Clinical information
The proband (II-1), a 26-year-old male (as of 2023), was first evaluated at the Pediatric Clinic at Peking University First Hospital at age 14 with symptoms of intermittent vomiting, episodic loss of consciousness, transient amnesia, and difficulty turning his eyes. The birth history was obtained from the parents, and the patient was born prematurely at 28 weeks with postnatal growth retardation. The patient had been diagnosed with "cerebral palsy" at one and a half years old and was unable to walk. Neurological examination found that the patient presented with hypertonia and torsion spastic dystonia (Table 1). Magnetic resonance imaging (MRI) revealed degeneration in the bilateral lateral ventricular cerebral white matter (Fig. 1a). Further laboratory tests showed elevated blood glucose (9.30 mM, normal 3.61-6.11 mM), elevated lactate (2.8 mM, normal 0.5-2 mM), and elevated pyruvate (160 mM, normal 30-100 mM) (Table S1). Throughout the following year, the patient reported intermittent episodes of headache, dizziness, and spasms. These symptoms were tolerable and showed improvement compared to the initial presentation. At age 16, the patient returned with vomiting and intermittent headache. Biochemical examination revealed elevated blood glucose (6.91 mM, normal 3.61-6.11 mM), blood triglycerides (2.78 mmol/L, normal 0.56-1.7 mmol/L), and blood cholesterol (6.0 mM, normal 3.4-5.2 mM) (Table S1), indicating hyperlipidemia. The patient was prescribed bezafibrate tablets for lipid-lowering therapy. After six months, blood glucose (6.21 mM, normal 3.61-6.11 mM) and triglycerides (3.22 mM, normal 0.56-1.7 mM) (Table S1) remained elevated, while cholesterol returned to normal. During the follow-up visit at age 24, the patient presented with visual impairment, including high myopia, vitreous turbidity, and retinal detachment (Table 1). He was also found to have increased muscle tone and walking disorder (Table 1). MRI revealed an abnormal signal in the punctate patch of the bilateral lateral ventricle. According to the comprehensive medical history, it is evident that the patient's condition exhibited progressive symptom deterioration in correlation with spastic paraplegia. The clinical manifestations observed encompass lower-extremity spasticity, intellectual disability, increased muscle tone, walking disorder, visual impairment, and lateral ventricular white matter lesions. Collectively, these findings are consistent with the characteristic clinical manifestations of spastic paraplegia.
Table. 1 Clinical manifestation of the patient
|
Age at
|
14 y
|
15.9 y
|
16.1 y
|
17.1 y
|
24.2 y
|
25.8 y
|
Dystonia
|
+
|
ND
|
ND
|
ND
|
ND
|
+
|
Spasticity
|
+
|
+
|
+
|
ND
|
+
|
+
|
gait impairment
|
|
|
|
|
Walking disorder
|
Walking disorder(wheelchair bound)
|
Visual impairment
|
|
|
|
|
+
|
+
|
Brain MRI
|
punctate white matter lesions of lateral ventricle
|
ND
|
ND
|
ND
|
ND
|
Abnormal signal in punctate patch of bilateral lateral ventricle
|
Abbreviations are as follows: ND, no data.
Discovery and verification of a DDX53 deletion mutation in the patient
Initially, comprehensive whole exome sequencing (WES) and mitochondrial DNA (mtDNA) sequencing were conducted using blood samples from the patient and his mother but failed to detect any causative mutations. We performed SNP array tests and found a 225,630 bp deletion in the X chromosome p22.11 region (GRCh37/hg39 Xp22.11 (23007778_23233408del)) of the patient's genetic spectrum (Fig. 1c). Interestingly, the deleted fragment contained only the DDX53 gene. To confirm the observed deletion mutation of DDX53, polymerase chain reaction (PCR) experiments were performed for the patient's DNA sample. The results unequivocally substantiated the complete absence of DDX53 in the patient (Fig. 1d). DDX53 encodes a protein consisting of 631 amino acids that contains an SF2 helicase core domain (222 aa-601 aa) at the C-terminus and a KH (K homology) domain at the N-terminus (Fig. 1e).
Patients carrying DDX53 variants were associated with neurological disorders
To gain insights into the mutational landscape of DDX53, a comprehensive DDX53 variant analysis was conducted by querying the gnomAD database. It revealed a total of 274 DDX53 variants, including 261 missense variants, 11 frameshift/nonsense variants, and 2 deletion variants. The majority of the missense variants had low allele frequencies, ranging from 0.0000055 to 0.0019949. In addition, only 4 variants surpassed a 1% allele frequency (Table S2), which indicated that DDX53 was highly conserved in different human populations and may play an important role. Intriguingly, the shift/nonsense variants displayed a remarkably low allele frequency, with values as low as 0.0000602, and DDX53 homozygous deletion variants were not reported in gnomAD (Table S2). These findings further emphasize the essential role of DDX53 and suggest that its deletion mutation may contribute to the development of the disease.
