Functional identification of two variants in unrelated Chinese patients with DNM1L-related mitochondrial disorders

Abstract

Background

The DRP1 protein, a member of the dynamin superfamily of GTPases, is encoded by the dynamin-1-like (DNM1L) gene and plays a critical role in mitochondrial fission. There was significant clinical heterogeneity in DNM1L-related disorders.

Methods

Whole exome sequencing (WES) was used to identify potential genetic causes of the phenotype in probands. Bioinformatics analysis was performed to analyze the pathogenicity of the identified variants, and 3D protein modeling was constructed to predict their effects on protein structure. Preliminary studies of the functional effects of the variant sites on the encoded proteins were performed by in vitro experiments.

Results

Two de novo variants, c.1049G>C (p.Gly350Ala) and c.2161C>T (p.Gln721*), were detected in affected individuals. One patient presented with severe epileptic encephalopathy while the other exhibited a distinctive clinical phenotype of hemiparesis. In silico analysis, conservative analysis, and 3D homology modelling indicated that the p.Gly350Ala and the p.Gln721* variants are deleterious. Furthermore, the results of the artificial transfection experiments demonstrated that the p.Gly350Ala variant resulted in a reduction in DNM1L expression at both the transcriptional and protein levels (p < 0.05). In contrast, the p.Gln721* variant exhibited no significant alteration in protein levels (p = 0.08), although it did result in a reduction in mRNA levels.

Conclusions

The present findings suggest that these variants may contribute to DRP1 deficiency, potentially triggering a range of DNM1L-related disease phenotypes. This study serves to expand the spectrum of variants associated with DNM1L-related disorders.

Introduction

Mitochondria undergo continuous fusion and division processes, which enable them to exhibit enhanced interaction behavior [1]. The balance of fusion and fission is crucial for the maintenance of mitochondrial reticular morphology and the regulation of mitochondrial activity and function [2]. Mitochondrial function is regulated by a dual genome comprising mitochondrial DNA (mtDNA) along with nuclear DNA (nDNA). The mtDNA encodes a small fraction of the mitochondrial proteins, whereas the nDNA encodes more than 99% of the mitochondrial proteins that are synthesized in the cytoplasm and then transferred to the mitochondria [3, 4]. Clinical manifestations and genetic features of mitochondrial diseases are highly diverse. Mutations in genes associated with mitochondrial fission and fusion have been documented as a causative factor for certain neurodegenerative diseases [5]. Furthermore, pathogenic variants that result in mitochondrial dysfunction have been linked to a spectrum of diseases, ranging from isolated optic atrophy to fatal encephalopathy [3, 6].

DNM1L, located on the chromosome 12p11.21, is responsible for encoding GTPase dynamin-1-like protein (also referred to as dynamin-related protein 1 or DRP1), an important central effector in mitochondrial fission [7]. DRP1 is a protein found in the cytoplasm. It moves to the outer mitochondrial membrane (OMM) and consists of a large GTPase domain, a middle domain, a non-conserved variable domain, as well as a GTPase effector domain (GED). These are involved in regulating GTPase activity and oligomerization [8]. Upon activation, DRP1 is recruited to the OMM from the cytoplasm via receptors. Once at the OMM, DRP1 oligomerizes and hydrolyzes guanosine triphosphate (GTP), assembling into helical filaments which wrap and narrow the membrane [5, 9]. This process enables organelle division. Additionally, DRP1 mediates peroxisome fission, which is crucial for the maintenance of normal peroxisome morphology and function [10].

DRP1 impairment is linked to de novo dominant and compound heterozygous variants in DNM1L, which can lead to a severe neurological disorder known as encephalopathy due to defective mitochondrial and peroxisomal fission-1 (EMPF1, MIM#614388), in addition to isolated autosomal dominant optic atrophy 5 (MIM#610708). EMPF1 is characterized by early onset with developmental delay, progressive encephalopathy, dystonia (hypertonia or hypotonia), ataxia, and nystagmus. Furthermore, numerous patients present with refractory seizures and followed by a neurological decline and/or potentially fatal encephalopathy during childhood, indicating consistency with a lethal condition [11]. On the whole, individuals with the DNM1L variant have a variable and unpredictable clinical course.

