Novel VRK1 Variants and a Founder Effect in Axonal Polyneuropathy

Abstract

Background and Objectives

Biallelic VRK1 variants are linked to a range of neurogenic disorders, including Charcot-Marie-Tooth and related disease (CMTR), motor neuron diseases, and spinal muscular atrophy (SMA). This study aimed to characterize the genetic and clinical features of VRK1-related peripheral neuropathy in a Chinese cohort with axonal CMTR.

Methods

Eight families were identified from a cohort of 351 families with axonal CMTR, including axonal CMT and distal hereditary motor neuropathies (dHMN). Clinical and genetic data were retrospectively analyzed. Haplotype analysis investigated the founder effect of the high-frequency variant, and RNA sequencing was conducted for the splice variant. A systematic literature review (PubMed/CNKI, 2009–2025) was performed to compile global cases.

Results

VRK1 was the third most common gene in our axonal recessive CMTR cohort. Ten VRK1 variants were identified, including 8 novel variants: c.7C>T, c.83T>G, c.215T>G, c.539C>T, c.879_882del, c.974T>A, c.1073_1076del, and c.1159+2T>A. The nonsense variant c.1124G>A (p.W375*) was present in 4 of 8 families with haplotype analysis supporting a founder effect. RNA sequencing confirmed c.1159+2T>A caused exon 12 skipping and frameshift (p.K357Vfs*39). The clinical phenotypes included dHMN in 7 cases and axonal recessive CMT in one case. Two patients with dHMN exhibited upper motor neuron signs. All patients presented with progressive distal lower limb weakness, without neurodevelopmental impairment. In total, 53 cases from 37 families reported to date (including the present cohort) revealed substantial clinical heterogeneity, with dHMN (28.8%) and SMA (26.9%) being most common. Early-onset cases exhibited rapid progression, whereas later-onset cases had slower courses.

Discussion

This study identifies 8 novel VRK1 variants and confirms the founder effect of p.W375*, expanding the genotypic and phenotypic spectrum of VRK1-related peripheral neuropathy. The clinical presentation in this cohort was predominantly motor-dominant, ranging from dHMN to axonal CMT, with notable phenotypic heterogeneity influenced by the specific protein domain involved. These findings refine the understanding of VRK1-related peripheral neuropathy and provide valuable insights for accurate diagnosis and management.

Introduction

Vaccinia-related kinase 1 (VRK1) is a serine/threonine kinase that plays a key role in cellular processes such as cell cycle regulation, DNA repair, and the assembly of Cajal bodies, all of which are essential for the survival and function of motor neurons.^1-3 Initially, VRK1 variants were identified in patients with pontocerebellar hypoplasia or microcephaly accompanied by axonal peripheral neuropathy (PN).⁴ Subsequent studies reported VRK1 variants in individuals with motor-dominant phenotypes, including distal hereditary motor neuropathies (dHMN) and axonal recessive Charcot-Marie-Tooth disease (axonal recessive CMT), both of which are part of Charcot-Marie-Tooth and related disease (CMTR), as well as in spinal muscular atrophy (SMA) and amyotrophic lateral sclerosis (ALS), all without the presence of pontocerebellar hypoplasia.^5-8 All reported VRK1 variants are inherited in an autosomal recessive (AR) manner, either as homozygous or compound heterozygous variants.⁹ Despite increasing case reports, the genotypic and phenotypic spectrum of VRK1 variants remains underexplored, especially in different populations.

In this study, we investigated 8 unrelated Chinese families with VRK1 variants and conducted a systematic review of global cases of VRK1-related neurogenic motor disorders to explore genotype-phenotype correlations. Our findings expand the understanding of VRK1-related PN and offer valuable insights for improving genetic diagnosis and clinical management.

Methods

Patient Enrollment

From 2012 to 2025, we recruited 351 unrelated Chinese families exhibiting distal limb muscle weakness and atrophy, clinically diagnosed with axonal CMTR. In this study, CMT was defined as a sensorimotor neuropathy, whereas dHMN was defined as a purely motor neuropathy. All patients were recruited at Department of Neurology, the Third Xiangya Hospital. All probands underwent thorough neurologic examinations by 2 neurologists. Routine hematologic, blood biochemical, and electrophysiologic examinations were performed for all probands and available family members. Brain and lower limb muscle MRI was conducted for some index probands. Muscle fatty infiltration was assessed on axial T1-weighted non–fat-suppressed images using the Mercuri visual grading system.¹⁰ Grades were defined as 0 (normal muscle), 1 (mild moth-eaten appearance), 2 (early confluent infiltration, <30% muscle area), 3 (moderate infiltration, 30%–60% muscle area), and 4 (severe infiltration, >60% muscle area). Disease severity was evaluated using the CMT Neuropathy Score (CMTNS).¹¹ Diagnosis of dHMN or axonal CMT was based on clinical criteria established by the European CMT Consortium.¹² It should be noted that portions of this cohort, specifically the subgroups of patients harboring SORD and IGHMBP2 variants, have been reported in our previous studies.^13,14

