All experiments were carried out in compliance with the USPHS Guide for Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committees of the University of Florida. Dyt1 KI mice (Tor1a^tm1Yql) and their littermate WT mice were prepared and genotyped by PCR (8, 69). Since male Dyt1 KI mice exhibited significant motor deficits in the previous study (8), only males were used for the present experiments. Mice were housed under a 12 h light and 12 h dark cycle with ad libitum access to food and water. All experiments were performed by investigators blind to the genotypes. This study followed the recommended heterogenization of study samples of various ages, and the data were analyzed with age as a covariate (70).

Electrophysiological recordings for spontaneous firing and evoked firing of the striatal ChIs were obtained from 25 WT and 25 KI littermate male mice (6–11 weeks old), as described previously (71, 72).

Mice were anesthetized by the inhalation of isoflurane and decapitated. The brains were rapidly removed and briefly chilled in the ice-cold cutting solution containing (in mM) 190 Sucrose, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 1 CaCl₂, 10 MgCl₂, and 10 D-glucose and was oxygenated with 95% O₂–5% CO₂ (pH 7.35–7.45). In the same ice-cold cutting solution, coronal brain slices (300 µm-thick) were obtained with a Vibratome (LEICA VT 1000S, Leica Microsystems, Wetzlar, Germany). Slices were first incubated on a brain slice keeper (AutoMate Scientific, Inc. Berkeley, CA) with a thin layer of artificial cerebrospinal fluid (ACSF) containing (in mM) 127 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 2 CaCl₂, 5 MgCl₂, and 10 D-glucose, saturated with 95% O₂ and 5% CO₂ (pH 7.35–7.45) at 35°C for 60 min, followed by incubation at room temperature. After a minimum of 60-min incubation, each slice was transferred to a submerged recording chamber with the continuous flow (2 ml/min) of ACSF containing (in mM) 127 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 2 CaCl₂, 1 MgCl₂, and 10 D-glucose and was constantly oxygenated with 95% O₂ and 5% CO₂ (pH 7.35–7.45). All experiments were carried out at 32 ± 0.5°C by a dual automatic temperature controller (TC-344B, Harvard Apparatus) under visual guidance using an inverted microscope equipped with infrared differential interference contrast (IR-DIC) videomicroscopy (Axioskop-FS; Carl Zeiss, Jena, Germany) and a 40× water-immersion lens.

The large neurons with ellipsoid-like soma were selected in the dorsal striatum under the microscope for the electrophysiological recordings. ChIs are further identified by whole-cell recording with step-current injection (73) and, in some cases, by a post hoc immunohistochemistry with an anti-ChAT antibody (71). Patch electrodes had a resistance of 5–10 MΩ when filled with a K-gluconate-based solution containing the following intracellular solution containing (in mM): 112.5 K-gluconate, 4 NaCl, 17.5 KCl, 0.5 CaCl₂, 1 MgCl₂, 5 K₂ATP, 1 NaGTP, 5 EGTA, 10 HEPES, pH 7.2 (270–280 mOsm/l). Biocytin (0.1%) was added to the recording electrode solution to allow post hoc immunohistochemical identification of the recorded cells. While approaching the cell, positive pressure was applied to the patch electrode. The seal (>5 GΩ) between the recording pipette and the cell membrane was obtained by suctioning the electrode. Action potential currents were recorded in the voltage-clamp mode, maintaining an average of 0 pA holding current. After baseline recording, some ChIs were investigated further with bath applications of muscarine [(+)-Muscarine chloride (Sigma Aldrich, M6532-5MG), 10 μM, 90 s (s)] or, THP [DL-Trihexyphenidyl hydrochloride (Sigma Aldrich, T1516-5G), 5 μM, 90 s)] for testing the effect of agonist and antagonist to the muscarinic acetylcholine receptors on the ChIs, respectively. Moreover, some other ChIs were investigated with bath applications of quinpirole [Quinpirole hydrochloride (Sigma Aldrich, Q102-25MG), 10 μM, 2 min] for testing the effect of the agonist to D2R on the ChIs. The effect of quinpirole on spontaneous firing was also examined in a different condition., the spontaneous firing was recorded for 30 s just before adding the drug into the recording bath. The drug was added for 90 s, and recording data in the last 30 s during the drug treatment period were used as “after treatment”. In another condition, the firing was recorded for 3 min at around 10 min after starting the cell-attached recording. Ten µM quinpirole was then applied into the recording chamber for 3 min, followed by a 2 min wash out with ACSF and 3 min recording.