The phenotypic and genetic heterogeneity observed in mitochondrial diseases has led to molecular genetic testing becoming a crucial testing modality [12]. However, novel sequence alterations may not be classified as either benign or pathogenic since there is a lack of functional evidence. In addition, there have been only 52 documented cases worldwide as of now, since Waterham et al. identified DNM1L in 2007 [13]. Further research is required in order to gain a more comprehensive understanding of this rare disease.

The present study describes a case series of two sporadic patients with de novo heterozygous variants (c.1049G>C, p.Gly350Ala and c.2161C>T, p.Gln721*) of the DNM1L gene. Although their clinical presentation is incompletely similar, we summarized their clinical features and follow-up information. Furthermore, bioinformatics analysis and in vitro functional experiments were employed to evaluate the effects of these two variants on the expression and stability of the DNM1L protein.

Materials and methods

Study subjects and ethical considerations

From 2020 to 2023, a total of two unrelated Chinese families were recruited to participate in this study. A comprehensive clinical data set was assembled, along with detailed family histories. The study was approved by the Ethics Committee of Children’s Hospital Affiliated to Zhengzhou University (2024-K-081) and was conducted in accordance with the principles outlined in the Declaration of Helsinki. Prior to the commencement of the study, written informed consent was obtained from both parents.

Whole-exome sequencing

As previously described, genomic DNA was extracted from the peripheral blood of the major members of both families [14]. The processing and analysis of the WES data were conducted as described previously [15]. Briefly, available DNA was sequenced on the Novaseq 6000 platform (Illumina, San Diego, USA) with 150 bp paired-end reads. The sequencing reads were aligned to the human reference genome (hg19/GRCh37). The use of the Picard v1.57 software (http://picard.sourceforge.net/) enabled the removal of PCR duplicate sequences. The variant annotation and interpretation were conducted in accordance with the previously described methodology [16, 17]. Sanger sequencing was then performed to confirm potential disease-causing variants.

Bioinformatics analysis

To investigate the potential functional consequences of the different variants on the corresponding protein, pathogenicity prediction was conducted using REVEL (https://sites.google.com/site/revelgenomics/), FATHMM (http://fathmm.biocompute.org.uk/), and CADD (http://cadd.gs.washington.edu/score), respectively. All suspected variants were classified in accordance with the guidelines of the American College of Medical Genetics and Genomics (ACMG) [18]. Amino acid sequences and 3D structure of the DNM1L1 protein (PDB: 5WP9, 4.22 Å; PDB: 3W6O, 1.90 Å) were obtained from the UniProt database (https://www.uniprot.org/) and Protein Data Bank (https://www.rcsb.org/). The MEGA11 software was utilized to analyse the degree of preservation [19], and the PyMOL software was employed to visualize the 3D structure of the protein.

Plasmid construction and transfection

The entire sequence of DNM1L cDNA was cloned into the phage expression vector, which was purchased from Bioeagle Biotech Company. Wild-type (WT) cDNA plasmid (pcDNA3.1-Flag-EFla-copGFP-wt) was acquired by double digestion and ligation of the amplification products with pcDNA3.1-EFla-copGFP using a pair of amplification primers (DNM1L-BamHI-F/DNM1L-NotI-R). The WT plasmid was employed as a template for targeted mutagenesis using the Mut Express II Rapid Mutagenesis Kit V2 (Vazyme, Nanjing, China). The sequences of primers used in these experiments are presented in Table 1. All the resulting recombination plasmids were verified via Sanger sequencing. For the transfection experiments, 293 T cells were incubated in six-well plates in Dulbecco’s Modified Eagle Medium (DMEM; Gibco, Grand Island, USA) supplemented with 10% fetal bovine serum at 37°C with 5% CO₂. When the cells reached 70% − 80% confluence, the DNM1L-wt/mut plasmids (wild-type or mutated) were transfected into 293 T cell lines using Lipofectamine™ 3000 (Invitrogen, Carlsbad, USA) according to the manufacturer’s instructions. Following a 48-hour incubation period, total RNA and proteins were extracted from transfected 293 T cells for subsequent real-time quantitative PCR (qRT-PCR) and western blotting (WB) experiments.

