ADCY5-related dyskinesia presenting as familial myoclonus-dystonia

Heterozygous ADCY5 mutations, encoding adenylyl cyclase 5, cause familial dyskinesia with facial myokymia (FDFM; OMIM:606703) and benign hereditary chorea (BHC; OMIM:118700), comprising ‘ADCY5-related dyskinesia’ [1‐5]. We report a three-generation family with ADCY5-related dyskinesia manifesting as myoclonus-dystonia, caused by a novel missense variant identified through whole-exome sequencing (WES) of a grandmother-granddaughter pair.

ADCY5-related dyskinesia is an autosomal dominant movement disorder where symptoms worsen with anxiety [9]. We describe a three-generation family with this condition presenting primarily as myoclonus-dystonia. In this study, contemporary sequencing technology and unbiased filtering in an affected grandmother-granddaughter pair identified a causative ADCY5 variant. The T>G substitution causes replacement of hydrophobic methionine with positively charged arginine at a highly conserved residue within the protein’s second cytoplasmic loop (Fig. 1b). No previous evidence of this variant has been found in local or public databases.

Myoclonus-dystonia is a condition characterised by the presence of both myoclonus (rapid muscle contractions of brief duration) and dystonia (sustained and often repetitive twisting movements resulting in abnormal posture) [10]. Abnormal movements tend to affect the upper limbs, trunk and neck, with relative sparing of the lower limbs. Myoclonus-dystonia classically shows a marked improvement of symptoms with alcohol and is also commonly associated with psychiatric comorbidities such as depression, anxiety and obsessive compulsive disorder. Onset is typically in the first two decades of life [11]. Classical myoclonus-dystonia is caused by heterozygous mutations in SGCE, and the condition can be inherited in an autosomal dominant fashion but with greatly reduced penetrance in the case of maternal transmission owing to maternal imprinting of SGCE. However, both myoclonus and dystonia have also been previously described in cases of ADCY5-related dyskinesia (Table 1), suggesting that this gene should also be considered in such cases.

Our proband presented with a childhood-onset jerky movement disorder where both myoclonus and dystonia were prominent features. Her mother displayed similar dystonic movements, while her maternal grandmother displayed a more ataxic phenotype with associated myoclonus. The family’s clinical features were judged to fall within the spectrum of myoclonus-dystonia at the time of investigation, although clearly their two generations of maternal transmission ran contrary to SGCE as a likely cause. There are a number of overlapping features between classical myoclonus-dystonia and ADCY5-related dyskinesia. These include the presence of both jerky myoclonic movements and dystonia, early onset, a predilection for upper body involvement with relative sparing of the legs and a relatively benign course. There are, however, also a number of notable differences. These include the very frequent involvement of the face and associated dysarthria in ADCY5 cases, together with more ataxia and disordered eye movements. It has also been reported that disturbed sleep and axial hypotonia are frequently found in ADCY5 cases [5, 12]. However, neither of these features were apparent in our family.

For further information regarding ADCY5 and its protein product, please see Supplementary Data. Previous reports are listed in Table 1. Chen et al. [1] described a missense ADCY5 mutation (c.2176G>A, p. A726T) co-segregating with disease in a large German-American FDFM kindred [1]. The same group reported a further de novo ADCY5 mutation (c.1252C>T, p. R418W) in two unrelated patients [2]. Carapito et al. [3] reported a patient with early-onset chorea and dystonia who was found to have a de novo splice site ADCY5 mutation (c.2088+1G>A) [3]. Mencacci et al. [4] found four separate ADCY5 mutations in a cohort of BHC patients negative for NKX2-1 mutations [4]. These mutations were c.1252C>T, p. R418W, detected in two unrelated cases; c.2117C>T, p. A706V; and the variant c.1-5G>C in the 5′ untranslated region, which occurred together with a missense variant c.29C>T, p. P10L, although the phase of the two mutations was unknown.

Recently, Chen et al. [5] reported genotype-phenotype correlations in 50 patients with ADCY5 mutations, including patients with somatic mosaicism [5]. A p. A726T mutation caused a relatively mild phenotype of hand and facial dystonia and chorea. A moderate-severe disorder was caused by p. R418W and p. R418Q mutations involving axial hypotonia, limb hypertonia, intermittent dyskinesias, myoclonus and chorea. Notably, a p. M1029K mutation caused severe dystonia, chorea and myoclonus. This is the same residue as that affected in this report. However, the substitution is for lysine instead of arginine in our family. Interestingly, both lysine and arginine are positively charged, and these substitutions may therefore have similar effects on the properties of the resulting mutant protein. In the family reported by Chen et al., symptoms were unusually severe, with developmental delay, contractures, reduced cognition and psychotic depression. In comparison, none of our affected family have exhibited any of these particularly severe symptoms, although II:2 has a relapsing mood disorder. This, along with the variable features exhibited between individuals in our report, suggests that alterations at residue 1029 may not necessarily lead to the most severe forms of ADCY5-related dyskinesia. Instead, it is likely a number of modifying factors such as specific modifier genes that are involved in determining phenotypic severity. Indeed, Chen et al. reported their family’s symptoms lessened with age in at least one case, suggesting the existence of complex regulatory influences.

In the last several months, a number of further reports of ADCY5 mutations have been published. Chang et al. [12] describe five more patients with the p. R418W mutation and two additional patients with p. R418G and p. R418Q mutations [12]. Notably, these patients all presented with motor milestone delay. Dy et al. [13] report two patients with p. R418W (one of whom was already reported by Chang et al.) and also a severely affected child with a c.2080_2088del (p. K694_M696del) deletion [13]. Interestingly, this severely affected nonverbal and nonambulant patient received no therapeutic benefit from propranolol or from any other drug. However, the patients in this study did appear to benefit from deep-brain stimulation (DBS), as did another p.418W patient reported by Meijer et al. [14]. Zech et al. [15] identified seven early-onset generalised dystonia cases with novel ADCY5 mutations (c.2180G>A; p. R727K; c.1378A>T, p. I460F; c.1196C>T, p. P399L; c.1400A>G, p. N467S; c.3177_3182delTGA, p. D1060del; c.3625A>G, p. M1209V) [14]. Notably, the authors of this report highlight that the ADCY5 clinical spectrum may extend to isolated and focal dystonia presentations. Finally, the phenotypic spectrum of ADCY5 disorders has recently been extended by Westenberger et al. [16], who describe two patients with ADCY5 mutations (c.3045C>A, p. D1015E; c.3074A>T, p. E1025V) and who display fetures of alternating hemiplegia of childhood [16].

This report illustrates the importance of considering ADCY5-related dyskinesia in cases that fall within the myoclonus-dystonia spectrum, particularly where SGCE analysis is normal. It also adds to the growing literature on the phenotype of individuals with ADCY5 mutations, which appears to cover an increasingly broad range of movement-related features and which varies greatly in terms of phenotypic severity. The treatment of this condition remains challenging, although DBS shows some promise [13, 14]. Although our proband initially appeared to respond to propranolol, this improvement was not sustained and may in retrospect have been partly attributable to a temporary reduction in anxiety relating to the identification of a definitve diagnosis. We note the original paper describing FDFM reported one affected individual who had modest symptomatic benefit using propranolol but that in at least one other report, there was no response to propranolol [9, 13]. Further research will be needed in order to identify the molecular and neurophysiological mechanisms influencing the phenotypic variability of ADCY5-related dyskinesia and to determine whether an effective pharmacological agent can be found to treat this condition.