Lisdexamfetamine Therapy in Paroxysmal Non‐kinesigenic Dyskinesia Associated with the KCNMA1‐N999S Variant

ABSTRACT Background KCNMA1‐linked channelopathy is a rare movement disorder first reported in 2005. Paroxysmal non‐kinesigenic dyskinesia (PNKD) in KCNMA1‐linked channelopathy is the most common symptom in patients harboring the KCNMA1‐N999S mutation. PNKD episodes occur up to hundreds of times daily with significant morbidity and limited treatment options, often in the context of epilepsy. Cases We report 6 cases with the KCNMA1‐N999S variant treated with lisdexamfetamine (0.7–1.25 mg/kg/day), a pro‐drug of dextroamphetamine. Data were collected retrospectively from interviews and chart review. Parent‐reported daily PNKD episode counts were reduced under treatment, ranging from a 10‐fold decrease to complete resolution. Conclusion Our findings suggest that lisdexamfetamine is an effective therapy for PNKD3 (KCNMA1‐associated PNKD). Treatment produced dramatic reductions in debilitating dyskinesia episodes, without provocation or exacerbation of other KCNMA1‐associated symptoms such as seizures.

The KCNMA1 gene (OMIM 600150) encodes the poreforming alpha-subunit of the voltage-and calcium-sensitive "BK" potassium channel. BK channels are broadly expressed in the brain and peripheral tissues, such as smooth muscle and neuroendocrine organs, typically suppressing neuronal excitability. 1 Neurologic abnormalities associated with KCNMA1 mutations or variants of unknown significance (VUS) are described as KCNMA1-linked channelopathy. 1 The neurologic phenotype in KCNMA1-linked channelopathy can include seizures, speech/ language and motor developmental delays, cerebellar atrophy, microcephaly, hypotonia, facial dysmorphisms, visceral malformations, and movement disorders such as ataxia, dystonia, and paroxysmal non-kinesigenic dyskinesia (PNKD). 2 Almost all patients with KCNMA1-linked channelopathy exhibit PNKD, epilepsy, or both. PNKD Type 3 (PNKD3) is defined as PNKD occurring in association with a pathogenic KCNMA1 variant, with or without epileptic seizures (OMIM 609446).
Currently, the KCNMA1-N999S variant [NM_002247.3 c.2984 A>G (p.N999S)] is the most common mutation reported in the literature, occurring in all known cases as a heterozygous de novo mutation. 2 Using de-identified survey data (University of Maryland School of Medicine IRB NHSR Protocols HP-00086440 and HP-00092434) and published literature, 2 we found the majority of patients harboring N999S present with debilitating PNKD (9 out of 12 subjects) and just over half also experience seizures (7 subjects). When introduced into BK channels, the N999S mutation produces strong gain-of-function (GOF) channel activity, 3 supporting the autosomal dominant allele designation. Two other KCNMA1 GOF mutations are also highly associated with PNKD, providing early phenotypic insights into GOF versus loss-of-function variants (LOF) 2,4 and hypothesizing a common mechanistic basis for PNKD3.
Dyskinesias in PNKD3 typically manifest as protracted sudden onset behavioral arrest or drop attacks, with equivocally   Due to phenomenological overlap with cataplexy, these events have also been referred to as cataplexy without narcolepsy. 5 PNKD is often the earliest presenting symptom of KCNMA1linked channelopathy and typically starts before the age of 24 months. In the authors' experience, PNKD episodes are frequently mistaken for seizures due to background epileptiform activity on EEG or co-morbid epileptic seizures, which are also prevalent in KCNMA1-linked channelopathy, 2 despite the lack of confirmatory abnormalities on scalp EEG during the dyskinesias. Treatment options in PNKD3 are limited. Episodes are refractory to a wide range of anticonvulsants, although acetazolamide can reduce but not eliminate the dyskinesias in some patients. 5 In 2019, the family of a young adult reported to one of the authors a 10-year history of daytime remission of dyskinesias after starting lisdexamfetamine, a prodrug of dextroamphetamine. 6 The episodes were phenotypically similar to those in PNKD3 and subsequent genetic testing revealed a KCNMA1-N999S variant. In addition, a prior case report presented in abstract form at the 2018 American Epilepsy society meeting 7 described an individual with a different KCNMA1 variant (N536H, also a GOF mutation) whose PNKDs were successfully treated with dextroamphetamine. 8 These anecdotal observations were shared with neurologists and families via the patient advocacy group KCNMA1 Channelopathy International Advocacy Foundation (KCIAF; www.kciaf.org), social media, and news media. 9 Several affected individuals were subsequently started on lisdexamfetamine by their treating physicians, with consistent, though anecdotal, reports of reduction or remission of PNKD. Here we provide our experience with lisdexamfetamine-responsive PNKD in the setting of KCNMA1-N999S. Study objectives were to describe the efficacy, dosage range, duration of effect, and side effects.

