Bi-allelic variations in CRB2, encoding the crumbs cell polarity complex component 2, lead to non-communicating hydrocephalus due to atresia of the aqueduct of sylvius and central canal of the medulla

CRB2 variations were first identified in patients with isolated steroid resistant nephrotic syndrome [10] and in patients with congenital nephrosis with cerebral ventriculomegaly [11, 13]. CRB2 was indeed originally largely associated to renal anomalies including renal tubular or glomerular microcysts detected at microscopic examination. Interestingly, one of our case did not show any renal features as already reported in other cases [11], making hydrocephalus one of the major signs associated to bi-allelic CRB2 variations and highlighting CRB2 as one of the major causative gene for hydrocephalus. Among the previously described cases, 5 were reported with Sylvius Aqueduct stenosis as the cause of hydrocephalus, but none of these cases had a neuropathological examination, only MRI. While imaging analysis such as ultrasound scans or MRI are very useful tools to evaluate brain malformations during pregnancy, only neuropathology allows to differentiate Sylvius Aqueduct stenosis from atresia, which are two different malformations with distinct etiologies. Our work thus clarifies the etiology of the cerebral phenotype associated to CRB2 variants, revealing that unlike Sylvius Aqueduct stenosis where the third ventricle is usually enlarged, in CRB2 mutated cases, only the lateral ventricles are dilated, the downstream tract being atretic from the third ventricle to the medulla. This neuropathological feature was highly suggestive of MPDZ variations [20] which have been associated with non-syndromic hydrocephalus due to Sylvius Aqueduct atresia, distinguishing them from cases with L1CAM variations, in which hydrocephalus was associated with Sylvius Aqueduct stenosis. More recently, the same team reported the neuropathological phenotype of fetuses with CCDC88C pathogenic variants. Both described fetuses had atresia of both the Sylvius Aqueduct and the canal of the medulla with additional hydrops of the choroid plexuses. Mesencephalic and medullar lesions were also very similar to our cases [21]. Thus, importantly, such specific features, in particular the absence of third ventricle dilatation, may be seen during prenatal US scan and could be used to guide prenatal genetic testing. In accordance with such similar neuropathological characteristics, MPDZ, CCDC88C and CRB2 are functionally linked, and are crucial for the establishment of apicobasal polarity as well as cell–cell junction formation and maintenance. In this view, our immunohistochemical analysis revealed normal localization and level of proteins of the PAR polarity complex as well as components of the tight and adherens junctions suggesting normal apicobasal polarity and adhesion of the epithelial cells.

In view of the proposed function of CRB2 in preventing neocortical anomalies, we next analyzed neocortical lamination and found no defect in our fetal cases as compared to controls. However, immunohistochemistry analysis using specific antibodies allowing to discriminate specific cortical layers would ensure the absence of cortical lamination defect. In addition, specific ablation of Crb2 in the mouse telencephalon leads to cortical lamination abnormalities that are transient [22], indicating that early transient anomalies cannot be excluded in patients with CRB2 variations that might partially underlie intellectual deficiency and seizures.

Concerning the eye, various functional analyses argue for an important role of CRB2 in the development and maintenance of the neural retina as revealed by the appearance of retinal rosette-like structures in Crb2-/- mice [23, 24]. We thus underwent a histopathological analysis of the eyes in all our cases that revealed no histological anomaly, especially no folds or rosettes of the neural retina. Besides the possibility of redundant functions of human CRB proteins in the retina, we cannot exclude late onset retinal involvement, implying the need for ophthalmological follow-up of CRB2 mutated patients. In the literature, only a few patients with bi-allelic CRB2 variations have been reported with retinitis pigmentosa, and they all share the same homozygous missense variation (p.Arg1249), that was shown in vitro to accelerate epithelial mesenchymal transition with subsequent degeneration of retinal pigment epithelium cells [15]. As this is the only variation reported so far in the intracellular region of the protein, a specific effect may be suspected. Interestingly, when compound with a variation affecting the extracellular domain, this variation leads to tubulopathy without retinitis pigmentosa (Fig. 4) [10]. Recently, a phenotype-genotype correlation has been suggested in view of the clustering of pathogenic variations in exons 8 and 10 in patients presenting with kidney anomalies associated to hydrocephalus, whereas variations in exons 12 and 13 seems to be associated with isolated renal disease [17]. Our data are consistent with this hypothesis, indeed our patients also had variants in exons 8 and 10, but one of them had isolated hydrocephalus.

