Propranolol Relieves L-Dopa-Induced Dyskinesia in Parkinsonian Mice

Background: Parkinsonism is caused by dopamine (DA) insufficiency and results in a hypokinetic movement disorder. Treatment with L-Dopa can restore DA availability and improve motor function, but patients can develop L-Dopa-induced dyskinesia (LID), a secondary hyperkinetic movement disorder. The mechanism underlying LID remains unknown, and new treatments are needed. Experiments in mice have shown that DA deficiency promotes an imbalance between striatal acetylcholine (ACh) and DA that contributes to motor dysfunction. While treatment with L-Dopa improves DA availability, it promotes a paradoxical rise in striatal ACh and a further increase in the ACh to DA ratio may promote LID. Methods: We used conditional Slc6a3DTR/+ mice to model progressive DA deficiency and the β-adrenergic receptor (β-AR) antagonist propranolol to limit the activity of striatal cholinergic interneurons (ChIs). DA-deficient mice were treated with L-Dopa and the dopa decarboxylase inhibitor benserazide. LID and motor performance were assessed by rotarod, balance beam, and open field testing. Electrophysiological experiments characterized the effects of β-AR ligands on striatal ChIs. Results: LID was observed in a subset of DA-deficient mice. Treatment with propranolol relieved LID and motor hyperactivity. Electrophysiological experiments showed that β-ARs can effectively modulate ChI firing. Conclusions: The work suggests that pharmacological modulation of ChIs by β-ARs might provide a therapeutic option for managing LID.


Introduction
Parkinsonism is caused by the lack of dopamine (DA) production or the death of DA-producing cells in the substantia nigra [1,2]. The reduction in DA availability to the motor striatum produces severe motor skill impairments, presenting as postural instability, rigidity, resting tremor, and bradykinesia. L-Dopa, the precursor of DA, increases DA availability to help restore motor function in patients with Parkinsonism. Despite its effectiveness, L-Dopa can produce a secondary hyperkinetic movement disorder called L-Dopa-induced dyskinesia (LID). LID occurs in 30% to 80% of patients with idiopathic Parkinson's disease (PD) [3][4][5], as well as those with genetically-acquired Parkinsonism [6,7]. LID consists of choreic and dystonic movements that drastically reduce a patient's quality of life [8]. The cause of LID is unknown but it is hypothesized to result from synaptic plasticity within the dorsal striatum [9], a brain structure that participates in motor learning and habit formation [10,11]. except during behavioral testing. For terminal procedures, mice were anesthetized using Beuthanasia (270 mg/kg i.p.) prior to sacrifice.
Prior investigations revealed that DAT(dT) mice develop Parkinsonism~2 weeks following their first dT injection and have limited survival up to 30 days [13]. As such, mice slated for behavioral experiments were treated with L-Dopa and the dopa decarboxylase inhibitor benserazide on treatment week (TW) 1, 14 days following their first dose of dT. Thereafter, L-Dopa and benserazide were administered daily until the end of the experiment on TW 18. Low-dose propranolol was administered during TWs 10 to 12, and a higher concentration of propranolol was administered during TWs 13 to 15. A washout period without propranolol was provided on TWs 16 to 18. Electrophysiology experiments in acute brain slices were performed to determine if progressive DA deficiency reduced the spontaneous firing of ChIs and modified responses to β-AR ligands in the absence of L-Dopa. Thirty-day-old Slc6a3 +/+ and Slc6a3 DTR/+ mice were treated with dT (as above) and were sacrificed 14 days after their first injection.
For all studies, the experimenters were blind to genotype. Sample-size estimates were determined by a power analysis based on prior work [23,36,37]. The studies included both male and female mice, but the number of mice prevented post hoc identification of outcome based on sex. During the electrophysiology experiments, "outliers" were defined as cells demonstrating an abrupt change in spontaneous activity or visual movement of the electrode across the surface of the cell. Outlier data were removed from all subsequent analyses. All electrophysiological recordings were replicated in four or more mice. The number of experimental repetitions is indicated in the results section.

