A homozygous frameshift variant in an alternatively spliced exon of DLG5 causes hydrocephalus and renal dysplasia
Based on the ambition to provide maximum diagnostic yield, CENTOGENE offers follow up of negative WES/WGS reports in a research setting. Recently, this approach resulted in the association of a special form of congenital hydrocephalus with bi-allelic inactivation of the DLG5 gene. The finding, which represents the definition of a novel rare disorder, was published in Clinical Genetics ahead of print on February 21st.
Clin Genet. 2019 Feb 21. doi: 10.1111/cge.13513. [Epub ahead of print]
Abstract
The DLG5 gene encodes the 1919-residue DLG5 protein which is composed of numerous functional domains, and has been implicated in a variety of cellular structures and processes. Homozygous knockout of the murine homologue Dlg5 results in fully penetrant but variably severe hydrocephalus and renal cysts. Certain DLG5 haplotypes have been suggested to predispose to inflammatory bowel disease, but a DLG5-linked monogenic phenotype has not previously been reported.
We investigated a pediatric female patient diagnosed with obstructive hydrocephalus, renal dysplasia, atrial septal defect type II, and cleft lip and palate. Bilateral ventriculomegaly and echogenic kidneys had been noticed at gestational week 22. The parents are first cousins and there is a history of intra-uterine fetal death due to similar congenital abnormalities. Whole exome sequencing (WES) of the index case in a routine diagnostic setting did not reveal potentially pathogenic variants in known disease genes. An extended analysis of the WES data in a research frame identified the homozygous DLG5 frameshift variant c.3081_3106del (p.Arg1027fs) in DLG5. Sanger sequencing confirmed this finding, and revealed heterozygosity in both healthy parents. We were therefore very surprised to note that data from the Exome Aggregation Consortium (ExAC) strongly suggest that heterozygous presence of a truncating DLG5 variant is not tolerated (probability of loss-of-function intolerance = 1.00). A more detailed analysis, however, revealed that our variant and the resulting pre-terminal stop codon localize to the alternatively spliced in-frame exon 15. The shorter of the corresponding isoforms is therefore predicted to escape nonsense-mediated decay, and this, in turn, implies that certain functions of DLG5 may be retained upon presence of the c.3081_3106del variant.
Our clinical-genetic findings, together with those on the murine Dlg5 knockout, characterize DLG5 as a novel recessive disease gene in humans. Theoretical considerations suggest that heterozygous truncating variants in exons other than exon 15 may be associated with a potentially distinct phenotype. Further studies are necessary to test this hypothesis