Clinical Genetics | 2019
A homozygous frameshift variant in an alternatively spliced exon of DLG5 causes hydrocephalus and renal dysplasia
Abstract
To the Editor The DLG5 gene encodes the 1919-residue DLG5 protein which is composed of numerous functional domains. The protein has been implicated in a variety of cellular structures and processes such as cell-cell contacts, cell polarity, cell migration, neuronal cytokinesis, and epithelial-mesenchymal transition. Some additional insights into its physiological function were obtained by studying Dlg5 knockout mice, which are characterized by defective closure of renal and neural tubes. Despite its large size and apparent multi-functionality, no monogenic conditions have yet been linked to DLG5. We report on a 2-months-old female who was born with obstructive hydrocephalus, atrial and ventricular septal defects, cleft lip and palate, and a renal phenotype including multi-cystic dysplasia with grade I hydronephrosis and increased cortical echogenicity (Figure 1A,B). The healthy parents are first degree cousins, and there is a history of intra-uterine fetal death associated with ventriculomegaly and echogenic kidneys (Figure 1C). Following ethics approval by the Ethical Commission of Rostock University (A2015-0102), informed consent was obtained from all individuals enrolled. Whole exome sequencing did not reveal potentially pathogenic variants in known disease genes. A subsequent extended analysis which focused on homozygous truncating variants in additional genes flagged the DLG5 frameshift variant c.3081_3106 del26 (ENST00000372391, p.Arg1027Argfs10*; genomic coordinates of deleted nucleotides: chr10:7,581,136-7,581,161 [hg19]) which was subsequently confirmed by Sanger sequencing (Figure 1D). Based on the high clinical overlap between the murine Dlg5 knockout and our patient, the suggestive genetic nature of the condition, and the absence of other candidate mutations, we considered the homozygous c.3081_3106del26 variant in DLG5 to be truly causative. Consultation of variation databases [eg, gnomAD] revealed a general rarity of truncating variants in DLG5, and the related probability of loss-of-function intolerance value of pLI = 1.00 is highly indicative of haploinsufficiency. Because these considerations seemed, at first glance, incompatible with the absence of a phenotype in heterozygous carriers of our family, we had a closer look at the consequences of c.3081_3106del26. We noticed that both the variant and the resulting pre-terminal stop codon localize to a large in-frame exon (exon 15 in the major isoform ENST00000372391) which is absent from several other isoforms (eg, ENST00000372388), and does not encode any of the many known functional domains of DLG5 (Figure 1E1). Our variant is therefore predicted to spare a subset of DLG5 transcripts from nonsense-mediated decay (Figure 1E2), and this is in contrast to truncating variants elsewhere in the gene (Figure 1E3). Residual DLG5 function based on a comparatively benign nature of our truncating variant may thus explain that heterozygous carriers in our family are unaffected. In summary, our study adds DLG5 to the already long list of genes associated with congenital syndromic hydrocephalus, and adds renal dysplasia as a rare co-morbidity. It further suggests that the clinical consequences of truncating DLG5 variants depend on localization to one of the many constitutively spliced exons vs to the alternatively spliced exon 15. Unique effects for variants in alternatively spliced exons have been recognized and discussed before, but usually in the context of detrimental effects for specific tissues or organs. Our variant, however, may more appropriately be interpreted to entail reduced severity, and to do so at an organismal level. We propose that this concept may be of potentially broader relevance for genetic disorders.