Fig. 4figure 4

Schematic representation of CRB2 protein showing its known functional domains and the position of all variations previously reported in the literature (above) and the three variations we report in this study (below)

In view of the association of renal anomalies, hydrocephalus/ventriculomegaly and retinal dystrophy, CRB2 has been proposed to lead to a phenotypic spectrum characteristic of a ciliopathy [13]. Ciliopathies refer to a wide range of genetic diseases due to altered function of proteins that localize to motile and/or primary (immotile) cilia or to centrosomes. Because of the specific properties of primary versus motile cilia which are respectively unique, largely ubiquitous and associated with signal transduction, versus multiple, present on highly specialized cells to move fluids or to propel cells, the physiological consequences of defects in motile and primary cilia are different [25, 26]. However, both motile and primary cilia have been involved in hydrocephalus, with different underlying mechanisms [27,28,29]. Here, we observed a priori normal motile cilia harboring the apical side of the ependymal cells lining the central canal of the medulla arguing for a distinct mechanism affecting the formation of the Sylvius aqueduct and central canal of the medulla, although potential anomalies of the primary cilium cannot be excluded [30]. Formation of this canal is going through a complex process with remodeling of the pseudostratified ventricular layer cells which undergo apical constriction, a cell shape change involving CRB2 as well as CCDC88C and MPDZ that have been shown to work cooperatively to drive apical constriction of the neural plate cells during neurulation [4, 5]. CRB2, through its intracellular FERM-binding domain, interacts with actin-binding proteins including moesin [31, 32] and is indeed connected to the actin-myosin network essential for the conversion of the primitive lumen into the central canal of the medulla. Excitingly, CRB2 has recently been proposed to act through interaction with a locally secreted CRB2 variant modifying cell polarity and cohesion of the ventricular layer to allow its remodeling [33]. Thus, CRB2 loss might lead to altered regulation of apical constriction of the ventricular layer that might lead to invagination and rosette formation instead of delamination normally enabling dorsal collapse. Together, these studies and our findings argue for a common pathological mechanism underlying hydrocephalus associated to variations in all 3 genes that indeed lead to very similar neuropathological outcomes.

To note, elevated levels of α-fetoprotein in maternal serum and in amniotic fluid were frequently reported during pregnancy of cases with CRB2 variations as observed here for the fetus 3. This data was not available for the other cases. While unexplained, this biological anomaly might be useful to guide the molecular diagnostic.

Overall, our work highlights the major benefit of performing neuropathological examination in fetuses and accurately describing the phenotype of rare diseases to improve the understanding of the underlying pathological mechanisms. Hydrocephalus associated to bi-allelic variations in CRB2, MPDZ and CCDC88C constitutes a separate pathogenic group of congenital non-communicating hydrocephalus with Sylvius aqueduct and central canal of the medulla atresia that might be linked to disturbed apical constriction of the ventricular layer cells, a complex remodeling process required for the formation of the central canal of the medulla.

Web resources

https://alphafold.ebi.ac.uk

https://www.cgl.ucsf.edu/chimera/

CADD (v1.6): https://cadd.gs.washington.edu

gnomAD: https://gnomad.broadinstitute.org

OMIM: https://omim.org

Polyphen2 (v.2.2.2): http://genetics.bwh.harvard.edu/pph2/

SIFT (v6.2.0): https://sift.bii.a-star.edu.sg/www/SIFT_aligned_seqs_submit.html

UniProtKB: https://www.uniprot.org

https://www.nextprot.org/entry/nx_q5ij48/structures

留言 (0)

沒有登入
gif