Whole-cell recordings were made by breaking through the membrane. The electrophysiological intrinsic membrane properties (capacitance, input resistance, and time constant) were measured while holding the membrane potential at −60 mV. The liquid junction potential was compensated. Electrode access resistances during all whole-cell recordings were maintained at <25 MΩ. The current steps were injected in multiple 200 pA from −0.8 to 0.6 nA, and the evoked-action-potentials were recorded in the whole-cell recording mode with the current clamp.

To further characterize the electrophysiological properties of μ-opioid receptors and BK channels of the striatal ChIs in Dyt1 KI (n = 5) and control WT littermate male mice (n = 4) at 13–21 weeks-old, the current-voltage relationship of the striatal ChIs was measured by whole-cell recording mode during the voltage ramp (50) with the glass recording electrode filled with a K-gluconate solution containing the following (in mM): 125 K-gluconate, 0.5 EGTA, 19 HEPES, 0.3 GTP, 1 Mg-ATP, 10 NaCl, 2 MgCl₂, 1 CaCl₂. The brain slices were prepared as described above and incubated until recording. Each slice was transferred to a submerged recording chamber with the continuous flow (2 ml/min) of ACSF. The dorsal striatal ChIs were identified from their shape, and the spontaneous firing was confirmed in cell-attached mode, and then the membrane current was recorded in whole-cell recording mode. The membrane potential was held at −70 mV with a voltage clamp. The voltage ramp was applied from −60 to −140 mV in 500 ms, and the membrane current was recorded during the ramp. After multiple recordings, voltage ramp protocols were repeated when the recorded neurons were exposed to 1 μM tetrodotoxin (TTX), 1 μM TTX + 1 μM D-Ala2-MePhe4-Gly[ol]5enkephaline (DAMGO; AdipoGen; AG-CP3-0005V-M005), 1 μM TTX + 1 μM DAMGO + 1 μM Paxilline (PAX; Alomone labs; P-450) sequentially.

The recording data were acquired using pClamp 10 software and further analyzed by Mini Analysis Program (Synaptosoft). Signals were filtered at 5 kHz, and digitized at 10 kHz with a DigiData 1440 (Molecular Devices). Investigators who were blind to the genotypes performed the electrophysiological recordings and analysis. The 30 s (s) before drug treatment and the last 30 s during the drug treatment were used to analyze the drug effect on the spontaneous firing. For the 10-min quinpirole recording analysis, the 3 min before quinpirole treatment and the 3 min after 2 min of quinpirole washout were quantified.

Since the neurons were recorded with the internal solution containing 0.1% biocytin, the recorded neurons were stained with fluorescent-conjugated streptavidin through biocytin-streptavidin binding. After the recordings, the brain slices were fixed overnight with 4% paraformaldehyde in 0.1M phosphate buffer (PB; pH 7.3) and stored in 0.1M PB. The slices were rinsed twice with 0.5% Triton X-100 in 0.02M PB for 10 min and then incubated for 2 hours protected from light in 0.5% (v/v) Triton X-100, 1% (w/v) bovine serum albumin (BSA), 0.02M PB, 0.2% (v/v) streptavidin Alexa Fluor 594 conjugate (Life technologies, S11227). The slices were rinsed with 0.5% Triton X-100 in 0.02M PB twice for 5 min each and then 0.02M PB once. The slices were washed in 10 mM glycine/PBS three times 5 min each and blocked in 2% gelatin/PBS for 15 min, 10 mM glycine/PBS for 5 min, and 0.1% BSA/PBS for 5 min. The blocked slices were incubated in goat anti-ChAT antibody (EMD Millipore, AB144P; 1:100 dilution) in 1% BSA/PBS for 2 h and washed in 0.1% BSA/PBS six times 5 min each. The slices were then incubated with Alexa Fluor 488 donkey anti-goat IgG (H+L) (Invitrogen, A11055; 1:200 dilution) in 1% BSA/PBS for 2 h and washed in 0.1% BSA/PBS six times 5 min each. The slices were mounted on glass slides with Vectashield Hard Set mounting medium for fluorescence (Vector Lab Inc., H-1400) and stored at 4°C overnight. The double-positive cells were confirmed using a ZEISS Axiophot RZGF-1 microscope with 2.5× or ×20 Plan-NEOFLUAR objective lens and FITC filter for Alexa Fluor 488 and Texas Red filter for Alexa Fluor 594, respectively. The dendrites were stained with streptavidin Alexa Fluor 594 conjugate and digitized at ×40 magnification using MBF Bioscience Neurolucida 7 and NeuroExplorer software (MicroBrightFields Bioscience). Sholl analysis (74) was performed using ImageJ software (NIH). Representative images were also taken by Olympus IX81-DSU Spinning Disk Confocal Microscope with ×60 Water immersion objective lens, FITC filter for Alexa Fluor 488, and Texas Red filter for Alexa Fluor 594, respectively.