Table 1.

The primers for plasmid construction, point mutation, and qRT-PCR

Name	Sequence (5′→3′)
DNM1L-BamHI-F	gctcggatccATGGAGGCGCTAATTCCTGT
DNM1L-NotI-R	tcgagcggccgcTCACCAAAGATGAGTCTC
DNM1L-mut1-F	TAACACTATTGAAGCAACTGCAAAATATAT
DNM1L-mut1-R	ATATATTTTGCAGTTGCTTCAATAGTGTTA
DNM1L-mut2-NotI-R	tcgagcggccgcTCACCAAAGATGAGTCTCCCGGATTTCAGCAATAATTTGACTGGCTCCTTATAATGCCTTTAGCA
DNM1L-qRT-PCR-F	ACCCGGAGACCTCTCATTCT
DNM1L-qRT-PCR-R	TTGACAACGTTGGGTGAAAA

mut1, c.1049G>C, p.Gly350Ala; mut2, c.2161C>T, p.Gln721*

Real-time quantitative PCR

Total RNA was extracted according to the manufacturer’s protocol from Takara, Japan. DNM1L mRNA expression was then detected by qRT-PCR. The primer sequences are listed in Table 1. The relative mRNA expression levels of each dataset were processed by the 2^−ddCT method and normalized to the housekeeping gene β-Actin.

Western blot

The proteins derived from the transfected cells were collected and quantified utilizing a BSA reagent (Thermo Fisher Scientific, Waltham, USA). Following standard operating procedures, polyacrylamide gel electrophoresis was performed. Then, the proteins were incubated with primary anti-DNM1L antibody (1:2000 dilution, Mabnus, Wuhan, China) and internal reference diluted with anti-GAPDH antibody (1:1000 dilution, Cell Signaling Technology, Danvers, USA) at 4 °C overnight. After washing, the blots were visualized with chemiluminescent substrate, followed by analysis via the Image J software.

Statistical analysis

All statistical analyses in this study were conducted using the Prism 8 software (GraphPad, La Jolla, USA). A t-test was employed to ascertain the statistical significance of the observed difference between the two groups, with p < 0.05 deemed to be the threshold for statistical significance.

Results

Clinical characteristics of patients

Patient 1 is the sole offspring of unrelated parents. She was born at full term following a pregnancy without incident and had a family history that presented no notable issues. The child was noted to have normal speech and motor development from the neonatal phase to three years of age. Nevertheless, her intellectual development was subpar in comparison to her peers (IQ = 62), and regular language developmental milestones were not achieved by the age of three (see Table 2). She was hospitalized at around 6.5 years old due to recurring nocturnal seizures caused by mountaineering. Following this, she exhibited persistent myoclonic status epilepticus in both lower limbs for two weeks, which improved with symptomatic treatment. Several tests were conducted to identify the cause of the epilepsy, but apart from a vitamin D deficiency, the other physical examinations and supporting tests were not indicative, including normal brain magnetic resonance imaging (MRI) and electroencephalogram (EEG). A seizure recurred about two months later, leading to admission to intensive care for refractory status epilepticus (RSE). Subsequently, she developed tremors, requiring assistance to walk even short distances, and experienced 1 to 2 seizures each day after waking from sleep.

Table 2.