Variant Screening in Axonal CMTR Cohort

Peripheral blood samples (10 mL) were collected from all participants, and genomic DNA was isolated using the standard phenol-chloroform method. Genomic DNA from index patients was subjected to whole exome sequencing (WES). Exome capture was conducted using the Sure Select Human All Exon V6 Kit (Agilent Technologies, Santa Clara, CA), followed by high-throughput sequencing on the Illumina Hiseq 2500 platform (Illumina, San Diego, CA). FASTQ files were aligned to the human reference genome (hg19) using BWA. Variant calling was performed using the Genome Analysis Toolkit, and variant annotation was performed using ANNOVAR. Variants with a minor allele frequency greater than 0.05 were excluded based on the Exome Aggregation Consortium.¹⁵ Furthermore, the detected VRK1 variants were filtered against several population database including 1000 Genomes Project (1000G),¹⁶ Genome Aggregation databases (gnomAD),¹⁷ and the single-nucleotide polymorphism Database (dbSNP).¹⁸ In silico analysis of the variants was conducted using MutationTaster, CADD, SIFT, and PolyPhen-2 to predict variant pathogenicity.^19-22 PCR and Sanger sequencing were used to verify WES results and perform cosegregation analysis for candidate variants. To investigate the structural implications of these variants, a clustering framework based on prior studies was applied, categorizing them into 3 regions within the protein: Cluster I (ATP-binding site), Cluster II (catalytic site), and Cluster III (C-terminal regulatory region).^1,23 Gene functions, inheritance patterns, clinical phenotypes, and variants pathogenicity were annotated and classified according to the Online Mendelian Inheritance in Man²⁴ and the American College of Medical Genetics and Genomics (ACMG) standards.²⁵

Haplotype Analysis

SNP microarray data were obtained from 4 unrelated individuals carrying the pathogenic Trp375* variant across 4 VRK1-related dHMN families. Genotyping was performed using the Illumina Infinium Asian Screening Array (ASA) kit, with raw data analyzed using DRAGEN and BCFtools software. Shared identical-by-descent (IBD) segments overlapping the Trp375* variant were analyzed to explore potential founder effects. In addition, 5 microsatellite markers near VRK1 Trp375* were selected from UCSC genome browser (D14S987, D14S1067, D14S611, D14S979, and D14S65). PCR and capillary electrophoresis were used to identify the shared DNA segment.

Transcript Analysis

To assess the impact of the c.1159+2T>A variant on VRK1 exon 12 splicing in the affected individual (F8:II-1), mRNA was extracted from peripheral blood mononuclear cells, reverse-transcribed into cDNA, and subjected to mRNA sequencing by XiangYin Biotechnology Co. Ltd. (Hangzhou, China). The resulting FASTQ files were aligned and quantified using the STAR aligner, and transcript exon junctions were visualized with Integrative Genomics Viewer. Alternative splicing events were analyzed using rMATS software (v4.2.0), and sashimi plots were generated using the rmats2sashimiplot script (v3.0.0). mRNA sequencing was also performed on the unaffected parent (F8:I-1) as a control.

Literature Review and Statistical Analysis

A systematic search was conducted on PubMed²⁶ and CNKI²⁷ databases for articles published from 2009 to 2025 regarding VRK1-related neurogenic motor disorders, using keywords (“VRK1,” “distal hereditary motor neuropathy,” “spinal muscular atrophy”, “motor neuron disease”, “Charcot-Marie-Tooth”). Studies involving unvalidated cases or duplicates were excluded. Data from 22 published studies were combined with 8 Chinese families identified in our cohort. Information on variants, pedigree information, demographics, clinical features (age at onset, phenotypes), and comorbidities were collected. Descriptive statistics was used to analyze ethnic distribution, age at onset, and genotype-phenotype correlations. The results were visualized using bar charts and stacked plots generated with GraphPad Prism 9.0 (GraphPad Software, Inc., San Diego, CA).

Standard Protocol Approvals, Registrations, and Patient Consents

The study was approved by the Ethics Committee of the Third Xiangya Hospital of Central South University (Ethics Approval No.024-S601). Written informed consent was obtained from all participants.

