CHD8-related disorders redefined: an expanding spectrum of dystonic phenotypes

Human CHD8 is located on chromosomal region 14q11.2. Its main transcript comprises 38 exons encoding for a 290-kDa Chromodomain-Helicase-DNA-binding protein, which acts as an ATP-dependent chromatin remodeler [1]. Through its dynamic interaction with different transcriptional cofactors, which include beta-catenin, CHD7, CTCF, E2F, and p53, the CHD8 protein is involved in upregulating or downregulating the expression of several genes playing pivotal roles in the central nervous system at various stages of development, as well as genes relevant to cell cycle control, epigenetic modifications, and to the proliferation and differentiation of neural progenitor cells and neuroglia [2‐4]. Furthermore, CHD8 dosage loss has been shown to be embryonically lethal in animal models [5]. It is therefore unsurprising that loss-of-function variants in CHD8 have been associated with a spectrum of syndromic neurodevelopmental disorders (CHD8-NDD), displaying a wide range of clinical variability, with a yet unexplained higher penetrance in male subjects, and allelic heterogeneity [6‐8]. Heterozygous truncating variants are the most common cause of disease, followed by missense variants and inframe insertions/deletions. Reported variants span across the whole length of the gene, without any clear genotype–phenotype correlation recognized to date. The vast majority arise de novo in the probands; however, several familial cases have been documented as well, proving that a proportion of individuals at the milder end of the clinical spectrum may preserve their fitness [9].

The core clinical manifestations reported in individuals harboring CHD8 pathogenic variants include autism spectrum disorders (ASD) (76% of recorded patients), psychomotor delay, cognitive impairment (68%), postnatal overgrowth with macrocephaly (52%), musculoskeletal defects (79%), abnormal neuroimaging findings (including ventriculomegaly and cerebellar vermis atrophy) (15%), gastrointestinal problems (53%), and behavioral (88%) and sleep-pattern abnormalities (28%) [8]. Seizures, hypotonia, and dysmorphic facial and skeletal features have also been described. However, the growing implementation of non-targeted high-throughput genomic profiling assays, in the forms of whole-exome and whole-genome sequencing analyses, for the diagnosis of patients not primarily referred for neurodevelopmental abnormalities, has been adding new layers to the complexity of CHD8-related phenotypes. In 2021, our group described two probands with pathogenic CHD8 truncating variants and childhood-onset progressive dystonia, suggesting that the spectrum of CHD8-related disorders may include a dystonic component, whose characteristics of progressive intensification and generalization are capable of exerting a major impact on the quality of life of affected individuals [10].

Here, we expand our case series by describing three additional, previously unidentified patients with variable dystonic phenotypes harboring heterozygous CHD8 (likely) pathogenic variants, broadening the clinical spectrum of CHD8-related movement disorders, and highlighting the implications in terms of diagnosis and clinical management.

In-depth clinical and radiological phenotyping was conducted to assess neurological and systemic abnormalities previously described in CHD8-NDD-affected patients from the literature [8]. Whole-exome and whole-genome sequencing was performed on genomic DNA extracted from peripheral leukocytes using previously validated protocols and analysis pipelines [11, 12]. CHD8 variants were reported according to RefSeq transcript NM_001170629.2; for the predicted protein change, RefSeq NP_001164100.1 was used. Variant interpretation was conducted according to international ACMG/AMP criteria [13]. All patients involved in this case series, or their legal representatives, provided written informed consent as appropriate and in line with ethical guidelines.

Since its identification as a morbid gene in 2012 [6], the role of CHD8 in neurodevelopmental disorders has been thoroughly investigated, with more than a hundred cases affected by pathogenic variants reported to date. An extreme phenotypic variability, in terms both of range and intensity of expressivity of clinical manifestations, has been highlighted to date. In this landscape, movement abnormalities are beginning to emerge as an under-recognized component. In a recent extensive review of CHD8-NDD cases, “involuntary movements” were reported in 17% of patients in whom this aspect could be assessed [8]. However, very few well-characterized instances have so far been described, limiting our understanding of the actual relevance of such manifestations in the overall clinical presentation and management of CHD8-NDD patients. The case series presented in this work validates the observation, made by Doummar et al. in 2021, that CHD8 heterozygous loss-of-function variants can be associated with phenotypes prominently characterized by young-onset dystonia, with mild-to-moderate, in some cases even barely detectable signs of neurocognitive disorders [10]. Despite the limited number of cases, a heterogeneous spectrum of dystonic manifestations has already been recorded, ranging from focal to generalized, from paroxysmal or action-induced to permanent, from apparently isolated to complex with spasticity and cerebral palsy-like pictures. CHD8-related dystonia appears to be progressive: focal, paroxysmal, or action-induced dystonic phenotypes were observed mostly in younger probands, often paired with mild neurobehavioral abnormalities and developmental milestones delay; on the other hand, patients with a sufficiently long documented history of disease, namely both individuals from the Doummar series and Patient 1 from this work, showed increasingly aggravating, painful, and poorly controlled manifestations, including tremor and spasticity. Although dystonic tremor and essential tremor are often difficult to discriminate [14], tremor in CHD8 patients was observed only in the context of generalized dystonia/cerebral palsy-like pictures and in body parts that were also affected by dystonia, and it was therefore classified as a dystonic tremor by movement disorder experts.