Case Series
Data from six cases treated with lisdexamfetamine were collected in 2020-2021. All cases were previously known to the authors as having reported a reduction in PNKDs after initiation of lisdexamfetamine. Data were obtained retrospectively via combination of chart review of medical records and from family interviews (Tables 1 and 2). Specifically, data on dyskinesia duration, daily frequency, and time of onset and duration of medication responses were obtained from interviews and based on estimates and approximations from parental recollection in an open manner. Subjects' dyskinesias started between 7 and 24 months of age. Lisdexamfetamine (0.71 to 1.25 mg/kg daily) led to a reduction of parentally-observed PNKD in all cases ( Table 1). The reported onset of this effect across cases ranged from 20 to 60 minutes after taking an oral dose, with a duration of 8 to 13 hours. This reported time of onset and duration is consistent with the known pharmacokinetic profile of the active metabolite of lisdexamfetamine, dextroamphetamine. 10 In three cases (Cases A, B, C), there were no observed dyskinesias during this therapeutic time window, down from a baseline of up to 300 daily events. Although most subjects take lisdexamfetamine once daily in the morning, one child (Case D) takes an additional dose of lisdexamfetamine immediately before bed, which led to cessation of previously prolonged, severe nocturnal events.
Side effects, including appetite suppression, insomnia, and irritability were reported (Table 2). These adverse effects did not result in discontinuation of lisdexamfetamine but in most cases prevented further increases in dosage. In four of the six cases, parents reported improvement in one or more non-motor related areas, such as speech, academic performance, concentration, or social skills, based on subjective family observation. PNKD did not resolve with commonly used antiepileptic medications. Two cases had a partial response to acetazolamide, and one case partially responded to a benzodiazepine.
Notably, none of the cases reported new-onset seizures after starting lisdexamfetamine (Table 2), including those with a prior history of seizures or myoclonic jerks. One case with active daily absence seizures (Case A) reported a subjective decrease in observed absence seizures.

Discussion
This case series suggests that lisdexamfetamine is well-tolerated and effective for PNKD3 in children as young as 3 years old. Our observation of the successful use of stimulants in KCNMA1-linked channelopathy corroborates the previously reported case treated with dextroamphetamine. 8 Although the long-term effects of using stimulants in PNKD patients is not yet known, stimulants are routinely used for the long-term management of ADHD in children. Dextroamphetamine is FDA-approved for ADHD and narcolepsy in children age 3 and older. At present, lisdexamfetamine is approved for the treatment of ADHD in patients 6 and older. While the patients in our series experienced some of the expected side-effects for stimulant therapy, such as insomnia and anorexia, none were severe enough to discontinue treatment. In addition, no subject had exacerbation of existing conditions, notably there was no worsening or new development of seizures. Due to the severe neurodevelopmental consequences of PNKD3 at the frequency of up to hundreds of episodes per day, the risk to benefit ratio may be favorable for this treatment option. This data provides support for initiation of a clinical trial to further characterize the efficacy of this treatment regime, which would include a control group that was not possible in this retrospective case series study.
While the mechanism remains to be determined, almost all observations of PNKD3 episodes express as hypokinetic movement. Interestingly, attacks of familial PNKD can be triggered by stimulants such as caffeine. 4 Of note, neither lisdexamfetamine nor dextroamphetamine alter KCNMA1-encoded BK channel activity in heterologous systems 8 (and ALM unpublished data), suggesting that drug effects on motor control may be mediated indirectly and not via specific modulation of BK channel activity. There is superficial similarity to freezing episodes in Parkinson's disease, in that PNKD3 patients have difficulty initiating lower limb movement such as walking, and which suggests a hypothesis that PNKD3 could be linked to basal ganglia dysfunction.
This case series is retrospective and relies on post hoc parental recollection of pre-and post-treatment PKND counts. However, all the individuals in our series had high-frequency, persistent symptoms that were consistent with other individuals harboring KCNMA1-N999S variants and other pathogenic KCNMA1 variants. 2,8 In all cases, the reported decrease in PNKD events was robust and matched the expected pharmacokinetics for lisdexamfetamine, arguing against reporting bias significant enough to invalidate the key observation. Despite the overall effectiveness of lisdexamfetamine, debilitating dyskinesias often continue to occur in the morning prior to medication onset and in the evenings when effectiveness wanes as blood concentrations decrease. As also seen with stimulants used to treat ADHD, insomnia and appetite suppression essentially prevents around-theclock use of lisdexamfetamine in most cases, and thus alternative treatments are needed.
KCNMA1-linked channelopathy is also associated with mild to severe developmental delay and intellectual disability. Several parents reported improvements in school, speech, or social interactions co-occurring with reductions in PNKD attacks. This may in part reflect treatment of a co-morbid deficit in executive function, or may provide evidence that the frequent motor pauses from PNKDs directly interfere with development. However, a more global beneficial effect of lisdexamfetamine on neurological function cannot be excluded.