Data Availability

The data that support the findings of this study are available from the corresponding author on reasonable request.

Results

Clinical Findings

Among the 8 VRK1-related PN families, one family was diagnosed with axonal recessive CMT based on both sensory symptoms and electrodiagnostic findings, while 7 families presented with a dHMN phenotype, 2 of whom exhibiting upper motor neuron signs (Figure 1A). The age at onset ranged from early childhood to adulthood. Earlier onset was associated with more severe and rapidly progressive phenotypes, while later-onset cases displayed relatively slower clinical progression (Figure 1B). The detailed clinical manifestations of the probands were summarized in Table 1.

Figure 1. Clinical Features and Lower Limb Muscle MRI in the Probands of Family 4 and Family 7.

(A) Clinical phenotype distribution of 8 VRK1-related PN probands. (B)Age-dependent phenotypic spectrum of the VRK1 cohort stratified by age at onset (years). (C) Patient F4:II-1. (a) Atrophy of the muscles in the distal lower limbs. (b–c) Axial T2-weighted and fat-suppressed T1-weighted images reveal mild atrophy and fat infiltration (Mercuri grade 1) in the semimembranosus, semitendinosus, biceps femoris, and soleus muscles. (d–e) Axial views reveal diffuse mild to moderate atrophy and fat infiltration in the calf muscles (Mercuri grade 2–3). (D) Patient F7:II-1. (a, c) Marked atrophy of lower limbs and intrinsic hand muscles. (b) pes cavus and foot drop; foot deformities corrected by orthopedic surgery. (d) Spinal deformities including scoliosis, lumbar hyperlordosis, and pelvic anteversion. (e–f) T2-weighted brain MRI showing normal brachium pontis width and pons diameter, excluding pontocerebellar hypoplasia. (g–h) Axial images demonstrate moderate to severe atrophy and fat infiltration (Mercuri grade 3–4), predominantly affecting the medial and posterior thigh muscles, including gracilis, adductor longus, adductor magnus, semimembranosus, semitendinosus, biceps femoris, and soleus. (i–j) Axial T2-weighted and fat-suppressed T1-weighted images of the calves demonstrate diffuse moderate to severe atrophy and fat infiltration (Mercuri grade 3–4). White, yellow, and red arrows indicate Mercuri grades 2, 3, and 4 fat infiltration, respectively.

Table 1.

Clinical Features of 8 VRK1-Related Peripheral Neuropathy Probands

Family No.	1	2	3	4	5	6	7	8
Individuals	Ⅳ-3	Ⅵ-1	Ⅱ-1	Ⅱ-1	Ⅱ-2	Ⅱ-1	Ⅱ-1	Ⅱ-1
Sex/age, y	M/25	M/39	M/39	M/23	M/21	F/21	M/11	M/11
AAO, y	22	16	18	15	15	7	3	4
Variants	c.1124G>A(hom)	c.1124G>A(hom)	c.1124G>A /c.706G>A	c.1124G>A /c.215T>G	c.539C>T (hom)	c.83T>G/c.1073_1076del	c.7C>T/c.974T>A	c.879_882del/c.1159+2T>A
Amino acid change	p.W375*	p.W375*	p.W375* /p.V236M	p.W375* /p.V72G	p.A180V	p.l28R /p.E361Lfs*65	p.R3C/p.L325*	p.K295Qfs*11/splicing
Initial symptoms	Weakness of lower limbs	Weakness of lower limbs	Weakness of lower limbs	Weakness of lower limbs	Weakness of lower limbs	Weakness of lower limbs	Unsteady gait	Weakness of lower limbs
UL muscle strength (proximal/distal)	5/5	5/3	5/3	5/5	5/5	5/1	5/3	5/1
LL muscle strength (proximal/distal)	5/3	4/2	3/2	5/3	5/2	4/0	3/2	2/0
aMuscle atrophy (UL/LL)	−/+	+/++	++/+++	−/+	−/+	+/++	+/+++	++/+++
bReflexes (UL/LL)	−/−	−/−	−/−	+/−	+++/+++	+++/+	+/+	NA
cGait disturbance	+	++	++	+	+	++	++	+++
dSensory findings(UL/LL)	−/−	−/−	−/−	−/−	−/−	−/−	+/+	−/−
Upper motor neuron signs	−	−	−	–	+	+	−	−
Brain MRI	Normal	Normal	Normal	Normal	Normal	Normal	Normal	NA
Other symptoms	Pes cavus; feet drop	Pes cavus; feet drop	Pes cavus; feet drop; fine motor disorder of both hands	Pes cavus; feet drop	Pes cavus; feet drop	Pes cavus; feet drop; muscle cramps	Pes cavus; clubfoot; fine motor disorder of both hands; scoliosis; waddling gait	Pes cavus; feet drop
CMTNS	4	7	12	3	3	12	21	14

Abbreviations: AAO = age at onset; CMTNS = Charcot-Marie-Tooth disease neuropathy score; LL = lower limbs; NA = not available; UL = upper limbs.

