Abstract
Activation of the primary motor cortex (M1) is important for the execution of skilled movements and motor learning, and its dysfunction contributes to the pathophysiology of Parkinson's disease (PD). A well-accepted idea in PD research, albeit not tested experimentally, is that the loss of midbrain dopamine leads to decreased activation of M1 by the motor thalamus. Here, we report that midbrain dopamine loss altered motor thalamus input in a laminar- and cell type-specific fashion and induced laminar-specific changes in intracortical synaptic transmission. Frequency-dependent changes in synaptic dynamics were also observed. Our results demonstrate that loss of midbrain dopaminergic neurons alters thalamocortical activation of M1 in both male and female mice, and provide novel insights into circuit mechanisms for motor cortex dysfunction in a mouse model of PD.
SIGNIFICANCE STATEMENT Loss of midbrain dopamine neurons increases inhibition from the basal ganglia to the motor thalamus, suggesting that it may ultimately lead to reduced activation of primary motor cortex (M1). In contrast with this line of thinking, analysis of M1 activity in patients and animal models of Parkinson's disease report hyperactivation of this region. Our results are the first report that midbrain dopamine loss alters the input–output function of M1 through laminar and cell type specific effects. These findings support and expand on the idea that loss of midbrain dopamine reduces motor cortex activation and provide experimental evidence that reconciles reduced thalamocortical input with reports of altered activation of motor cortex in patients with Parkinson's disease.
Introduction
Skilled voluntary movement is essential for nearly all behaviors. It requires coordinated communication across the motor pathway, which includes the basal ganglia, cerebellum, thalamus, and cortex, to produce the intended movement. The primary motor cortex (M1) is the output center of the motor pathway. M1 directly controls movement via corticospinal- (Lemon, 1993) and corticostriatal- (Shepherd, 2013) projecting neurons in deep layers, while cells in superficial layers provide feedback to other cortical regions and the striatum (Oswald et al., 2013; Shepherd, 2013). The activity of M1 pyramidal neurons is regulated by neighboring inhibitory neurons, including parvalbumin-expressing (PV+) cells. Engagement of M1 excitatory and inhibitory cells is required for many aspects of motor function, including movement execution (Kaufman et al., 2013; Melzer et al., 2017), motor planning (Svoboda and Li, 2018), and learning (Hosp et al., 2011; Chen et al., 2015; Biane et al., 2016; Kida et al., 2016).
A major long-range input to M1 arises from the ventroanterior/ventrolateral nuclei of the thalamus (Mthal). Mthal projections to M1 are most dense in L2/3 and L5 (Hooks et al., 2013; Hunnicutt et al., 2014) and make glutamatergic synapses with excitatory and inhibitory neurons (Biane et al., 2016; Shigematsu et al., 2016). The synaptic properties of this input are not as well understood as thalamic projections to sensory cortices. Studies of similar thalamocortical driver pathways predict that activation of the Mthal-M1 projection will elicit large, depressing postsynaptic currents in both excitatory and inhibitory cells (Swanson and Maffei, 2019), and that the excitation of M1 will be regulated through feedforward inhibition. While feedforward inhibition mediated through PV+ GABAergic neurons is an important component of cortical processing (Gabernet et al., 2005; Wang et al., 2010; Swanson and Maffei, 2019), it is not well studied in M1. Thus, determining the engagement of excitatory and inhibitory neurons following thalamocortical stimulation is crucial in understanding how M1 gates voluntary movement.
Transient disruption of M1 activity diminishes control over voluntary movement (Schieber and Poliakov, 1998; Stepniewska et al., 2014; Chen et al., 2019), and chronic changes in M1 activity have been associated with movement disorders, including Parkinson's disease (PD). PD is a highly prevalent movement disorder characterized by progressive tremor, rigidity, akinesia, and postural instability (Wenning et al., 2005; Ostrem and Galifianakis, 2010; Hirsch et al., 2016). While it is well established that the loss of dopaminergic neurons in the SNc leads to dramatic shifts in synaptic transmission in many motor areas (Day et al., 2006; Bagetta et al., 2010; Fan et al., 2012), the effects on M1 remains understudied. Emerging reports show evidence of abnormal neural activity within M1 in PD (Lindenbach and Bishop, 2013; Calabresi and Di Filippo, 2015). A working hypothesis in the field is that loss of midbrain dopamine impacts thalamocortical excitation of M1, yet the synaptic underpinnings of this dysfunction remain unclear.
Here, we performed whole-cell recordings of M1 pyramidal neurons and inhibitory PV+ interneurons to determine how loss of midbrain dopaminergic neurons impacts Mthal synaptic transmission. We used optogenetic stimulation of the Mthal-M1 pathway in a mouse model of PD and measured synaptic drive and dynamics in a dopamine-depleted state. Our results indicate that midbrain dopamine depletion leads to layer-specific reduction of Mthal input to M1 pyramidal neurons, while Mthal-evoked responses in PV+ cells are overall preserved. Furthermore, analysis of synaptic dynamics reveals frequency-dependent changes in Mthal-evoked responses selectively on pyramidal neurons. The ratio between excitation and inhibition evoked onto pyramidal neurons was preserved, suggesting that the circuit in M1 recruits additional mechanisms to preserve the E/I ratio of Mthal activation. This is the first direct experimental evidence that Mthal transmission is reduced because of loss of dopaminergic input to the motor pathway, and provide a mechanistic framework for M1 dysfunction in PD.
Materials and Methods
Animals
All experimental procedures were designed and executed following the guidelines of the National Institute of Health and were approved by Stony Brook University's Institutional Animal Care and Use Committee. Mice of both sexes were used in all experiments. Most recordings of excitatory neurons were collected from C57BL/6NCrl (#027, Charles River) animals. Recordings of fast-spiking interneurons, as well as a subset of pyramidal neurons, were made from acute slice preparations obtained from the progeny (Pvcretdtomato) of female PV-Cre (B6;129P2-Pvalbtm1(cre)Arbr/J, #00809, The Jackson Laboratory) and male Ai14 (B6.Cg-Gt(ROSA)26Sortm14(CAG-tdTomato)Hze/J, #007914, The Jackson Laboratory) mice. The number of animals and number of cells within each experimental group are reported as N and n, respectively.
Surgical procedures
Animals at P36-P50 were anesthetized with a ketamine/xylazine cocktail prepared in sterile saline (100 mg/kg ketamine, 10 mg/kg xylazine), and received an intraperitoneal injection of desipramine (1.25 mg/ml, 20 ml/kg) 30 min before surgery. Each animal received a unilateral injection of 6-OHDA centered on the SNc and an injection of AAV9-CAG-ChR2-Venus (diluted to final titer 1.127 × 1012 GC/ml, Penn Vector Core; ChR2) (Petreanu et al., 2007) in the VA/VL of the same hemisphere. Following the opening of craniotomies over each injection site, a pressure injection system (Nanoinject, Drummond) was used to inject 50 nl of virus into the VA/VL (bregma –0.7, midline ±1.0, surface –3.75) at 4.6 nl intervals. The injection pipette was left in place for 5 min to prevent virus spread up the needle tract. 6-OHDA was prepared fresh (15 mg/ml in 0.02% ascorbate saline) and injected at each of two injection sites (250 nl, 2.75 µg at each site) centered along the anterior-posterior extent of the SNc (bregma –3.1/2.8, midline –1.2, surface –3.93). Control animals were injected with equivalent volumes of vehicle solution. Following surgery, animals recovered on a heating pad and were monitored daily for food and water intake.
Cylinder motor task
Before preparing slices, animals were assessed for motor impairment via a cylinder motor task (Iancu et al., 2005). Each animal was placed in a clear acrylic cylinder surrounded by mirrors, with a camera positioned overhead, and allowed to freely explore for 10 min. Exploratory mouse movements were filmed and analyzed by an experimenter blind to experimental conditions. Forelimb impairment was expressed as the ratio of weight-bearing wall touches made by the limb contralateral to the SNc injection and total wall touches made by both forelimbs.