Clinical features of the probands

Patients (sex)	Age at onset	Outcomes	Clinical features or relevant biochemical findings						Other findings
			Intellectual disability	Language retardation	Seizures	MRI findings	EEG findings	Lactate level
P.t 1 (F)	3 y	Alive, 9 y	+	+	+	Abnormal	Abnormal	+	Dyskinesia, encephalopathy
P.t 2 (F)	1.5 y	Alive, 5.5 y	-	+	-	Normal	Normal	+	Hemiplegia, strabismus

P.t patient, F female, y years, (+) presence, (-) absence

When she was 6 years and 9 months old, a generalized tonic-clonic seizure (GTCS) associated with multifocal generalized epileptiform abnormalities were recorded through polymyographic video/EEG. Basal ganglia damage was shown in the brain MRI (data not shown), and laboratory tests suggested mildly elevated plasma lactate (2.58 mmol/L; reference range 0.7–2.1 mmol/L) as well as increased aspartate aminotransferase (49.05 U/L; reference range 13–35 U/L). However, hereditary aminoacidopathies and organic aciduria were ruled out. She had been administered coenzyme Q₁₀ (100 mg/day), L-carnitine (1 g/day), vitamin B₁ (300 mg/day), vitamin B₂ (50 mg/day), vitamin C (600 mg/day), benzhexol hydrochloride (2 mg/day), and antiepileptic therapy. It is regrettable that, despite the absence of further deterioration in the patient’s condition, the symptoms did not appear to improve with Coenzyme Q₁₀ and L-carnitine treatment, as observed in the other patients mentioned in our previous article [15]. At a recent assessment (7 years and 2 months), the child demonstrated comprehension skills and could recite stories, though her speech appeared to be slurred. Furthermore, motor regression with significant tremors was detected, and she was unable to sit on her own. Minimal sleep was observed; however, it seemed to be relatively more restful than previously reported.

Patient 2, a second child born to non-consanguineous Chinese parents, was delivered at full term with regular birth parameters. The detailed clinical features of the proband are as previously described [15]. Briefly, she was delayed in speech development and developed left-sided limb dysfunction at the age of 1.5 years, which progressed to hemiparesis. Physical and ancillary examinations were unremarkable (see Table 2). Furthermore, a two-year follow-up study revealed that the administration of coenzyme Q₁₀ (20 mg/d) and L-carnitine (1 g/d) was effective in addressing the patient’s hemiplegic phenotype. Her current intelligence quotient (IQ) score is 75.

Identification of two novel variants in DNM1L

To identify potential pathogenic variants in patients, we performed parallel sequencing techniques. Sequencing of the mitochondrial DNA genome from blood-derived DNA of P.t 1 did not detect any possible pathogenic variants. Diagnostic analysis of the patient/parent trio using WES revealed a novel heterozygous c.1049G>C DNM1L (NM_012062.5) variant (see Fig. 1), resulting in the amino acid substitution NP_036192.2: p.Gly350Ala. The occurrence was confirmed as de novo through Sanger sequencing analysis. The de novo DNM1L c.1049G>C, p.Gly350Ala variant was classified as “Pathogenic” (see Table 3) according to ACMG criteria PS2, PM1, PM5, PS3_supporting, PM2_supporting, PP2, and PP3.

Fig. 1.

Family pedigrees and results of variants of the probands. The presence of the c.1049G>C (p.Gly350Ala) variant in P.t 1 and the c.2161C>T (p.Gln721*) variant in P.t 2 was confirmed by Sanger sequencing. Both variants were de novo heterozygous. The arrow indicates the proband. Females and males are denoted by circles and squares, respectively

Table 3.

 Variant identified in P.t 1 and prediction of pathogenicity of c.1049G>C

Table 3 Variant identified in the patient 1 and prediction of pathogenicity of c.1049G>C
Gene	Variant	Variant loci	Source	REVEL		FATHMM		CADD		Variant classification
				Score	Prediction	Score	Prediction	Score	Prediction
DNM1L (NM_012062.5)	c.1049G>C, p.Gly350Ala	Exon 9, Middle domain	De novo	0.92	D	0.99	D	26.9	D	P

D, Damaging; P, Pathogenic

No pathologically significant variants were observed throughout the mtDNA genome of P.t 2. Trio-WES analysis of P.t 2 revealed an additional stop-gain variant, c.2161C>T, p.Gln721* in DNM1L (see Fig. 1). And the de novo DNM1L c.2161C>T, p.Gln721* nonsense variant was categorized as “VUS” according to the ACMG guidelines, using the criteria: PS2_Moderate, PM2_supporting, and PP3.

