Abstract
Pederick DT, Richards KL, Piltz SG, Kumar R, Mincheva-Tasheva S, Mandelstam SA, Dale RC, Scheffer IE, Gecz J, Petrou S, Hughes JN, Thomas PQ. Neuron 2018;97:59–66.e5.
X-linked diseases typically exhibit more severe phenotypes in males than females. In contrast, protocadherin 19 (PCDH19) mutations cause epilepsy in heterozygous females but spare hemizygous males. The cellular mechanism responsible for this unique pattern of X-linked inheritance is unknown. We show that PCDH19 contributes to adhesion specificity in a combinatorial manner such that mosaic expression of Pcdh19 in heterozygous female mice leads to striking sorting between cells expressing wild-type (WT) PCDH19 and null PCDH19 in the developing cortex, correlating with altered network activity. Complete deletion of PCDH19 in heterozygous mice abolishes abnormal cell sorting and restores normal network activity. Furthermore, we identify variable cortical malformations in PCDH19 epilepsy patients. Our results highlight the role of PCDH19 in determining cell adhesion affinities during cortical development and the way segregation of WT and null PCDH19 cells is associated with the unique X-linked inheritance of PCDH19 epilepsy.
Commentary
How can an X-linked gene mutation be devastating in the presence of a wild-type allele but be benign in its hemizygous form? This enigma of the gene protocadherin-19 (PCDH19) girls clustering epilepsy (PCDH19-GCE) inheritance has, for many years, puzzled scientists and clinicians alike. Recent work by Pederick and colleagues has now begun to shed light on the potential underlying cellular mechanisms.
PCDH19-GCE, an epilepsy disorder associated with intellectual disability and autism, is caused by mutations in the X-linked gene PCDH19. PCDH19 mutations lead to epilepsy in heterozygous girls, but not in hemizygous boys (1). This inheritance scheme is highly unusual, as in most cases X-linked mutations evoke more severe phenotypes in males than females, or can even be lethal in males, as in the case of Rett syndrome (2). This phenomenon of “female protection” in X-linked disorders is believed to be mediated by compensatory effects of the healthy allele on the second X chromosome present only in females. Autism is an example where a higher prevalence in males is hypothesized to be at least partially caused by the frequent occurrence of autism-causing mutations on the X chromosome, and by the existence of “protective genes” expressed exclusively on X chromosomes (3).
It has long been speculated that the highly unusual X-linked inheritance pattern of PCDH19-GCE is due to a mechanism vaguely described as “cellular interference.” Originally named “metabolic interference,” cellular interference is defined as the detrimental effect of interactions between mutated and wild-type cells in the same organism (4). PCDH19 belongs to the family of protocadherins, a group of cell adhesion proteins important for many aspects of cellular and neuronal function, such as cell migration and axon outgrowth (5). Mutations in PCDH19 are thus likely to mediate defects in cell–cell interaction and communication that may underlie the suspected cellular interference in PCDH19-GCE. Yet, until the recently published work by Pederick and co-authors, no study has provided proof of concept for this hypothesis or revealed the underlying mechanism.
A simple but powerful “mix and match” assay enabled Pederick et al. to test the concept of cellular interference in vitro by analyzing how the expression pattern of PCDH19 and associated protocadherins influences cell sorting. Using two different populations of fluorescently labeled cells the researchers assessed how expression of varying combinations of PCDH19, PCDH10, and PCDH17 affected cell mixing and clustering. The authors showed that functional or complete loss of PCDH19 on one cell population led to incomplete mixing of the two cell populations despite the fact that expression patterns of other protocadherins on these cells remained identical. In contrast, homozygous loss of PCDH19 on both cell populations had no effect. These results provided a mechanism of PCDH19-mediated cellular interference based on alterations in cell adhesion patterns. Previous studies have identified the minimal binding domain of PCDH19 (6), and showed that it interacts with the cell adhesion molecule, N-cadherin (7). However, the structural underpinnings of how interactions of PCDH19 with itself, other protocadherins or even different families of cell adhesion molecules affect intercellular interactions remain largely unknown. Deciphering these details will be important to fully understand the functional consequences of loss of PCDH19.