Slice electrophysiology
Slices were prepared ∼2 weeks after 6-OHDA injection (days following surgery: Vehicle = 15.66 ± 0.73, 6-OHDA = 16.36 ± 0.79, p = 0.52). Animals were anesthetized with isoflurane using the bell jar method and rapidly decapitated. The brain was dissected and sectioned in ice-cold oxygenated ACSF using a vibrating blade microtome (Leica VT1000S); 300 µm slices containing forelimb M1 (Tennant et al., 2011) were transferred into 37°C ACSF to recover for 30 min, then moved into room temperature ACSF to stabilize for 1 h. Whole-cell patch clamp was performed at room temperature and guided by DIC optics. Recordings were obtained with pulled borosilicate glass pipettes with a resistance of 3-4 mΩ, filled with internal solution containing the following (in mm): 100 K-Glu, 20 KCl, 10 K-HEPES, 4 Mg-ATP, 0.3 Na-GTP, 10 Na-phosphocreatine, and 0.4% biocytin, pH 7.35 titrated with KOH and adjusted to 295 mOsm with sucrose (Erev[Cl–] = −49.8 mV), unless otherwise noted. PV+ interneurons were targeted under fluorescent light in Pvcretdtomato animals. Mthal terminal fields were stimulated with 1 ms pulses of blue light (470 nm), produced by a high-powered LED lamp (X-cite 120LEDMini, Excitalas), passed through a filter and delivered through a 40× water immersion objective. Intensity of ChR2 stimulation was controlled by the LED dial, and the intensity of light emitted from the objective was determined using a power meter (PM100D, Thorlabs). Light pulses were delivered at 1%, 5%, 10%, 15%, and 20% LED intensity (in mW emitted from objective: 0.47, 1.04, 1.61, 2.20, and 2.78, respectively). To assess input–output curves for evoked Mthal currents, stimulation was delivered for 5 sweeps (30s between each sweep, at each intensity), while EPSCs were recorded in voltage clamp. Short-term dynamics of Mthal-evoked responses were assessed with optical stimulation at 5 and 10 Hz at 5% LED intensity, and responses were recorded in voltage clamp. Current-clamp recordings of responses to Mthal input were performed using 5% LED intensity. We measured evoked Mthal-EPSCs and disynaptic Mthal-IPSCs as well as spontaneous synaptic currents onto M1 neurons by recording neurons in voltage clamp with an internal solution containing the following (in mm): 20 KCl, 100 Cs-sulfate, 10 K-HEPES, 4 Mg-ATP, 0.3 Na-GTP, 10 Na-phosphocreatine, 3 QX-314 (Tocris Bioscience), 0.2% biocytin (Erev[Cl–] = −49.8 mV)). Evoked Mthal-EPSCs and sEPSCs were recorded while holding neurons at the reversal potential for chloride (−50 mV), while evoked Mthal-IPSCs and sIPSCs were isolated at the reversal potential for excitatory cations (10 mV). ACSF used in all electrophysiology experiments contained the following (in mm): 126 NaCl, 3 KCl, 25 NaHCO3, 1 NaH2PO4, 2 MgSO4, 2 CaCl2, and 14 dextrose. Series resistance (Rs) was tracked throughout the experiment, and data from cells whose Rs was >10% of their input resistance or changed >20% over the course of the experiment were excluded.
Immunohistochemistry
Tissue containing injection sites, and all slices used for recordings were postfixed in 4% PFA (0.01 m PBS, pH 7.4) for 1 week. Brain tissue containing injection sites were sectioned in the coronal plane at 50 µm thickness with a vibrating blade microtome (Leica VT1000S). Tissue sections processed with immunofluorescence protocols were rinsed in PBS 3 times, for 10 min each (3 × 10 min), then incubated in an antigen retrieval solution (10 mm sodium citrate in dH2O, pH 8.5, at 45°C, 30 min for 50 µm sections and 45 min for 300 µm sections). Tissue was rinsed again in PBS 3 × 10 min, then incubated in glycine (50 mm, in PBS, 1 h for 50 µm sections, 2 h for 300 µm sections). Following additional 3 × 10 min PBS rinses, the tissue was incubated in a preblock solution [in PBS: 5% BSA (Sigma), 5% normal goat serum (Vector Laboratories), 0.2%/1% Triton-X (VWR) for 50 µm/300 µm sections] for 1 h or 3 h at room temperature for 50 or 300 µm sections, respectively. Following preblock, sections were incubated overnight at 4°C in an antibody incubation solution [in PBS: 1% BSA (Sigma), 1% normal goat serum (Vector), 0.1% Triton-X (Sigma)] containing the appropriate primary antibodies and streptavidin reagents (Table 1). The following day, tissue was rinsed in 3 × 10 min in PBS and transferred into the antibody incubation solution containing secondary antibodies (Table 1); 300 µm slices were incubated in secondary antibodies for 6 h at room temperature, and 50 µm slices were incubated for 4 h. Following a 3 × 10 min PBS rinse, 300 µm slices were counterstained with Hoechst33342 (1:5000, Invitrogen H3570) for 20 min and 50 µm sections containing ChR2 injection sites were counterstained with Neurotrace (435/455) for 30 min, then rinsed in 0.1 m PB, mounted, and coverslipped with fluorescent mounting medium (Fluoromount-G, Fisher Scientific).
Details for antibodies used in immunohistochemistry
Sections through the SNc and VTA, containing 6-OHDA injection sites, were processed and visualized in DAB. This tissue underwent the described protocol, with the following modifications: following glycine incubation, sections were rinsed 3 × 10 min in PBS and then incubated in 0.3% H2O2 (in PBS, 30 min incubation at 4°C). Sections were then rinsed 3 × 5 min in PBS, then blocked for endogenous avidin-biotin reactivity (Avidin/Biotin Blocking Kit, Vector Laboratories). Following 3 × 1 min rinses, tissue was placed in the preblocking solution, then primary antibodies (Table 1). After secondary antibody incubation, sections were rinsed 3 × 10 min in PBS and then incubated in avidin-biotin HRP (Vectastain Elite ABC kit, Vector) for 1 h at room temperature. Following an additional 3 × 10 min rinse, sections were developed for 60s in DAB (DAB Peroxidase Substrate Kit, Vector). Sections were rinsed one more time for 3 × 10 min in 0.1 m PBS, pH 7.4, mounted on gelatin-coated slides and air-dried for 1 week. Slides were then dehydrated in a series of alcohols (70%, 95%, 100%), cleared in xlyenes, and coverslipped with Entellan mounting medium. Imaging of fluorescently labeled sections was performed on a laser-scanning confocal microscope (Olympus), and brightfield images were obtained using a widefield microscope (Olympus).
Animal inclusion criteria
Sections containing the SNc, as well as more anterior sections containing the striatum, were processed for TH immunoreactivity and imaged. These images were used as inclusion criteria for an animal to remain in the study. Lesioned animals included in this dataset showed severe dopaminergic cell loss along the extent of the SNc, as well as significant loss of TH+ afferents in the striatum. Previous work using this same mouse model of PD determined that 6-OHDA-lesioned mice showed >90% dopaminergic cell loss in the injected SNc, and 30% dopaminergic cell loss in the ipsilateral VTA (Swanson et al., 2021). Animals with injection sites too medial, leading to complete loss of VTA dopaminergic neurons, or too posterior, leading to loss of noradrenergic neurons in the locus coeruleus, were excluded from the study. All animals in this study showed significant forelimb motor impairment as tested with the cylinder motor task before slice preparation for patch-clamp recording. Sections containing the thalamus were processed for GFP immunoreactivity to enhance the visualization of ChR2 expression and imaged. ChR2 injection site expression that extended into adjacent thalamic nuclei, particularly the ventroposteromedial nucleus or the posteromedial nucleus, or expression in the needle tract in forelimb motor cortex, were criteria for exclusion from the study.
Data analysis
Neurons were sorted by cortical layer as previously described (Swanson et al., 2021). Measurements from collected electrophysiological recordings were performed using custom-made procedures in Igor (Wavemetrics), as well as by using a template-matching system in Clampfit (Molecular Devices). For analysis of input–output curves for Mthal-evoked responses, EPSC amplitude was measured as the absolute difference between the prelight pulse holding current (averaged across 100 ms) and the absolute minima of the EPSC. The squared inverse of EPSC coefficient of variation (1/CV2) and variance to mean ratio (VMR) were measured using EPSCs evoked at 5% LED intensity, as previously described (van Huijstee and Kessels, 2020). Briefly, 1/CV2 was calculated as the squared inverse of the SD of EPSC amplitude, divided by the mean amplitude (averaged across 5 sweeps/cell), for each cell, and VMR was calculated as the variance of the EPSC amplitude (averaged across 5 sweeps/cell) divided by the mean. EPSC charge and decay tau at 5% LED intensity were calculated in Clampfit. Analysis of short-term dynamics was performed in Igor by measuring the amplitude of each event. Evoked E/I ratio measurements were analyzed as follows: 5 sweeps of evoked EPSCs (or IPSCs) were averaged in Igor, the baseline was subtracted, and the charge for each average trace was measured as the area under the curve. Evoked EPSC and IPSC charge measurements were then compiled as an EPSC/IPSC ratio. For current-clamp traces, the number of action potentials occurring in the 100 ms window following the light pulse were counted.
