Loss of Cajal bodies in motor neurons from patients with novel mutations in VRK1

Abstract

Distal hereditary motor neuropathies (dHMNs) are a heterogeneous group of diseases, resembling Charcot–Marie–Tooth syndromes, but characterized by an exclusive involvement of the motor part of the peripheral nervous system.

Here, we describe two new compound heterozygous mutations in VRK1, the vaccinia-related kinase 1 gene, in two siblings from a Lebanese family, affected with dHMN associated with upper motor neurons (MNs) signs. The mutations lead to severely reduced levels of VRK1 by impairing its stability, and to a shift of nuclear VRK1 to cytoplasm. Depletion of VRK1 from the nucleus alters the dynamics of coilin, a phosphorylation target of VRK1, by reducing its stability through increased proteasomal degradation. In human-induced pluripotent stem cell-derived MNs from patients, we demonstrate that this drop in VRK1 levels leads to Cajal bodies (CBs) disassembly and to defects in neurite outgrowth and branching. Mutations in VRK1 have been previously reported in several neurological diseases affecting lower or both upper and lower MNs. Here, we describe a new phenotype linked to VRK1 mutations, presenting as a classical slowly progressive motor neuropathy, beginning in the second decade of life, with associated upper MN signs. We provide, for the first time, evidence for a role of VRK1 in regulating CB assembly in MNs. The observed MN defects are consistent with a length dependent axonopathy affecting lower and upper MNs, and we propose that diseases due to mutations in VRK1 should be grouped under a unique entity named `VRK1-related motor neuron disease’.

Introduction

Hereditary motor and sensory neuropathy (HMSN), or Charcot–Marie–Tooth (CMT) disease, are the most common group of inherited peripheral neuropathies (IPN), with an overall prevalence of 1/2500 (1). These diseases are characterized by length-dependent progressive degeneration of the peripheral nervous system (PNS) and extensive phenotypic and genetic heterogeneity, with alterations in over 80 genes reported to date (2,3). Clinically most patients present with distal weakness and muscle amyotrophy predominant at the lower limbs, resulting in gait difficulties. The onset usually occurs within the first two decades, and the disease then progresses slowly. Progression to the upper limbs is frequent but occurs later in the disease course. Skeletal abnormalities such as pes cavus, reduced or absent deep tendon reflexes are also present, while sensory loss is also often, but not always described. Patients presenting with exclusive motor neuropathy are affected with distal hereditary motor neuropathy (dHMN), also known as neuronopathy, distal spinal muscular atrophy (SMA) or spinal CMT. It is not always clear if dHMN must be considered as a separate clinical entity; indeed, many forms of dHMN have minor sensory abnormalities or begin as pure motor neuropathies and progress toward an axonal CMT (CMT2) with time. As a consequence, there is a genetic overlap between CMT2 and dHMN, and mutations in the same gene, even the same mutation in a gene, may cause both phenotypes (4). The same kind of overlap also exists between CMT/dHMN and upper motor neuron (MN) diseases, such as hereditary spastic paraplegias (5). Similarly, minor sensory involvement might also be present in close clinical conditions affecting the MNs, such as amyotrophic lateral sclerosis (ALS) and SMA (4,6).

Here, we describe two novel compound heterozygous missense mutations in the Vaccinia-related Kinase 1 gene (VRK1), in two siblings affected with a distal dHMN, associated with upper MN signs. Since the first description of VRK1 as the culprit gene in SMA with pontocerebellar hypoplasia (SMA-PCH/PCH1A, MIM 607596) (7), mutations in this gene have been reported in a spectrum of neurological diseases affecting MNs (either lower only or lower and upper) (8–11), or motor and sensory nerves (12).

VRK1 encodes the vaccinia-related kinase 1, a ubiquitously expressed, mainly nuclear, serine/threonine kinase, playing a crucial role in regulating cell cycle (13,14). Several substrates are known to date for VRK1, including VRK1 (15), the transcription factors p53 (15), c-jun (16), ATF2 (17) and CREB (18), proteins involved in DNA replication and repair, such as histone H2B (15), histone H3 (19) and NBS1 (20), coilin, the main component of Cajal bodies (CBs) (21), myelin basic protein MBP (15), a component of central myelin and peripheral myelin, or barrier-to-autointegration factor (BAF), a protein required for nuclear envelope assembly (22,23). VRK1 encodes a 396 amino acid protein containing a N-terminal serine/threonine kinase domain (residues 37-275) (24) and a C-terminal nuclear localization signal (NLS) (residues 356–360). VRK1 is mainly a nuclear protein, although the presence of a small fraction in the cytoplasm and membrane compartments has been described (24,25).

In this study, we describe bi-allelic mutations in VRK1 in a new form of dHMN associated with upper MN signs, leading to decreased VRK1 expression. We show that loss of VRK1 likely contributes to altered coilin dynamics in fibroblasts. In patient’s human-induced pluripotent stem cells (hiPSC)-derived MNs, we describe altered neurite length and branching and depletion of CBs. The loss of CBs, center for the assembly of ribonucleoproteins required for splicing, suggests that VRK1 has a major role in regulating splicing activity in MNs.