The spontaneous firing frequency and the drug effects were analyzed by a linear mixed model (lme), generalized linear mixed model (glmmTMB), and emmeans program in R version 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria) for the animal-based nested data, or SAS GENMOD Procedure GEE model. The distribution of the data was checked by R shapiro.test. The coefficient of variation (CV), which is defined by standard deviations (SD) of the interspike intervals (ISI) per mean ISI (73), was analyzed by a generalized linear mixed model (R glmer) for the animal-based nested SD of ISI in gamma distribution concerning the offset log mean ISI. The number of paradoxically-excited ChIs was analyzed by R fisher.test. Wilcoxon rank-sum exact test was performed by R wilcox.test.

Median and confidence interval was analyzed by R MedianCI. The resting membrane property and I _H current were analyzed by R lme with the animal-based nested data. The recording order per ChI was also used as a variable for I _H . The current-step-evoked firing was analyzed by the SAS GENMOD Procedure GEE model with a negative binomial distribution. The membrane currents produced by voltage ramps were analyzed by the SAS GENMOD with gamma distribution concerning age. The number of intersections in Sholl analysis and the soma size of the ChIs were analyzed by Student’s t-test (75). The length of the longest traced dendrites was analyzed by glmmTMB. Significance was assigned at p < 0.05.

The striatal ChIs play a vital role in the pathogenesis of dystonia (76). The striatal ChIs in the Dyt1 KI mice were characterized by the electrophysiological recording of acute brain slices. The spontaneous firing of the striatal ChIs was recorded by cell-attached recording mode with a voltage clamp (WT, 25 cells/14 mice; KI, 27 cells/15 mice). As shown in the representative traces of the striatal ChIs (Figure 1A), the firing frequency was significantly decreased in Dyt1 KI mice compared to WT mice [mean ± standard errors Hz; WT, 5.7 ± 0.3; KI, 4.3 ± 0.4; t(DF: 27) = −2.17, p = 0.039; Figure 1B]. Dyt1 KI mice also showed significantly increased CV [WT, 0.19 ± 0.02; KI, 0.29 ± 0.04; t(27) = 2.02, p = 0.044; Figure 1C]. Increased CV suggests a high irregularity of firing.