^aMuscle atrophy: normal: (−); mild:(+); moderate: (++); severe (involved in proximal muscle):(+++).

^bReflexes: loss: (−); reduced: (+); normal (++); brisk: (+++).

^cGait disturbance: unable to walk: (+++); need walking stick:(++); drop-foot gait: (+), normal:(−).

^dSensory findings: hyperpathia: (+); normal sense: (−).

Five families (F1-5) had adolescent or adult onset. The probands in these families showed similar clinical presentations, initially manifesting as progressive distal lower limb weakness. Neurologic findings included varying degrees of distal limb muscle atrophy, pes cavus, foot drop, reduced or absent tendon reflexes, and steppage gait. Two patients (F2:VI-1 and F3:II-1) became nonambulatory after approximately 15–20 years of disease progression. CMTNS in this group ranged from 3 to 12.

By contrast, the remaining 3 patients (F6-8) experienced symptom onset during infancy or early childhood with rapidly progressive distal muscle weakness and atrophy, which later extended proximally. The most severe case in this cohort (F7:II-1) had onset at age 3 and became nonambulatory by age 10. Notably, 2 patients exhibited clinical signs, including brisk upper limb reflexes in F5:II-2 and bilateral Babinski signs in F6:II-1, suggestive of upper motor neuron involvement. The early-onset patient group had higher CMTNS scores ranging from 12 to 21.

Two patients (F4:II-1 and F7:II-1) underwent muscle MRI. F4:II-1 showed diffuse calf muscle atrophy and fatty infiltration (Mercuri grade 2), while F7:II-1 showed diffuse muscle atrophy and fatty infiltration in the calf and thigh (Mercuri grade 3–4). In addition, brain MRI was performed or reviewed for 7 probands. All examinations revealed normal brain structures, with no evidence of pontocerebellar hypoplasia or atrophy (Table 1 and Figure 1D).

Nerve conduction studies revealed consistent motor axonal neuropathy across all patients, with sensory involvement identified only in one case (F7:II-1). Needle EMG (EMG) showed chronic denervation in the tibialis anterior, gastrocnemius, rectus femoris, and interosseous muscles, with active denervation changes observed in most patients. Patient F5:II-2 demonstrated EMG abnormalities in the thoracic paraspinal muscles (T8, T10). The detailed electrophysiologic findings are presented in Table 2. These findings indicate that while lower motor neuron dysfunction is the predominant feature in VRK1-related PN, a subset of patients may also exhibit upper motor neuron signs and occasional sensory involvement, reflecting the broad clinical heterogeneity associated with VRK1 variants.

Table 2.

Electrophysiologic Findings of the 8 Families

Patient	Sex/Age (y)	Variants	Side	Motor nerve								Sensory nerve				Needle EMG findings
				Median nerve		Ulnar nerve		Peroneal nerve		Tibial nerve		Median nerve		Peroneus super nerve
				Motor amplitude (mV)	MNCV (m/s)	Motor amplitude (mV)	MNCV (m/s)	Motor amplitude (mV)	MNCV (m/s)	Motor amplitude (mV)	MNCV (m/s)	Sensory amplitude (uV)	SNCV (m/s)	Sensory amplitude (uV)	SNCV (m/s)
F1:Ⅳ-3	M/25	c.1124G>A	L	8	—	8.4		1.8	43	1.9	—	38.1	51.1	16.7	45.5	Chronic neurogenic changes
			R	7.1	—	7.1		2.5	45	2.4	42.9	46.9	51.7	8.4	42.4
F2:Ⅵ-1	M/45	c.1124G>A	L	—	—			0.4	47.9	1.5	49.7	51.1	65	29	49.5	Active denervation and chronic reinnervation
			R	7.4	67	5.4	52.2	0.1	43.2	1.1	52.8	57.5	62.5	42	51.2
F3:Ⅱ-1	M/39	c.1124G>A/c.706G>A	L	—	—			1.9	47.7	1.5	50.4	44.8	57.9	15.1	55	Active denervation and chronic reinnervation
			R	3.0	57.3	5.7	53.5	1.4	48.1	1.1	47	—	—	—	—
F4:Ⅱ-1	M/23	c.1124G>A/c.215T>G	L	8.9	64.3	7.7	56.5	0.63	45.5	4.6	45	12.8	71.8	14.8	51	Active denervation and chronic reinnervation
			R	9	62.3	10	57	1.96	46.6	1.86	49.5	17.6	70.3	12.2	48
F5:Ⅱ-2	M/21	c.539C>T	L	8.1	58.8	8.7	63.1	—	—	—	—	47.7	61.9	—	—	Active denervation and chronic reinnervation
			R	7.6	56.6	8.5	60.4	1.32	47.1	9.1	45.8	41.1	67.1	30.6	53.4
F6:Ⅱ-1	F/31	c.83T>G/c.1073_1076del	L	—	—			1.1	51.6	2.4	54.2	—	—	12.4	52.3	Active denervation and chronic reinnervation
			R	0.6	62.7	4	56.4	—	—	4.1	53.2	—	—	—	—
F7:Ⅱ-1	M/11	c.7C>T /c.974T>A	L	—	—			0.14	34.6	0.04	30.5	—	—	12.8	65.6	Active denervation and chronic reinnervation
			R	1.76	53.2	0.4	54.5	0.19	33.9	0.08	32.4	21	69.3	NP
F8:Ⅱ-1	M/11	c.879_882del /c.1159+2T>A	L	—	—			0.93	50.8	2.80	51.8	—	—	23.2	53.3	Active denervation and chronic reinnervation
			R	6.3	59.7	2.7	59.5	0.77	49.2	1.45	58.1	16.8	54.3	19.4	59.7

