Whole exome sequencing study identifies candidate loss of function variants and locus heterogeneity in familial cholesteatoma

Abstract

Cholesteatoma is a rare progressive disease of the middle ear. Most cases are sporadic, but some patients report a positive family history. Identifying functionally important gene variants associated with this disease has the potential to uncover the molecular basis of cholesteatoma pathology with implications for disease prevention, surveillance, or management. We performed an observational WES study of 21 individuals treated for cholesteatoma who were recruited from ten multiply affected families. These family studies were complemented with gene-level mutational burden analysis. We also applied functional enrichment analyses to identify shared properties and pathways for candidate genes and their products. Filtered data collected from pairs and trios of participants within the ten families revealed 398 rare, loss of function (LOF) variants co-segregating with cholesteatoma in 389 genes. We identified six genes DENND2C, DNAH7, NBEAL1, NEB, PRRC2C, and SHC2, for which we found LOF variants in two or more families. The parallel gene-level analysis of mutation burden identified a significant mutation burden for the genes in the DNAH gene family, which encode products involved in ciliary structure. Functional enrichment analyses identified common pathways for the candidate genes which included GTPase regulator activity, calcium ion binding, and degradation of the extracellular matrix. The number of candidate genes identified and the locus heterogeneity that we describe within and between multiply affected families suggest that the genetic architecture for familial cholesteatoma is complex.

Fig 1. Analysis overview.

Variants were called using GATK and FreeBayes, then filtered using a hard filter. High confidence variants were selected based on those that were detected by both variant callers. Variants were further filtered according to population allele frequency (retaining those < 1%) and predicted functional impact. Two distinct analyses were performed to identify potentially important genes, pathways, and ontology terms: 1) Identification of genes that have deleterious variants in multiple families; 2) A gene-based mutational burden analysis.

Fig 2. Gene ontology and pathway analysis.

Performed on genes from filtered variants detected by the family overlap analysis in at least one family (A) and the TRAPD mutational burden analysis (B). Colours indicate the database used; (red) CORUM: the comprehensive resource of mammalian protein complexes, (green) GO MF: gene ontology for molecular function, and (blue) REAC: Reactome: the comprehensive resource of mammalian protein complexes. Dot size inversely indicates p-value. Only those terms with a p < 0.01 are shown (hypergeometric test). See S2 Data.

Fig 3. Gene-based mutational burden analysis was performed on individual samples.

Based on allele frequencies from the cholesteatoma and control (gnomAD) cohort variants were split into dominant (A) and recessive (B) groups. The dot colour indicates the number of variants counted across the total cholesteatoma cohort, blue indicates a variant count of 0, and orange with a maximum count of 16. Statistical differences were determined using a two-sided exact Fisher’s exact test (p<0.05). Points labelled with gene names have greater than 5 candidate variants in common across all samples. Refer to S3 Data for a comprehensive list of TRAPD genes.

Fig 4. Common pathways enriched.

Common pathway and ontology terms were found to be enriched for genes containing deleterious variants (p < 0.01; Hypergeometric test) in both the family overlap (red) and TRAPD (blue) analysis. The number of genes with deleterious variants in each pathway or ontology term is shown. Pathway and ontology terms where there is a significant increase in the genes associated with that pathway in the TRAPD analysis compared to the overlap analysis are highlighted (p<0.05; one-sided 2-sample test for equality of proportions with continuity correction).