Discussion
The present girl came to our observation at the age of 3 months for severe epileptic seizures, started 3 weeks early with episodes of loss-of-contact of brief duration, then evolved in severe drug-resistant epilepsy mainly characterized by spasms in flexion; brain MRI disclosed a complex brain malformation, with thin corpus callosum, polymicrogyria, cortical dysplasia, heterotopias and asymmetric ventricles. These findings, together with a chorioretinal atrophy and optic disc coloboma, allowed to perform a clinical diagnosis of Aicardi syndrome (Aicardi, 2005). Submitted to a NGS-gene panel for epileptic encephalopathies and brain malformation syndromes, the girl was found as harboring two variants in trans (compound heterozygosis) of DCHS1 gene: c.7408G>T: and c.3512G>A. In a in silico evaluation, both showed pathogenic significance, even with different grade of pathogenicity, as demonstrated by the analysis through VarCards server, which reports the results from 23 distinct prediction tools, including SIFT, Polyphen2 and CADD (Varcards Database, 2020). These variants are unlikely to affect splicing (only one splice site is altered by the variant c.7408G>T) and are both missense variants, therefore unlikely to result in a truncated protein. However they could affect stability and localization of the protein, or its interaction with other proteins of the Hippo-pathway; in this aspect, western blots and/or immunofluorescence microscopy would be, in future, helpful to clarify more the pathogenic role of such variants.
DCHS1 gene encodes a transmembrane cell adhesion molecule that belongs to the protocadherin superfamily. DCHS1 is a ligand for FAT4, which is another protocadherin; DCHS1 and FAT4 form an apically located adhesive complex in the developing brain, but their functions encompasses also an important role in cell-cell adhesion of fibroblasts, which is thought to be necessary for wound healing (Cappello et al, 2013). It has been shown that mutations in genes encoding the receptor-ligand cadherin pair DCHS1 and FAT4 lead to periventricular neuronal heterotopia, both in humans and in experimental mice models, with reduced expression of Dchs1 or Fat4 (Beste et al, 2016). In particular, within mouse embryonic neuroepithelium, it has been observed an increased progenitor cell numbers and a reduction of their differentiation into neurons, resulting in the heterotopic accumulation of cells below the neuronal layers in the neocortex, reminiscent of the human phenotype. A concurrent knockdown of Yap, a transcriptional effector of the Hippo signaling pathway, was also observed, thus confirming that Dchs1 and Fat4 proteins are upstream of Yap as key regulators of mammalian neurogenesis (Cappello et al, 2013) (Beste et al, 2016).
DCHS1 gene is the causative factor of a rare (less than 20 patients reported so far) and complex malformation syndrome, Van Maldergem type 1, which is characterized, in its classical form, by intellectual disability, typical craniofacial features, auditory malformations, hearing loss, skeletal and limb malformations, brain abnormalities with periventricular neuronal heterotopia and other variable anomalies (van Maldergem et al, 1992) (Sotos et al, 2017) (Ulubas et al, 2018) (Ivanovski et al, 2018). The function ofDCHS1 gene product in fibroblast adhesion can explain why some allelic variants of this gene have been found in a familiar form of mitral valve prolapse, with high penetrance (Durst et al, 2015).
An exact cause of Aicardi Syndrome has never been precisely elucidated and its pursuit is still matter of debate (Wong et al, 2018). In the last 15 years, many studies have suggested a potential role of X-chromosome (Xp22) or other autosomal (1p, 3q, 6q, 12q) translocations. At the same time, male (46XY) patients have been reported as affected by severe microcephaly (that is not a distinctive sign of the disease), thus suggesting that the (possible) mutation in X-chromosome may not be always lethal but (in some cases) related to a very severe phenotype in males or that other genes may be involved in the pathogenesis of the disease (Chappelow et al, 2008). Modern array-CGH studies, however, have failed in reporting (small or large) CNVs in X-Chromosome, with only one patient out of a total of 156 subjects found to have a de novo 157-kb deleted region in Xq25, in which no known codifying exons of any known gene were found (Wong et al, 2018).
A monogenic origin of the disease has been also investigated with potential candidate genes. In particular FLNA gene, which encodes for a protein (Filamin A) involved in brain cytoskeleton organization and whose mutations are related to brain heterotopias; however, no disease-causing mutations for this variants have been found in AS patients (Anderson et al, 2009). Another gene whose variants were potentially associated with AS was CDKL5 , which plays a role in synaptic formations and is implicated in the etiology of early onset drug-resistant epilepsies (Nemos et al, 2009), but a genetic analysis of this gene in 10 patients affected by AS gave negative results. With the development of NGS gene panels and the use of whole exome sequencing (WES) or Whole genome sequencing (WGS) in large cohort of AS patients, two novel variants in previously unidentified genes have been found (Schrauwen et al, 2015): the genes involved, in two unrelated girls, were TEAD1 (11p15.3) and OCEL1 (19p13.11): however a further Sanger sequencing of these two genes in other 38 AS patients failed to find pathogenic variants of these genes (Wong et al, 2017). Hypomethylation of KCNAB3 (17p13.1) was also proposed among the causes of the disease (Piras et al. 2017), but no consistent finding has been found. More recently, a WES-based study on 11 patients affected by AS has failed to find any significant variant in these patients (the Authors analyzed in particular TEAD1 and OCEL1 ) (Lund et al, 2016).
It is worthwhile to note that the genetic location of the mutated gene of the present patient (DCHS1 ) is close to TEAD1. Given the proximity of TEAD1 and DCHS1 , it cannot be excluded that an imbalance in this region may be among the causes of AS, which could be hypothetically caused by DCHS1 variants and not byTEAD1 mutations. If DCHS1 has a pivotal role, however, it cannot be excluded that some (still unknown) gene regulators or modifiers localized in X-chromosome may contribute to the phenotype.
By contrast, there are more than 50 genes in the region containingDCHS1 and TEAD1 , many of which are involved in autoimmunity, therefore it is unlikely that mutations in all genes in this region will share the same phenotype. A further in-silicoanalysis through geneMANIA tool (Warde-Farley et al, 2017) to assess potential interactions between the two proteins suggested no physical interaction, co-expression, co-localisation, shared pathway, protein domains or genetic interactions. The only evidence found for any links between the two proteins is that they are both involved in the Hippo signaling pathway, causing (similar) structural brain changes.
Another hypothesis is that this patient could be the first report of a new syndrome or a variant of Van Maldergem syndrome, presenting with the same clinical characteristic triad of the AS syndrome, absence of facial dysmorphism and only the dystonic posture of the hand as the clinical symptoms of Van Maldergem syndrome (apart from some brain anomalies present both in AS and Van Maldergem syndrome, as periventricular heterotopia, brain asymmetry, corpus callosum hypoplasia) (table 1). This hypothesis may be supported by the lower pathogenicity of c.7408G>T variant, which resulted “pathogenic” only in 7 out of 23 prediction tools. Lastly, it cannot be excluded that this variants in DCHS1 can be an incidental finding, and the phenotype of the girl caused by other genetic anomalies.
Further researchers of variants in DCHS1 gene in patients affected by Aicardi or Van Maldergem syndromes may contribute to elucidate the role of this and other genes involved in the hippo-pathway in the pathogenesis of the diseases, together with new studies on low-level mosaicism and balanced rearrangements, as well as platforms examining changes at the DNA and chromatin level affecting regulatory regions, in particular for AS (Wong et al, 2018).