Drug Administration
Diphtheria toxin (List Biological Laboratories, Campbell, CA, USA), with a half-life of~18 h, was dissolved in 0.9% saline and administered intramuscularly [13]. L-Dopa (10 mg; Abcam #Ab142497), with an estimated half-life of 1 h, was dissolved in phosphate-buffered saline (PBS) (with 25 mg of ascorbic acid) for a concentration of 1 mg/mL [38]. The decarboxylase inhibitor benserazide (Abcam #Ab143181) was combined with the L-Dopa solution for a concentration of 0.48 mg/mL. Benserazide was used reduce some of L-Dopa's untoward effects such as nausea. It may also prolong the half-life of L-Dopa. Propranolol HCl (Tocris # 0624), with a half-life of 3-6 h, was dissolved in 0.9% saline at concentrations of 0.4 mg/mL and 1.6 mg/mL. Propranolol and L-Dopa/benserazide were administered by separate i.p. injections. The vehicle consisted of an equal volume of 0.9% saline (0.01 mL/gm mouse). Unless specified in the text, chemicals and drugs were obtained from Sigma (St. Louis, MO, USA).

Behavior
Dyskinesia was scored by visual inspection each week and motor function was periodically assessed by rotarod, balance beam, and open field tests (Tables 1 and 2). Mice were acclimated to the behavioral chamber for 90 min prior to treatment. Treatment was provided, and the animal was returned to the behavioral chamber for another 120 min. As the half-life of L-Dopa is~1 h and that of propranolol is 3-6 h, the behavior of each subject was monitored by video recording for 1 min before treatment and then in 1 min intervals every 20 min for 100 min. Recordings were later reviewed and scored by a blinded experimenter. A modified dyskinesia scoring scale was developed, based on mice treated with 6-hydroxydopamine (6-OHDA) [39]. Two types of dyskinesia were identified and scored-limb dyskinesia was quick, repetitive, and uncontrolled limb movements; and general dyskinesia was shaking and involuntary movements of the arms. To separate the effects of novelty from the pharmacological effects of the drug, mice were acclimated to the entire procedure for 2 days prior to the weekly test.

Motor Skill Learning
We used the rotarod test to detect alterations in basal ganglia-mediated motor coordination and learning [23]. We recorded the time that a mouse remained on a 10.5-cm rotating rod (Rotamex 4/8 system; Columbus Instruments, Columbus, OH, USA), accelerating from 5.4 rpm to 40 rpm over 4 min. The maximum time allowed for each trial was 240 s. A grip, where the mouse held on for one full rotation, was considered a fall. Mice were allowed three trials, with a 15-min inter-trial interval, and the results were averaged. To control for motor learning, the rotarod test was conducted before each new treatment phase and prior to their daily treatment. Mice were not exposed to the rotarod prior to the first test. The rotarod was disinfected with Clidox (VWR) following each trial, and no animals were tested during the inter-trial interval.

Motor Coordination
Learning-independent motor coordination was measured using the balance beam [23]. Mice traversed a 60-cm-long, 15-mm-diameter cylindrical rod that was elevated 30 cm above a cushioned table. The mice were placed on one end of the beam and allowed to walk to the other side. Mice that fell were placed back on the beam at the position from which they had fallen and allowed to continue. The average velocity was recorded. The test was performed once a week and prior to their daily treatment. Mice were not exposed to the balance beam prior to the first test. The beam was disinfected with Clidox following each trial.

Novelty Locomotion
We used animal activity monitor cages (San Diego Instruments, San Diego, CA, USA) to detect deficits in locomotion [23]. Four infrared beams separated by 8.8 cm that crossed the width of each chamber were connected to an International Business Machines computer, which recorded the number of times each beam was broken. Locomotor activity was measured by ambulations (two consecutive beam interruptions) summated over 5 min intervals. On each test day, animals were acclimated to individual activity chambers for 90 min. After the injection(s), mice were returned to their respective activity chambers. To avoid measuring an aversive response to handling and pain, and to capture the peak DA release [40], recordings were made for 80 min prior to treatment and for 60 min following the injection. Locomotor ambulations were measured three times each week, and results were averaged over the treatment interval. Following each experiment, locomotor chambers were cleaned with water, dried, and disinfected with Clidox.