Results

Subjects

Two siblings (patients II.2 and II.3), aged 47 and 38 years, respectively, from a Lebanese non consanguineous family presented with dHMN (dCMT/dHMN) associated with upper MN signs (Fig. 1A). Clinically, both patients share slowly progressive distal muscular weakness and atrophy of the four limbs, which started around age 10 for both patients. At age 38, patient II.3 is slightly more severely affected than his sister (II.2) and uses sometimes a walking stick. Other features include hyperreflexia and Babinski bilateral signs. Physical examination showed no sensory component. Brain and spinal cord magnetic resonance imaging (MRI) performed at age 34 in patient II.3 showed no abnormalities (Fig. 1B). Cranial nerves were normal. Nerve conduction studies revealed some drop in the amplitudes of the compound muscle action potentials (CMAPs) of all motor nerves in the upper and lower extremities with mild slowing in their conduction velocities (Table 1). The electromyography (EMG) revealed mild neurogenic motor units action potentials with preserved recruitment after insertion of a bipolar needle electrode (Table 1). These findings are suggestive of mild peripheral motor neuropathy. Genealogical data indicate an autosomal recessive mode of inheritance (Fig. 1A).

Figure 1

Genealogical, clinical and molecular characteristics of patients with VRK1 mutations. (A) Pedigree from the studied family. Affected individuals are shaded in black. (B) Normal MRI images of patient II.3: Axial FLAIR (B1); sagittal T1-weighted image (B2); sagittal T2-weighted image (B3). (C) Chromatograms of patient and control showing the two compound heterozygous mutations identified in VRK1 (NM_003384): c.656G>T (p.Arg219Ile) and c.761G>T (p.Trp254Leu) in exons 8 and 9, respectively. (D) RT-PCR amplification of vrk1 transcripts from mouse neuromuscular tissues. Expected sizes of amplicons are 168 pb, 228 pb and 300 pb for the three vrk1 isoforms; NM_001029844, NM_001029843 and NM_011705, respectively. (E) Alignment of VRK1 amino acid sequences between different species using CLUSTAL 2.1. Arginine 219 and Tryptophan 254 highlighted in green and in red, respectively, (NP_003375) are highly conserved across vertebrates. (F) Schematic representation of the VRK1 gene and protein. The human VRK1 gene is covering a genomic region of 84 268 bp at chromosome 14q32.2 and is composed of 13 exons. The transcript (NM_003384) is 1745 bp long, with a coding sequence of 1191 bp (exons 2–13), encoding a 396 amino acid protein (NP_003375). Untranslated regions (5'- and 3'-) are hashed. Kinase domain is highlighted in purple and the putative NLS in blue. All mutations identified to date in VRK1 are indicated. ^(*)Our study, Renbaum et al. (2009) (7), Najmabadi et al. (2011) (53), Gonzaga-Jauregui et al. (2013) (12), Nguyen et al. (2015) (8), Stoll et al. (2016) (9), Tomaselli et al. (2016) (10) and Li et al. (2019) (11).

Identification of compound heterozygous mutations in VRK1

Although we searched in priority for apparently compound heterozygous variants shared by Patient II.2 and Patient II.3, a search for homozygous variants was performed, which showed the absence of shared homozygous variants between the two siblings (Supplementary Material, Table S1). In contrast, whole exome sequencing (WES) sequencing led to the identification of 4265 possibly compound heterozygous variants, in the coding sequences of genes shared by the two siblings, of which 134 had a frequency above 1% in the Genome Aggregation Database (gnomAD dataset; http://gnomad.broadinstitute.org/). Additional filtering narrowed down the number of candidates to 10 variants in four genes (LAMA1, TTN, EVPL and VRK1). Details of the filtering steps are available in Supplementary Material, Table S1. EVPL and TTN variants have been eliminated by segregation studies and LAMA1 variants were also excluded because one of the two variants was described three times at the homozygous state in the gnomAD dataset. We focused on two compound heterozygous variants in the VRK1 gene, because of their complete absence from the gnomAD dataset and a prediction for pathogenicity by several prediction tools. Moreover, at that time, VRK1 was described as the causative gene for SMA with pontocerebellar hypoplasia (PCH1A, MIM 607596) (7), a neurological disease, sharing clinical signs with the disease affecting our patients. VRK1 (NM_003384) comprises 13 exons and encodes a 396 amino acid protein (NP_003375), containing an N-terminal serine/threonine kinase domain (residues 37–275) (24) and a C-terminal NLS (residues 356–360). Both variants are novel missense heterozygous transversions: c.656G>T (p.Arg219Ile) and c.761G>T (p.Trp254Leu) in VRK1 exons 8 and 9, respectively. Segregation studies confirmed that the mutations are, indeed, compound heterozygous, and segregate with the disease in the family (Fig. 1A and C). Both the arginine at amino acid 219 and the Tryptophan at amino acid 254 are highly conserved in vertebrates (Fig. 1E). RT-PCR performed between exons 7 and 9 of VRK1 in both patients showed that both mutations are true missense variations with no effect on splicing (data not shown). Finally, we showed, in mouse, that vrk1 is expressed in the tissues affected in our patients’ disease: the central (spinal cord) and peripheral (sciatic nerve) mouse nervous system (Fig. 1D). In consequence, we propose that the two novel missense variations c.656G>T (p.Arg219Ile) and c.761G>T (p.Trp254Leu) in VRK1 (Clinvar accession numbers SCV000882440 and SCV000882740, respectively) are the molecular defects underlying this particular form of distal hereditary motor neuropathy associated with upper MN signs, bringing to twelve the number of mutations described to date in VRK1 (Fig. 1F).