Motor amplitude (normal range: median nerve ≥5.0 mV; peroneal nerve ≥2.0 mV; tibial nerve ≥4.8 mV); MNCV = motor nerve conduction velocity (normal range: median nerve ≥50.0 m/s; peroneal nerve ≥37.0 m/s; tibial nerve ≥37.0 m/s); sensory amplitude (normal range: median nerve ≥20.0 μV; peroneus superficial nerve ≥6.0 μV); SNCV = sensory nerve conduction velocity (normal range: median nerve ≥44.0 m/s; peroneus superficial nerve ≥41.0 m/s); — = not done; NP = no potential; needle EMG findings, indicating needle EMG results from concentric needle EMG in distal lower limb muscles (e.g., tibialis anterior, gastrocnemius). “Chronic neurogenic changes” indicate reinnervation patterns with large-amplitude, polyphasic motor unit action potentials (MUAPs) and reduced recruitment; “active denervation and chronic reinnervation” includes spontaneous activity (fibrillation potentials, positive sharp waves) plus chronic changes, consistent with axonal motor neuropathy. All examinations were performed using standard protocols.

Genetic Findings and Analysis

In our axonal CMTR cohort, AR families and undiagnosed sporadic cases accounted for 45.8% (161/351) of the cases. Among these, 10 distinct VRK1 (NM_003384.3) variants were identified in 8 unrelated Chinese families, representing 5% (8/161) of total. VRK1 ranks as the third most common causative gene in this cohort after SORD and IGHMBP2 (Figure 2A).^14,28 The variants include c.7C>T(p.Arg3Cys), c.83T>G(p.Ile28Arg), c.215T>G(p.Val72Gly), c.974T>A (p.Leu325*), c.879_882del (p.Lys295Glnfs*11), c.1073_1076del(p.Glu361Leufs*65), c.1159+2T>A (splicing), and c.539C>T(p.Ala180Val), which were reported for the first time in VRK1-related PN. The variants c.1124G>A (p.Trp375*) and c.706G>A (p.Val236Met) have been previously described.^7,29 The pedigrees and sequencing results of 8 VRK1-related PN families are shown in Figure 2, B and C.

Figure 2. Genetic Analysis of VRK1 Variants in 8 Unrelated Families With Neurogenic Motor Disorders.

(A) Causative gene distribution in our in-house cohort of 351 axonal CMTR families. AR families and undiagnosed sporadic cases accounted for 45.8% (161/351) of the cases. VRK1 variants were identified in 5% of AR families and undiagnosed sporadic cases (8/161), ranking third after SORD and IGHMBP2. (B) Pedigrees of 8 families with VRK1-related peripheral neuropathy. Squares and circles indicate male patients and female patients, respectively; the black arrow indicates the proband. (C) Sanger sequencing chromatograms of 8 identified novel variants. The upper chromatograms represent normal sequences and lower ones represent mutant sequences (marked by red arrow).

