Resolving SLC6A1 variable expressivity with deep clinical phenotyping and Drosophila models

Abstract

Variants in SLC6A1 result in a rare neurodevelopmental disorder characterized by a variable clinical presentation of symptoms including developmental delay, epilepsy, motor dysfunction, and autism spectrum disorder. SLC6A1 haploinsufficiency has been confirmed as the predominant pathway of SLC6A1-related neurodevelopmental disorder (SLC6A1-NDD); however, the molecular mechanism underlying the variable clinical presentation remains unclear. Here, through work of the Undiagnosed Diseases Network, we identify an individual with an inherited p.A334S variant of uncertain significance. To resolve this variant and better understand the variable expressivity associated with SLC6A1, we assess the phenotypes of the proband in comparison with a cohort of 13 individuals diagnosed with SLC6A1-NDD. We then create an allelic series in Drosophila melanogaster to functionally characterize these variants. Informatic clustering based on these clinical findings points to significant clinical overlap between the unsolved individual and confirmed SLC6A1-NDD. We confirm phenotypes in flies expressing SLC6A1 variants consistent with a partial loss-of-function mechanism. We conclude that the p.A334S variant is a hypomorphic allele and begin to elucidate the underlying variability in SLC6A1-NDD. These insights will inform clinical diagnosis, prognosis, intervention, and inform therapeutic design for those living with SLC6A1-NDD.

Keywords: Drosophila melanogaster model; GABA; SLC6A1; deep clinical phenotyping; functional characterization; neurodevelopmental disorder; variable expressivity.

Figure 1

Comprehensive pedigrees and their genotypes for families with SLC6A1 variants Standard pedigree structures are utilized—filled circles and squares denote clinically affected individuals and probands are indicated by black arrows. (A) The UDN family pedigree is denoted with the phenotypes as noted in family members, respectively. (B) Thirteen families diagnosed with SLC6A1-neurodevelopmental disorders recruited through the Brain Gene Registry. (C) Schematic diagram of SLC6A1 protein structure from conceptual translation of transcript NM_003042.3 with mapping location (circles) of the amino acid variants observed in this study. In total, we identified 11 different SNVs and indels that are distributed throughout the protein. Blue circles represent missense variants, red circles represent frameshift variants, and yellow circles represent the splice variant observed in this study. Image created in Biorender.

Figure 2

Clinical phenotypic analyses (A) SLC6A1 phenotypic grid of all individuals in the study. HPO terms from all individuals were categorized into 11 key phenotypic domains described in the draft SLC6A1 disease model. The frequency of terms per phenotypic domain is plotted. Neurological and cognitive domains show the highest frequency of terms. (B and C) Rapid neurodevelopmental assessment protocol (RNAP) profiles of individuals. The 804-item questionnaire in the RNAP was categorized into 7 domains namely: cognitive ability, adaptive behavioral function, social and emotional function, sensory-motor function, sleep symptoms, communication ability, physical features, and neurologic symptoms. The RNAP domain score was generated for each individual, which was then normalized based on the number of questions answered by the individual in each domain. These scores are plotted to visualize the RNAP profile for each individual. (D) Phenotypic similarity heatmap for individuals with SLC6A1 variants. A heatmap was generated using proband phenotypic similarity scores and ordered based on hierarchical agglomerative clustering of proband phenotypic similarity. Dendrograms showing clusters are present at the left and top sides of the heatmap. Green box highlights the two frameshift variants that cluster together with similar severe phenotypes. The UDN p.A334S variant clusters with other confirmed cases of SLC6A1.

Figure 3

Gat-specific GAL-4 construct design (A) The Trojan GAL-4 cassette is a strong hypomorphic allele that simultaneously truncates the Drosophila Gat protein and expresses a GAL-4 protein under the regulatory control of the fly gene promoter. (B) Current lineage tracing of Gat^TG4 expression in the fly larval brain. (C) Anti-Repo immunostaining labels glial cells. (D) Gat and Repo merge identifies that Gat^TG4 is expressed in Repo-positive glial cells. (E) The candidate UDN p.A334S and known BGR p.S295L, p.G297R variants are conserved across species including in humans and flies.

Figure 4

Variant-specific sleep profiles (A) Expression of a heterozygous hypomorphic loss-of-function allele Gat^TG4/+ results in a severe decrease in sleep latency compared with SLC6A1^Ref;;Gat^TG4/+ (p < 0.0001). Sleep latency is suppressed by expression of SLC6A1^Ref driven by Gat^TG4. Expression of variant lines also induces a significant decrease in sleep latency in SLC6A1^S295L (p = 0.0005), SLC6A1^G297R (p = 0.0001), and SLC6A1^A334S (p = 0.0007) compared with SLC6A1^Ref. (B) Total sleep is reduced in SLC6A1^G297R (p = 0.0118) and SLC6A1^A334S (p = 0.0053). (C) Sleep bout length is reduced in SLC6A1^S295L (p < 0.0001), SLC6A1^G297R (p < 0.0001), and SLC6A1^A334S (p < 0.0001). (D) Sleep bout number is increased in SLC6A1^S295L (p < 0.0001), SLC6A1^G297R (p < 0.0001), and SLC6A1^A334S (p < 0.0001). (E) Total activity is increased in SLC6A1^S295L (p = 0.0003), SLC6A1^G297R (p = 0.00063), and SLC6A1^A334S (p = 0.0002). Welch’s ANOVA with Dunnett’s T3 multiple comparisons test. Error bars represent SD.