Table 1

Nerve conduction studies in Patient II.3 at age 16 years

Motor conduction studies
Nerve		DL	CMAP	CV
Peroneal L	Ankle	6.2	3.2
	Below fibula	15.5	2	40
	Above Fibula	17.3	1.6	53
Peroneal R	Ankle	6.5	4.6
	Below fibula	16.7	3.5	39
	Above fibula	18	3.2	46
Posterior tibial L	Ankle	6.3	1.07
	Knee	16.6	0.1	38
Posterior tibial L	Ankle	5.5	2.2
	Knee	17.3	1.3	36
Median R	Wrist	4	8.7
	Elbow	8.6	8.5	57
Ulnar R	Wrist	2.9	7.8
	Below elbow	7.5	7.3	50
	Above elbow	8.5	7.4	43
Sensory conduction studies
Nerve		DL	SNAP	CV
Median R		3.1	18.9	50
Sural L		4.6	16.8	40
Sural R		2.6	18.7	42

DL, distal latencies; CMAP, compound motor action potential.

SNAP, sensory nerve action potential; CV, conduction velocities.

Normal values: DL < 4 ms, CMAP > 4, CV > 40 m/s

Decrease of VRK1 levels in patients’ cells due to post-translational defects

In order to further demonstrate the pathogenicity of the mutations, we studied the expression of VRK1 in patients’ cells by western blot and qRT-PCR. We demonstrated a decrease in VRK1 protein levels in patients as compared to controls in both lymphocytes and fibroblasts: 59% and 68% decrease, respectively (Fig. 2A and B). To test whether this decrease is a result of an impairment of the transcriptional process, VRK1 mRNA levels were measured by qRT-PCR: no significant changes in VRK1 transcript levels were observed between controls and patients (Fig. 2C). These results evidenced that the decrease in VRK1 levels observed in patients cells is due to a post-translational defect. Interestingly, blockade of the proteasome by MG132 treatment increased VRK1 protein levels by 44% in patients’ fibroblasts, showing that the decrease in VRK1 levels in patients’ cells is due to its degradation by the proteasome (Fig. 2D).

Figure 2

Decrease of VRK1 levels in patients’ cells due to post-translational defects. (A) Western blotting of VRK1 in fibroblasts and immortalized lymphoblastoid cells’ lysates from controls and patients shows a strong decrease of VRK1 protein levels in patients (both cell types). VRK1 is detected at 45 Kda (lower band in fibroblasts, unique band in lymphoblastoid cells) and GAPDH at 37 KDa. (B) This decrease was quantified as VRK1/GAPDH ratio. Bars represent the mean of three independent experiments ±SEM. Statistical significance was evaluated by two-way ANOVA test, ^*P < 0.05, ^**P < 0.005. (C) qRT-PCR analyses did not show any significant differences in VRK1 mRNA levels in patients’ lymphocytes and fibroblasts versus controls. VRK1 expression levels were normalized to three different reference genes: GAPDH, GUS-B and B2M. Relative expression of VRK1 was calculated as fold-differences using the comparative ∆∆C_T method. Two calibrator samples were used (Ctrl 1F and Ctrl 2F for fibroblasts and Ctrl 1L and Ctrl 2L for lymphocytes), and each experiment was repeated three times. Data are expressed as geometric mean ±SEM of ∆∆C_T values for each sample. Statistical significance was evaluated by one-way ANOVA test, n.s: not significant. (D) Western blot shows restoration of normal VRK1 levels in patients’ fibroblasts after proteasome blockade by MG132. VRK1/GAPDH levels are quantified in (E) before and after MG132 treatment. After MG132 treatment, VRK1 levels in patients are restored to controls’ values. The results are statistically significant as evaluated by two-way ANOVA test, ^**P < 0.005, n.s: not significant. The graph represents the mean of three independent experiments ±SEM. NT: not treated, Ctrl: control, Ctrl1F/2F: fibroblasts controls, Ctrl1L/2L: lymphocytes controls.

Mislocalization of VRK1 in patients’ fibroblasts

Although small fractions of VRK1 can be present in the cytoplasm and other cytoplasmic compartments, VRK1 is mainly a nuclear protein. Protein subcellular localization was evaluated in patients and controls’ fibroblasts; three possible VRK1 locations were observed (Fig. 3). Interestingly, VRK1 was mislocalized in more than 70% of analyzed patients’ cells. In the remaining 30%, VRK1 was either both in the nucleus and the cytoplasm or completely absent. However, VRK1 was never strictly nuclear, unlike in 50% of controls’ cells. Nuclear localization was restored in at least 97% of patients’ fibroblasts after MG132 treatment suggesting that the missense mutations lead to the export of VRK1 to the cytoplasm for degradation by the proteasome (Fig. 3).

Figure 3

Mislocalization of VRK1 in patients’ fibroblasts is corrected by proteasome blockade. (A) Immunolabeling of VRK1 in fibroblasts from patient II.2 shows a shift of VRK1 from the nucleus to the cytoplasm as compared to control (Control 1F). Nuclear localization is restored in patient’s fibroblasts after MG132 treatment. Arrows, arrowheads and cross signs indicate examples of strictly nuclear, strictly cytoplasmic, nuclear and cytoplasmic localization respectively. (B) These results have been quantified by measurement of R, the ratio of fluorescence intensity of nuclear VRK1/cytoplasmic VRK1, by ImageJ software. MG132 treatment induces relocalization of VRK1 to the nucleus in patients’ cells as attested by the drastic decrease of R to values below 0.9 in both patients II.2 and II.3. Three range values were determined for R, corresponding to three possible VRK1 subcellular localizations: (R > 2) = strictly nuclear, (0.9 < R < 2) = nuclear and cytoplasmic, (0.1 < R <0.9) = strictly cytoplasmic. Mean percentages of cells in each range are presented in the table for controls and patients. For each individual, we scored 300 cells from three independent experiments. NT: not treated.