To assess possible changes in the input–output function of PV+ neurons, frequency versus current injection (fI) curves were obtained by calculating the number of action potentials for each suprathreshold current step. Somatic currents of increasing amplitude were injected while holding neurons at a baseline potential of −70 mV (step duration: 700 ms, step amplitude: −100 to 450 pA in 50 pA increments). Rheobase was determined by injecting depolarizing current at 1 pA increments until a single action potential was elicited. Action potential threshold was measured at rheobase as previously described (Swanson et al., 2021). Max frequency was calculated as the maximum firing rate across all current injections. Input resistance was calculated as the slope of the linear regression line fit to the I–V plot of each cell. This line was fit through the linear range of amplitudes of steady-state voltage in response to somatic current injection.
The amplitude and instantaneous frequency of sEPSCs and sIPSCs were measured in Clampfit using custom template matching criteria, and these data were organized into cumulative probability plots in Microsoft Excel using the statistics software add-on XLSTAT (100 events per cell). Following initial data analyses in Igor and Clampfit, data were compiled in Microsoft Excel. For figures where each data point was representing measurements from a single animal rather than a single cell, data collected from neurons located within each layer of the given animal were averaged, and the mean from each animal was then used in any statistical tests performed.
Statistical analyses
Statistical tests were performed in Microsoft Excel and the add-in statistical program XLSTAT. All data are shown as mean ± SEM for the number of neurons (n) and the number of animals (N) indicated. Unpaired two-tailed Student's t tests were used to test for statistically significant differences in mean between experimental groups. Two-way repeated-measures ANOVAs were used to test significance in experiments where the same cells were measured across a continuous dependent variable, and pairwise comparisons were made post hoc with Tukey HSD tests. The differences between cumulative distributions were assessed with the Kolmogorov–Smirnov test. The Fisher's exact test was used to determine differences between groups regarding the proportions of firing neurons in current-clamp data (see Fig. 5). p values ≤0.05 were considered significant.
Results
Targeting Mthal input onto M1 neurons in a mouse model of PD
Each animal received a unilateral injection of AAV9-CAG-ChR2-Venus into VA/VL (Mthal), as well as two ipsilateral injections of 6-OHDA, centered along the anterior-posterior extent of the SNc (Fig. 1A). Immunohistochemical enhancement of the injection site and terminal fields revealed targeted ChR2 injections of Mthal and the expected double-banded thalamocortical axon distribution in M1 (Fig. 1B) (Hooks et al., 2013; Biane et al., 2016). Recorded M1 neurons were filled with biocytin and stained with fluorescently conjugated streptavidin to confirm pyramidal neuron morphology and laminar location (Fig. 1B). At the time of recording, all lesioned animals showed significant motor impairment, expressed as biased use of the forelimb ipsilateral to the 6-OHDA injection (Fig. 1C, wall touch ratio: Vehicle = 0.48 ± 0.015, 6-OHDA = 0.24 ± 0.027, p = 9.8E−10). Sections containing midbrain dopaminergic nuclei were stained for TH; animals with absence of TH+ neurons within the SNc of the injected hemisphere were considered successfully lesioned (Fig. 1D).
Experimental setup. A, Schematic represents dual, ipsilateral injections of ChR2 into Mthal and 6-OHDA/vehicle into the SNc. B, Left, 50 µm coronal section of thalamus (red represents Neurotrace) containing ChR2 injection site (green) within Mthal. Top right, 300 µm recorded slice with ChR2 (green) expression producing a double-banded terminal field pattern in M1. Bottom right, 300 µm recorded slice of containing two recorded pyramidal neurons (red represents Streptavidin; blue represents Hoechst). C, Cylinder motor assessment results. Wall touch ratio calculated as touches with forelimb contralateral to SNc injection divided by total forelimb use. Vehicle N = 31, 6-OHDA N = 29. ***p ≤ 0.001. D, Coronal midbrain sections stained for TH immunoreactivity and visualized with DAB. Scale bars, 500 µm.
Reduced thalamocortical drive onto L2/3 pyramidal cells following dopamine depletion
Expression of ChR2 within Mthal produces a double-banded terminal field pattern within M1, with the densest ChR2-expressing axons in L2/3 as well as L5. Each layer plays a unique role in the circuit of M1, with L2/3 acting primarily as an input and integration layer while L5 serves more as an output layer (Weiler et al., 2008; Oswald et al., 2013; Shepherd, 2013). With their distinct functions in mind, we asked whether loss of dopaminergic signaling in the motor pathway may have laminar-specific effects on Mthal input to pyramidal neurons in L2/3 and L5.
First, we examined the effect of dopamine depletion on the Mthal-M1 synapses onto L2/3 neurons. All L2/3 cells included in these experiments exhibited a typical pyramidal cell firing pattern (Fig. 2A) and were negative for GAD67 immunoreactivity (Fig. 2B). To first examine the baseline parameters of this input, single light pulses (1 ms) were delivered at 5% LED intensity and responses were recorded in voltage clamp. Optical activation of Mthal axons elicited EPSCs in L2/3 neurons of both vehicle-injected and lesioned animals (Fig. 2C). Mthal-EPSCs were monophasic, showed no jitter, and had delay from stimulus consistent with monosynaptic connections, similar to our previous recordings of thalamocortical inputs to other cortical circuits using the same approach and viral construct (Wang et al., 2013, 2019; Kloc and Maffei, 2014). A small number of sweeps across all neurons showed synaptic events on the tail of the Mthal-EPSC, suggesting possible polysynaptic events on the decay phase of the evoked current (Kloc and Maffei, 2014). These sweeps were not included in the analysis of Mthal-EPSC kinetics. At a 5% LED intensity (see Materials and Methods), the evoked EPSC amplitude was significantly reduced in L2/3 neurons of lesioned animals (Fig. 2C and second data point in Fig. 2H, in pA: Vehicle = 563.92 ± 75.41, 6-OHDA = 183.65 ± 34.56, p = 0.0002). Further reflecting the reduction in EPSC magnitude, the evoked EPSC charge was significantly reduced in L2/3 neurons in the lesioned group relative to controls (Fig. 2D, in pA·ms: Vehicle = 12,416.55 ± 1918.97, 6-OHDA = 3986.25 ± 669.38, p = 0.0011). The kinetics of the evoked EPSC were also altered: EPSC decay time constant (τ), when normalized to the respective EPSC amplitude, was significantly longer in lesioned animals (Fig. 2E, in ms/pA: Vehicle = 0.032 ± 0.0087, 6-OHDA = 0.086 ± 0.021, p = 0.014). These data suggest depletion of the dopaminergic activity in the motor pathway leads to reduced baseline strength and slower current decay at the Mthal input to M1 L2/3 excitatory neurons.