Immunohistochemistry
Brains were incubated in 4% paraformaldehyde with 0.1 M phosphate-buffered saline (PBS; pH 7.4) for 24 h, 30% sucrose in PBS for 24 h, and then held in PBS (all at 4 • C). Next, 50 µm coronal sections were prepared with a vibratome and washed × 3 in PBS for 5 min. Sections were incubated in a blocking solution containing PBS with 2% bovine serum albumen and 0.2% Triton X-100 for 1 h before incubation with anti-choline acetyltransferase antibody (1:800 in blocking solution; AB144P, Millipore) overnight at 4 • C. Sections were washed × 4 in PBS for 5 min and incubated in Alexa Flour 488 antibody (1:200; A-11001, Invitrogen) for 2 h at room temperature. Following another round of washes in PBS (5 min × 4), sections were placed on a glass slide and cover-slipped with Fluoromount (Electron Microscopy Sciences) before imaging on a fluorescent microscope (Zeiss).

Statistical Analysis
Values given in the text and in the figures are mean ± standard error (SE). "n" represents the number of mice or cells indicated in the text. Differences in mean values were assessed by paired or un-paired 2-tailed t-tests. A Welch's t-test was used for comparing unpaired data when there was a difference in variance between two population variances. The Holm-Sidak method with alpha = 0.05 was used for data requiring multiple comparisons. Repeated measures (rm) ANOVAs were used for data with multiple groups and a mixed-effects model was employed for data with missing values. Non-parametric data, as determined by an F-test, was compared using two-tailed Wilcoxon matched-pairs signed rank test (for paired groups) or the Kolmogorov-Smirnov (KS) test (for unpaired groups). ANOVA with Tukey's multiple comparisons test was used to detect differences in groups due to treatment. Statistical analyses were performed with GraphPad Prism (v.8.3.1). Differences were considered significant if p < 0.05.

L-Dopa Promotes Limb Dyskinesia in DAT(dT) Mice
We used L-Dopa-responsive, dopamine-deficient Slc6a3 DTR/+ mice to determine if propranolol, a β-AR antagonist, can modify LID, improve motor function, and stabilize ChI firing in the dorsolateral (DL) motor striatum. Male and female Slc6a3 DTR/+ received dT (50 µg/kg, i.m.; half-life~18 h) at 30 and 32 days of life [13,33] and were termed DAT(dT) mice ( Figure 1A). This treatment protocol produces a progressive reduction in striatal DA, ACh, and motor function [13]. Controls were similarly-treated Slc6a3 +/+ mice named WT(dT) mice. Under the treatment paradigm, dT had no effect on WT mice and genotype had no effect on motor performance [13]. As Slc6a3 DTR/+ transgenic mice became symptomatic~2 weeks following their first injection of dT, mice received daily treatment with L-Dopa (25 mg/kg/day, i.p.) [38] and the peripherally-acting dopa decarboxylase inhibitor benserazide (12 mg/kg/day, i.p.) [46] beginning on day of life 42 (TW 1) and terminating on day 168 (TW 18). Mice used for electrophysiological experiments received dT at 30 and 32 days of life and were sacrificed 2 weeks later ( Figure 1B). Treatment had no effect on the animal's ability to feed, as DAT(dT) and WT(dT) mice demonstrated similar weight over the course of treatment ( Figure 1C).
Dyskinesia were assessed weekly. We adapted and modified a dyskinesia scoring system used previously [5,39]-limb dyskinesia was repetitive leg movements whose severity depended on the frequency and number of limbs involved, while general dyskinesia included twitching and shaking of the trunk and the arms (Table 1). Motor function was periodically assessed by the latency to fall from an accelerating rotarod, the time required to traverse a balance beam, and ambulations in the open field (Table 2). Dyskinesia and behavioral measures were averaged over each treatment interval.