ClinVar and Decipher databases recorded 15 patients with DDX53 deletion mutations, of which 14 patients exhibited HSP-like human conditions. These patients presented neurological abnormalities accompanied by additional neurological lesions, such as intellectual disabilities, microcephaly, optic neuropathy, and agenesis of the corpus callosum (Table S3). In the absence of concrete research findings, the role of DDX53 in neurological diseases remains unclear. Additionally, due to the open nature of the ClinVar and Decipher databases, there was limited evidence supporting the correlation between genotypes and clinical phenotypes. To further elucidate the pathogenicity of DDX53, we conducted an in-depth analysis using the Human Gene Mutation Database (HGMD) and found that DDX53 variants might be associated with three human diseases (Table S3). First, a study examining rare copy number variants in autism spectrum disorder (ASD) identified three cases of autism with DDX53 deletion mutations, implicating the DDX53-PTCHD1 locus as a potential risk factor for ASD[20]. Additionally, another study proposed DDX53 as a novel candidate causative gene for disorders resembling steroid-resistant nephrotic syndrome (SRNS)[27]. Furthermore, a study on neurodevelopmental disorders identified a patient with a DDX53 mutation (772C>A)[28]. Collectively, these findings suggested that DDX53 mutations may be associated with neurological disorders.
DDX53 had exclusive expression restricted to the cerebellum and testis
According to the Human Protein Atlas (HPA) database, the expression level of DDX53 in human tissues was remarkably low. Specifically, DDX53 showed specific expression limited to normal testicular and brain tissue (Fig. 2a). Further examination of DDX53 expression within testicular tissue showed prominent expression in the seminiferous ducts. In contrast, within the brain, DDX53 expression reached its highest levels in the cerebellum, surpassing other brain tissues by more than tenfold (Fig. 2b-c). DDX53 was mainly expressed in the vermis, flocculonodular lobe, and cerebellar cortex in the cerebellum. Although trace amounts of DDX53 expression were also detected in other brain regions, such as the midbrain, cerebral cortex, and amygdala, the levels were much lower than those in the cerebellum (Fig. 2d-g). The tissue-specific expression pattern suggests that DDX53 may have crucial physiological functions, primarily in the testis and brain.
Bioinformatics analysis of DDX53
To assess the evolutionary significance of DDX53, a comprehensive conservation analysis was undertaken. The DDX protein family, known for its vital role in various biological processes, exhibits remarkable conservation throughout biological evolution[29]. Accordingly, we conducted a phylogenetic analysis to trace the evolutionary history of DDX53. The phylogenetic tree revealed that DDX53 first appeared in placental mammals and shared a high degree of sequence homology with most primate species. DDX53 also exhibited a closer evolutionary relationship with rodent species (Fig. 3a). Interestingly, DDX53 was not expressed in mice, suggesting that there may be species-specific differences in the regulation or function of DDX53.
To elucidate the biological function of DDX53, we used the BioGRID database to explore its potential interacting proteins and found that DDX53 may interact with EGFR, DDX43, HDAC2, CAPZB, and PPM1B (Fig. 3b). Notably, the interaction between DDX53 and DDX43 was strong, with an interaction score of over 0.99 (Table S4). DDX53 and DDX43 showed significant homology, amounting to 62%[30]. Intriguingly, both proteins exhibited a conserved KH (K homology) domain at their N-terminus with an impressive sequence homology of 88%. The KH domain spans 70 amino acids and encompasses the conserved VIGXXGXXI motif. Mostly, this unique domain appeared to be exclusive to DDX53 and DDX43 within the larger DDX protein family[31], suggesting that they may have potential overlapping biological functions facilitated by the KH domain. To gain deeper insights into the putative biological functions of DDX53, we conducted a comprehensive GO (Gene Ontology) analysis and found that DDX53 might have helicase activity and thus affect the binding process of RNA and ATP (Table S5).