Although it is possible that dystonia is an overall rare manifestation across the spectrum of CHD8-related disorders, the low rate of cases so far reported in the literature could also partly be attributable to referral biases, as most case series involved patients primarily referred for ASD, the severity of which could have hindered proper assessment of movement patterns and possible associated clinical comorbidities. Furthermore, as highlighted by our series, some of the most debilitating motor manifestations arose in late childhood/early adulthood, when patients may have already been lost for follow-up. At the same time, our series suffers from a bias on its own, as all our patients were referred primarily for their dystonic manifestations. Further studies on larger, unbiased case series are necessary to determine the actual prevalence of CHD8-related movement disorders. Careful, focused motor system re-assessment of previously published cases may also provide further insight into the topic.

Establishing valid genotype–phenotype correlations in CHD8-NDD has so far proved to be a complex effort, due to the apparently even distribution of causative variants across the whole gene and the overall higher prevalence of truncating variants [8, 15]. This observation seems consistent with our study group, as all identified variants were truncating and were predicted to result in haploinsufficiency, leading to comparable effects on protein expression. Hence, it could be hypothesized that factors other than allelic variations in CHD8, whether genetic or non-genetic, could have more significant relevance in this context.

An interesting observation in this regard can be made about the gender of our CHD8-dystonic probands. As opposed to the overall prevalence of CHD8-related ASD, which is characterized by a solid 2:1 male-to-female ratio despite the autosomal dominant mode of transmission of the causative gene [8], all dystonic CHD8 patients from our case series and the one from Doummar et al. happened to be females. With all due caution relative to the small sample size and the uncertainty ascribable to the possible incomplete phenotyping of male CHD8-NDD patients reported so far in the literature, the available data suggest a possible gender unbalance with regards to CHD8-related dystonia. The hypothesis that dystonia is more common in CHD8 female patients than in their male counterparts would be further corroborated by the anecdotal observation that neither of the two male sons of the adult dystonic patient described by Doummar et al. showed any sign of movement disorder, despite carrying the same causative variant of the mother and exhibiting an otherwise penetrant phenotype, with manifest signs of neurodevelopmental impairment. If confirmed, such an oppositely skewed distribution of ASD and dystonia between males and females in CHD8 patients could represent an interesting model for both the long speculated female protection effect in ASD [16], and the increased prevalence of specific forms of idiopathic dystonia commonly observed in female patients from the general population [17].

The confirmation of movement disorders as a part of the CHD8 phenotypic spectrum, to a point where dystonia can even be the most prominent and earliest manifestation of CHD8-related disorder, prompts a series of further considerations. First it strongly raises the question of whether a neurologic evaluation focused on the assessment of possible movement disorders should be included in the first-line clinical management of all newly diagnosed CHD8 patients and be proposed at least once to the ones already in follow-up. Second, it remarks the potential diagnostic weakness of targeted custom gene panels in the investigation of neurologic disorders, as even apparently isolated forms could mask more complex, syndromic conditions, leading to possible delays in providing the correct diagnosis for the disease and the subsequent counseling to the proband and their family. Finally, it highlights the importance of continuing pursuing a more and more comprehensive neurologic phenotyping of patients harboring genomic variations which have so far been primarily associated with neurodevelopmental disorders. This consideration is especially relevant regarding genes involved in regulatory processes or showing a higher expression in the central nervous system even in postnatal life [18]. The pathophysiological bases of motor system involvement in CHD8-related disorders are still unclear; however, the characterization of important CHD8 targets relevant to movement disorders such as beta-catenin and SCN2A [19], the recurrent identification of cerebellar structural abnormalities in probands [8, 10], and recent animal experiments highlighting a crucial role for CHD8 in extrapyramidal nervous structures’ development [20], represent promising research prospects for investigating the etiological plausibility of this connection.

In conclusion, we characterized three additional unrelated cases of syndromic and non-syndromic movement disorder patients harboring CHD8 pathogenic or likely pathogenic variants, confirming our previous observation that movement disorders should be considered a non-neglectable component of the CHD8-NDD spectrum and emphasizing the increasingly evident connection between movement and neurodevelopmental disorders.