VRK1 depletion in patients’ fibroblasts leads to reduced coilin levels by facilitating its proteasomal degradation

We analyzed the effect of the absence of VRK1 from the nucleus of patients’ fibroblasts on coilin expression. We observed a significant decrease of coilin levels in patients’ cells as compared to controls’ (Fig. 4A and B). Having demonstrated above that MG132 blockade of the proteasome is able to restore VRK1 levels and nuclear localization, we checked whether it would also restore coilin levels. Indeed, MG132 treatment of patients’ fibroblasts resulted in the restoration of normal coilin levels (Fig. 4). Therefore, we conclude that in the absence of VRK1, coilin is likely degraded in the proteasome.

Figure 4

Depletion of VRK1 in patients’ fibroblasts leads to altered coilin expression. (A) Immunolabeling of coilin in fibroblasts from patients and controls shows a drastic decrease of coilin levels in the nucleus of patients II.2 and II.3 as compared to controls 1F and 2F. Here, we have presented the results for one patient (II.2) and one control (1F). Normal expression levels are restored after MG132 treatment. (B) Relative coilin fluorescence intensity was measured in 300 cells for each sample (scored from three independent experiments); the median for each sample is shown on the graph. Statistical significance was evaluated by Mann–Whitney test, ^****P < 0.0001, n.s: not significant. NT: not treated.

VRK1 is necessary for CBs assembly in hiPSC-derived MNs

Given the upper and lower MN signs described in our patients, we sought to investigate the effects of VRK1 and coilin abnormalities in a more relevant cellular model, and therefore, we produced spinal MNs by differentiation of hiPSCs from patients II.2 and one control hiPSC-derived MNs. Differentiation efficiency was assessed, in each hiPSC cell line, by immunostaining of HB9 and ISLET1, two transcription factors of spinal MNs. As shown in Supplementary Material, Fig. S1, we observed no differences in the differentiation efficiency between patient II.2 and the control. Also, there were no differences in the iPS proliferation rates between patient II.2 and control (data not shown). In hiPSC-derived MNs, immunostaining and western blot experiments confirmed the decrease of VRK1 levels in patient versus control (Fig. 5A). Quantification by western blot evidenced a 65% decrease of VRK1 levels in patient II.2 as compared to control (Fig. 5B and C). Interestingly, we showed that this decrease is accompanied by disassembly of CBs in patient’s MNs; indeed, while in controls, more than 80% of MNs have nicely assembled CBs, in patient II.2, CBs are present in only 20% of MNs. Moreover, when present, CBs have a smaller size than control’s (Fig. 5D and E).

Figure 5

VRK1 is essential for CBs assembly and for neurite elongation and branching in MNs. (A) Immunolabeling of VRK1 and coilin shows reduced VRK1 expression and CBs disintegration in hiPSC-derived MNs from patient II.2 as compared to control hiPSC-derived MNs (1M). NFM, neurofilament medium chain. (B–C) This reduction was confirmed and quantified by western blot experiments and statistically significant decrease of 64% was found in patient’s MNs. Quantification was performed on data from three independent experiments. Statistical significance was evaluated by unpaired t-test, ^*P < 0.05. (D) Percentage of MNs containing CBs was evaluated in at least 800 cells counted from three independent experiments (828 cells for control 1M and 1269 cells for patient II.2). Statistical significance was evaluated by unpaired t-test, ^***P < 0.0005. (E) In CBs positive cells (CBs⁺), a tendency to decreased CBs size was observed in hiPSC-derived MNs from patient when compared to control. CBs size was calculated, for each individual, by measuring CBs area on 50 CBs⁺ hiPSC-derived MNs, using ImageJ software. Data were recorded from three independent experiments. (F) Images of AAV 2/6-CMV-GFP infected hiPSC-derived MNs from patient II.2 and control 1M. (G–I) Graphs presenting neurite length and branching studies in AAV 2/6-CMV-GFP infected hiPSC-derived MNs in patient II.2 versus control (median values from two independent experiments are presented). Neurite length was calculated by summing up the length of all neuritic branches per cell (G). For branching studies, we evaluated two parameters: number of branches (H) and number of branching points (I) per cell. All three parameters are decreased in patient II.2’s hiPSC-derived MNs: (i) neurite length is decreased by 36%, (ii) number of branches per cell drops from 5 (in control) to 3 (in patient) and (iii) number of branching points decreases from 2 (in control) to 1 (in patient) (median values from two independent experiments). 50 MNs cells were counted for each individual. Statistical significance was evaluated by Mann–Whitney test, ^*P < 0.1, ^**P < 0.01, ^***P < 0.001.

Impaired neurite length and branching in patient’s hiPSC-derived MNs

In order to further investigate the effects of reduced VRK1 levels, and the resulting CB loss and disruption, on the function of MNs, we studied neurite length and branching in hiPSC-derived MNs from patient II.2 as compared to a control. To this end, patient and control’s hiPSC-derived MNs were infected by low levels of adeno-associated virus (AAV) 2/6-CMV-GFP virus. Indeed the use of low virus titers led to weak infection and GFP expression by few MNs, thereby allowing tracing of GFP-expressing neurites in isolated MNs. Interestingly, patient’s MNs exhibited a significant decrease in (i) total neurite length, (ii) number of branching points and (iii) number of branches per cell, as compared to control (Fig. 5F-I).

Discussion

In this work, we identified two novel compound heterozygous missense mutations in VRK1, the vaccinia-related kinase 1 gene, in two patients from a non-consanguineous Lebanese family, affected with a slowly progressive distal form of CMT (dHMN) associated with upper MN signs. The mutations, c.656G>T (p.Arg219Ile) and c.761G>T (p.Trp254Leu), located in the kinase domain of VRK1, segregate in the pedigree, following an autosomal recessive mode, are absent from reference datasets, are well conserved among species and are considered to be damaging by pathogenicity prediction algorithms (Fig. 1).

