The 5‐HT2A/2C inverse agonist nelotanserin alleviates L‐DOPA‐induced dyskinesia in the MPTP‐lesioned marmoset

Nelotanserin is a serotonin 2A and 2C (5‐HT2A/2C) inverse agonist that was previously tested in the clinic for rapid‐eye movement sleep behaviour disorder and psychosis in patients with Parkinson's disease (PD) dementia. Its effect on L‐3,4‐dihydroxyphenylalanine (L‐DOPA)‐induced dyskinesia has however not been investigated. As 5‐HT2A antagonism/inverse agonism is a validated approach to alleviate dyskinesia, we undertook the current study to evaluate the anti‐dyskinetic potential of nelotanserin in the 1‐methyl‐4‐phenyl‐1,2,3,6‐tetrahydropyridine (MPTP)‐lesioned marmoset. Parkinsonism was induced in six common marmosets (Callithrix jacchus, three females and three males) that were then chronically treated with L‐DOPA to induce dyskinesia. On experimental days, they were administered L‐DOPA in combination with vehicle or nelotanserin (0.1, 0.3 and 1 mg/kg) subcutaneously, in a randomised fashion. Dyskinesia and parkinsonism were rated post hoc by a blinded observer. In comparison to vehicle, the addition of nelotanserin 0.3 and 1 mg/kg to L‐DOPA diminished peak dose dyskinesia by 47% (P < 0.05) and 69% (P < 0.001). Nelotanserin 0.3 and 1 mg/kg also reduced the severity of global dyskinesia, by 40% (P < 0.01) and 55% (P < 0.001), when compared to vehicle. Nelotanserin 0.1 mg/kg did not alleviate peak dose or global dyskinesia severity. Nelotanserin had no impact on the anti‐parkinsonian action of L‐DOPA. Our results highlight that nelotanserin may represent an efficacious anti‐dyskinetic drug and provide incremental evidence of the potential benefit of 5‐HT2A/2C antagonism/inverse agonism for drug‐induced dyskinesia in PD.

dyskinesia severity.Nelotanserin had no impact on the anti-parkinsonian action of L-DOPA.Our results highlight that nelotanserin may represent an efficacious anti-dyskinetic drug and provide incremental evidence of the potential benefit of 5-HT 2A/2C antagonism/inverse agonism for drug-induced dyskinesia in PD.

| INTRODUCTION
Although L-3,4-dihydroxyphenylalanine (L-DOPA) is efficacious at relieving the motor symptoms of Parkinson's disease (PD), its long-term use almost invariably leads to motor complications (Hely et al., 2005) such as abnormal involuntary movements, that is, dyskinesia, as well as motor fluctuations, which, in its simplest form, consists of a reduction of the duration of the antiparkinsonian action of L-DOPA.Only one drug was granted approval by the US Food and Drug for the treatment of L-DOPA-induced dyskinesia, amantadine extended release (Perez-Lloret & Rascol, 2018), while the atypical anti-psychotic clozapine is regarded as clinically efficacious for the treatment of dyskinesia by the International Parkinson and Movement Disorder Society (Fox et al., 2018), although the risk of agranulocytosis (Alvir et al., 1993) limits the use of clozapine in practice.
Nelotanserin is a 5-HT 2A/2C inverse agonist (al-Shamma et al., 2010;Teegarden et al., 2010).Nelotanserin was tested in three phase II trials in PD patients with dementia (NCT02708186, NCT02640729, NCT02871427) for visual hallucinations and rapid-eye movement sleep behaviour disorder (Axovant Sciences Ltd, Therapies SG, 2015;Axovant Sciences Ltd, Therapies SG, 2016a;Axovant Sciences Ltd, Therapies SG, 2016b).Nelotanserin appeared to be well tolerated based on the results disclosed to date (Stefani et al., 2021) but has shown little efficacy, and its development has since been discontinued.To date, the effects of nelotanserin on dyskinesias have not been reported.We undertook the current study to investigate whether nelotanserin would alleviate L-DOPA-induced dyskinesia in the MPTP-lesioned marmoset to provide incremental evidence of 5-HT 2A antagonism/inverse agonism as an effective anti-dyskinetic approach.As the marmoset has high predictive value of the efficacy of drugs at the phase II level (Beaudry & Huot, 2020;Veyres et al., 2018), if nelotanserin is demonstrated to be an efficacious antidyskinetic molecule in this animal model of PD, it could conceivably be repurposed for the treatment of druginduced dyskinesia.