The above described in vitro assay was an elegant way to quantitatively determine cell interactions, but was carried out in a bone marrow–derived cell line, leaving some uncertainty if similar mechanisms are present in neurons. With a series of in vivo experiments in mouse models the authors convincingly showed that the disturbance of cell adhesion patterns had profound effects on neuronal sorting in the brain. Using CRISPR/Cas9-technology to tag the wild-type PCDH19 protein, the authors found a striking column-like pattern of segregation of PCDH19-positive and -negative cells in the developing cortex of mice. Random X chromosome inactivation in wild-type mice, in contrast, generated a completely different pattern, elegantly shown through heterozygous expression of tagged PCDH19 in a wild-type background. With yet another transgenic mouse, in which both PCDH19 alleles were replaced with two different reporter genes, the authors also demonstrated that homozygous loss of PCDH19 led to the normal cellular distribution pattern observed in wild-type mice. The abnormal cell pattern in the cortex of heterozygous, but not homozygous mice was reflected in an increased occurrence of spike-wave discharge events in cortical EEGs in these animals. In summary, these experiments provided evidence for cellular interference mediated by heterozygous loss of PCDH19 that may underlie the increased brain network excitability in PCDH19-GCE.
The study by Pederick et al. also represents a striking example of the power and shortcomings of mouse models for studying human brain disorders. The in vitro and in vivo results strongly implied that heterozygous loss of PCDH19 during development will lead to morphologic abnormalities in the brain. Supporting this assumption, MRI analyses of brains from girls with PCDH19-CGE revealed defects in cortical folding and thickening in some patients. This was at odds with the previously reported absence of cortical malformations in the mouse model (8). Of note, in contrast to the gyrencephalic (folded) human brains, mouse brains are lissencephalic (smooth). The difference in brain structure and development between the two species may explain the absence of detectable morphologic changes in the mouse model. Therefore, while the mouse model enabled the authors to identify the remarkable consequences of altered cell adhesion on cortical cellular organization, the same model was unsuitable to show the effects on brain morphology, initially suggesting that PCDH19 is not necessary for proper brain development. Moreover, the Pcdh19 mutant mice did not exhibit spontaneous seizures. These issues highlight the importance of careful consideration of strengths and limitations of mouse models for the functional study of epilepsy-causing mutations.
A recent report adds complexity to the expected consequences of altered neuronal sorting in the cortex of PCDH19 heterozygous females: Bassani et al. (9) showed that PCDH19 interacts with GABAA receptors and can influence GABAergic transmission. The clustering of PCDH19-postive and -negative neurons in the cortex of heterozygous girls is therefore likely to disrupt cortical GABAergic microcircuits, which may lead to an imbalance in excitatory and inhibitory brain activity causing epilepsy. It will be interesting to assess GABAA receptor function and expression in relation to PCDH19 expression in the PCDH19-GCE mouse model.
The phenomenon of cellular interference is most likely not limited to rare cases of X-linked diseases. In fact, the majority of non–X-linked epilepsy-associated mutations are disease causing in the heterozygous form. This raises the possibility that some of the phenotypes associated with a certain mutation may be mediated by an indirect effect of cells carrying a mutated protein on neighboring wild-type cells. In line with this hypothesis, mosaic loss of the epilepsy-associated phosphatase and tensin homolog on chromosome 10 in cultured neurons alters synaptic function of adjacent wild-type cells, whereas intracellular signaling and cell body size are only affected in the mutated cells (10). Together, with the findings reported by Pederick et al., this highlights the importance to study cellular interference in disorders like epilepsy, where abnormal cells often exist alongside their healthy counterparts.