Mutations in VRK1 were first described in SMA with pontocerebellar hypoplasia (SMA-PCH) (7), but then proved to be responsible for several other neurological diseases, affecting MNs (lower or lower and upper) (8–11), with or without sensory abnormalities (12). In total, including ours, 12 mutations are now reported in VRK1 in 14 patients from 10 families (Fig. 1), described under five different clinical entities (Table 2). There is no clear phenotype–genotype correlation, although the p.Arg358^* mutation is often associated with more severe phenotypes, in particular when it is found at the homozygous state. The presence of pontocerebellar hypoplasia (PCH), first described as a cardinal clinical sign of the disease, should however not be seen as a distinctive clinical manifestation, because most patients do not have PCH. The same is true for microcephaly, which has been described in only five patients from three families (Table 2), of whom three are homozygous for the p.Arg358^* mutation. In contrast, the involvement of upper MN is almost constant, as manifested by brisk tendon reflexes, even in patients described as affected with distal SMA (Table 2).

Table 2

Comparison of the genetic, clinical and electrophysiology findings of all patients described to date with VRK1 mutations

Reference		This study	This study	Renbaum et al.(2009) (7)	Renbaum et al. (2009) (7)	Najmabadi et al. (2011) (53)	Gonzaga-Jauregui et al. (2013) (12)	Gonzaga-Jauregui et al. (2013) (12)	Gonzaga-Jauregui et al. (2013) (12)
Family		Family 10	Family 10	Family 1	Family 1	Family 2	Family 3	Family 3	Family 4
Patients		Patient II.2	Patient II.3	Proband IV.12	Patient IV.6	M017N	BAB3022	BAB3280	BAB5311
Clinical diagnosis		dHMN+upper MN signs	dHMN+upper MN signs	SMA with PCH	SMA with PCH	PCH	Complex motor and sensory axonal neuropathy + microcephaly	Complex motor and sensory axonal neuropathy + microcephaly	Complex motor and sensory axonal neuropathy + microcephaly
Mutations	Allele 1	c.656G>T	c.656G>T	c.1072C>T	n.d. (deceased)		c.266G>A	c.266G>A	c.1072C>T
		p.Arg219Ile	p.Arg219Ile	p.Arg358*		p.Arg133Cys	p.Arg89Gln	p.Arg89Gln	p.Arg358*
	Allele 2	c.761G>T	c.761G>T	c.1072C>T			c.706G>A	c.706G>A	c.1072C>T
		p.Trp254Leu	p.Trp254Leu	p.Arg358*	n.d. (deceased)	p.Arg133Cys	p.Val236Met	p.Val236Met	p.Arg358*
Genotype		Compound heterozygous	Compound heterozygous	Homozygous	n.d. (deceased)	Homozygous	Compound heterozygous	Compound heterozygous	Homozygous
Type of mutation		Missense	Missense	Nonsense	Nonsense	Missense	Missense	Missense	Nonsense
Inheritance pattern		AR	AR	AR		AR	AR	AR	AR
Consanguinity		−	−	+	+	+	−	−	−
Origin		Lebanese	Lebanese	Ashkenazi Jewish	Ashkenazi Jewish	Iranian	n.a.	n.a.	Ashkenazi Jewish
Sex/age, Y		F/47	M/38	F/died at 11,5 Y	F/died at 9,5 Y (sisters)	n.a.	F/10	F/2	M/9
Onset		Childhood (~10 years)	Childhood (~10 years)	Early	Early	n.a.	Early	Early	Early
Initial symptoms		Motor	Motor	Motor +early-onset ataxia	Motor +early-onset ataxia	n.a.	Motor	Motor	Motor
Progression		Slow	Slow	Rapid	Rapid	n.a.	Rapid	Rapid	Rapid
Functional impairment		+. Minor	+. Minor. Sometimes use a walking stick.	++. Unable to sit at 1 year or walk without support at age 2 years. Wheelchaired.	++	n.a.	++. Never able to walk alone.Wheelchaired at 4 years, limited use of arms and hands at 6 years	++. Unable to sit or walk without support at age 20 months	++. Wheelchaired from 6 years
Family		Family 5	Family 6	Family 6	Family 7	Family 8	Family 9	Family 9
Patients			Family 1. Patient II.4	Family 1. Patient II.1	Family 2. Patient III.3		Patient II.7	Patient II.9
Clinical diagnosis		Possible ALS	Adult onset distal SMA	Adult onset distal SMA. Similar to patient II.4, but later onset, more progressive	Childhood MN disease, juvenile ALS	dHMN+upper MN signs	Adult-onset dSMA	Adult-onset dSMA
Mutations	Allele 1	c.961C>T	c.356A>G	c.356A>G	c.403G>A	c.683C>T	c.1124G >A	c.1124G >A
		p.Arg321Cys	p.His119Arg	p.His119Arg	p.Gly135Arg	p.Thr228Met	p.W375*	p.W375*
	Allele 2	c.356A>G	c.1072C>T	c.1072C>T	c.583T>G	c.1072C>T	c.1124G >A	c.1124G >A
		p.His119Arg	p.Arg358*	p.Arg358*	p.Lys195Val	p.Arg358*	p.W375*	p.W375*
Genotype		Compound heterozygous	Compound heterozygous	Compound heterozygous	Compound heterozygous	Compound heterozygous	Homozygous	Homozygous
Type of mutation		Missense	Missense	Missense	Missense	Missense	Nonsense	Nonsense
Inheritance pattern		Sporadic	AR	AR	AR	AR	AR	AR
Consanguinity		−	−	−	−	−
Origin		Hispanic	Ashkenazi Jewish	Ashkenazi Jewish	n.a.	n.a.	Chinese	Chinese
Sex/age, Y		M/32 Y	F/35 Y	M/older	F/3Y	F/34Y	M/47	F/
Onset		Late	Teenage (15 years)	Late teenage	Late	Childhood (~10 years)	Adulthood	Adulthood
Initial symptoms		Motor	Motor	Motor	Motor	Motor	Motor. difficulties in walking on tiptoe	Wasting and weakness of distal symmetrical lower limbs
Progression		Slow	Moderate	Moderate	Rapid	Slow	Slow	Slow
Functional impairment		+. Minor. Falls, walking difficulties	+. Long arm crutch and mobility scooter	+.Wheelchair at 40 years	++. Loss of ambulation at 10 years. Complete dependance for daily activities at 18 years	+. Minor	++. Requires wheelchair to mobilize at age 42 years	+. High steppage gait due to bilateral foot drop at age 40 years
