Mutations in DDX3X Are a Common Cause of Unexplained Intellectual Disability with Gender-Specific Effects on Wnt Signaling

Abstract

Intellectual disability (ID) affects approximately 1%-3% of humans with a gender bias toward males. Previous studies have identified mutations in more than 100 genes on the X chromosome in males with ID, but there is less evidence for de novo mutations on the X chromosome causing ID in females. In this study we present 35 unique deleterious de novo mutations in DDX3X identified by whole exome sequencing in 38 females with ID and various other features including hypotonia, movement disorders, behavior problems, corpus callosum hypoplasia, and epilepsy. Based on our findings, mutations in DDX3X are one of the more common causes of ID, accounting for 1%-3% of unexplained ID in females. Although no de novo DDX3X mutations were identified in males, we present three families with segregating missense mutations in DDX3X, suggestive of an X-linked recessive inheritance pattern. In these families, all males with the DDX3X variant had ID, whereas carrier females were unaffected. To explore the pathogenic mechanisms accounting for the differences in disease transmission and phenotype between affected females and affected males with DDX3X missense variants, we used canonical Wnt defects in zebrafish as a surrogate measure of DDX3X function in vivo. We demonstrate a consistent loss-of-function effect of all tested de novo mutations on the Wnt pathway, and we further show a differential effect by gender. The differential activity possibly reflects a dose-dependent effect of DDX3X expression in the context of functional mosaic females versus one-copy males, which reflects the complex biological nature of DDX3X mutations.

Figure 1

Location of Amino Acid Substitutions in DDX3X Schematic view of DDX3X with the 15 different amino acid substitutions (and one in-frame deletion) found in affected females and the 3 different amino acid substitutions found in affected males. DDX3X consists of two subdomains: a helicase ATP-binding domain and a helicase C-terminal domain. All amino acid substitutions found in affected females are located within these two protein domains. The three amino acid substitutions found in affected males are all located within the helicase ATP-binding domain.

Figure 2

Facial Profiles of Females with De Novo DDX3X Mutations Facial features of 30 of 38 females with a de novo variant in DDX3X. Common facial features include a long and/or hypotonic face (e.g., individuals 2, 4, 5, 12, 22, 32), a high and/or broad forehead (e.g., individuals 1, 7, 9, 23, 24, 26), a wide nasal bridge and/or bulbous nasal tip (e.g., individuals 11, 13, 15, 16), narrow alae nasi and/or anteverted nostrils (e.g., individuals 2, 8, 9, 12, 14, 18, 24, 27, 32, 35), and hypertelorism (e.g., individuals 5, 7, 8, 20, 27). Informed consent was obtained for all 30 individuals shown. The individual numbers correspond to the numbers mentioned in Tables 1 and S1.

Figure 3

The Effect of Wild-Type and Variant DDX3X mRNA on WNT3A-Mediated Ventralization (A–C) Zebrafish embryos at 2 dpf either uninjected (A) or injected with human WNT3A without (B) or with (C) human DDX3X show a range of ventralized phenotypes. These were scored according to Kelly et al. as normal, class I, or class II ventralization. No injection condition resulted in severe ventralization (class III or IV). (D) Co-injection of WNT3A and DDX3X shows an augmentation of WNT3A-mediated ventralization with increasing dose up to 15 pg. (E) Individual variants of DDX3X were tested for their effect on increasing WNT3A-mediated ventralization using 200 fg of WNT3A mRNA and 15 pg of DDX3X mRNA per embryo (dose of maximal response from D). Scoring is the same as in (D). The female de novo variants and the male familial variants are shown together. The substitution p.Glu196Lys is a rare allele from the Exome Variant Server (EVS). The red horizontal line delineates a division between those variants that behave as “wild-type” (e.g., the male variants and p.Glu196Lys) and those that do not. All female de novo variants are significantly different from wild-type but not different from WNT3A alone. Male variants are significantly different from WNT3A alone but not different from DDX3X + WT. ^∗p < 0.001 compared with WNT3A + wild-type DDX3X co-injection. #p < 0.0001 versus WNT3A alone. (F) Graph illustrating the effect of combining the results from the female de novo variants and comparing to the male familial variants and to the p.Glu196Lys variant from the EVS. There is clear segregation of effect based on the source of the genetic variant. ^∗∗p < 0.0001 versus WNT3A + wild-type DDX3X. ##p < 0.0001 versus WNT3A alone. For each graph, total (n) is shown at the bottom of each bar.