Genotype-phenotype correlations and novel molecular insights into the DHX30-associated neurodevelopmental disorders

Abstract

Background: We aimed to define the clinical and variant spectrum and to provide novel molecular insights into the DHX30-associated neurodevelopmental disorder.

Methods: Clinical and genetic data from affected individuals were collected through Facebook-based family support group, GeneMatcher, and our network of collaborators. We investigated the impact of novel missense variants with respect to ATPase and helicase activity, stress granule (SG) formation, global translation, and their effect on embryonic development in zebrafish. SG formation was additionally analyzed in CRISPR/Cas9-mediated DHX30-deficient HEK293T and zebrafish models, along with in vivo behavioral assays.

Results: We identified 25 previously unreported individuals, ten of whom carry novel variants, two of which are recurrent, and provide evidence of gonadal mosaicism in one family. All 19 individuals harboring heterozygous missense variants within helicase core motifs (HCMs) have global developmental delay, intellectual disability, severe speech impairment, and gait abnormalities. These variants impair the ATPase and helicase activity of DHX30, trigger SG formation, interfere with global translation, and cause developmental defects in a zebrafish model. Notably, 4 individuals harboring heterozygous variants resulting either in haploinsufficiency or truncated proteins presented with a milder clinical course, similar to an individual harboring a de novo mosaic HCM missense variant. Functionally, we established DHX30 as an ATP-dependent RNA helicase and as an evolutionary conserved factor in SG assembly. Based on the clinical course, the variant location, and type we establish two distinct clinical subtypes. DHX30 loss-of-function variants cause a milder phenotype whereas a severe phenotype is caused by HCM missense variants that, in addition to the loss of ATPase and helicase activity, lead to a detrimental gain-of-function with respect to SG formation. Behavioral characterization of dhx30-deficient zebrafish revealed altered sleep-wake activity and social interaction, partially resembling the human phenotype.

Conclusions: Our study highlights the usefulness of social media to define novel Mendelian disorders and exemplifies how functional analyses accompanied by clinical and genetic findings can define clinically distinct subtypes for ultra-rare disorders. Such approaches require close interdisciplinary collaboration between families/legal representatives of the affected individuals, clinicians, molecular genetics diagnostic laboratories, and research laboratories.

Fig. 1

Location of identified DHX30 germline variants. a Highly conserved sequence motifs within the helicase core region are shown with color coding that corresponds to the primary function of the motif (as previously described by Lessel et al., 2017). Double-stranded RNA-binding domains (dsRBD) 1 and 2 at the N-terminus (N-) are shown in gray. A winged helix domain (WHD), a ratchet-like (RL) domain, and an oligosaccharide binding (OB) domain are shown in yellow at the C-terminus (-C). The position of the first and last amino acid within each motif/domain is indicated below. Previously reported heterozygous missense variants and newly identified DHX30 variants are denoted in gray and black, respectively. Frameshift and nonsense variants are denoted in blue. Mutated amino acid residues within the helicase core region are marked in red. The position of previously and newly identified variants are indicated with red arrows. b Genomic region, chr3.hg19:g.(47098509-48109065)del, of the ~ 1 Mb deletion identified in case 24

Fig. 2

Protein variants of DHX30 affect ATPase and helicase activity. a, b ATPase assays were performed for DHX30-WT, novel DHX30 missense variants (a), and two common polypmorphisms, p.(Val556Ile) and p.(Glu948Lys) (b) in the presence of exogenous RNA. ATPase activity was calculated by subtracting phosphate values obtained with GFP alone from those obtained with GFP-tagged DHX30-WT and mutants. These figures were then normalized on precipitated protein amounts using the intensities of the GFP signal in the western blot. Means ± standard deviation values are based on 3 replications. **,***: significantly different from DHX30-WT, ns: not significantly different from DHX30-WT (**p< 0.01;***p< 0.001; n=3; One-Way ANOVA, followed by Dunnett’s multiple comparisons test). Values were normalized on DHX30-WT ATPase activity obtained in the presence of RNA. c Increasing amounts of His₆-SUMO-tagged DHX30 WT protein were incubated with a ³²P-labeled RNA substrate in the presence (lane 3–7) or absence (lane 8) of ATP and analyzed by native PAGE. The position of the RNA duplex and the single-stranded RNA are indicated in the first and second lanes, respectively. Their schematic representation is shown at the right side. d Helicase assay was repeated for selected DHX30 missense variants affecting either conserved motifs within the helicase core region (lane 4–8) or the auxiliary RL domain (lane 9)

Fig. 3

Missense variants in DHX30 initiate the formation of cytoplasmic aggregates and impair global translation. Puromycin incorporation assay in U2OS cells expressing DHX30-GFP fusion proteins (green). Translation was monitored by staining against puromycin (red), SGs were detected by ATXN2 (magenta) and nuclei via DAPI staining (blue). Arrows indicate transfected cells. Note the correlation between formation of clusters and lack of puromycin staining. Scale bars indicate 10 μm