Mthal input to M1 L2/3 neurons is reduced in dopamine-depleted animals. A, Whole-cell recordings of L2/3 Pyr neurons in M1, with optical activation of ChR2-expressing Mthal axons. Right, Firing pattern of a L2/3 neuron in vehicle-injected and lesioned animal. Calibration: 20 mV, 100 ms. B, Confocal images at 60× of a recorded L2/3 neuron. Top, GAD67– recorded neuron (open arrow) with adjacent unrecorded GAD67+ neurons (filled arrows) at a single z-plane depth. Bottom left, L2/3 recorded neuron shown as a collapsed stack spanning cell's full structure. Bottom right, Merge of L2/3 neuron and surrounding ChR2-expressing axons. Scale bars, 50 µm. C, Superimposed example traces of Mthal-EPSCs from a vehicle-injected (black) and lesioned (green) animal. Calibration: 50 pA, 50 ms. D, Calculated EPSC charges, and group averages (black circles) using 5% LED intensity. E, Calculated EPSC decay tau, normalized to EPSC amplitude, and group averages (black circles) using 5% LED intensity. F, Average 1/CV2 of EPSCs in response to 5% LED intensity. G, Average VMR of EPSCs in response to 5% LED intensity. H, I, Average input–output curves of EPSC amplitude evoked by increasing LED intensity. J, Left, Example traces of EPSCs from vehicle-injected and lesioned animal in response to 5 Hz stimulus train. Calibration: 100 pA, 125 ms. Right, Average EPSC amplitudes, normalized to the first EPSC, in response to 5 Hz stimulation. K, Left, Example traces of EPSCs from vehicle-injected and lesioned animal in response to 10 Hz stimulus train. Calibration: 100 pA, 125 ms. Right, Average EPSC amplitudes, normalized to the first EPSC, in response to 10 Hz stimulation. Vehicle L2/3 neurons N = 7, n = 16, 6-OHDA L2/3 neurons N = 8, n = 13. Data are mean ± SEM. *p ≤ 0.05. **p ≤ 0.01. ***p ≤ 0.001.
To begin investigating the synaptic mechanisms underlying this reduction in Mthal-EPSC amplitude, we calculated the inverse square of the CV (1/CV2) and VMR in each group. The 1/CV2 coefficient from lesioned animals was reduced compared with controls (Fig. 2F, 1/CV2: Vehicle = 356.94 ± 97.96, 6-OHDA = 71.04 ± 16.53, p = 0.015), while there was no significant difference in VMR (Fig. 2G, VMR: Vehicle = 2.86 ± 0.56, 6-OHDA = 3.78 ± 0.76, p = 0.33). Previous studies interpreted a shift in 1/CV2, with no change in VMR, as indicative of a presynaptic origin for the observed changes in synaptic current, possibly dependent on a change in release sites (van Huijstee and Kessels, 2020). We next compared the input–output relationship at this synapse by varying stimulation intensity. Mthal-EPSCs onto L2/3 neurons were reduced at all intensities in lesioned animals compared with their control counterparts (Fig. 2H, two-way repeated-measures ANOVA: effect of group, F(1,27) = 16.30, p = 0.00040, repetition, F(4,108) = 39.75, p < 0.0001, and interaction, F(4108) = 2.45, p = 0.050, Tukey HSD post hoc test). This effect persisted when data were compared across animals within each experimental group (Fig. 2I, two-way repeated-measures ANOVA: effect of group, F(1,13) = 7.24, p = 0.019, repetition, F(4,52) = 27.38, p < 0.0001, Tukey HSD post hoc test).
To further assess the consequence of dopamine depletion on Mthal transmission, we delivered trains of light stimuli at 5 and 10 Hz (Fig. 2J,K) and quantified the short-term dynamics of Mthal-EPSCs. Consistent with the driver role of Mthal input (Sherman, 2007), evoked EPSCs showed short-term depression in both vehicle and lesioned mice. While there was no significant difference in short-term depression between groups in response to 5 Hz stimulation (Fig. 2J, two-way repeated-measures ANOVA: effect of group, F(1,26) = 0.41, p = 0.53, repetition, F(3,78) = 72.62, p < 0.0001, and interaction, F(3,78) = 1.49, p = 0.22, Tukey HSD post hoc test), 10 Hz trains of stimuli unveiled a significantly larger synaptic depression in the second and third EPSC in lesioned animals (Fig. 2K, two-way repeated-measures ANOVA: effect of group, F(1,24) = 3.66, p = 0.068, repetition, F(3,72) = 337.09, p < 0.0001, and interaction, F(3,72) = 5.11, p = 0.003, Tukey HSD post hoc test). These data support the interpretation that midbrain dopamine depletion alters synaptic transmission at the Mthal-M1 input onto L2/3 pyramidal neurons, and further support the interpretation that these changes are expressed presynaptically. Together, our findings demonstrate reduced baseline transmission and frequency-dependent increase in short-term depression of Mthal-M1 transmission following midbrain dopamine depletion.
Dopamine depletion impacts Mthal synaptic short-term depression in L5
To assess whether the effect of midbrain dopamine depletion on the Mthal-M1 connection may be layer-specific, we performed whole-cell recordings of L5 excitatory neurons and optically stimulated Mthal terminal fields (Fig. 3). L5 neurons included in the analysis showed regular firing pattern, pyramidal morphology, and lacked GAD67 expression (Fig. 3A,B). In contrast to L2/3, there was no significant difference in the amplitude of baseline Mthal-EPSCs onto L5 pyramidal neurons (Fig. 3C, and second data point in Fig. 3H, in pA: Vehicle = 450.82 ± 64.65, 6-OHDA = 433.70 ± 55.00, p = 0.85). Similarly, there was no difference in either EPSC charge (Fig. 3D, in pA·ms: Vehicle = 9077.54 ± 1566.94, 6-OHDA = 7417.00 ± 992.31, p = 0.43), or decay τ of scaled responses (Fig. 3E, in ms/pA: Vehicle = 0.042 ± 0.0071, 6-OHDA = 0.033 ± 0.010, p = 0.46). In accordance with unaltered Mthal-L5 synaptic strength, EPSC 1/CV2 and VMR were comparable between groups (Fig. 3F, 1/CV2: Vehicle = 265.11 ± 77.24, 6-OHDA = 146.44 ± 25.32, p = 0.22) (Fig. 3G, VMR: Vehicle = 3.39 ± 0.46, 6-OHDA = 4.00 ± 0.65, p = 0.44). Furthermore, increasing LED intensity did not reveal any differences in Mthal EPSCs onto L5 cells between experimental groups, whether quantified across neurons (Fig. 3H, two-way repeated-measures ANOVA: effect of group, F(1,34) = 0.013, p = 0.91, repetition, F(4,136) = 56.35, p < 0.0001, and interaction, F(4136) = 0.27, p = 0.90, Tukey HSD post hoc test) or across animals (Fig. 3I, two-way repeated-measures ANOVA: effect of group, F(1,13) = 0.003, p = 0.96, repetition, F(4,52) = 33.57, p < 0.0001, and interaction, F(4,52) = 0.46, p = 0.76, Tukey HSD post hoc test). These results indicate that, in contrast to Mthal input onto L2/3 cells, the baseline Mthal input strength to L5 neurons is unaffected by 6-OHDA. The laminar-specific effects of midbrain dopamine depletion may disrupt balanced thalamocortical activation of superficial and deep layers and affect circuit computations, effects that could have significant functional consequences for the output of each circuit and their downstream targets.
Mthal input to M1 L5 neurons is largely preserved in dopamine-depleted animals. A, Whole-cell recordings of L5 pyramidal neurons in M1, with optical activation of ChR2-expressing Mthal axons. Right, Typical firing pattern of a L5 neuron in vehicle-injected and lesioned animal. Calibration: 20 mV, 100 ms. B, Confocal images at 60× of an example recorded L5 neuron. Top, GAD67– recorded neuron (open arrow) with adjacent unrecorded GAD67+ neurons (filled arrows) at a single z-plane depth. Bottom left, L5 recorded neuron shown as a collapsed stack spanning cell's full structure. Bottom right, Merge of L5 neuron and surrounding ChR2-expressing axons. Scale bars, 50 µm. C, Superimposed example traces of Mthal-EPSCs from a vehicle-injected (black) and lesioned (teal) animal. Calibration: 50 pA, 50 ms. D, Calculated EPSC charges, and group averages (black circles) using 5% LED intensity. E, Calculated EPSC decay tau, normalized to EPSC amplitude, and group averages (black circles) using 5% LED intensity. F, Average 1/CV2 of EPSCs in response to 5% LED intensity. G, Average VMR of EPSCs in response to 5% LED intensity. H, I, Average input–output curves of EPSC amplitude evoked by increasing LED intensity. J, Left, Example traces of EPSCs from vehicle-injected and lesioned animal in response to 5 Hz stimulus train. Calibration: 100 pA, 250 ms. Right, Average EPSC amplitudes, normalized to the first EPSC, in response to 5 Hz stimulation. K, Left, Example traces of EPSCs from vehicle-injected and lesioned animal in response to 10 Hz stimulus train. Calibration: 100 pA, 250 ms. Right, Average EPSC amplitudes, normalized to the first EPSC, in response to 10 Hz stimulation. Vehicle L5 neurons N = 9, n = 21, 6-OHDA L2/3 neurons N = 6, n = 15. Data are mean ± SEM. *p ≤ 0.05. **p ≤ 0.01. ***p ≤ 0.001.