D2R is reduced in the striatum of KI mice (23), but it is not known whether it is specifically reduced in striatal ChIs or not. The effect of quinpirole, a D2R agonist, on the spontaneous firing of striatal ChIs was further analyzed (WT, 10 cells/6 mice; KI, 12 cells/6 mice). A representative trace is shown in Figure 2A. The firing frequencies were decreased by quinpirole in both WT [Hz; before, 4.4 ± 0.8; after, 2.6 ± 0.8; t(13) = −2.9, p = 0.014; Figure 2B] and Dyt1 KI mice [Hz; before, 3.2 ± 0.4; after, 2.2 ± 0.4; t(17) = −3.0, p = 0.0087; Figure 2C]. After the quinpirole application, the recording chamber solution was changed to ACSF for more than 2 minutes. Most ChIs showed recovery of the spontaneous action potentials in both WT and Dyt1 KI mice. To compare the quinpirole inhibitory effect, we divided the frequency after the treatment by that before. Quinpirole showed less inhibitory effect in Dyt1 KI mice compared to WT mice [WT, 54 ± 6%; KI, 80 ± 5%; Chi-Square (1) = 4.47, p = 0.034; Figure 2D]. Since the inhibitory D2R is expressed on the striatal ChIs, the results suggest that the inhibition through D2R may be less effective in Dyt1 KI mice. As shown in Figures 2B,C, none of the ten ChIs in WT mice and two out of twelve ChIs in Dyt1 KI mice showed paradoxical excitation, which is a reversed excitatory response to quinpirole. Although paradoxically-excited ChIs were observed only in Dyt1 KI mice, Fisher’s exact test showed no statistically significant difference in the number of paradoxically-excited ChIs between WT and Dyt1 KI mice (p = 0.48). On the other hand, there was no significant difference in the CV after the quinpirole treatment in WT (before, 0.30 ± 0.07; after, 0.34 ± 0.08; z = 0.70, p = 0.49) and Dyt1 KI mice (before, 0.33 ± 0.07; after, 0.42 ± 0.09; z = 0.88, p = 0.38). Moreover, there was no significant difference between WT and Dyt1 KI mice in the effect of quinpirole on the CV (WT, 125 ± 20%; KI, 146 ± 34%; z = 0.60, p = 0.55).

Paradoxical activation of D2R by quinpirole was reported in several mouse and rat models of DYT1 dystonia (77–79). However, the paradoxical activation was not reproduced as detailed above. These published studies examined the quinpirole effect after extended baseline recording. Therefore we repeated the quinpirole experiment to mimic their recording condition. The firing frequency was compared before and after quinpirole treatment (WT, 14 cells/7 mice; KI, 14 cells/8 mice). The representative traces are shown in Figure 2E. There was no significant long-term effect of quinpirole on the firing frequencies in both WT [Hz; before, 2.6 ± 0.4; after, 2.4 ± 0.4; t(20) = −0.37 p = 0.72; Figure 2F] and Dyt1 KI mice [Hz; before, 2.6 ± 0.4; after, 2.8 ± 0.4; t(20) = 0.31, p = 0.76; Figure 2G]. As shown in Figures 2F,G, five out of fourteen ChIs in WT mice and seven out of fourteen ChIs in Dyt1 KI mice showed increased frequency compared to those before quinpirole treatment, which would qualify as paradoxical excitation. Fisher’s exact test showed no significant difference in the number of paradoxically-excited ChIs between WT and Dyt1 KI mice (p = 0.70). These results suggest no significant difference in the appearance of paradoxical excitation cells between the genotypes. The frequency after the treatment was divided by that before the treatment. Shapiro test showed that the ratio data were not normally distributed (WT, p = 0.030; KI, p = 1.7 × 10⁻⁶; all, p = 5.6 × 10⁻¹⁰). Wilcoxon rank-sum exact test showed that there was no significant difference in the long-term effect of quinpirole between WT and Dyt1 KI mice [median, (lower, upper 95% confidence interval); WT, 88, (44, 134); KI, 106, (50, 147); W = 111, p = 0.57; Figure 2H].

Muscarine, a muscarinic acetylcholine receptor agonist, inhibits the striatal ChI firing, and this effect was found to be absent in ChKO mice (35) but was found to be normal Ch2KO mice (36). The effect of muscarine on the spontaneous firing of striatal ChIs in KI mice has not been examined and is explored here. We compared the firing before and after muscarine application (WT, 11 cells/6 mice; KI, 13 cells/6 mice). The firing frequencies were significantly decreased by muscarine in both WT [Hz; before, 4.9 ± 0.5; after, 2.6 ± 0.5; t(15) = −3.15, p = 0.0066; Figure 3A] and Dyt1 KI mice [Hz; before, 4.6 ± 0.7; after, 2.2 ± 0.5; z = −3.2, p = 0.0012; Figure 3B]. To compare the inhibitory effect, we divided the frequency after the treatment by before the treatment. There was no significant difference in the inhibitory effect of muscarine between WT and Dyt1 KI mice [WT, 53 ± 3%; KI, 54 ± 5%; Chi-Square (1) = 0.02, p = 0.90; Figure 2C]. Since inhibitory muscarinic acetylcholine receptors, M2/M4, are expressed on the striatal ChIs, these results suggest that the inhibition through M2/M4 is not altered in Dyt1 KI mice.

