De novo protein-coding gene variants in developmental stuttering

Abstract

Developmental stuttering is a common childhood condition characterized by disfluencies in speech, such as blocks, prolongations, and repetitions. While most children who stutter do so only transiently, there are some for whom stuttering persists into adulthood. Rare-variant screens in families including multiple relatives with persistent stuttering have so far identified six genes carrying putative pathogenic variants hypothesized to act in a monogenic fashion. Here, we applied a complementary study design, searching instead for de novo variants in exomes of 85 independent parent-child trios, each with a child with transient or persistent stuttering. Exome sequencing analysis yielded a pathogenic variant in SPTBN1 as well as likely pathogenic variants in PRPF8, TRIO, and ZBTB7A - four genes previously implicated in neurodevelopmental disorders with or without speech problems. Our results also highlighted two further genes of interest for stuttering: FLT3 and IREB2. We used extensive bioinformatic approaches to investigate overlaps in brain-related processes among the twelve genes associated with monogenic forms of stuttering. Analyses of gene-expression datasets of the developing and adult human brain, and data from a genome-wide association study of human brain structural connectivity, did not find links of monogenic stuttering to specific brain processes. Overall, our results provide the first direct genetic link between stuttering and other neurodevelopmental disorders, including speech delay and aphasia. In addition, we systematically demonstrate a dissimilarity in biological pathways associated with the genes thus far implicated in monogenic forms of stuttering, indicating heterogeneity in the etiological basis of this condition.

Fig. 1. Locations of identified (likely) pathogenic variants in stuttering and published pathogenic neurodevelopmental disorder variants in the same genes.

Pathogenic and likely pathogenic variants identified in this study are visualized above the linear protein schematics. The variants previously published as causal for monogenic neurodevelopmental disorders related to SPTBN1 [63, 64], PRPF8 [71], TRIO [65, 73] and ZBTB7A [66, 67] are visualized below the schematics. Missense variants are indicated in purple and pLoF variants in red. Protein domains are represented with yellow squares: CH calponin homology domain, SPEC spectrin repeats, PH Pleckstrin homology domain, PRO8NT PrP8 N-terminal domain, PROCN central domain in pre-mRNA splicing factors of PRO8 family, RRM RNA recognition motif, U5/6BDG U5/6-snRNA binding site, RNase-HH RNase-H homology domain, SEC14 protein structural domain that binds small lipophilic molecules, RhoGEF guanine nucleotide exchange factor, SH3 Src homology 3, S_TKc Serine/Threonine protein kinases, catalytic domain, BTB Broad-Complex, Tramtrack and Bric a brac, ZNF Zinc finger.

Fig. 2. Neural gene expression patterns of genes associated with monogenic forms of stuttering.

A. Developmental brain expression pattern of the twelve stuttering-associated genes across eight developmental periods spanning from eight post conception weeks (pcw) to ten months (mos) of age. Grey circles depict expression levels in individual brain samples collected from the cerebellum, cortex, hippocampus, amygdala, striatum and thalamus. The trendlines in yellow-orange (estimated with locally estimated scatterplot smoothing) visualize the overall pattern of gene expression change over time in the different regions of the brain. The vertical dashed lines represent time of birth. B. Gene expression levels of these stuttering-associated genes and the background gene set in spatial gene expression data of the adult human cortex. Spatial gene expression data of 48 human cortex tissue sections were clustered into 12 data-driven clusters, of which seven represent cortical layers and five were located in the white matter [59]. Violin plots and grey box plots show the distribution of gene expression levels of the background gene set in the 12 clusters. Yellow box plots show the gene expression levels of the stuttering-associated genes. Box plots show median and first and third quartiles, with whiskers extending to 1.5 times the interquartile range.