| Animals
Experiments were conducted in six common marmosets (Callithrix jacchus; McGill University colony; three males and three females), weighing 350-450 g and aged 4-6 years at the time of the experiments.Animals were pair-housed in under conditions of stable temperature (24 ± 1 C), humidity (50%) and a 12-h light/dark cycle (07:15 lights on).Animals were cared for in accordance to a protocol approved by McGill University and the Montreal Neurological Institute-Hospital (The Neuro) Animal Care Committees (Animal Use Protocol 2017-7922), in agreement with the regulations defined by the Canadian Council on Animal Care.They had unlimited access to water and were fed with food (Mazuri ® diet, boiled eggs, pasta, nuts, etc.) and fresh fruits twice daily.Home cages were enriched with primate toys and perches.Before commencing the studies, marmosets were acclimated to handling, subcutaneous (s.c.) injections and transfer to the experimental room and observation cages.The animals used in the current experiments had previously been administered other experimental drugs, but had been drug-free, with the exception of L-DOPA/ benserazide (from this point forward referred to as L-DOPA), for at least 30 days at the beginning of the study.

| Treatment administration
On experimental days, marmosets were administered L-DOPA (15/3.75mg/kg s.c.) in combination with nelotanserin (0.1, 0.3 and 1 mg/kg s.c.) or vehicle (10% DMSO in 0.9% NaCl s.c.), in a randomised within-subject design, in which all animals received all treatments, in a random order.
Following treatment administration, marmosets were placed individually into observation cages (36 Â 33 Â 22 in) containing food, water and a wooden perch and left undisturbed for 6 h.Behaviour was recorded via webcam and analysed post hoc by a trained rater blinded to the treatment.At least 72 h were left between each treatment in any animal.
L-DOPA-induced dyskinesias were rated from 0 to 4, where 0 = absent, whereas 4 = severe, continuous, replacing normal activity, present more than 70% of the observation period.In each 5-min period of assessment, choreic and dystonic dyskinesias were graded separately on a 0-4 scale and the dyskinesia score given reflected the most disabling dyskinesia observed.

| Statistical analysis
To calculate the global parkinsonism and dyskinesia scores, we computed the area under the curve (AUC) of the time courses, following which we performed one-way analysis of variance (ANOVA) followed by Tukey's multiple comparisons test.When making comparisons between different treatment groups, the score of the group against which the others were compared was used as the denominator.Time course data are graphed as the median, while AUC data are presented as the mean with standard error (SEM).Statistical significance was set to P < 0.05.Statistical analysis was performed using GraphPad Prism 8.3.1 (GraphPad Software Inc, San Diego, CA, USA).

| RESULTS
Nelotanserin was well tolerated by marmosets.No adverse events or sedation were observed.
Peak dose dyskinesia corresponds to the time period during which dyskinesia severity was maximal and the reversal of parkinsonism was optimal (here, from 90 to 150 min after treatment administration).As depicted in Figure 1c, nelotanserin also lowered the severity of peak dose dyskinesia (Friedman statistic [FS] = 18.00,P < 0.001, Friedman test).Thus, nelotanserin 0.3 and 1 mg/kg both reduced peak dose dyskinesia, by 47% and 69% (P < 0.05 and P < 0.001, Dunn's post test), compared to vehicle.As it was the case for global dyskinesia, nelotanserin 1 mg/kg diminished peak dose dyskinesia severity to a greater extent than the 0.1 mg/kg dose (by 57%, P < 0.05, Dunn's post test).Nelotanserin 0.1 mg/kg did not significantly alleviate the intensity of peak dose dyskinesia.