Repetitive stimulation of thalamocortical inputs revealed a frequency-dependent change in short-term depression at Mthal synapses onto L5 pyramidal neurons. Short-term synaptic depression was comparable between experimental groups for trains of stimuli at 5 Hz (Fig. 3J, two-way repeated-measures ANOVA: effect of group, F(1,32) = 0.048, p = 0.83, repetition, F(3,96) = 76.28, p < 0.0001, and interaction, F(3,96) = 1.45, p = 0.23, Tukey HSD post hoc test), while there was a significant increase in short-term depression in response to 10 Hz stimulus trains in lesioned animals (Fig. 3K, two-way repeated-measures ANOVA: effect of group, F(1,31) = 13.56, p = 0.001, repetition, F(3,93) = 173.97, p < 0.0001, and interaction, F(3,93) = 0.78, p = 0.51, Tukey HSD post hoc test). This effect suggests that, while baseline Mthal drive onto L5 neurons is preserved following midbrain dopamine depletion, the activation of these synapses by patterned stimulation is impaired. The frequency-dependent increase in synaptic depression in the absence of change in 1/CV2 suggests that this impairment engages distinct mechanisms from those we reported for L2/3.
Mthal drive onto PV+ interneurons in dopamine-depleted animals
A core feature of thalamocortical drive is the engagement of cortical inhibitory interneurons. While thalamocortical afferents excite multiple types of inhibitory interneurons, in many cortices the largest thalamic drive is onto PV+ cells (Swanson and Maffei, 2019). PV+ interneurons make perisomatic synapses onto glutamatergic neurons and are known for mediating powerful feedforward inhibition (Neske et al., 2015). We investigated how chronic midbrain dopaminergic cell loss impacts Mthal input onto M1 PV+ interneurons by performing whole-cell recordings in PVcretdtomato mice (Fig. 4). Recorded cells in L2/3 and L5 exhibited the fast-spiking firing phenotype typical of PV+ neurons (Fig. 4A) (Nassar et al., 2015). Tdtomato expression was preserved after immunohistochemical procedures, and all the neurons included in the analysis showed nonpyramidal morphology (Fig. 4B). Dopamine depletion had no effect on baseline Mthal-EPSC charge (Fig. 4D, in pA·ms: Vehicle = 4679.78 ± 1041.98, 6-OHDA = 2763.34 ± 1093.03, p = 0.33), scaled decay tau (Fig. 4E, in ms/pA: Vehicle = 0.041 ± 0.012, 6-OHDA = 0.072 ± 0.013, p = 0.10), or input–output relationship (Fig. 4H, two-way repeated-measures ANOVA: effect of group, F(1,27) = 1.10, p = 0.30, repetition, F(4108) = 38.63, p < 0.0001, and interaction, F(4108) = 2.66, p = 0.035, Tukey HSD post hoc test) in L2/3 PV+ neurons. There were also no significant changes in CV or VMR of Mthal-EPSCs onto L2/3 PV+ neurons (Fig. 4F, 1/CV2: Vehicle = 141.27 ± 38.64, 6-OHDA = 130.90 ± 60.62, p = 0.89; Fig. 4G, VMR: Vehicle = 3.85 ± 1.27, 6-OHDA = 4.39 ± 1.01, p = 0.74).
Mthal input on PV+ interneurons is not affected by dopamine depletion. A, Whole-cell recordings of L2/3 and L5 PV+ neurons in M1, with optical activation of ChR2-expressing Mthal axons. Right, Firing pattern of L2/3 and L5 PV+ neurons in vehicle-injected and lesioned animals. Calibration: 20 mV, 100 ms. B, Confocal images at 60× of a recorded L5 PV+ neuron. Left panels, Tdtomato (PV) expression in a recorded neuron and adjacent unrecorded neuron (filled arrows), and merge image with biocytin labeling of recorded neuron GAD67-recorded neuron at a single z-plane depth. Right, L5 PV+ recorded neuron shown as a collapsed stack spanning cell's full structure, and merged image with surrounding ChR2-expressing axons. Scale bars, 50 µm. C, Superimposed example traces of Mthal-EPSCs in L2/3 PV+ neurons from a vehicle-injected (black) and lesioned (orange) animal. Calibration: 50 pA, 50 ms. D, EPSC charges, and group averages (black circles) measured on responses evoked using 5% LED intensity. E, EPSC decay tau, normalized to EPSC amplitude, and group averages (black circles) using 5% LED intensity. F, Average 1/CV2 of EPSCs in response to 5% LED intensity. G, Average VMR of EPSCs in response to 5% LED intensity. H, Average input–output curve of Mthal-EPSC amplitudes in L2/3 PV+ neurons. I, Example traces of thalamocortical EPSCs in L5 PV+ neurons from a vehicle-injected (black) and lesioned (magenta) animal. Calibration: 50 pA, 50 ms. J, EPSC charges, and group averages (black circles) measured on responses evoked 5% LED intensity. K, EPSC decay tau, normalized to EPSC amplitude, and group averages (black circles) using 5% LED intensity. L, Average 1/CV2 of EPSCs in response to 5% LED intensity. M, Average VMR of EPSCs in response to 5% LED intensity. N, Average input–output curve of Mthal-EPSC amplitudes in L5 PV+ neurons. O, Example traces of L2/3 PV+ EPSCs from a vehicle-injected and lesioned animal in response to 10 Hz stimulus train. Calibration: 50 pA, 125 ms. P, Average EPSC amplitudes, normalized to the first EPSC, in response to 10 Hz stimulation. Q, Example traces of L5 PV+ Mthal-EPSCs from vehicle-injected and lesioned animal in response to 10 Hz stimulus train. Calibration: 50 pA, 125 ms. R, Average EPSC amplitudes, normalized to the first EPSC, in response to 10 Hz stimulation. Vehicle L2/3 PV+ neurons N = 6, n = 13, 6-OHDA L2/3 PV+ neurons N = 11, n = 16; Vehicle L5 PV+ neurons N = 9, n = 15, 6-OHDA L5 PV+ neurons N = 10, n = 15. Data are mean ± SEM. *p ≤ 0.05. **p ≤ 0.01. ***p ≤ 0.001.
Midbrain dopamine depletion spares M1 excitatory neuron output. A, Example traces of L2/3 excitatory neuron output in response to 5% LED stimulation of Mthal inputs. Pie charts represent the percentage of L2/3 excitatory neurons that fired in response to Mthal stimulation. B, Example traces of L5 excitatory neuron output in response to 5% LED stimulation of Mthal inputs. Pie charts represent the percentage of L5 excitatory neurons that fired in response to Mthal input. C, Example traces of L2/3 PV+ neuron output in response to 5% LED stimulation of Mthal inputs. Pie charts represent the percentage of L2/3 PV+ neurons that fired in response to Mthal input. D, Example traces of L5 PV+ neuron output in response to 5% LED stimulation of Mthal inputs. Pie charts represent the percentage of L5 PV+ neurons that fired in response to Mthal input. Calibration: 20 mV, 250 ms. Vehicle L2/3 Pyr neurons N = 6, n = 12, 6-OHDA L2/3 Pyr neurons N = 10, n = 12; Vehicle L5 Pyr neurons N = 7, n = 13, 6-OHDA L5 Pyr neurons N = 7, n = 10; Vehicle L2/3 PV+ neurons N = 7, n = 9, 6-OHDA L2/3 PV+ neurons N = 7, n = 10; Vehicle L5 PV+ neurons N = 5, n = 8, 6-OHDA L5 PV+ neurons N = 8, n = 12. *p ≤ 0.05. **p ≤ 0.01. ***p ≤ 0.001.