The 8 novel variants were absent in gnomAD, 1000 G, and dbSNP databases. The c.7C>T variant has been reported in ClinVar with an association to Pontocerebellar hypoplasia type 1A, as well as an unclear phenotypic link. All variants were predicted to be disease-causing by bioinformatics software (eTable 1). According to the ACMG guidelines, the variants c.1073_1076del, c.974T>A, c.879_882del, and c.1159+2T>A were classified as pathogenic, while c.215T>G, c.83T>G and c. 7C>T were classified as likely pathogenic. The variant c.539C>T was classified as a Variant of Uncertain Significance (PM1_suopporting+PM2_supporting+PM3_supporting+PP3_moderate). Although currently designated as a VUS based on ACMG criteria, we included this homozygous variant because it segregated with the disease, was absent from population databases, showed deleterious predictions in multiple in silico tools, and the patient's phenotype was consistent with confirmed VRK1 cases. The bioinformatics analysis and ACMG classifications of the 8 novel variants were listed in eTable 1. RNA sequencing analysis of the VRK1 splice variant c.1159+2T>A revealed aberrant exon connectivity, as demonstrated by a sashimi plot showing 2 reads skipping exon 12 (Figure 3A, red arrow). In contrast to the normal exon 12 inclusion observed in a healthy control (Figure 3B), the patient's BAM plot confirmed exon 12 skipping (Figure 3C), leading to a frameshift and premature stop codon (p.Lys357Valfs*39).

Figure 3. RNA-seq Analysis of the VRK1 Splice-Site Variant c.1159+2T>A.

(A) Sashimi plot demonstrating aberrant exon connectivity with 2 sequencing reads skipping exon 12 in the patient (red arrow). (B) Integrative Genomics Viewer (IGV) BAM plot from a healthy control showing normal inclusion of exon 12. (C) IGV BAM plot from the patient carrying the c.1159+2T>A variant, confirming 2 reads skipping exon 12 (red arrow). This splicing defect results in a frameshift and premature stop codon (p.Lys357Valfs*39).

The c.1124G>A (p.Trp375*) is the most prevalent, identified in 4 unrelated families in either a homozygous or compound heterozygous state. Given that the c.1124G>A variant was also reported in other 3 additional Chinese families,^29-31 we performed haplotype analysis using SNP microarray and microsatellite markers. Haplotype analysis revealed a 300 kb shared IBD segment (Chr14:97202161-97533412) among these families, indicating that this high-frequency variant might be associated with the founder effect (Figure 4A). In the predicted structural model, Trp375 is located in the low complexity disordered regulatory region at the C-terminus. Truncated variants in this region may retain certain protein functions and lead to a milder clinical phenotype (Figure 4B).

Figure 4. Haplotype Analysis of the High-Frequency Variant c.1124G>A (p.Trp375*) in 4 Unrelated Families.

(A) Haplotype analysis revealed a 300 kb shared IBD segment (Chr14:97202161-97533412) among these families, indicating that this high-frequency variant was associated with the founder effect. (B) Molecular modeling of VRK1 p.Trp375* variant showed that Trp375 is located in the low complexity disordered regulatory region at the C-terminus.

To further elucidate the genotype-phenotype correlation, we applied a structural clustering framework to the variants identified in our cohort and the literature. Variants were categorized based on their location: Cluster I (ATP-binding site) and Cluster II (catalytic site) within the kinase domain and Cluster III in the C-terminal regulatory region (Figure 5A). Our cluster analysis revealed a distinct stratification in disease severity (Figure 5, B and C). Variants located in Cluster I and II were predominantly associated with severe, early-onset phenotypes (often <5 years). By contrast, the presence of at least one Cluster III variant was significantly linked to a milder clinical course, slower progression, and later onset, as exemplified by the recurrent p.Trp375 variant in our cohort. Notably, an exception within Cluster III involves variants disrupting the nuclear localization signal (NLS), such as p.Arg358, which are associated with severe neurodevelopmental outcomes despite their C-terminal location (eFigure 1).

Figure 5. Variant Distribution and Genotype-Phenotype Correlations in VRK1-Related Motor Neuron Syndromes.

(A) Schematic representation of the VRK1 protein structure, showing the distribution of all reported pathogenic variants associated with VRK1-related motor neuron syndromes. Variants identified in this cohort are boxed in red, with novel variants highlighted in red font. (B) Clinical phenotypes associated with different VRK1 variant cluster combinations in this cohort. (C) Correlation between VRK1 variant cluster combinations stratified by age at onset (years) in this cohort.