Neither variant was identified in the 1000 Genomes Project, the Genome Aggregation Database (gnomAD), or the dbSNP database. The c.1049G>C (p.Gly350Ala) variant was predicted to be deleterious by online bioinformatics tools (see Table 3). A conservation analysis was conducted, which demonstrated that the positions of the c.1049G>C (p.Gly350Ala) and c.2161C>T (p.Gln721*) variants are highly conserved across different species (see Fig. 2B). Additionally, Fig. 2C presents the protein three-dimensional (3D) modeling of the DNM1L-wild and identified variants. As shown in Fig. 2C, the P.t 1 involves a G350 residue that adopts a φ angle of 89.7° and a ψ angle of 19.6°, thus potentially disfavoring residues other than glycine. The substitution of glycine by a charged amino acid or other amino acids may result in unfavorable spatial conflicts with neighboring residues. Furthermore, the variant c.2161C>T (p.Gln721*) alters the spatial relationships of the DNM1L protein, resulting in the truncated protein losing its interaction with neighbouring amino acid residues, which may reduce the stability of the DNM1L protein.

Fig. 2.

Location diagram and conservation analysis of DNM1L variants. A Schematic representation of DNM1L protein domains with localization of identified variants in patients (red font). B A comparison of the amino acid sequences of DNM1L proteins from different species. The amino acid sequences of the variant sites (c.1049G>C, p.Gly350Ala and c.2161C>T, p.Gln721*) are evolutionarily conserved and highlighted with a red box. C The 3D modeling of the protein of the DNM1L-wild and identified variants

Variants altered the transcript and protein levels of DNM1L

In order to investigate the functional impact of c.1049G>C (p.Gly350Ala) and c.2161C>T (p.Gln721*) variants, an artificial transfection system was constructed, followed by protein expression and qRT-PCR analysis. Sequencing in Fig. 3A and B confirmed the successful integration of the c.1049G>C (p.Gly350Ala) and c.2161C>T (p.Gln721*) variants. The results of qRT-PCR demonstrated a notable reduction in mRNA transcript levels for the p.Gly350Ala and p.Gln721* variants in comparison to the WT (p < 0.05, see Fig. 3C). Furthermore, as illustrated in Figs. 3D and E, WB assays demonstrated that the p.Gly350Ala variant resulted in decreased protein levels, in contrast to the p.Gln721* variant, which exhibited no significant difference compared to WT (p = 0.08).

Fig. 3.

Functional consequences of the DNM1L variant in vitro. A and B Sequencing results showed that the c.1049G>C (p.Gly350Ala) and c.2161C>T (p.Gln721*) variants were successfully introduced; C The results of qRT-PCR revealed that the p.Gly350Ala and p.Gln721* variants caused a decreased mRNA transcript level compared to the WT (t-test, *:p < 0.05; **:p < 0.01); D and E The results of the WB analysis demonstrated that the p.Gly350Ala variant was associated with a reduction in protein levels, while there was no significant difference between the p.Gln721* variant and the WT (t-test, p = 0.08). All experimental results were based on three independent biological replicates

Discussion

Since Waterham et al. reported in 2007 [13] that DNM1L deficiency could cause mitochondrial disease characterized by global developmental delay and brain developmental abnormalities in infancy, significant progress has been made in comprehending DNM1L-related disorders. Subsequent studies have further revealed the phenotypic heterogeneity of DNM1L variants, with reported disease manifestations including early death, seizures, and dominant optic atrophy [20–22]. Given the broad spectrum of disease outcomes observed in affected individuals, comprehensive characterization of novel disease-related genes and systematic pathogenicity studies are of critical academic and clinical significance. These efforts are essential for the precise interpretation and diagnosis of DNM1L variants, as well as for understanding how DNM1L variants alter physiology.