While nigral dopamine depletion reduced Mthal-EPSC charge in L5 PV+ neurons (Fig. 4J, in pA·ms: Vehicle = 4853.98 ± 1001.79, 6-OHDA = 2622.75 ± 430.52, p = 0.050), there was no effect of 6-OHDA lesion on the input–output curve of Mthal-EPSCs in this layer (Fig. 4N, two-way repeated-measures ANOVA: effect of group, F(1,28) = 1.45, p = 0.24, repetition, F(4,112) = 71.70, p < 0.0001, and interaction, F(4112) = 0.14, p = 0.97, Tukey HSD post hoc test). Additionally, dopamine depletion did not impact the scaled decay tau of Mthal-EPSCs in L5 PV+ neurons (Fig. 4K, in ms/pA: Vehicle = 0.046 ± 0.0083, 6-OHDA = 0.056 ± 0.0080, p = 0.39), suggesting that the decrease in charge may depend on the variability of the responses and possibly on the presence of residual asynchronous events on the tail of the EPSC. No changes were observed in the inverse squared CV (Fig. 4L, 1/CV2: Vehicle = 118.58 ± 34.94, 6-OHDA = 64.29 ± 13.91, p = 0.16) or VMR (Fig. 4M, VMR: Vehicle = 7.44 ± 1.77, 6-OHDA = 4.03 ± 0.71, p = 0.085). As 10 Hz stimulation frequency revealed an effect of lesion on Mthal input short-term dynamics, we chose this frequency when repeating these experiments in PV+ neurons. While 10 Hz stimulation of Mthal terminal fields evoked depressing EPSCs in both lesioned and control animals, the degree of synaptic depression was not affected by a loss of midbrain dopamine for either L2/3 (Fig. 4O,P, two-way repeated-measures ANOVA: effect of group, F(1,23) = 0.131, p = 0.72, repetition, F(3,69) = 38.60, p < 0.0001, and interaction, F(3,69) = 1.52, p = 0.22, Tukey HSD post hoc test) or L5 PV+ cells (Fig. 4Q,R, two-way repeated-measures ANOVA: effect of group, F(1,23) = 0.19, p = 0.67, repetition, F(3,69) = 86.14, p < 0.0001, and interaction, F(3,69) = 1.05, p = 0.38, Tukey HSD post hoc test). These results indicate that, while Mthal input onto L2/3 excitatory neurons is reduced following dopamine depletion, the magnitude and properties of Mthal responses onto PV+ inhibitory neurons in M1 are largely preserved.
M1 excitatory neuron output is maintained in dopamine-depleted animals
The magnitude of Mthal-EPSCs onto L2/3 excitatory neurons and the short-term dynamics of Mthal-M1 synapses in both L2/3 and L5 are impacted by midbrain dopamine depletion. With reduced Mthal input to M1 excitatory neurons in mind, we investigated whether these changes coincided with shifts in the input–output transformation of these cells (Fig. 5). We addressed this question by stimulating Mthal terminal fields at 5% LED intensity while recording evoked responses of L2/3 and L5 neurons in current clamp. In both layers, neurons showed a variety of responses to Mthal stimulation: while some fired an action potential, others showed only subthreshold EPSPs. We compared the fraction of neurons firing at least one action potential following Mthal axon stimulation but found no significant difference between control and lesioned animals in excitatory L2/3 neurons (Fig. 5A, Fisher's exact test: L2/3 Excitatory Neurons p = 1). These data suggest that, despite reduced Mthal drive to these neurons, the transformation from synaptic input to output was preserved. Our published work shows that the intrinsic properties of M1 pyramidal neurons are not affected by the 6-OHDA manipulation, excluding the possibility that the maintenance of evoked firing despite the decrease in Mthal-EPSC amplitude onto L2/3 neurons may depend on changes in voltage-gated conductance (Swanson et al., 2021). Thus, we posited that the maintenance of L2/3 neuron output may result from some compensatory mechanism within the cortical synaptic circuit.
One such mechanism could arise from changes in the input–output transformation of Mthal drive onto L2/3 PV+ neurons, which could in turn alter inhibitory drive onto pyramidal neurons. To test this possibility, we delivered light stimuli to Mthal axonal fields while recording PV+ L2/3 neurons and asked whether midbrain dopamine depletion impacted their evoked output. The fraction of L2/3 PV+ neurons firing in response to Mthal stimulation trended toward a reduction in lesioned mice; however, these results were not statistically significant (Fig. 5C, Fisher's exact test: L2/3 PV+ Neurons p = 0.37). Together with the lack of changes in Mthal-EPSCs onto PV+ neurons, these results suggest that the preservation of L2/3 excitatory neuron output does not rely on reduced output of PV+ neurons during Mthal stimulation.
In L5 lesioned animals, there were no differences in the proportion of excitatory neurons that fired following Mthal stimulation (Fig. 5B, Fisher's exact test: L5 Excitatory Neurons p = 0.22), and in the output of L5 PV+ neurons (Fig. 5D, Fisher's exact test: L5 PV+ Neurons p = 1).
Altered PV+ neuron excitability could explain the preserved output of M1 pyramidal neurons in the face of reduced Mthal drive. Our previous work highlighted the effects of nigral dopamine depletion on synaptic transmission-dependent excitability of M1 pyramidal neurons, and we predicted that these effects may extend across multiple cell types within M1 (Swanson et al., 2021). We recorded subthreshold and suprathreshold voltage responses during somatic current injection (Fig. 6) from PV+ neurons in L2/3 and L5. L2/3 PV+ neurons showed no change in input resistance (Fig. 6B, in mΩ: Vehicle = 203.76 ± 14.37, 6-OHDA = 209.31 ± 15.83, p = 0.78), rheobase (Fig. 6C, in pA: Vehicle = 94.92 ± 14.31, 6-OHDA = 92.50 ± 13.24, p = 0.90), or action potential threshold (Fig. 6D, in mV: Vehicle = −41.66 ± 0.84, 6-OHDA = −41.96 ± 1.26, p = 0.84). However, PV+ neurons from lesioned animals exhibited a reduced frequency-current relationship for large current injections (Fig. 6E, two-way repeated-measures ANOVA: effect of group, F(1,27) = 2.88, p = 0.10, repetition, F(8216) = 100.50, p < 0.0001, and interaction, F(8216) = 4.481, p < 0.0001, Tukey HSD post hoc test), and a reduced max frequency (Fig. 6F, in Hz: Vehicle = 109.29 ± 10.07, 6-OHDA = 76.57 ± 10.01, p = 0.029).
6-OHDA lesion shift M1 PV+ neuron synapse-dependent excitability. A, G, Superimposed responses to hyperpolarizing and depolarizing current steps in individual L2/3 and L5 PV+ neurons of vehicle and 6-OHDA-injected animals. Calibration: Left, 10 mV, 100 ms; Right, 20 mV, 100 ms. B, H, Input resistance across subthreshold current steps. C, I, Rheobase to elicit single action potential. D, J, Action potential threshold at rheobase. E, K, Action potential frequency during suprathreshold current injections. F, L, Max action potential frequency across all current steps. n in parentheses indicates sample size for rheobase and threshold measurements: L2/3 PV+ vehicle neurons N = 8, n = 14(12), ACSF L2/3 PV+ 6-OHDA neurons N = 9, n = 15(10), L5 PV+ vehicle neurons N = 9, n = 16(9), L5 PV+ 6-OHDA neurons N = 11, n = 17(13). Data are mean ± SEM. *p ≤ 0.05. **p ≤ 0.01. ***p ≤ 0.001.
In contrast, measurements in L5 PV+ neurons revealed that 6-OHDA lesion induces an increase in input resistance (Fig. 6H, in mΩ: Vehicle = 207.15 ± 9.33, 6-OHDA = 251.02 ± 18.46, p = 0.046), with no change in rheobase (Fig. 6I, in pA: Vehicle = 81.78 ± 8.52, 6-OHDA = 83.31 ± 11.43, p = 0.92) or action potential threshold (Fig. 6J, in mV: Vehicle = −42.12 ± 0.76, 6-OHDA = −40.95 ± 0.73, p = 0.30). Additionally, there was no effect on L5 PV+ neuron frequency-current relationship (Fig. 6K, two-way repeated-measures ANOVA: effect of group, F(1,31) = 0.62, p = 0.44, repetition, F(8248) = 59.93, p < 0.0001, and interaction, F(8248) = 1.16, p = 0.33), or max frequency (Fig. 6L, in Hz: Vehicle = 100.18 ± 7.61, 6-OHDA = 90.42 ± 10.28, p = 0.45). Together with the results shown in Figures 4 and 5, these data strongly suggest that loss of midbrain dopamine has subtle effects on Mthal drive onto PV+ neurons and on PV+ neuron excitability that may not fully explain the preservation of L2/3 neurons output in the face of reduced Mthal drive.