Muscle atrophy	+. Distally in hands and feet	+. Distally in hands and feet	+. Distal muscle atrophy, progressive muscle wasting	+. Distal muscle atrophy, progressive muscle wasting	n.a.	+. Distal progressive muscle wasting progressing to proximal	-	+
Sensory disturbance	Normal in all modalities	Normal in all modalities	+/−	+/−	n.a.	−	−	−
DTR	Brisk	Brisk	Brisk	Brisk	n.a.	Loss	Normal	Decreased
Other upper MN signs	+. Bilateral Babinski signs	+. Bilateral Babinski signs	−	−	−	−	−	−
Intellectual disability	−	−	Mild	+	n.a.	−	−	−
Ataxia	−	−	++	++	n.a.	−	−	−
Hypotonia	−	−	+	+	n.a.	+	+ (mild)	+ (worsening)
Speech problems	−	−	++ (dysarthria)	++ (dysarthria)	n.a.	−	−	+ (hypophonia and dysarthria)
Skeletal deformations	−	−	+. Bilateral equinovarus	+. Bilateral equinovarus	n.a.	+. Scoliosis (worsening)	−	+. Scoliosis (worsening)
Feeding problems	−	−	++ (gastrostomy)	++ (gastrostomy)	n.a.	Gastric feeding tube	−	Gastric feeding tube
Respiratory problems	−	−	−	−	n.a.	+	−	−
Microcephaly	−	−	+ (progressive)	+ (progressive)	n.a.	+	+(in utero, nonprogressive)	+ ( in utero, severe)
MRI (Brain and spinal cord)	Normal	Normal	Cerebellar hypoplasia. Normal spine	Cerebellar hypoplasia. Mild cerebral atrophy	n.a.	Microcephaly with simplified gyral pattern	Microcephaly with simplified gyral pattern	Microcephaly with simplified gyral pattern and underdevelopped cerebellar vermis
NCS measurements motor nerves*	Decreased amplitudes. Slightly reduced MNCVs: upper and lower	Decreased amplitudes. Slightly reduced MNCVs: upper and lower	Axonal	Demyelinating?	n.a.	Amplitudes and MNCVS n.r. (upper and lower)	Upper: normal to slighlty decreased MNCVs. Decreased amplitudes Lower: normal MNCVs. Decreased amplitudes	Upper: normal MNCVs. Decreased amplitudes Lower: normal MNCVs. Decreased amplitudes
Muscle atrophy		+. Profound proximal and distal muscle atrophy with fat replacement (seen at muscle MRI)	+. Profound symmetric distal muscle wasting at upper and lower limbs. Weakness more severe at lower limbs.	n.a.	+. Distal amyotrophy of the lower limbs	+. Distal amyotrophy of the lower limbs	++. Severe distal muscle wasting in legs and arms	+. Severe and moderate wasting and weakness in lower extremities. Slight wasting and weakness of the ulnar intrinsic muscle of both hands
Sensory disturbance		+ (pain)	-	n.a.	−	−	−	−
DTR		Brisk	Brisk in the upper limbs. Absent in the lower limbs	n.a.	Brisk	Brisk	Absent in lower and upper limbs	Absent in lower limbs. Brisk in upper limbs.
Other upper MN signs		−	+	n.a.	+	n.a.	−	−
Intellectual disability		−	−		−	−	−	−
Ataxia		−	−	n.a.		−	−	−
Hypotonia		−	−	n.a.		+	+(all limbs)	+(all limbs)
Speech problems		−	−	n.a.	+ (hoarse and soft)	−	−	−
Skeletal deformations		+. Pes cavus and hammer toes	−	n.a.	+. Severe scoliosis	n.a.	n.a.	+.Pes arcuatus
Feeding problems		−	−	n.a.	−	−	−	−
Respiratory problems		−	+(mild)	n.a.	+ (dyspnea and weak cough)	−	−	−
Microcephaly		−	−	n.a.	−	−	−	−
MRI (Brain and spinal cord)		Normal	Mild to moderate generalized atrophy+atrophy of the spinal cord. No PCH. No Cerebellar Ataxia	n.a.	Normal	Normal	Normal	Normal
NCS measurements motor nerves*		Upper and lower: normal MNCVS and amplitudes	Severely decreased amplitudes. Relatively preserved MNCVs.	n.a.	Severely decreased amplitudes.	Motor	Upper and lower limbs: MNCVs not recordable, severe reduction in CMAP amplitudes.	Normal MNCVs at upper and lower limbs. Normal CMAPs at upper limbs. Reduction of CMAPs at lower limbs, which remain normal except at the peroneal nerve
NCS measurements sensory nerves*		Normal	Normal	Axonal	Demyelinating?	n.a.	Upper: normal NCVs. Slighltly decreased amplitudes Lower: n.r. NCVs. amplitudes n.a.	Upper: slighlty decreased NCVs. Normal amplitudes Lower: n.r.	Upper: normal MNCVs. Decreased amplitudes Lower: n.r.
EMG		Mild neurogenic motor units action potentials with preserved recruitment. Denervation of calf muscles	Mild neurogenic motor units action potentials with preserved recruitment	Fasciculations and fibrillations	Fasciculations and fibrillations	n.a.	n.a.	n.a.	n.a.
Nerve biospy		n.a.	n.a.	Normal (sural)	n.a.	n.a.	Axonal neuropathy	n.a.	n.a.
Muscle biospy		n.a.	n.a.	Neurogenic atrophy	Neurogenic atrophy	n.a.	Neurogenic myopathy	n.a.	n.a.
NCS measurements sensory nerves*		Upper and lower: normal NCVS and amplitudes	Upper and lower: normal MNCVS and amplitudes	n.a.	Initially normal. From 19 years: decreased amplitudes	n.a.	Normal sensory NCVs and amplitudes.	Normal sensory NCVs and amplitudes at upper and lower limbs.
EMG		Fasciculations and fibrillations +chronic denervation/re-innervation in lower muscles	Chronic and active denervation	Chronic and active denervation	Chronic and active denervation	n.a.	Chronic neurogenic changes of high amplitude, long duration motor units	Chronic neurogenic changes
Nerve biospy		n.a.	n.a.	n.a.	n.a.	n.a.	n.a.	n.a.
Muscle biospy		Neurogenic atrophy without evidence of reinnervation	n.a.	n.a.	Neurogenic atrophy	n.a.	n.a.	Neurogenic atrophy (right quadriceps)