Pathogenic variants of DNM1L have been linked to a neurological disorder known as “EMPF1”, with the core clinical features of developmental delay/regression, hypotonia, encephalopathy, seizures, ataxia, persistent lactic acidosis, and peripheral neuropathy [9, 23–25]. The present study investigated two patients with DNM1L variants, revealing phenotypic heterogeneity in disease severity. While both patients achieved normal developmental milestones in infancy, their subsequent clinical courses differed markedly. P.t 1 developed progressive severe epileptic encephalopathy, accompanied by global developmental delay, RSE, and basal ganglia damage. These symptoms are closely resembling those previously reported in DNM1L cases. In contrast, P.t 2 exhibited hemiparesis, a distinct clinical manifestation not previously described in DNM1L-associated disorders, thus expanding the phenotypic spectrum of DNM1L variants. Furthermore, treatment with coenzyme Q₁₀ (20 mg/d) and L-carnitine (1 g/d) has been demonstrated to be efficacious in the management of the hemiplegic phenotype.

To date, over thirty pathogenic or likely pathogenic DNM1L variants have been reported, with more than 70% identified as de novo variants and over 60% localized to the middle domain [2, 7, 23, 26, 27]. Clinical phenotype-genotype correlation studies have revealed domain-specific associations: variants in the middle domain are more frequently linked to severe manifestations such as RSE, as exemplified by Patient 1, who harbored a middle domain variant and presented with the RSE phenotype consistent with prior reports [15, 23]. In contrast, milder clinical phenotypes including optic atrophy and ataxia are linked to the GTPase domain [9, 22, 28, 29]. Consequently, the severity of the clinical phenotype might be related to the domain in which the DNM1L variants are located. Herein, we present our findings of two potentially pathogenic variants in the DNM1L gene, comprising a novel missense variant in the middle domain (p.Gly350Ala) and a novel stop-gain variant in the GED domain (p.Gln721*). These variants result in amino acid alterations at positions 350 and 721, respectively, which are highly conserved across the dynamin family and have not been documented in general population databases.

Pathogenic variants in the middle domain are typically associated with severe clinical and functional outcomes. For instance, the lethal variant in the middle domain of DRP1 (p.Ala395Asp) has been documented in neonates with microcephaly, lactic acidemia, and abnormal brain development [30]. Mechanistically, missense variants in conserved amino acids, such as p.Ala395Asp and p.Gly350Asp, cause mitochondrial elongation in HeLa cells, resulting in impaired higher-order DRP1 protein assembly and function [30]. Furthermore, Chao et al. [31] identified a heterozygous variant (c.1048G>A) in the DNM1L gene leading to a p.Gly350Arg substitution at a greatly conserved residue within the middle domain. Functional analyses demonstrate that the introduction of the p.Gly350Arg variant in Drosophila lead to a dominant-negative effect, resulting in an increase in peroxisome size, an altered cellular distribution, and aberrant mitochondrial morphology and transport. Notably, P.t 1 in our study possessed a distinct missense substitution that impacted this amino acid (p.Gly350Ala). Although her clinical manifestations were consistent with the aforementioned case, follow-up records indicate she is currently alive. The different outcome between the two patients could be explained by different amino acid changes (p.Gly350Arg vs. p.Gly350Ala) or different inherited genetic backgrounds. Additionally, long-term monitoring of P.t 1 is essential to determine whether the p.Gly350Ala variant leads to other clinical outcomes.

Intramolecular interactions between the GEDs and the GTPase-middle domain, as well as the interactions between adjacent GEDs, are critical in regulating DRP1 GTP hydrolysis [32]. Studies have demonstrated that pathogenic GED variants (p.Tyr691Cys, and p.Arg710Gly) could disrupt the GED-GTPase interactions, thereby impeding GED-GED assembly and reducing GTPase activity [2, 11]. Clinically, patients with GED variants generally display milder phenotypes compared to those with middle domain variants, such as seizures, impaired mobility, and optic atrophy. We have identified a de novo variant p.Gln721* in P.t 2, which is very close to the missense change reported previously [11]. The patient did not exhibit the typical clinical manifestations observed in known GED variant cases; instead, a distinct phenotype was detected, which raises questions about the pathogenicity of p.Gln721*.