Midbrain dopamine loss does not shift the E/I ratio of Mthal inputs onto M1 neurons
Simultaneous engagement of excitatory and inhibitory cortical neurons is a key feature of thalamocortical transmission (Isaacson and Scanziani, 2011). In pyramidal neurons, the integration of the monosynaptic excitatory input with the disynaptic feedforward inhibitory input is fundamental to how the cortex processes thalamic signals. To address how midbrain dopamine depletion impacts the integration of these synaptic events, we measured the ratio of Mthal-evoked excitation and inhibition (E/I ratio) onto M1 pyramidal neurons. We used a cesium-based internal solution and recorded Mthal-EPSCs by holding neurons at –50 mV, the reversal potential for Cl– in our experimental conditions, and disynaptic IPSCs by holding neurons at 10 mV, the reversal potential for cation-mediated currents within the same pyramidal neuron in M1. The ratio of Mthal-EPSC and disynaptic IPSC charge was compared between vehicle and 6-OHDA experimental conditions (Fig. 7).
Mthal-evoked E/I ratio is preserved following 6-OHDA lesion. A, C, Example traces of isolated Mthal-EPSC and disynaptic IPSC onto a Pyr neuron evoked by 5% LED. Calibration: 500 pA, 100 ms. B, Average EPSC/IPSC charge ratio for L2/3 Pyr neurons between experimental groups. D, Average EPSC/IPSC charge ratio for L5 Pyr neurons between experimental groups. Vehicle L2/3 neurons N = 3, n = 6, 6-OHDA L2/3 neurons N = 3, n = 7; Vehicle L5 neurons N = 5, n = 9, 6-OHDA L5 neurons N = 4, n = 8. Data are mean ± SEM. *p ≤ 0.05. **p ≤ 0.01. ***p ≤ 0.001.
Despite the decrease in Mthal-EPSC amplitude in L2/3 of 6-OHDA mice, the E/I ratio of Mthal-evoked responses was preserved (Fig. 7A,B, EPSC charge/IPSC charge: Vehicle = 0.15 ± 0.076, 6-OHDA = 0.18 ± 0.084, p = 0.79). The same result was observed when comparing the ratio of Mthal-EPSCs to disynaptic IPSCs onto L5 pyramidal neurons in vehicle and 6-OHDA conditions (Fig. 7C,D, EPSC charge/IPSC charge: Vehicle = 0.22 ± 0.066, 6-OHDA = 0.14 ± 0.019, p = 0.31). In view of the results reported in Figures 4–6, showing no changes in Mthal drive onto PV+ neurons and no changes in PV+ neurons excitability aside from depolarization block above 50 Hz, the reduced disynaptic Mthal-IPSC in L2/3 is unlikely to depend on a reduced recruitment of PV+ neurons by Mthal, but may depend on synaptic plasticity at inhibitory inputs onto pyramidal neurons. The coordinated reduction of Mthal-EPSC and disynaptic Mthal-IPSC may explain the preserved proportion of M1 L2/3 neurons firing action potentials in response to Mthal stimulation. Overall, loss of midbrain dopamine does not alter the E/I ratio of evoked thalamocortical responses in M1.
Dopamine depletion drives broad shifts in M1 recurrent synaptic activity
Previous work has shown that unilateral 6-OHDA lesions of the nigra impact spontaneous synaptic activity in other motor areas, including the substantia nigra reticulata (Faynveitz et al., 2019). In view of the reduced feedforward inhibition onto pyramidal neurons in L2/3, we asked whether loss of midbrain dopamine may modulate synaptic drive in M1. We isolated sEPSCs and sIPSCs onto L2/3 and L5 excitatory neurons (Fig. 8) by recording at the reversal potential for Cl– (−50 mV), and at the reversal potential for cations mediating excitatory currents (10 mV), respectively. In L2/3, the amplitude of sEPSC recorded from pyramidal neurons was significantly reduced in 6-OHDA animals (Fig. 8C1, Kolmogorov–Smirnov test p < 0.0001), while there was no change in EPSC frequency (Fig. 8C2, Kolmogorov–Smirnov test p = 0.47). These changes were accompanied by a reduction in sIPSC amplitude (Fig. 8C3, Kolmogorov–Smirnov test p < 0.0001), and modest increase in sIPSC frequency (Fig. 8C4, Kolmogorov–Smirnov test p = 0.002). These results suggest that L2/3 pyramidal neurons received reduced excitatory and inhibitory drive in 6-OHDA mice. This effect may explain the preserved E/I ratio reported in Figure 7.
6-OHDA lesions shifts the intracortical E/I balance in M1 pyramidal neurons. A, B, Example traces of sEPSCs and sIPSCs recorded in control and 6-OHDA animals. Calibration: 100 pA, 500 ms. C, Cumulative probability plots of L2/3 sEPSC amplitude (C1), sEPSC instantaneous frequency (C2), sIPSC amplitude (C3), and sIPSC instantaneous frequency (C4). D, Cumulative probability plots of L5 sEPSC amplitude (D1), sEPSC instantaneous frequency (D2), sIPSC amplitude (D3), and sIPSC instantaneous frequency (D4). Vehicle L2/3 neurons N = 6, n = 9, 6-OHDA L2/3 neurons N = 6, n = 9, Vehicle L5 neurons N = 9, n = 11, 6-OHDA L5 neurons N = 7, n = 11. *p ≤ 0.05. **p ≤ 0.01. ***p ≤ 0.001.
In L5, 6-OHDA animals showed a significant increase in both magnitude and frequency of sEPSCs onto L5 excitatory neurons (Fig. 8D1, Kolmogorov–Smirnov test p < 0.0001; Fig. 8D2, Kolmogorov–Smirnov test p < 0.0001), and a comparable increase in spontaneous inhibitory drive (Fig. 8D3, Kolmogorov–Smirnov test p < 0.0001; Fig. 8D4, Kolmogorov–Smirnov test p < 0.0001). These data suggest that, while loss of midbrain dopamine alters synaptic activity onto L5 pyramidal neurons in M1, these shifts are balanced, resulting in no net effect on input–output transformations.
Discussion
Mthal (Planetta et al., 2013; Du et al., 2018) and M1 (Lefaucheur, 2005; Calabresi and Di Filippo, 2015; Guo et al., 2015) have been previously implicated as sites of dysfunction in PD. The accepted view of PD pathophysiology predicts that dopaminergic cell loss triggers widespread shifts in synaptic transmission along the motor pathway, resulting in abnormal thalamocortical excitation (Albin et al., 1989; Blandini et al., 2000; Barroso-Chinea et al., 2008; Braak and Del Tredici, 2008). However, the impact of dopamine loss on the synaptic physiology of the Mthal-M1 pathway had not been directly tested. Our findings corroborate these predictions and provide the first synaptic and circuit mechanisms demonstrating how cell loss within midbrain dopamine centers influences thalamocortical activation of M1 and its output.
Layer-specific reduction in Mthal-M1 drive
6-OHDA lesions of the SNc diminish Mthal drive onto excitatory L2/3, but not L5, excitatory neurons, indicating that the impact in M1 is layer-specific. This distinction is important when considering the differential role of neurons in L2/3 versus L5 in this cortical region. In vivo imaging studies show that while both L2/3 and L5 neurons are active during movement, the activity of L2/3 excitatory cells is modulated by shifts in sensory experience, while L5 neuron activity correlated with the magnitude of motor response (Huber et al., 2012; Masamizu et al., 2014; Heindorf et al., 2018). The evidence that L2/3 receives major input from sensory cortices (Hooks et al., 2013), while L5 contains a high density of pyramidal tract and corticostriatal neurons (Oswald et al., 2013; Shepherd, 2013; Jara et al., 2014), suggests that L2/3 neurons are wired to integrate sensory cues into the motor response, while L5 activity underlies the “output” function of M1. Our data indicate that reduced Mthal drive to L2/3 in 6-OHDA animals may impair sensorimotor integration in M1, while Mthal drive within the output layer is preserved.
Analysis of Mthal-evoked EPSC amplitudes in L2/3 revealed a reduction in the CV with no change in VMR. These parameters provide insights into the site of expression and possible mechanisms underlying a change in synaptic strength (Brock et al., 2020). A previous study reported that decreased EPSC amplitude, coupled with increased CV and unchanged VMR, indicates that reduction of the number of release sites is the likely mechanism (van Huijstee and Kessels, 2020). According to this analysis, the reduction in Mthal-EPSC amplitude in L2/3 following midbrain dopamine depletion depends on presynaptic changes.