Fig. 4

Protein variants of DHX30 lead to embryonal developmental defects in zebrafish. In vivo analyses of selected DHX30 missense variants. Assessment of embryonic development after injection of DHX30 WT and mutant cDNAs in a zebrafish model. Bar graph indicating the percentage of cmcl2-GFP positive zebrafish embryos assessed 4–7 days post fertilization (dpf). The presented data are derived from three independent studies. The total number of embryos assessed are 45, 23, 21, 30, 34, 33, and 29 for WT, V556I, E948K, R493H, R725H, R785C, and R908Q, respectively. ****: significantly different from WT (****p< 0.0001; χ² test)

Fig. 5

Recombinant protein variants of DHX30 induce translocation of the DHX30-WT in the cytoplasmic clusters. Immunocytochemical detection of RFP-DHX30 WT (red) and ATXN2 (magenta) after co-expression of DHX30-GFP mutants (green) in U2OS cells. Bar graph indicating the percentage of cells where RFP-DHX30 WT co-localizes with DHX30-GFP mutants within cytoplasmic clusters identified as SGs via co-staining with ATXN2 (****: significantly different form DHX30-WT: **** p < 0.0001; n > 100 from 3 independent transfections; one-way ANOVA followed by Dunnett’s multiple comparisons test). Scale bars indicate 10 μm

Fig. 6

Analyses of the nature of missense variants within the helicase core motifs (HCM). a RNA unwinding activity of purified DHX30 R493H, H562R, and R785C mutants was analyzed upon addition of DHX30 WT protein. Increasing amounts of mutant proteins were incubated with 20 ng of WT protein and assayed for their ability to unwind a radiolabeled RNA duplex in the presence of ATP. b Assessment of embryonic development after co-injection of DHX30 R493H and R785C with DHX30 WT cDNA in a zebrafish model. Bar graph indicating the percentage of cmcl2-GFP positive zebrafish embryos 4–7 days postfertilization (dpf). The presented data are derived from three independent studies. The total number of embryos assessed are 58, 43, and 51 for WT, R493H+WT, and R785C+WT, respectively. ****: significantly different from WT (****p< 0.0001; χ² test)

Fig. 7

DHX30 deficiency in HEK293T cells leads to reduced formation of stress granules. a Immunocytochemical detection of endogenous ATXN2 (magenta) in WT HEK293T cells and DHX30-deficient HEK293T cells before (left panel) and after (right panel) heat shock at 43.5 °C for 1 h. Note that, upon heat stress and depletion of DHX30 (right hand, lower panel), ATXN2 does not alter its diffuse cytoplasmic distribution to accumulate in cytoplasmic foci, as observed in WT HEK293T cells (right hand, upper panel). Nuclei are identified via DAPI staining (blue). Scale bars indicate 10 μm. b Bar graph indicating the percentage of cells containing stress granules. (*: significantly different from WT HEK293T cells: *p< 0.05; n > 200 from 3 independent experiments; unpaired t test ). c Western blotting detection of DHX30 knock-out efficiency in HEK293T cells. Expression of DHX30 was reduced by 90% as detected by a DHX30 specific antibody. Tubulin was used as loading control

Fig. 8

DHX30 deficiency in zebrafish cells leads to reduced formation of stress granules. a Representative confocal images of TIAL-1-labeled stress granules (green) in dhx30 wild-type (+/+) and homozygous mutants (−/−). Zebrafish underwent normal conditions or heat shock treatment at 42 °C. Nuclei were counterstained with DAPI (blue). b Analyses of TIAL-1-labeled stress granules per 50 nuclei. The total number of embryos assessed are 8, 9, 8, and 8 for dhx30 +/+ (28 °C), dhx30 −/− (28 °C), dhx30 +/+ (42 °C), and dhx30 −/− (42 °C), respectively. Data are presented as means ± standard error of mean based on the indicated number of embryos. ***: significantly different from DHX30+/+ (***p< 0.001; unpaired Student’s t test)

Fig. 9

Behavioral analyses of dhx30 mutant zebrafish. a Distance moved of dhx30 mutants and wild-type sibling controls measured at 5 days post fertilization. b Average of distance moved during 14-h daytime. c Average of distance moved during 10-h nighttime. N = 15, 18, and 25 for +/+, +/−, and −/− animals, respectively. d Social preference index (SPI) calculated during 10-min baseline and post-baseline period. SPI = 1 indicates a fish that spends 100% of its time near a conspecific, SPI = − 1 indicates a fish that spends 100% of its time near the empty well, and SPI = 0 indicates a fish that spends equal amounts of time near the conspecific and near the empty well. e The change in SPI between baseline and post-baseline, indicating the preference of zebrafish to stay close to conspecific fish. N = 13, 6, and 17 for +/+, +/−, and −/− animals, respectively. Data are presented as means ± standard error of mean based on the indicated number of embryos. *: significantly different from dhx30+/+ (*p< 0.05; unpaired Student’s t test)