To further examine the functional effects resulting from these variants, we assessed the impact of DNM1L variants on mRNA and protein expression levels in vitro. The qRT-PCR analysis demonstrated that both variants cause a significant decrease in DNM1L expression at the transcriptional level compared to the WT (p < 0.05). Furthermore, the WB results revealed that the p.Gly350Ala variant resulted in decreased protein levels, whereas the p.Gln721* variant exhibited no significant difference in protein abundance relative to WT (Fig. 3D and E). Regarding the “nonlinear” relationship between mRNA and protein expression of the p.Gln721* variant, two hypotheses were proposed to explain this phenomenon: (1) As a very late truncation variant, p.Gln721* is expected to escape nonsense-mediated decay (NMD) [33], leading to the production of stable truncated proteins; (2) DNM1L proteins have been demonstrated to induce dominant negative gain-of-function in DNM1L-related mitochondrial diseases, where sufficient protein levels exert an effect on complex function [31]. While mRNA levels are decreased (~ 30%), this reduction is insufficient to lower protein abundance below a functional threshold. Consequently, we speculate that the variant’s pathogenicity likely stems from the dominant negative effect of truncated DNM1L proteins (e.g., impaired GTPase activity and oligomerization with wild-type monomers), rather than mRNA reduction itself [30]. However, the results of functional experiments also cast doubt on the pathogenicity of the p.Gln721* variant. For instance, under the assumption that p.Gln721* is a loss-of-function (LOF) variant, this heterozygous mutation would result in a 50% reduction in protein expression [34]. However, this finding is inconsistent with the experimental results obtained. Consequently, we hypothesise that this variant may not meet the PVS1 criteria, which would consequently necessitate its reclassification as a VUS.

Nevertheless, it should be noted that this study is not without its limitations. Firstly, the absence of residual material from patients renders the applicability of the artificial transfection system we have created to human disease states uncertain. Therefore, the effects of the p.Gly350Ala and p.Gln721* variants on mitochondrial and peroxisomal morphology in patient cells and/or artificially transfected cells warrant further investigation. Secondly, Chang et al. reported that mutations at position 350 of DNM1L disrupt DRP1 oligomerization and the formation of fission complexes, potentially leading to DRP1 degradation in isolated mitochondria [30]. Importantly, however, total cellular protein expression levels of the mutant DRP1 did not differ significantly from those of the wild-type [30]. In contrast, the present study observed a reduced expression of the p.Gly350Ala variant at the total cellular protein level, a phenomenon distinct from the above-described scenario. This discrepancy is speculated to result from decreased mRNA stability, such as degradation caused by structural abnormalities, although this hypothesis awaits further experimental validation.

In conclusion, two de novo DNM1L variants were identified in this study. The results of in silico analyses and functional experiments suggest that the variants may contribute to some extent to DRP1 defects, which may trigger the emergence of a range of DNM1L-related disease phenotypes. Encouragingly, P.t 2 appears to exhibit the mildest clinical phenotype to date, and long-term follow-up will be conducted to understand the progression and outcome of DNM1L-related disorders. Furthermore, considering the clinical heterogeneity of mitochondrial disorders, more clinical data and functional studies of DNM1L variants are needed to elucidate the genotype-phenotype correlations of DNM1L-related disorders. It is also recommended that DNM1L variants be considered in cases of developmental delay, epilepsy, dystonia, and/or hemiparesis.

Authors’ contributions

Z.Z conceived the study and drafted the manuscript. Z.C and X.B collected the clinical data. Z.X and X.L conducted the survey and follow-up. J.L, M.X, Q.Z and Y.Z contributed to the analysis of data. D.L funded the project. Y.Y and D.L reviewed and edited manuscript.

Funding

This study was funded by the National Natural Science Foundation of China (Project Number: 82000850) and the Henan Provincial Medical Science and Technology Research Joint Venture Project (Project Number: LHGJ20250618).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of the Children’s Hospital Affiliated to Zhengzhou University. Signed informed consent was obtained from the patient’s parents.

Consent for publication

The written informed consent for publication of clinical details were obtained from all parents.

Competing interests

The authors declare no competing interests.