Propranolol Reduces Limb Dyskinesia in DAT(dT) Mice
On TW 10, WT(dT) and DAT(dT) mice began receiving additional daily treatments of low-dose propranolol (5 mg/kg/d; i.p.), extending for 3 weeks (TWs 10 to 12). On TW 13, the daily dose of Brain Sci. 2020, 10, 903 8 of 17 propranolol was increased to the target dose of 20 mg/kg/d; i.p. [47]. After three additional weeks of treatment, propranolol was stopped, providing a 3-week wash out (TWs 16 to 18). This treatment paradigm was used to establish dose-dependency on LID and motor activity. Low-dose propranolol had no effect, but the larger dose led to a sharp reduction of limb dyskinesia in the DAT(dT) responders (p = 0.04, ANOVA with Tukey's multiple comparisons test; Supplementary Materials Videos 2 and 3). After propranolol treatment was discontinued, we observed no significant changes in dyskinesia scores across the treatment groups. Overall, treatment had a significant impact on limb dyskinesia in DAT(dT) responder mice, compared to WT(dT) mice (F (4,32) = 22.96; p < 0.001, rm-ANOVA) and DAT(dT) non-responders (F (4,32) = 24.41; p < 0.001, two-way rm-ANOVA, interaction between groups and treatment).

Propranolol Reduces Limb Dyskinesia in DAT(dT) Mice
On TW 10, WT(dT) and DAT(dT) mice began receiving additional daily treatments of low-dose propranolol (5 mg/kg/d; i.p.), extending for 3 weeks (TWs 10 to 12). On TW 13, the daily dose of propranolol was increased to the target dose of 20 mg/kg/d; i.p. [47]. After three additional weeks of treatment, propranolol was stopped, providing a 3-week wash out (TWs 16 to 18). This treatment paradigm was used to establish dose-dependency on LID and motor activity. Low-dose propranolol had no effect, but the larger dose led to a sharp reduction of limb dyskinesia in the DAT(dT) responders (p = 0.04, ANOVA with Tukey's multiple comparisons test; Supplementary Materials Videos 2 and 3). After propranolol treatment was discontinued, we observed no significant changes in dyskinesia scores across the treatment groups. Overall, treatment had a significant impact on limb dyskinesia in DAT(dT) responder mice, compared to WT(dT) mice (F(4,32) = 22.96; p < 0.001, rm-ANOVA) and DAT(dT) non-responders (F(4,32) = 24.41; p < 0.001, two-way rm-ANOVA, interaction between groups and treatment).

β-Adrenergic Receptors Modulate ChI Firing
The mechanism underlying LID is unknown, but animal studies suggest that the disorder develops through an L-Dopa-dependent increase in ACh production that heightens an imbalance between ACh and DA [12,13]. Striatal ACh is predominantly produced by ChIs that are putatively regulated by β-ARs [25,26]. To examine the effect of β-AR on the spontaneous firing rate of ChIs in the DL striatum, we treated 30-day-old Slc6a3 DTR/+ (n = 30) and Slc6a3 +/+ (n = 26) mice with dT and obtained cell-attached recordings of ChIs 2 weeks following treatment ( Figure 6A-C) [13]. Experiments in 6-week-old C57B/6 mice allowed comparisons with our prior work in dopaminedeficient mice [13,38]. WT(dT) mice were similar when measured on TW 1, before and following treatment (arrow) with L-Dopa and benserazide, (B) Ambulations in DAT(dT) and WT(dT) mice were similar across treatment weeks. (C) Ambulations in WT(dT) mice were similar to DAT(dT) responders and non-responders when measured on TW 1, before and after treatment. (D) DAT(dT) responder mice demonstrated a higher number of ambulations following treatment. Ambulations in WT(dT) and DAT(dT) non-responders were similar. (E) Average ambulations over time following treatment. @ p < 0.05, @@@ p < 0.001, t-test or KS test, DAT(dT) responder or DAT(dT) non-responder vs. WT(dT); # p < 0.05, t-test, DAT(dT) non-responder vs. WT(dt) mice; && p < 0.01, two-way rm-ANOVA, for interaction between group(s) and treatment.