Motor stands for distal lower limbs weakness/atrophy or gait disturbance. n.a., not available; n.r., not recordable; n.d., not determined

dHMN, distal hereditary motor neuropathy; AR, autosomal recessive; F, female; M, male; DTR, deep tendon reflexes; CMAP, compound muscle action potential; SMA, spinal muscular atrophy; PCH, pontocerebellar hypoplasia;

NCS, Nerve Conduction Studies; MNCV, Motor Nerve Conduction Velocities; EMG, Electromyography studies. Genbank accession number for VRK1: NM_003384.

Although the phenotype that we describe here is not new, this is the first time that patients with VRK1 mutations present with a classical slowly progressive motor neuropathy, beginning in the second decade of life and associated to upper MN involvement. The description of this more classical `dHMN’ phenotype further expands the range of diseases related to defects in VRK1, and underlines the heterogeneity of phenotypes resulting from VRK1 mutations, as emphasized by the diversity of clinical terms used to describe the disease and its progression: SMA, distal SMA, ALS, juvenile ALS, HMSN, pure motor neuropathy (distal HMN). Considering the possible overlap among all these diseases, and the fact that the feature, common to all patients, is the involvement of lower MNs, we propose that patients with mutations in VRK1 should be grouped under a unique entity named `VRK1-related motor neuron disease’, and that the diagnosis might be evoked by slowly progressive distal motor neuropathy associated or not with upper MN signs.

VRK1 encodes an ubiquitously expressed, mainly nuclear, serine/threonine kinase, from the vaccinia-related kinase (VRK) family, playing a crucial role in many cellular processes like cell division and cell cycle progression (13,14). In nervous tissues, VRK1 has been described to be expressed in fetal and adult brain (26). Additionally, we show that vrk1 is expressed in mouse tissues from the central (brain, cerebellum and spinal cord) and PNS (sciatic nerve; Fig. 1C), the latter two being affected in patients with VRK1 mutations. While the role of VRK1 in cell cycle and cellular proliferation has been widely studied, its function in the nervous system has only been addressed in one study, where Vrk1 has been shown to be important for neuronal migration, through amyloid β precursor protein-dependent mechanisms (27). Here, we show that the two identified missense mutations lead to a significant decrease (more than 50%) of VRK1 levels in patients in all tested cell types (fibroblasts, immortalized lymphoblastoid cells and hiPSC-derived MNs, thereafter termed MNs) due to post-translational degradation of the mutated proteins (Figs 2A and 5B). Additionally to decreased protein levels, we also demonstrate that the remaining mutated VRK1 proteins are mislocalized, and shift from the nucleus to the cytoplasm (Fig. 3), although this phenomenon is less striking in hiPSC-derived MNs (Fig. 5A). Both protein levels and nuclear localization are restored to normal after MG132 treatment (Fig. 2C–E), thereby suggesting that VRK1 is likely transported to the cytoplasm for ubiquitination and degradation in the proteasome. Knowing that VRK1 is a nuclear kinase, the absence, or reduced levels, of VRK1 from the nucleus in patients’ cells strongly suggest abnormal phosphorylation of VRK1 nuclear targets. Several substrates are known to date for VRK1, including VRK1 (15), the transcription factors p53 (15), c-jun (16), ATF2 (17) and CREB (19), proteins involved in DNA replication and repair, such as histone H2B (15), histone H3 (19) and NBS1 (20), myelin basic protein MBP (15), a component of central myelin and peripheral myelin, BAF, a protein required for nuclear envelope assembly (22) and coilin (21). Among all VRK1 phosphorylation targets, we focused on coilin, the main component of CBs, due to the fact that defects in CBs’ size and quantity have been described in MNs from patients with SMA (MIM 253300) (28). CBs were first discovered by Ramón y Cajal in 1903 in pyramidal neurons from the cerebral cortex. They are dynamic, membrane-free, nuclear organelles enriched in several nuclear proteins and RNA-protein complexes (29), and playing an important role in RNA processing, particularly splicing, by guiding the modification of the snRNA moiety of snRNP proteins (small nuclear ribonucleoproteins) from the spliceosome (30,31). CBs also participate in the biogenesis and delivery of telomerases to telomeres (32). CBs assemble on coilin, which acts as a scaffold protein, but they do not exist in all cells. Indeed, CBs are most prominent in cells that are transcriptionally active (33), such as post-mitotic neurons. In dividing cells, dynamic assembly/disassembly of CBs is regulated, during cell cycle, by post-translational modifications, including phosphorylation (34,35).