DDX53 exhibited minimal impact on mitochondrial function
The DDX53-deficient patients in our study exhibited abnormal nervous system function along with metabolic abnormalities, such as elevated lactic acid, impaired glucose oxidation, and lipoprotein and cholesterol accumulation. These metabolic features are commonly associated with diseases caused by impaired mitochondrial function[32], and certain subtypes of HSP have been reported to be involved in mitochondrial dysfunction [33]. Based on these findings, we hypothesized that DDX53 may play a role in the development of HSP-like diseases by affecting mitochondrial function. To test our hypothesis, we initially aimed to establish a neuronal cell model of DDX53 knockdown (KD) for mitochondrial function studies. Previous studies have demonstrated that SH-SY5Y cells can differentiate into neuron-like cells with mature morphology and biochemical properties[34, 35]. Therefore, we attempted to knock down DDX53 in SH-SY5Y cells. However, the expression level of DDX53 in SH-SY5Y cells was remarkably low, failing to achieve the desired effective knockdown. As an alternative, we proceeded with K562 cells that exhibited a moderate level of DDX53 expression, for subsequent experiments. Compared to those in control cells, the mRNA levels of DDX53 in K562 DDX53-KD1 and K562 DDX53-KD2 cells were markedly reduced by approximately 95% and 90%, respectively (Fig. 4a), and the mRNA levels in DDX53-overexpressed (OE) cells were increased to 600% (OE) cells (Fig. 4b). These findings indicate that the DDX53 KD and OE cell models were successfully constructed. We found that DDX53 KD and OE cells had no significant difference in cellular respiratory capacity, including basal, phosphorylation-coupled, maximal respiration, and spare respiration capacity, compared to controls (Fig. 4c-d). We then investigated mitochondrial respiratory chain complex assembly. As shown in Fig. 4e and f, the steady-state level of all assembled complexes did not show apparent variations between the KD or OE cells and control cells. However, transcriptomic analysis revealed that the expression levels of most nuclear genome-encoded OXPHOS genes were increased in DDX53 KD cells compared to control cells (Fig. 4h), but the expression levels of mitochondrial genome-encoded OXPHOS genes were not significantly different in KD cells (Fig. 4g). Altogether, our findings showed that DDX53 deficiency had a minimal impact on mitochondrial function.
DDX53 impaired RNA metabolism in the nervous system
To investigate the regulatory effect of DDX53 on gene expression, RNA sequencing (RNA-seq) was conducted using control and DDX53-KD1 cells. A total of 1295 differentially expressed genes were identified; of which, 857 genes were downregulated, and 438 genes were upregulated in KD cells compared to control cells (Fig. 5a). GO biological process (GO-BP) analysis revealed enrichment of 158 pathways among the downregulated genes in KD cells; these pathways were predominantly associated with synaptic activity, neuron projection development, and neuron differentiation (Fig. 5b). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway cluster analysis highlighted the ECM-receptor pathway, GABAergic synapse, and relaxin signaling pathway as the top potential biological pathways affected by DDX53 loss (Fig. 5c). Additionally, disease ontology cluster analysis revealed significant enrichment of peripheral nervous system diseases, epilepsy syndrome, and psychotic disorder (Fig. 5d). These findings suggest that DDX53 is likely involved in synaptic organization, neuron function, and neuromuscular junction regulation. Its deficiency may disrupt the expression of essential genes involved in neuronal function, potentially contributing to the development of neurological disorders.
To further explore the molecular mechanism of DDX53 in gene regulation, RIP-seq was employed using DDX53 OE cells. After filtering known rRNA sequences, a total of 58.97 million reads were obtained, with an average length of 200 bp, and 77.58% of the reads were successfully mapped to the reference genome. MEME analysis revealed a preference of DDX53 binding to sequences containing consecutive U bases (Fig. 6a). Mapping the reads to the human genome showed that 66.62% of the 717 binding peaks were located in exon regions, 32.56% were located in introns, and a small fraction was located in stop codons and intergenic regions (Fig. 6b). A total of 284 peak-associated RNAs were annotated, including mRNAs (86.15%), misc RNAs (11.61%), and a small number of lncRNAs and snoRNAs (Fig. 6c). GO-BP analysis identified histone modification, mRNA processing, and histone methylation as the top potential biological processes associated with DDX53 binding (Fig. 6d). KEGG analysis revealed mRNA surveillance pathway, nucleoplasmic transport, and lysine degradation as the top biological pathways involved (Fig. 6e). These findings suggest that DDX53 is involved in RNA metabolism and has a significant correlation with mRNA binding.
Through a combined analysis of RNA-seq and RIP-seq, 176 genes out of the 4724 genes altered after DDX53 KD were found to have binding peaks in RIP-seq (Fig. 6f). Ninety-six of these genes were found to be associated with neurofunction (Table S6). The overlapping genes were subjected to GO-BP analysis, revealing significant enrichment of the positive regulation of cell projection organization (Fig. 6h). Additionally, KEGG pathway analysis indicated a significant enrichment of the nucleocytoplasmic transport pathway (Fig. 6i). Among the 30 downregulated and upregulated genes selected on fold change (FC), 21 genes were related to the nervous system, among which MAP1B, GABBR1 and RELN exhibited the most significant differences (Fig. 6g). MAP1B encodes a protein crucial for maintaining synaptic function[36], and GABBR1 encodes GABA receptors involved in inhibitory synaptic transmission[37]. RELN is essential for cell localization, neuron migration, and synaptic plasticity during brain development[38]. These findings suggest that DDX53 likely plays a crucial role in regulating the expression of genes involved in synaptic function through direct mRNA binding.