Discussion
Parkinsonism is a disabling condition that is produced by a reduction in DA availability. While L-Dopa remains the mainstay of treatment, evidence from human and animal studies suggests that DA deficiency creates a subsequent decline in ACh availability that contributes to the motor deficit [13,19,20]. LID is a major barrier to treatment in some patients with dopa-responsive dystonia [6,7], and eventually occurs in 30-80% of PD patients [3][4][5]. The cause of LID remains unclear, but recent evidence points toward the imbalance in the ACh to DA ratio [12][13][14][15]. If so, targeting the ChI, the main source of striatal ACh might prevent or alleviate this secondary movement disorder.
We used a new transgenic mouse as a model for Parkinsonism. While no animal model replicates the pathology of PD in humans [50], the exposure of Slc6a3 DTR/+ mice to dT produces a progressive reduction in striatal DA-by~90% over 4 weeks-and a decline in motor control [13]. We found that daily treatment with L-Dopa and benserazide rescued the DAT(dT) mice, increasing their survival rate from 2 to 4 weeks [13] to greater than 17 weeks. Weight and locomotor activity were preserved, while performance on the balance beam and rotarod was only marginally depressed when compared with the controls. Under the treatment parameters, a subset of mice (30%) developed LID, characterized by abnormal repetitive limb movements. This is consistent with estimates of LID in humans [3][4][5], but it is possible that the β-AR blocker prevented the emergence of LID in some of the mice. These responder mice also manifested higher locomotor ambulations and lower motor skill learning and coordination on behavioral testing, consistent with the development of LID in humans, where a higher incidence and/or early development of LID is related to the severity and duration of disease [9,32].
LID observed in DAT(dT) mice was more symmetric with relative lack of spinning and unilateral movements that may be seen following unilateral ablation of striatal DA [39]. This LID was suppressed by the higher dose of the β-AR blocker propranolol, suggesting its potential therapeutic use in PD. Consistent with prior studies [47,51,52], propranolol reduced LID without adversely affecting either motor performance or exploratory behavior in the open field, indicating its anti-dyskinetic effect is not due to a reduction in general motor activity. Interestingly, we observed an increase in the locomotor ambulations of DAT(dT) non-responder mice, relative to WT(dT) controls, that was suppressed by propranolol. The reduction of locomotor activity with propranolol was less evident in DAT(dT) responder mice. However, while there was a slight (non-significant) increase in limb dyskinesia during the washout, we observed a significant increase in locomotor ambulations in both responder and non-responder mice during this period, indicating an additional anti-dyskinetic effect of propranolol. Propranolol has been shown to reduce LID in monkeys [52] and rats [47,51] following 6-OHDA. Propranolol also reduced LID by 40% in a small clinical trial [31]. Propranolol may ameliorate LID in a dose-dependent way without affecting L-Dopa's efficacy, while high doses may reduce LID at the expense of exacerbating Parkinsonian symptomatology [47,51,52].
Our study demonstrates the development of LID in an L-Dopa-responsive mouse model of Parkinsonism with progressive DA deficiency. In Slc6a3 DTR/+ mice, dT targets only cells that contain the DAT and progressively destroys both axon terminals in the striatum and DA-producing cell bodies in the substantia nigra [13,33]. This work confirms that spontaneous ChI firing is lower in DA-deficient mice [13], that β-ARs modulate ChI in the DL striatum [25,26], and it demonstrates the utility of targeting adrenergic receptors to restore the balance between ACh and DA. A limitation of the study is our use of young mice rather than the older rodents that have typically been used to analyze the effects of L-Dopa in Parkinsonian mice. We used adolescent mice at time points that corresponded to our earlier work in the DA-depleted dorsal striatum [13,38,53] and may better represent juvenile forms of PD. The electrophysiology data in the slice preparation may differ from plasticity that occurs following treatment with L-Dopa in vivo, presenting an avenue for future research. WT mice without dT treatment were not used as a control, as prior data indicated that the dose of dT used here has no effect on the behavior or physiology of non-transgenic mice [13] and L-Dopa does not promote LID in WT mice [40]. The dose of L-Dopa was similar to that used to rescue DA-deficient transgenic mice that lacked tyrosine hydroxylase [38] and not specifically optimized to induce LID. The doses of propranolol were chosen to allow comparison with other studies [47]. Higher doses might promote further improvement in LID but possibly at the expense of motor function [47,51,52]. We would predict that a β-AR agonist would heighten ChI firing and increase LID in DAT(dT) mice, but β-AR agonists that cross the blood-brain barrier are generally unavailable due to their toxicity in vivo.
