuronal subpopulations, and circuit connections. So far, the ZI has been reported to be involved in sleep [8], fear [9, 10], pain [11], binge-like eating and weight gain [12], defensive behaviors [13, 14], predatory hunting [15, 16], and neural development [17], demonstrating its great functional heterogeneity. Interestingly, some of these features such as sleep, fear, and anxiety-related disorders, pain, and impulse control disorders are present among the non-motor symptoms of PD [40, 41]. These studies concentrated on GABAergic or subpopulations of GABAergic neurons in rostral or central regions of the ZI. In developmental terms, the ZI is part of prosomere 3 of the diencephalon. Prosomere 3 is dominated by GABAergic neurons, as are the ZIr, ZId, and ZIv [18]. However, in the ZIc, the expression of the GABA gene GAD67 is very low, while the VgluT2 gene Slc17a6 is specifically enriched. The role of the ZIc and its neuronal subpopulations remains unknown. Here, our data revealed for the first time the important function of ZIc glutamatergic but not GABAergic neurons in the regulation of parkinsonian motor symptoms, which not only helps to expand and deepen the understanding of the pathogenesis of PD, but also provides support for the discovery of new therapeutic strategies and methods for PD.

Classical BG models postulate that two parallel pathways, the direct and indirect pathways, exert opposing control over movement [32, 42]. Recent studies on the function of the striatum, STN, and GP have elegantly verified, expanded, and updated our understanding of the classical model [20, 28, 43, 44]. Overactivity of the indirect pathway is thought to underlie the hypokinetic motor symptoms in PD [45]. Of note, the STN is the only glutamatergic nucleus within the BG. Abnormally-elevated firing of the STN in PD may contribute to the generation of motor symptoms, via increased glutamate release into the BG output regions [29–31]. Here, we found that abnormally-elevated firing of glutamatergic neurons in the ZIc, a structure outside the BG, also contributed to the generation of parkinsonian motor symptoms, and suppression of this abnormally-elevated activity ameliorated the motor symptoms in DD mice.

Supplementary Information

Below is the link to the electronic supplementary material.

Conflict of interest

The authors declare no conflicts of interest.