| Nelotanserin does not hinder the anti-parkinsonian action of L-DOPA
Nelotanserin did not have any effect on the anti-parkinsonian effect of L-DOPA (Figure 2).Indeed, parkinsonian disability was similar whether vehicle or nelotanserin were added to L-DOPA (F (3,20) = 0.14, P > 0.05, one-way ANOVA).
F I G U R E 1 Dyskinesia in six MPTP-lesioned marmosets treated with L-DOPA in combination with nelotanserin or vehicle.(a) Time course of dyskinesia scores over a 6-h observation period.Each time point represents the median cumulated dyskinesia score over a 30-min observation period (three periods of 5 min, every 10).The maximal dyskinesia score at any time point is 12.(b) Mean (± standard error and individual values) area under the curve (AUC) of the time course of dyskinesia.Nelotanserin 0.3 and 1 mg/kg both significantly reduced the severity of global dyskinesia.(c) Median peak dose dyskinesia scores with individual dyskinesia scores (90-150 min after treatment injection).Nelotanserin 0.3 and 1 mg/kg both significantly reduced the severity of peak dose dyskinesia.The maximal dyskinesia score during this 1-h observation period is 24.*P < 0.05; **P < 0.01; ***P < 0.001.Veh, vehicle.
Here, we have shown that 5-HT 2A/2C inverse agonism with nelotanserin is an effective way to diminish both global and peak dose dyskinesia severity.This antidyskinetic benefit was obtained without compromising the anti-parkinsonian effect of L-DOPA.As we have stated in Section 1, 5-HT 2A/2C inverse agonism is already used in the clinic to treat PD psychosis (Seppi et al., 2019), and there is a wealth of pre-clinical data showing the potential of the approach for the treatment of dyskinesia.Importantly, nelotanserin and especially pimavanserin have been well tolerated by the PD population, and the latter is standard of care for PD psychosis, while the development of the former has been discontinued for PD.That being said, it is encouraging that inverse agonism of 5-HT 2A/2C receptors is well tolerated by PD patients, while being effective for the treatment of dyskinesia, at least in pre-clinical models, as drugs exhibiting this mechanism of action could eventually be developed to address drug-induced dyskinesia in PD.
The mechanisms underlying the anti-dyskinetic action of 5-HT 2A antagonism and inverse agonism remain poorly understood.Autoradiographic binding studies performed in the MPTP-lesioned primate discovered higher 5-HT 2A receptor densities in the motor cortex and striatum in L-DOPA-treated dyskinetic animals (Huot et al., 2012;Riahi et al., 2011).These findings suggest an upregulation of 5-HT 2A receptors along the cortico-striatal pathway.Upon activation, 5-HT 2A receptors enhance glutamatergic transmission (Neuman & Rahman, 1996).Over-active glutamatergic transmission in the striatum is a tenet of our understanding of dyskinesia pathophysiology (Cenci, 2014).Thus, diminishing activity at 5-HT 2A receptors along the cortico-striatal pathway would dampen this excessive glutamatergic transmission within the striatum, thereby reducing dyskinesia.Of note, this mechanism of action has been demonstrated with 5-HT 1A agonists (Dupre et al., 2011) in the 6-OHDA-lesioned rat, but such mechanistic experiments have yet to be conducted with 5-HT 2A agonists and inverse agonists.
A limitation of our experiments is that, whereas the marmosets exhibited behavioural evidence of parkinsonism, the extent of nigro-striatal degeneration was not confirmed at the post-mortem level.In addition, we did not determine the pharmacokinetic (PK) profile of nelotanserin prior to administering it to marmosets.Whether the doses administered would lead to plasma exposure comparable to that achieved after administration in humans is therefore unknown.In addition, whether higher doses of nelotanserin would have elicited higher anti-dyskinetic benefit is also unknown, although we have previously demonstrated that there might be a ceiling to the reduction of dyskinesia conferred by 5-HT 2A blockade et al., 2019).Indeed, in previous experiments conducted by our group in the MPTP-lesioned marmoset with the highly selective neutral 5-HT 2A antagonist pruvanserin, we showed that there likely is a limit to the reduction dyskinesia that can be achieved through 5-HT 2A neutral antagonism, approximately ≈65% (Kwan et al., 2019).Here, the highest dose of the inverse agonist nelotanserin lowered dyskinesia severity by 69%, which is comparable to the results obtained with pruvanserin.Whereas these results suggest that 5-HT 2A inverse agonism may not be more efficacious than 5-HT 2A neutral antagonism at alleviating dyskinesia, two caveats to this comparison need to be discussed.The first one pertains to the absence of PK experiments with nelotanserin in the marmoset, as acknowledged above.The second one relates to the pharmacology of nelotanserin, more specifically its effect at 5-HT 2A receptors.Nelotanserin is a 5-HT 2A/2C inverse agonist; the role of modulation of 5-HT 2C receptors on dyskinesia severity is not fully understood (Huot, Fox, & Brotchie, 2011), and it is therefore unclear whether it influenced, or not, the severity of dyskinesia.
In summary, our results provide incremental evidence that 5-HT 2A antagonism/inverse agonism is an effective anti-dyskinetic strategy that alleviates the abnormal movements without hindering the anti-parkinsonian action of L-DOPA.As nelotanserin has entered clinical trials in PD patients and was shown to be well tolerated, it could potentially be repurposed for the treatment of L-DOPA-induced dyskinesia.

F
I G U R E 2 Parkinsonism in six MPTP-lesioned marmosets treated with L-DOPA in combination with nelotanserin or vehicle.(a) Time course of parkinsonism scores over a 6-h observation period.Each time point represents the median cumulated parkinsonism score over a 30-min observation period (three periods of 5 min, every 10).The maximal parkinsonism score at any time point is 108.(b) Mean (± standard error and individual values) area under the curve (AUC) of the time course of parkinsonism.The global anti-parkinsonian effect conferred by L-DOPA was not altered by nelotanserin.Veh, vehicle.