Clinical and Genetic Features of All Reported VRK1-Related Neurogenic Motor Disorders Cases

To contextualize these findings, we reviewed 23 globally reported studies (2009–2025) and included our cohort, compiling data from 37 families encompassing 53 patients with VRK1-related neurogenic motor disorders.^1,4-9,29-44 Among these, 38 distinct VRK1 variants were identified, predominantly missense (71.05%), nonsense (21.05%), and splice-site (7.90%). All, except one ambiguous heterozygous variant, were biallelic, comprising 9 homozygous and 21 compound heterozygous combinations. Based on prior studies,^1,23 the pathogenic variants were classified into 3 clusters within the protein sequence: Cluster I near the ATP-binding site, Cluster II adjacent to the catalytic site (both within the kinase domain), and Cluster III near the terminus of the low-complexity C-terminal regulatory region. The distribution of these variants in the protein domain is shown in Figure 5A, and the distribution of variant combinations across the 3 VRK1 variant clusters is shown in eFigure 1B and E. Consistent with previous findings, no unrelated compound heterozygous cases shared identical variant combinations except sibling pairs.⁹

Clinically, VRK1-related neurogenic motor disorders exhibit significant phenotypic heterogeneity, primarily affecting lower motor neurons, with overlapping features across the disease spectrum. The most frequent phenotypes were dHMN (28.3%) and SMA/dSMA (26.42%), followed by juvenile ALS (7.55%), axonal recessive CMT (5.66%), and hereditary spastic paraplegia (HSP, 3.77%). Associated comorbidities included congenital microcephaly (CM), pontocerebellar hypoplasia (PCH), intellectual disability (ID), and global developmental delay (eFigure 1A).

Among the 53 cases, 14 (26.4%) were from China, exhibiting a similar distribution of phenotypes: 9 with dHMN, 2 with dSMA, 1 with axonal recessive CMT, and 2 with dHMN plus upper motor neuron signs (eFigure 1D,G). In addition to the 8 patients reported in our study, the other 6 reported Chinese cases all carried the pathogenic c.1124G>A variant, further supporting a potential founder effect of this variant. Age at onset ranged from prenatal to 49 years (median: 14), with peaks during middle childhood (6–11 years) and adolescence (12–19 years) (eFigure 1C). Early-onset cases (<5 years) often exhibited severe phenotypes: infantile onset (<2 years, 7 cases) presented with profound motor and neurodevelopmental impairment (e.g., CM/PCH), and early childhood onset (3–5 years, 6 cases) demonstrated rapid progression with diverse phenotypes (dHMN, SMA, HSP, JALS). By contrast, later-onset cases (≥6 years) predominantly showed dHMN, SMA/dSMA, axonal recessive CMT, or JALS with slower progression, particularly in adolescence and adulthood (eFigure 1F).

Genotype-phenotype correlation analysis indicated that Cluster I/II (ATP-binding/catalytic sites) variants were associated with severe, early-onset forms. By contrast, those harboring at least one Cluster III variant demonstrated slower progression and survival into middle age, exemplified by the Chinese hotspot variant p.Trp375* (eFigure 1H–I). An exception is the Cluster III pathogenic variant p.Arg358*, which disrupts the nuclear localization signal and produces a truncated protein prone to cytoplasmic mislocalization, resulting in severe neurodevelopmental disorders.⁴⁵

Discussion

In this study, we identified 10 VRK1 variants across 8 Chinese families diagnosed with autosomal recessive or sporadic axonal CMTR. Notably, 8 of these variants are novel, significantly expanding the known pathogenic variants spectrum of the VRK1 gene. In our axonal recessive CMT cohort, VRK1 ranked as the third most common gene, after SORD and IGHMBP2,^14,28 with earlier onset generally correlating with more severe, rapidly progressing phenotypes, while later onset tends to reflect a milder disease course. One truncating pathogenic variant, c.1124G>A (p.Trp375*), was found in half of the families (4/8), indicating a high-frequency distribution in the Chinese population. Haplotype analysis showed that carriers of this variant shared a 300 kb ancestral segment, suggesting a founder effect. The identification of the likely founder variant p.Trp375* further emphasizes the need for population-tailored diagnostic strategies.

All 8 patients in our cohort presented with dHMN or axonal recessive CMT. Notably, brain MRI, which was available for 7 of the 8 probands, confirmed the absence of pontocerebellar hypoplasia or other structural malformations. Furthermore, detailed neurologic examinations and clinical history revealed no evidence of cognitive impairment or developmental delay in this cohort. This contrasts with previous reports that linked VRK1 variants to more complex neurodevelopmental phenotypes. For example, homozygous pathogenic variant p.Arg358* were associated with pontocerebellar hypoplasia and intellectual disability,⁴ while compound heterozygous pathogenic variants (V236M and R89Q) were linked to microcephaly and simplified gyral patterns on MRI.³⁵ Most of the patients in these studies also exhibited both upper and lower motor neuron involvement,^4,5,44 suggesting that the clinical presentation of VRK1 variants can vary widely. The lack of neurodevelopmental features in our cohort suggests that VRK1 variants may present with a more focused peripheral neuropathy phenotype in some populations.

