Exome variant prioritization in a large cohort of hearing-impaired individuals indicates IKZF2 to be associated with non-syndromic hearing loss and guides future research of unsolved cases

Abstract

Although more than 140 genes have been associated with non-syndromic hereditary hearing loss (HL), at least half of the cases remain unexplained in medical genetic testing. One reason is that pathogenic variants are located in 'novel' deafness genes. A variant prioritization approach was used to identify novel (candidate) genes for HL. Exome-wide sequencing data were assessed for subjects with presumed hereditary HL that remained unexplained in medical genetic testing by gene-panel analysis. Cases in group AD had presumed autosomal dominantly inherited HL (n = 124), and in group AR, presumed autosomal recessive HL (n = 337). Variants in known and candidate deafness genes were prioritized based on allele frequencies and predicted effects. Selected variants were tested for their co-segregation with HL. Two cases were solved by variants in recently identified deafness genes (ABHD12, TRRAP). Variant prioritization also revealed potentially causative variants in candidate genes associated with recessive and X-linked HL. Importantly, missense variants in IKZF2 were found to co-segregate with dominantly inherited non-syndromic HL in three families. These variants specifically affected Zn²⁺-coordinating cysteine or histidine residues of the zinc finger motifs 2 and 3 of the encoded protein Helios. This finding indicates a complex genotype-phenotype correlation for IKZF2 defects, as this gene was previously associated with non-syndromic dysfunction of the immune system and ICHAD syndrome, including HL. The designed strategy for variant prioritization revealed that IKZF2 variants can underlie non-syndromic HL. The large number of candidate genes for HL and variants therein stress the importance of inclusion of family members for variant prioritization.

Fig. 1

Flowchart of subject inclusion and group allocation. AD, autosomal dominant; AR, autosomal recessive

Fig. 2

Schematic representation of the Helios protein and its functional domains. Adapted and updated version of a previously published figure by Mohajeri et al. (2023). The variants reported in this study are shown above the protein, while previously reported variants are displayed below the protein. The underlined variant was found to by present homozygously. Grey arrows indicate missense variants, and black arrows indicate stop gain variants. The legend for the symbols used is provided within the figure: ‘Hearing loss’ signifies non-syndromic hearing loss and pertains to the subjects in this study; ‘Immune abnormality’ refers to variants associated with immune deficiency or dysregulation (Hetemäki et al. ; Mayr et al. ; Shahin et al. 2022); and ‘Multiple symptoms’ indicates syndromic disease, including both hearing loss and immune abnormalities, referencing two patients described by Mohajeri et al. (2023). Exon numbers are according to the reference sequence NM_001387220.1. ZF, zinc-finger motif

Fig. 3

Pedigrees of families W16-0482, W22-1907 and W22-2757 and co-segregation of the identified IKZF2 variants with HL. The clinical status of the individuals is indicated by the filling of the symbols: black indicates that the subject has HL, and white indicates that the subject has no HL. The clinical status is based on audiograms in all subjects marked with an asterisk. A Family W16-0482. V: c.485A>C p.(His162Pro). Subjects II:1 and II:4 (marked with a hashtag) did have HL, but a significantly milder phenotype than other affected family members. The variant (V) was found to be de novo in II:2 (Supplemental Fig. 2). B Family W22-1907. V: c.509G>A p.(Cys170Tyr). C Family W22-2757. V: c.434G>T p.(Cys145Phe). + , wildtype; square, male; circle, female; slash through symbol, deceased; arrow, proband

Fig. 4

Audiological features of subjects with IKZF2 variants. The first ten panels show the pure tone air conduction thresholds in dB HL of 0.25–8 kHz of subjects identified with an IKZF2 variant (c.485A>C p.(His162Pro) in family W16-0482; c.509G>A p.(Cys170Tyr) in family W22-1907; c.434G>T p.(Cys145Phe) in family W22-2757) from whom audiometric data was available. Black lines with circles represent the right ear, black lines with crosses represent the left ear, grey lines and dots represent the age- and gender-specific 95th percentile. The last panel shows the age-related typical audiograms (ARTA), derived from cross-sectional linear regression analysis of the most recent audiograms of individuals identified with IKZF2 variants. Each blue line represents a ten-year age span. dB HL, decibel hearing level; F, female; kHz, kilo Hertz; M, male; y, years

Fig. 5

Structural modeling of IKZF2 variants. A Crystal structure of the second N-terminal C2H2 domain (PDB: 7LPS) is illustrated, with the variant positions of interest His162 and Cys145 highlighted in sticks at the Zn⁺² coordination site. Atomic distances shown are in angstroms (Å). Missense variants c.485A>C p.(His162Pro) and c.434G>T p.(Cys145Phe) represented as grey sticks are depicted in (B, C), respectively. D The third variant position of interest, Cys170, is located at the third N-terminal C2H2 domains Zn⁺² coordination site, as depicted. The figure shows superposition of the modelled structue of the third C2H2 domain (yellow, with side chains optimized) onto the second C2H2 domain (light red) cocrystallized with Zn⁺². E The missense variant c.509G>A p.(Cys170Tyr) is predicted to cause steric clashes in its neighbourhood, indicated by the non-bonded distances that are less than the sum of their van der Waals radii (< 1.8 Å)

Fig. 6

Functional evaluation of the identified IKZF2 variants. Variant c.485A>C p.(His162Pro) in family W16-0482, variant c.509G>A p.(Cys170Tyr) in family W22-1907 and variant c.434G>T p.(Cys145Phe) in family W22-2757. A Assessment of variant protein expression in HEK293 cells. Immunoblot analysis of cells transfected with plasmids coding for 3xFLAG-tagged WT Helios or Helios with the amino acid substitutions alone, or a combination of WT and variant plasmids. There is a statistically significant decrease in Helios protein levels when the variants are expressed alone but no significant decrease when the Helios variants are co-expressed with WT Helios (right panel). A parametric one-way ANOVA was used for statistical analysis: *p < 0.05; **p < 0.01. B Luciferase reporter assays to assess the variants’ ability to repress IL2 promotor activity, demonstrating statistically significant decrease in repression of the IL2 promotor by the variant Helios proteins when expressed alone. Upon co-expression of WT and variant Helios in increasing ratios of transfected variant plasmids, a statistically significant decrease of IL2 promoter repression is only observed for ratios of 2:1 and higher. A parametric one-way ANOVA was used for statistical analysis: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. EV, empty vector control; IL2, interleukin-2; ns, not statistically significant; WT, wildtype