THP, a muscarinic acetylcholine receptor M1 antagonist, is effective in treating DYT1 patients and can reverse motor, electrophysiological, and EMG deficits in KI mice (23, 24, 26). The effect of THP on the spontaneous firing of striatal ChIs was examined to explore whether THP reverses the deficits by acting on striatal ChIs. We compared the firing before and after THP application (WT, 11 cells/3 mice; KI, 11 cells/4 mice). There was no significant alteration in the firing frequencies by THP in both WT [Hz; before, 4.6 ± 0.7; after, 3.8 ± 0.7; Chi-Square (1) = 2.76, p = 0.097; Figure 3D] and Dyt1 KI mice [Hz; before, 5.5 ± 1.0; after, 4.1 ± 1.0; Chi-Square(1) = 3.74, p = 0.053; Figure 3E]. To compare the THP effect, we divided the frequency after the treatment by that of before and compared WT and Dyt1 KI mice. There was no significant difference in the effect of THP between WT and Dyt1 KI mice [WT, 83 ± 2%; KI, 71 ± 7%; Chi-Square(1) = 1.80, p = 0.18; Figure 3F]. These results suggest that THP does not affect the spontaneous firing of the ChIs in KI mice.

This series of experiments is to determine whether the membrane property and intrinsic excitability contribute to the reduced spontaneous firing of striatal ChIs in the KI mice or not. After recording the spontaneous firing by cell-attached mode, the intrinsic membrane properties were measured in whole-cell recording mode. The resting membrane property of the striatal ChIs was characterized in the brain slices from 11 WT (29 cells) and 15 Dyt1 KI mice (26 cells). There was no significant difference in the resting membrane potential (RMP), the membrane capacitance, the input resistance (IR), or the time constant between WT and Dyt1 KI mice (Table 1).

The intrinsic excitability of the striatal ChIs in the brain slices was measured with current step injections (Figures 4A–C). The recorded neurons showed typical electrophysiological properties of the striatal ChIs (51). The hyperpolarization and cyclic nucleotide activated cation current (I _H ) was calculated (80). There was no significant alteration in I _H between WT and Dyt1 KI mice (Table 1), suggesting that HCN channels were normal in Dyt1 KI mice.

Moreover, the frequency-current relationship showed that there was no significant alteration in the firing frequencies between WT and Dyt1 KI mice (0.2 nA injection; WT, 34.4 ± 3.5 Hz; KI, 35.0 ± 3.4 Hz; z = 0.32, p = 0.75; 0.4 nA injection; WT, 53.1 ± 3.4 Hz; KI, 51.0 ± 3.4 Hz; z = −0.97, p = 0.33; 0.6 nA injection; WT, 57.3 ± 3.5 Hz; KI, 59.8 ± 3.5 Hz; z = 0.46, p = 0.65; Figure 4D). These results suggest that the intrinsic excitability of the striatal ChIs is not altered in Dyt1 KI mice.

Hyperpolarization of the membrane potential by voltage ramp induces an influx of cation ions through multiple voltage-gated ion channels and causes inward currents. Changes in opioid receptor signaling and BK channels have been reported in DYT1 mouse models (50). We decided to validate the findings in our KI mice. The membrane potential of the striatal ChIs was hyperpolarized by voltage ramp (−60 to −140 mV in 500 ms; Figure 4E) with 1 μM TTX (voltage-dependent Na⁺ channel blocker) in the brain slices from 4 WT (15 cells) and 5 Dyt1 KI mice (16 cells). The inward currents induced by the voltage ramp were recorded in whole-cell recording mode (Figure 4F). The recorded currents with TTX at −140 mV are shown in Figure 4G. There was no significant alteration in the inward currents with TTX between WT and Dyt1 KI mice [WT, −0.81 ± 0.04 nA; KI, −0.80 ± 0.07 nA; z = 0.12, p = 0.91; Figure 4G], suggesting that the overall hyperpolarization-activated ion channel function is normal in Dyt1 KI mice.