VRK1 encodes a serine/threonine kinase that is predominantly localized in the nucleus, with partial cytoplasmic presence.¹ It is involved in crucial cellular processes, such as chromatin remodeling, DNA damage response, cell cycle regulation, and Cajal body assembly, by phosphorylating key targets like histone H3, H2AX, p53, 53BP1, BAF, and Coilin.^2,46-49 Indeed, studies suggest that VRK1-associated motor neuron diseases are mainly related to altered Cajal body assembly and stability, impairing axonal RNA transport and preferentially affecting long-axon neurons, while defects in DNA damage response and chromatin remodeling also contribute to the phenotypic manifestation.¹ This functional role underscores VRK1's essential contribution to cellular integrity. Our genotype-phenotype analysis suggests that the position of variants within functional domains of the protein influences disease onset and severity. VRK1 variants may affect CNS development through either impaired kinase activity or dysfunction of the C-terminal domain, but the exact mechanism remains unclear. A mouse model lacking the C-terminal domain and the nuclear localization signal (NLS) showed reduced brain mass and neuronal tissue, and further research indicating a potential mechanism involving disrupted p53-dependent neurotrophic signaling and axonal guidance.⁵⁰

To better interpret the pathogenicity of different variants, we adopted a previously proposed structural clustering framework that groups variants into 3 regions: Clusters I and II, located within the kinase domain, and Cluster III, located within the flexible, low-complexity C-terminal domain. Cluster III is involved in interactions with nucleosomal histones and DNA, facilitating chromatin association.⁵⁰ Its structural flexibility may also regulate kinase activity, suggesting that C-terminal deletions could affect enzymatic function. By contrast, variants in Clusters I and II tend to directly impair kinase activity, which explains their association with earlier onset and more severe phenotypes. Cluster III variants, on the other hand, more commonly disrupt histone acetylation or interactions with regulatory protein complexes, leading to milder clinical presentations. It is important that the C-terminal domain contains the NLS, which is critical for proper protein localization. For instance, the p.Arg358* variant results in cytoplasmic mislocalization of the truncated protein because of loss of the NLS.⁴⁵ These findings emphasize the diverse molecular mechanisms by which VRK1 variants can alter cellular processes and contribute to varying clinical outcomes.

This study expands the VRK1 mutational spectrum by identifying 8 novel variants and suggests that the c.1124G>A (p.Trp375*) variant may represent a founder effect in the Chinese population. Affected individuals primarily displayed motor dominant axonal CMTR phenotypes with variable upper motor neuron signs, highlighting clinical heterogeneity. Our findings also indicate that variant location within specific protein domains affects disease severity and progression. These results support the inclusion of VRK1 variant screening in the diagnostic workup for inherited peripheral neuropathies, especially in populations with potential founder variant.

Glossary

ACMG

American College of Medical Genetics and Genomics

ALS

amyotrophic lateral sclerosis

AR

autosomal recessive

CM

congenital microcephaly

CMTR

Charcot-Marie-Tooth and related disease

dHMN

distal hereditary motor neuropathies

ID

intellectual disability

PCH

pontocerebellar hypoplasia

PN

peripheral neuropathy

SMA

spinal muscular atrophy

VRK1

vaccinia-related kinase 1

WES

whole exome sequencing

Author Contributions

K. Xu: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. S. Zeng: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. X. Li: analysis or interpretation of data. Z. Li: analysis or interpretation of data. Y. Liu: Analysis or interpretation of data. M. Wang: drafting/revision of the manuscript for content, including medical writing for content. Y. Xie: analysis or interpretation of data. L. Liu: drafting/revision of the manuscript for content, including medical writing for content. W. Cao: analysis or interpretation of data. S. Huang: analysis or interpretation of data. H. Zhao: analysis or interpretation of data. B. Tang: drafting/revision of the manuscript for content, including medical writing for content. R. Zhang: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data.

Study Funding

This study is funded by the National Natural Science Foundation of China (No. 82171172), the Hunan Provincial Key Research and Development Program (No. 2024JK2116), the Major Scientific Research Project of National Clinical Key Specialty in Hunan Province (No. Z2023012), and the Major Scientific Research Project for High-level Health Talents in Hunan Province (No. R2023020).

Disclosure

The authors report no relevant disclosures. Go to Neurology.org/NG for full disclosures.