CHD8 Variant and Rett Syndrome: Overlapping Phenotypes, Molecular Convergence, and Expanding the Genetic Spectrum

Abstract

Rett syndrome (RTT) is a rare, X-linked, severe neurodevelopmental disorder, predominantly associated with pathogenic variants in the methyl-CpG-binding protein-2 (MECP2) gene, with an increasing number of atypical RTT or RTT-like individuals having pathogenic variants in other genes, such as cyclin-dependent kinase-like 5 (CDKL5) or forkhead box G1 (FOXG1). However, ~20% of individuals with a clinical diagnosis of RTT remain genetically undiagnosed, highlighting the importance of ongoing genomic and functional studies to expand the genetic spectrum of RTT. We present a female who was born to healthy nonconsanguineous parents and presented with severe intellectual disability, macrocephaly, ataxia, absent speech, and poor eye contact. The affected individual was clinically diagnosed with atypical RTT, but genetic testing showed no pathogenic variants in MECP2, CDKL5, or FOXG1. Singleton whole genome sequencing was conducted, which identified a heterozygous stop-gain variant [NM_001170629.2: c.5017C>T, p.(Arg1673⁣^∗)], in the chromodomain-helicase-DNA-binding protein 8 (CHD8) gene. Variant curation revealed its absence in unaffected populations, in silico predictions of pathogenicity, and an existing association with intellectual developmental disorder with autism and macrocephaly (IDDAM) (OMIM #615032). In vitro functional analyses, including Western blots, quantitative reverse transcription polymerase chain reaction (qRT-PCR), and proteomic analyses, demonstrated a significant reduction of the CHD8 transcript and two CHD8 protein isoforms in the proband's skin fibroblasts relative to control fibroblasts. Additionally, proteomic analysis indicated a significant reduction of the MeCP2 protein, indicating a possible molecular link between CHD8 and MeCP2 and thus clinically between IDDAM and RTT. As the affected individual's phenotype is consistent with atypical RTT, our results suggest that CHD8 could be considered in the expanding genetic spectrum of atypical RTT, which may assist the diagnosis of other MECP2-negative RTT individuals.

Figure 1

Isoforms and functional domains of the chromodomain-helicase-DNA-binding protein 8 (CHD8) protein. The three isoforms of CHD8 protein, including (1) CHD8-S, a short isoform; (2) CHD8-L1, a long isoform; and (3) CHD8-L2, a long isoform [17]. CHD8-L1 and CHD8-L2 are composed of two histone-binding chromodomains (C1 and C2, yellow), a chromatin-remodeling helicase domain (helicase, cyan), multiple protein-interacting chromatin organization modifier domains (CR, magenta), and a DNA-binding brahma and kismet domain (BRK, pink) [17]. The position of the identified variant relative to CHD8-L1 and CHD8-L2 isoforms is indicated in red.

Figure 2

Variant validation using Sanger sequencing and quantitative reverse transcription polymerase chain reaction (qRT-PCR). (a) The Sanger chromatograms indicate the absence of the variant in the maternal DNA and presence in the proband fibroblasts and blood DNA, indicating a nonmaternal inheritance of the variant. (b) Two sets of cDNA primers, including a set of primers upstream of the variant and another downstream of the variant (Table S3 and Figure S2), were used to conduct qRT-PCR on CHD8 cDNA in the proband line versus the control lines. (c) CHD8 transcripts captured by both upstream and downstream cDNA primers showed significant reduction (upstream primers: ~42%, downstream primers: ~33%) in the CHX− proband samples relative to that of controls (Wilcoxon test: p = 0.0313 for both primer sets). CHX+ samples of both the proband and the controls showed equivalent levels of CHD8 transcripts.

Figure 3

Immunoblotting and mass spectrometry–based proteomic analysis. (a) Western blots indicating the level of CHD8 protein detected from controls (C1, C2) and proband (P) samples. Three technical repeats (n = 3) of Western blotting using the CHD8 C-terminal antibody (Cell Signaling Technologies #11891, 1:1000) showed the relative quantities of CHD8-L1 and CHD8-L2 against GAPDH (loading control). (b) Protein band quantification of the Western blots showed a significant reduction of the CHD8-L1 and CHD8-L2 isoform levels in the proband (P) (L1: ~51%, L2: ~48%) compared to those of the controls (C) (Mann–Whitney test: p = 0.0089, p = 0.0238, respectively). (c) The abundance of CHD8 is ranked significantly lower in the proteome of the proband compared to the controls. (d) The abundance of CHD8 is significantly lower in proband fibroblasts (70%, red dot) and lies outside of the control range (80%–104%, n = 5). (e) Volcano plot showed the relative amount of proteins in the proband line compared to the controls, with vertical lines indicating +/−1.5 log₂-fold change and the horizontal line indicating statistical significance. CHD8 is reduced significantly by ~30% (p < 0.001) in the proband line compared to the controls. MeCP2 (green) is significantly reduced by ~43% (p < 0.01), whereas bromodomain adjacent to zinc finger domain 1A (BAZ1A) encoding the accessory subunit of the ATP-dependent chromatin assembly factor (ACF) (orange) is significantly increased by ~72% (p < 0.001). CHD8-regulated proteins (purple), including acylglycerol kinase (AGK), CDC42-binding protein kinase (CDC42BPB), phosphatase and tensin homolog (PTEN), and dual-specificity tyrosine phosphorylation-regulated kinase 1A (DYRK1A), showed a reduction in their corresponding protein abundance, with AGK being the highest at ~55% (p < 0.001). Transportin 3 (TNPO3), nuclear receptor corepressor 1 (NCOR1), and proteasome assembly chaperone 2 (PSMG2) showed an increase of abundance with TNPO3 being the highest at ~39% (p < 0.001). (f) STRING network analysis revealed coexpression (black), interactions (magenta), and comentions in literature (lime green) between CHD8, MeCP2, CDKL5, FOXG1, and ACF.

Figure 4

Enrichment analysis on mass spectrometry–based proteomic analysis data. Enrichment analysis was conducted on proteins with a statistically significant change of abundance in the proband fibroblasts compared to the controls (Table S5). PhenoExamWeb was utilized to identify associated Human Phenotype Ontology (HPO) terms of the significantly (a) increased and (b) reduced proteins [14], and Metascape was utilized to investigate the associated Gene Ontology (GO) terms of the significantly (c) increased and (d) reduced proteins [15].