Stimulation of μ-opioid receptors produces outward currents by inducing the outflux of potassium ions (K⁺) and inhibiting the influx of calcium ions (Ca²⁺) through the G-protein-coupling mechanism (81). DAMGO stimulates μ-opioid receptors and attenuates the inward currents (82). DAMGO (μ-opioid receptor agonist; 1 μM) was used to analyze the property of μ-opioid receptors during the voltage ramp (50). DAMGO attenuated the inward currents of the hyperpolarized ChI membranes in WT and Dyt1 KI mice (Figures 4E,F). The DAMGO-induced outward currents were calculated by digital subtraction of the currents recorded with TTX and DAMGO from those with only TTX (Figure 4H). The currents recorded with TTX and DAMGO at −140 mV are shown in Figure 4I. There was no significant alteration in the μ-opioid receptor-induced outward currents between WT and Dyt1 KI mice [WT, 0.14 ± 0.04 nA; KI, 0.14 ± 0.03 nA; z = 0.0, p = 1.0; Figure 4I], suggesting normal μ-opioid receptor function in Dyt1 KI mice.

The opening of multiple ion channels induces hyperpolarization-activated inward currents in this recording condition. Among the ion channels, the flow of potassium ions through the BK channel is bidirectional, depending on the membrane potential (83, 84). In the depolarized membrane potential, the opening of the BK channel produces an outflux of potassium ions (K⁺). It causes outward currents during the falling phase of the action potential (85). When the membrane potential is hyperpolarized artificially by voltage ramp, the opening of the BK channel induces an influx of potassium ions (K⁺), which causes inward currents. Therefore, blocking the BK channel attenuates the inward currents of the artificially hyperpolarized membrane potential.

The BK channel activity was characterized by adding 1 μM PAX (BK channel blocker) during the voltage ramp. PAX blocked the BK channel and attenuated the BK-channel-derived inward currents (Figures 4E,F). The BK-channel-derived inward currents were calculated by digital subtraction of the currents recorded with TTX, DAMGO, and PAX from those with only TTX and DAMGO (Figure 4H). The currents recorded with TTX, DAMGO, and PAX at −140 mV are shown in Figure 4J. There was no significant alteration in the inward current through the BK channels between WT and Dyt1 KI mice [WT, −0.10 ± 0.02 nA; KI, −0.12 ± 0.02 nA; z = −0.65, p = 0.52; Figure 4J], suggesting normal ChI BK-channel function in Dyt1 KI mice.

The dendrite structures and soma size of the recorded ChIs were examined by Sholl analysis (74). The morphology of the ChI dendrites was analyzed by digital tracing. Dorsal striatal ChIs were filled with biocytin during the whole-cell patch-clamp recording and labeled with streptavidin Alexa Fluor 594 (Figure 5A). The brain slices were stained with goat anti-ChAT/anti-goat IgG Alexa Fluor 488 to verify cholinergic identity (Figure 5B). The representative dendrites of the striatal ChIs labeled with biocytin/streptavidin Alexa Fluor 594 (Figures 5C,D) and their digitized dendrites are shown (Figures 5E,F). The dendritic branch numbers in 8 ChIs from 6 WT mice and 17 ChIs from 10 Dyt1 KI mice were quantified by Sholl analysis (74) and the ImageJ program (NIH). There was no significant alteration in the number of intersections between WT and Dyt1 KI mice (p > 0.05 at each comparable data point; Figure 5G). Moreover, there was no significant alteration in the length of the longest traced dendrites between WT and Dyt1 KI mice (length ± standard errors; WT: 125 ± 9 μm; KI: 116 ± 6 μm; p = 0.38; Figure 5G). There was no significant alteration in the soma size between WT and Dyt1 KI mice (cell body area ± standard errors; WT: 261 ± 27 μm²; KI: 281 ± 14 μm²; p = 0.47; Figure 5H). The results suggest no significant morphological alteration in the striatal ChI dendrites between WT and Dyt1 KI mice.