Exome sequencing in an admixed isolated population indicates NFXL1 variants confer a risk for specific language impairment

Abstract

Children affected by Specific Language Impairment (SLI) fail to acquire age appropriate language skills despite adequate intelligence and opportunity. SLI is highly heritable, but the understanding of underlying genetic mechanisms has proved challenging. In this study, we use molecular genetic techniques to investigate an admixed isolated founder population from the Robinson Crusoe Island (Chile), who are affected by a high incidence of SLI, increasing the power to discover contributory genetic factors. We utilize exome sequencing in selected individuals from this population to identify eight coding variants that are of putative significance. We then apply association analyses across the wider population to highlight a single rare coding variant (rs144169475, Minor Allele Frequency of 4.1% in admixed South American populations) in the NFXL1 gene that confers a nonsynonymous change (N150K) and is significantly associated with language impairment in the Robinson Crusoe population (p = 2.04 × 10-4, 8 variants tested). Subsequent sequencing of NFXL1 in 117 UK SLI cases identified four individuals with heterozygous variants predicted to be of functional consequence. We conclude that coding variants within NFXL1 confer an increased risk of SLI within a complex genetic model.

Fig 1. Pedigree showing direct lines of descent between founder brothers and children in Robinson Crusoe validation cohort.

Founder brothers are individuals on the second line of the pedigree. Individuals with language impairment are colored in black. Individuals with typical language are denoted in white. Individuals with unknown phenotype are shaded grey. Genotypes at rs144169475 are represented by small circles; blue circles represent homozygote reference allele, red circles represent variant carriers, grey circles represent unknown genotype. Note that each individual may be represented through multiple lines of descent and so might appear more than once on this diagram. Children are labelled according to affection status—SLI1 to SLI15 and TLD1 to TLD17. Cases whose exomes were sequenced are indicated by asterisks. Three children (1 affected, 2 unaffected, none of whom carried the rs144169475 variant) are not represented on this figure since they were related to alternative founder families. SLI15 is known to be related to one of the founder brothers but the exact line of descent is unknown.

Fig 2. Putative contributory coding variants identified in NFXL1 by this study.

Position of putative NFXL1 coding variants with respect to exons and protein coding sequence. Genomic coding exons (exons 2–23) are shown by pink bands at the top. Protein motifs are represented by colored bands in the lower boxes. The red box represents a Znf RING motif, the yellow boxes represent Znf NFX1 motifs, the blue box represents a coiled-coil domain and the green box a transmembrane domain. Putative contributory coding variants are shown by arrows. Blue arrows denote synonymous changes, red arrows nonsynonymous changes. Sanger sequencing plots are given for all variants identified. Conservation of amino acid sequences across 11 species shown for all variants identified. The ref row shows the human reference allele and the variant row shows the observed variant in our samples. All sequences that differ from the reference sequence are shown in red.

Fig 3. Coding variants observed in SLIC probands and their families.

Pedigrees are shown for nuclear families of SLIC individuals carrying three coding variations in NFXL1. Individuals carrying the variants are identified with a black circle. Sequencing traces of each variant is shown. SLIC probands are colored in red and other family members with SLI (defined as expressive and/or receptive language skills >1.5SD below that expected for their age) are colored in orange. In pedigree 3 (rs151113647), the youngest sibling (colored in yellow) did not meet the criteria for SLI but had expressive and receptive language scores ∼1SD below that expected for his age. Individuals with no shading have typical language ability. DNA was not available for individuals colored in grey.