Variable Clinical and Molecular Expressivity of PCDH19 Variants and Girls Clustering Epilepsy A disorder of cellular "mosaics"

Abstract

Girls Clustering Epilepsy is the second most common developmental and epileptic encephalopathy. GCE is due to variants in the X chromosome gene PCDH19 and is underpinned by cellular mosaicism due to X-chromosome inactivation (XCI) in females or somatic variant in males. The hallmark feature is that seizures occur in clusters and mainly affect females. At the time of thesis submission, GCE was re-named “X-linked clustering epilepsy” (XCE) to accommodate the growing number of affected male cases. Seizures typically present as generalized tonic-clonic and/or focal, which may evolve to bilateral, tonic-clonic. The clinical profile includes variable cognitive impairment and psychiatric features. The prevalence of these comorbidities and cause of the variable clinical expressivity is unknown. We performed a systematic review and meta-analysis to identify the comorbidities associated with PCDH19 variants and examine phenotype- and genotype-phenotype associations. Data from 38 peer-reviewed original articles were used and included 271 individual cases. We found that seizure onset ≤ 12 months was significantly associated (p = 4.127 x 10-7) with more severe intellectual disability compared with onset > 12 months. We identified two recurrent variants p.Asn340Ser and p.Tyr366Leufs*10, occurring in 25 (20 unrelated) and 30 (11 unrelated) cases, respectively. PCDH19 variants were associated with psychiatric comorbidities in approximately 60% females, 80% affected mosaic males, and reported in nine hemizygous males. Executive dysfunction, and hyperactive, autistic, and obsessive-compulsive features were most frequently associated with PCDH19 variants. We developed a PCDH19 survey to systematically examine the comorbidities identified in our review using standardized neuropsychiatric assessments. The survey was completed by 122/186 (66%) participants diagnosed with GCE or with a confirmed likely pathogenic PCDH19 variant. Executive functions were measured using the Behavior Rating Inventory of Executive Function. Psychiatric comorbidities were assessed via the Social Responsiveness Scale or Social Communication Questionnaire, the Strengths and Difficulties Questionnaire, and the Dimensional Obsessive-Compulsive Scale. Genetic, seizure, and developmental information were also collected. Of the 112 evaluated participants (15 males), there were 70 unique variants. Thirty-five variants were novel and included a newly identified recurrent variant Ile781Asnfs*3. There were no phenotypic differences between published and unpublished cases. Seizures occurred in clusters in 94% individuals, with seizures resolving in 28% at an average age of 17.5 years. Developmental delay prior to seizure onset occurred in 18% of our cohort. Executive dysfunction and autism spectrum disorder (ASD) occurred in approximately 60% of individuals. The ASD profile included features of attention-deficit hyperactivity disorder. Obsessive-compulsive symptomology was observed in 21% individuals. There were no phenotypic differences between heterozygous females and mosaic males. We describe a mosaic male and two hemizygous males with atypical clinical profiles. Earlier seizure onset age and increased number of seizures within a cluster were associated with more severe clinical outcomes. No clinical profile was observed for transmitting males. The penetrance of GCE is incomplete; estimated to be around 80-90%, and might be explained by cellular interference. Cellular interference postulates that the coexistence of PCDH19 wild-type and variant cells would be pathogenic, whereas a homogenous cell population would be tolerated; an idea supported by the presence of asymptomatic PCDH19-negative hemizygous and symptomatic PCDH19-mosaic males. The cellular interference hypothesis was tested through analyses of GCE penetrant and non-penetrant female fibroblast cell lines using assays to determine XCI patterns and relative PCDH19 cDNA expression. Specifically, we hypothesized that XCI and PCDH19 cDNA expression will be skewed towards complete wild-type or variant expression in non-penetrant females. We have shown that XCI patterns do not correlate with relative PCDH19 cDNA expression in fibroblasts, thus invalidating use of this assay to infer PCDH19 expression. No clear association was observed between penetrance in XCE and the degree of variant and wild-type PCDH19 mRNA expression in skin fibroblasts. Although we were able to identify three non-penetrant females with 100% wild-type PCDH19 expression, we were unable to provide support for the mechanism of cellular interference through our finding clinical phenotypes in individuals with markedly skewed XCI. Neuropsychiatric disorders can be very responsive to early intervention; therefore, a better understanding of these comorbidities may help to inform treatment and ultimately lead to better developmental outcomes for individuals affected by GCE. We show that both seizure onset age and activity are associated with clinical outcomes. Clinicians can use this information to inform prognosis and provide targeted intervention and guidance for patients and their families.