Exome sequencing in rare neurological disorders

Abstract

Neurological disorders are complex traits, manifesting as a range of diverse phenotypes. The current diagnostic approach involves either stepwise testing, which is expensive and time consuming, or targeted next generation sequencing with a limited portfolio of genes. Both of these approaches have a lower diagnostic yield. Whole exome sequencing may be a more advantageous and faster method to discover disease causing gene mutations. This study evaluates the use of whole exome sequencing for diagnostic purposes in neurological disorders. Whole exome sequencing was performed in a heterogeneous cohort of patients with suspected inherited ataxia as an example of a neurological disorder, with the aim to identify candidate gene mutations. The study cohort consisted of 35 affected individuals from 22 randomly selected families of white European descent with no known consanguinity. All common sporadic, inherited and metabolic causes were excluded on routine clinical investigations prior to inclusion in this study. Whole exome sequencing was performed on 30 affected individuals. In-house bioinformatic analysis was based on previously published tools. A variant filtering algorithm excluded synonymous variants and focused on protein altering variants, nonsense mutations, exonic insertions/deletions and splice site variants. Minor allele frequency (MAF) was set at 1% in dbSNP137, 1000 genomes (April 2012 data release) and NHLBI-ESP6500 databases as well as in 286 unrelated in-house controls. Selection of the remaining variants was based on mode of inheritance. The variants were prioritized for brain and nerve cell expression and defined using carefully selected criteria. Genetic analysis was supported further by molecular genetic approaches (Sanger sequencing, reverse transcription PCR, quantitative pyrosequencing, cloning for allelic cis-trans study) and proteomics (Western blotting, immunohistochemistry). Confirmed pathogenic variants were found in 9/22 probands (41%) implicating 6 genes (KCNC3, SPG7, TUBB4A, SLC1A3, SACS and NPC1). Likely de novo dominant TUBB4A mutations were found in two families. In one family quantitative pyrosequencing revealed varying degrees of mosaicism in the mildly affected mother and heterozygosity in the severely affected offspring. In silico analysis further supported pathogenicity of the mutation and revealed that it could potentially disrupt ! ! ii! polymerizations of αβ-tubulin heterodimers. Possible pathogenic variants were identified in 5/22 probands (23%) implicating 5 genes (ZFYVE26, ZFYVE27, WFS1, WNK1 and FASTKD2). A predicted splice site mutation was detected in three members of an autosomal dominant pedigree in the previously described gene ZFYVE27. The ZFYVE27 protein (protrudin) levels were increased approximately 2.5 fold in the cerebellum but not in the frontal cortex of the affected individual. Protrudin is an endoplasmic reticulum (ER) protein and its anomalies have previously been shown to cause ER stress. In this study levels of the master regulator of ER stress, BiP/GRP78, were significantly increased in the patient’s cerebellum, which may indicate the ER pathology. In one family with possible pathogenic compound heterozygous FASTKD2 mutations, the in silico splice-site prediction was validated by sequencing analysis of cDNA clones. Likely de novo compound heterozygous mutations in ZFYVE26 (SPG15) in one family was validated with sequencing of cloned alleles and the result confirmed occurrence of the mutations in trans, therefore supporting their autosomal recessive inheritance. In conclusion, the likely molecular diagnosis in 14 out of 22 families (64%) was defined; a total of 11 genes were implicated. Disease genes previously described in isolated families were validated and the clinical phenotypes of known disease genes was broadened. This study has also demonstrated genetic heterogeneity of hereditary ataxias but shows the impact of exome sequencing in a group of patients difficult to diagnose genetically.