Monoamine action plays an essential role in striatal physiology and motor control [11,22]. Although research has often focused on DA, adrenergic receptors also contribute to striatal regulation [25,26,54]. Our data show that β-ARs modify the tonic firing rate of striatal ChIs in the DL striatum. β-AR are known to excite ChIs through a 3',5'-cyclic adenosine monophosphate (cAMP)-dependent, but protein kinase A (PKA)-independent pathway [26], a similar mechanism used by DA receptors to modulate the tonic firing of ChIs that is driven by HCN channels [27] (Figure 7). Our data shows that prolonged DA deficiency induces a larger increase in ChI firing by the β-AR agonist isoproterenol. Prolonged DA deficiency is also known to heighten the response of D1R agonists on ChIs [13], suggesting sensitivity changes in catecholamine receptors or down-stream alterations in adenylyl cyclase, cAMP, or components of the HCN channel. However, additional mechanistic studies will be required as the increase in cAMP production also leads to diverse intracellular changes via the rapid phosphorylation of cAMP response element binding protein (CREB), a transcription factor implicated as a molecular switch underlying long-term changes in brain function [25].
to modulate the tonic firing of ChIs that is driven by HCN channels [27] (Figure 7). Our data shows that prolonged DA deficiency induces a larger increase in ChI firing by the β-AR agonist isoproterenol. Prolonged DA deficiency is also known to heighten the response of D1R agonists on ChIs [13], suggesting sensitivity changes in catecholamine receptors or down-stream alterations in adenylyl cyclase, cAMP, or components of the HCN channel. However, additional mechanistic studies will be required as the increase in cAMP production also leads to diverse intracellular changes via the rapid phosphorylation of cAMP response element binding protein (CREB), a transcription factor implicated as a molecular switch underlying long-term changes in brain function [25].  . β-adrenergic receptors (β-ARs) enhance the spontaneous firing of ChIs through hyperpolarization-activated cation (HCN) channels. The HCN channel opens when the cell becomes hyperpolarized after each action potential and the flow of cations through the channel and into the cell provides rapid depolarization of the ChI, leading to repetitive firing. D1 (D1R) and D2 receptors (D2R) are coupled to the G-protein, adenylyl cyclase (AC) signal transduction pathway that modifies the rate of ChI firing via cAMP, independently of PKA [26,27]. β1-ARs and β2-ARs also enhance cAMP [25,30], which binds to the HCN cyclic nucleotide-binding domain (CNBD) [27] to reduce inhibition of HCN channel gating. HCN channels consist of four α-subunits and the β-subunit, TRIP8b, an auxiliary HCN tetratricopeptide repeat containing Rab8b-interacting protein [27]. TRIP8b modulates HCN surface expression and negatively affects activation of CNBD by cAMP [55].
While β-ARs are located on cells throughout the brain, the antagonist's effect on LID likely occurs within the striatum since an intrastriatal infusion of propranolol acutely reduces LID [51]. Propranolol seems to have little effect on dyskinesia produced by D1 or D2 receptor agonists [47], suggesting a presynaptic mechanism such as modulation of L-Dopa-mediated DA efflux [47], perhaps mediated through ACh regulation of DA release [55,56]. Therefore, in addition to its anticholinergic effect, propranolol's anti-dyskinetic properties may also be mediated via attenuation of L-Dopa-induced extra-physiological efflux of DA [47,55], thereby acting to improve the ACh and DA ratio during treatment of DA deficiency. Amantadine has been used most frequently in the treatment of LID and likely modifies the ACh to DA ratio. While being a weak N-Methyl-D-aspartate (NMDA) receptor antagonist [57], this anticholinergic drug inhibits nicotine currents [58] and modifies DA release [59,60]. This report provides additional evidence that β-ARs may be a potential target for novel treatment of LID. While our study has focused on propranolol, the anti-dyskinetic effect of this drug is likely related to its efficacy as a β-blocker that promotes a secondary anticholinergic action. Propranolol has been traditionally used to treat a variety of neurological and systemic symptoms [29,30], but with broad anticholinergic properties, propranolol may not be suitable for the elderly. However, centrally-acting β-blockers or drugs that might specifically target β-ARs or other G-protein coupled receptors on ChIs might also be expected to improve LID. By reducing intracellular cAMP, these drugs would diminish the abnormal firing rate of ChIs and reduce the heightened release of ACh that occurs in Parkinsonism following treatment with L-Dopa.