The PCDH19 gene is located at Xq22.1 and consists of six exons. The gene encodes a 1148 amino acid protein with typical features of the δ2-protocadherin sub-family, with 23 amino acid signal peptides, six conserved cadherin repeats in the extracellular domain, a transmembrane domain, and conserved motifs (CM1-CM2) in the C-terminal region. The first exon encodes the extracellular and transmembrane domains, as well as a small portion of the C-terminal region. While the rest of the C-terminal region is encoded by exons 2–6, the second, and likely the third exon are subject to alternative splicing. Exons 5 and 6 encode for CM1 and CM2 domains, respectively. More than 80% of the reported CE pathogenic variants are observed in the extracellular domain of the protein encoded by exon 1. Of the reported variants in this region, almost half are in the EC3 and EC4 domains (20% and 23%, respectively). Missense variants are most frequently reported (45%), followed by frameshift (27%), and nonsense variants (20%). In total, 145 unique germline PCDH19 pathogenic variants have been identified in CE, both in large families as well as singleton cases. Most PCDH19 variants are non-recurrent (exclusive to that individual or family) with the exception of p.Asn340Ser and p.Tyr366Leufs*10, which have been reported in 25 and 30 individuals, respectively.
Several mechanisms have been suggested to account for the unusual mode of inheritance. Of these, cellular interference has received the most support. Cellular interference is a mechanism reminiscent of metabolic interference and postulates that random inactivation of one X chromosome in females with a PCDH19 pathogenic variant generates cellular mosaicism in PCDH19-expressing tissue (i.e. co-existence of PCDH19-normal or PCDH19-abnormal cells). Such cellular mosaicism causes the condition by altering cell-cell interactions, function, and therefore neural networks in the brain. Cellular interference is consistent with the clinical finding that males hemizygous for a PCDH19 pathogenic variant in all their cells are typically unaffected as they have only one population of cells albeit with the pathogenic variant, whereas males with somatic mosaicism are affected similarly to heterozygous females.
The identification of affected males who are mosaic for PCDH19, and therefore have a mixture of PCDH19-normal and PCDH19-abnormal cells, strongly supports the hypothesis of cellular interference as the main pathogenic mechanism associated with PCDH19 pathogenic variants. The co-existence of normal and abnormal cells and the proportion of each population in the brain of these males cannot, however, be extrapolated from available tissues i.e., skin fibroblasts or lymphocytes. To establish that cellular interference is the pathogenic mechanism, it is necessary i) to demonstrate that neuronal cells are mosaic, but also that ii) females who are homozygous for PCDH19 pathogenic variants or deletions are also unaffected, akin to hemizygous males. Some support for the first point lies in the findings from a Pcdh19 knockout mouse model. Simultaneous labelling of wild-type Pcdh19 and null Pcdh19 cells in Pcdh19 heterozygous mice revealed a striking pattern of alternating Pcdh19 +ve and Pcdh19 -ve regions. This pattern was particularly obvious in the developing cortex where it resembled “tiger stripes”. Although pathogenesis in cells that express the abnormal allele corresponds to a loss-of-function, cellular interference would result in a gain-of-function at the tissue level, because of abnormal interactions between normal and abnormal cells. This hypothesis supposes that the loss of PCDH19 is compensated for, but by a mechanism that is relatively independent of gender. For the latter, there is yet to be a report of a female homozygous for loss of function pathogenic variant in PCDH19.
Molecular diagnosis
Testing for CE requires sequencing of the PCDH19 gene, either by traditional Sanger sequencing or massively parallel sequencing methodology (panel testing, exome, or genome). If no pathogenic variants are identified, then microarrays for copy number testing i.e., deletions or duplications should be performed in individuals presenting with the typical phenotype or inheritance pattern.
Determining the pathogenicity of novel variants requires the use of in silico tools and segregation in additional family members. In some cases, molecular/functional studies may be required to build evidence that a detected variant is causal of the individual’s condition. Whether the phenotype is consistent with the disease should also be considered.
Repository of the variant and levels of evidence for causality in publicly available international databases like ClinVar https://www.ncbi.nlm.nih.gov/clinvar/ or DECIPHER https://decipher.sanger.ac.uk/ is recommended to aid interpretation of novel variants.