Loss of dopamine and Mthal-M1 synapse dynamics
Short-term plasticity regulates the response of cortical neurons to fast, patterned stimulation (Hennig, 2013). Synaptic depression is a form of short-term plasticity common to driver thalamocortical pathways (Sherman, 2007), which aids in cortical processing by decorrelating coinciding inputs to reliably follow patterned activity (Blackman et al., 2013). The magnitude of short-term depression of synaptic responses is influenced by changes in release probability, receptor desensitization, or calcium dynamics (Fioravante and Regehr, 2011). Our results show that 6-OHDA lesions of the SNc lead to increased synaptic depression at Mthal synapses onto M1 pyramidal neurons in both L2/3 and L5. In L2/3, this phenomenon was accompanied by decreased 1/CV2, further supporting the interpretation that in this layer, the shift in short-term dynamics was because of a presynaptic change. In L5, the increased synaptic depression occurred independent of any changes to 1/CV2, suggesting that a different mechanism, likely postsynaptic, may be driving the increase in short-term depression at these synapses (Schneggenburger et al., 2002). In both layers, the changes in short-term depression induced by loss of midbrain dopamine were frequency-dependent as they were detected at 10 Hz, but not at 5 Hz. Together, these results demonstrate that midbrain dopamine loss alters thalamocortical synaptic dynamics, an effect that is likely to profoundly affect M1 responsiveness to patterned stimuli.
Mthal-M1 PV+ activation is preserved in PD model
Thalamocortical activation of PV+ interneurons provides the primary source of feedforward inhibition onto M1 pyramidal neurons. Feedforward inhibition is important for sculpting the tuning of adjacent pyramidal neurons (Merchant et al., 2008), the initiation of movements (Estebanez et al., 2017), and motor learning (Chen et al., 2015). We report that dopamine depletion did not alter Mthal drive or short-term plasticity at Mthal synapses onto PV+ interneurons. In addition, there was no significant difference in the input–output transformation of Mthal inputs, and no substantial changes in PV+ neurons excitability within the linear range of the input–output curve. Interestingly, for large input currents, L2/3 PV+ neurons showed depolarization block and decreased maximum firing frequency, suggesting that 6-OHDA has reduced the ability of PV+ neurons to track activity at high frequency. This effect is comparable to that previously reported for L2/3 pyramidal neurons (Swanson et al., 2021), pointing at impaired circuit excitability in response to high-frequency patterns of activity. In a low-frequency range, in L2/3, reduced thalamocortical activation of excitatory neurons is coupled with preserved activation of neighboring PV+ interneurons. This combination of effects is paired with no changes in either pyramidal or PV+ neurons in L5. These results point to possible alterations of M1 circuit computations because of differential activation of excitatory and inhibitory neurons by Mthal afferents in distinct activity regimes.
Preserved Mthal-evoked output of M1 pyramidal neurons
6-OHDA lesions have caused imbalances between excitatory and inhibitory circuits in other motor areas, including the basal ganglia. Like Mthal input to the cortex, glutamatergic cortical afferents in the striatum synapse onto both MSNs and PV+ cells (Ramanathan et al., 2002), the latter mediating powerful feedforward inhibition onto neighboring MSNs (Gittis et al., 2010). In a 6-OHDA model of PD, the activation of corticostriatal MSNs was affected by dopamine loss, while the activity of striatal fast-spiking interneurons (putatively PV+) was unchanged worsening an imbalance between striatopallidal and striatonigral (indirect and direct pathway, respectively) neurons (Mallet et al., 2006). Furthermore, recent work showed selective changes in thalamostriatal feedforward drive onto MSN neurons in the 6-OHDA model of PD (Parker et al., 2016), similar to the selective thalamocortical changes we observed in M1. The effects of 6-OHDA in striatal circuits have been studied more deeply compared with M1. In addition to the analysis of thalamostriatal and corticostriatal inputs, examination of local striatal circuitry reported a 6-OHDA-induced increase in the connectivity of fast spiking neurons and MSNs of the indirect pathway (Gittis et al., 2011). The experimental findings have informed computational models predicting that the imbalanced activation of striatal direct and indirect pathways depends diminished coordination of activity between fast spiking neurons and MSNs (Damodaran et al., 2015).
In M1, the information about the functional and circuit effects of loss of midbrain dopamine is not as extensive. Similar to the striatum, we observe an imbalance in Mthal activation of pyramidal neurons and PV+ inhibitory neurons, the primary drivers of feedforward inhibition. Interestingly, 6-OHDA did not significantly alter the ability of pyramidal neurons and PV+ neurons to generate action potentials in response to single Mthal stimuli, suggesting the engagement of compensatory mechanisms that may provide stability in the face of reduced Mthal drive to L2/3 neurons. Our previous work excluded changes in intrinsic excitability of pyramidal neurons in either layer as the underlying compensatory mechanism (Swanson et al., 2021). Reduced Mthal drive onto PV+ neurons or reduced excitability of PV+ neurons are also excluded by the results we report here, at least for firing frequencies < 50 Hz.
Analysis of spontaneous excitatory and inhibitory currents onto pyramidal neurons revealed modulation of both components of synaptic drive by 6-OHDA. In L2/3 pyramidal neurons, the amplitude of both sEPSCs and sIPSCs was reduced. The decrease in sEPSCs is likely dependent on the reduced Mthal input but could also include changes in activity of intracortical excitatory synapses. The reduction in inhibitory drive is unlikely to depend on reduced firing of PV+ neurons, as they show depolarization block only when activated at high frequency. However, other factors, including plasticity at inhibitory synapses onto pyramidal neurons from PV+ neurons or other GABAergic neurons in M1, may contribute to this effect. The coordinated reduction in excitatory and inhibitory drive may explain the preserved E/I ratio of Mthal-evoked responses onto L2/3 neurons. In high-frequency regimens, the 6-OHDA-induced depolarization block for both pyramidal neurons (Swanson et al., 2021) and PV+ neurons (Fig. 6) in L2/3 of M1 may alter the recruitment of intracortical circuits in the superficial layers, affecting cortical computations.
In contrast, in L5, Mthal evoked EPSCs were not affected by 6-OHDA, although there was a frequency-dependent increase in short-term depression that could reflect altered synaptic transmission. Consistent with this possibility, analysis of spontaneous synaptic drive revealed a comparable increase in excitatory and inhibitory synaptic drive onto L5 pyramidal neurons. This effect did not alter the E/I balance of evoked responses to Mthal single stimuli. Thus, loss of midbrain dopamine results in impaired Mthal activation of the superficial layers of M1, while inducing compensatory changes that provide stability to the E/I ratio of Mthal-evoked responses, at least for single stimuli and possibly to patterned stimuli at low frequencies. The changes in short-term dynamics of Mthal-EPSCs selectively onto pyramidal neurons in superficial and deep layers and the occurrence of depolarization block in L2/3 pyramidal and PV+ neurons at high frequencies of activation point to a loss of robustness in M1 circuit computations following loss of midbrain dopamine.
In conclusion, the engagement of M1 neurons is crucial for many aspects of movement, including action observation (Vigneswaran et al., 2013), motor learning (Molina-Luna et al., 2009; Hosp et al., 2011; Kida et al., 2016), and skilled movement execution (Kaufman et al., 2013; Ueno et al., 2018). Studies focused on the circuit consequences of PD predicted that basal ganglia dysfunction leads to abnormal synaptic transmission along the Mthal-M1 pathway. Our findings that dopamine depletion shifts the drive and dynamics of Mthal-M1 synapses are a novel contribution to the involvement of M1 in PD pathophysiology. Further, we found that, while the thalamocortical engagement of PV+ interneurons remained intact, intracortical inhibitory synaptic drive was altered, an effect that opens new questions concerning the role of M1 inhibitory circuits during movement in a Parkinsonian state. This study corroborates existing models of PD and emphasizes how a loss of dopamine impacts the output center for voluntary movement.
Footnotes
This work was supported by the Hartman Foundation for Parkinson's Disease. We thank Drs. Craig Evinger, Joshua Plotkin and Alfredo Fontanini for valuable feedback and discussions.
The authors declare no competing financial interests.
- Correspondence should be addressed to Arianna Maffei at Arianna.maffei{at}stonybrook.edu