VRK1 being a nuclear kinase, also regulated during cell cycle progression, is a good candidate to regulate CBs dynamics (14). Indeed, recent work has suggested that VRK1 regulates CBs dynamics during cell cycle, by phosphorylating coilin and protecting coilin from ubiquitination and degradation in the proteasome (34). In the same study, the authors showed that depletion of VRK1 by knockdown in three cell lines (Hela, MCF7 and SH-SY5Y) resulted in a drop of nuclear coilin and subsequently in the loss of CBs, and that coilin’s proteasomal degradation is prevented by VRK1 (34). Here, we confirm that VRK1 is crucial for coilin expression and stability. Indeed, in dividing cells (fibroblasts) of patients with bi-allelic mutations in VRK1, we show that the resulting depletion of VRK1 from the nucleus leads to reduced coilin levels by facilitating its proteasomal degradation (Fig. 4). However, in these cells, as expected in primary cells (36), coilin is not assembled in CBs, even in normal conditions. In order to investigate CBs organization and function in the disease context, independently of cell-cycle stage and ploidy, we developed an in vitro model of patient’s MNs, by producing hiPSCs-derived spinal MNs mimicking post-mitotic MNs from the spinal cord. Most importantly, we show that, while CBs are well assembled in control MNs, decreased VRK1 levels lead to a disintegration of CBs in patients’ MNs. Moreover, when CBs are present, they have a smaller size (Fig. 5) suggestive of their abnormal assembly. Several studies have shown that CBs number correlates with neuronal size and global transcriptional activity (33). For the first time, we bring evidence that VRK1 is necessary for CBs assembly in human MNs. Although we provide no experimental evidence that coilin phosphorylation is altered in patients, our results strongly suggest that the remaining mutant VRK1 proteins in our patient’s MNs are not sufficient to maintain the phosphorylation levels of coilin required for CBs assembly.

Considering the fact that CBs number positively correlates with global transcriptional activity and the cellular demand for pre-mRNA and pre-rRNA processing (29,33), our results strongly suggest that altered RNA metabolism due to CBs depletion is an essential component of the pathophysiological mechanisms of this VRK1-related disease. In post-mitotic neurons, RNA metabolism is essential to maintain metabolic and electrical activity, and indeed, we demonstrate that the patient’s hiPSC-derived MNs have shorter neurite length and altered branching, which are consistent with a length dependent axonopathy affecting both lower and upper MNs.

There are few studies about the contribution of CBs to the pathophysiology of neurological disorders; however, the depletion of CBs and the presence of smaller CBs, which we describe here in VRK1 patients’ MNs, are also hallmarks of SMA (MIM 253300) (28), another MN disease (lower MN) caused by loss or mutations in the survival MN (SMN) protein. SMN is part of the SMN complex, which localizes to the nucleus, where it accumulates in CBs. The SMN protein shuttles between the cytoplasm and the nucleus, and its subcellular localization is regulated by complex regulatory cues mediated by phosphorylation of serine/threonine and tyrosine residues (37). Knowing the function of SMN in pre-mRNA splicing, it is therefore tempting to speculate that VRK1 might be involved in the regulation of SMN function through its nuclear phosphorylation and that the defects observed in our patient’s MNs are due to impaired RNA metabolism.

In conclusion, we describe new VRK1 mutations in two siblings affected with distal hereditary motor neuropathy associated with upper MN signs. We provide, for the first time, evidence for a role of VRK1 in regulating CBs assembly in MNs. Many questions remain to be addressed regarding the interactions between components of the CBs, such as coilin and SMN, but our study provides strong evidence that VRK1 is a crucial protein in maintaining CBs integrity in MNs and might be a key factor in pathophysiology of lower and/or upper MN diseases. Taken together, our results strongly suggest that defective RNA metabolism might be a unifying theme in diseases affecting MNs, such as SMA and the one that we describe here.

Targeting VRK1 in MN diseases might therefore open up new research areas for therapy of these diseases.

Materials and Methods

Genetic analyses

Samples

After informed consent was obtained from all individuals included in this study, EDTA blood samples were collected, and genomic DNA was extracted from lymphocytes with the use of standard methods. All protocols performed in this study complied with the ethics guidelines of the institutions involved.

WES

WES was carried out on both patients II.2 and II.3 (Fig. 1A). Library preparation, capture and sequencing were performed by the French National Genotyping Center (CNG, Evry, France), using the in solution Agilent SureSelect Human All Exon kit v3.0 (38) and then sequenced on an Illumina HiSeq2000, using a paired-end 100 bp read sequencing protocol. Image analysis and base calling were performed using the Illumina Data Analysis Pipeline Software 1.5 with default parameters. Raw data were mapped to the built of the human genome (hg19) by using Burrows-Wheeler Aligner (BWA) v0.7.5 (39). Variant calling was subsequently performed using GATK (40) and annotation was done with ANNOVAR (41). All subsequent steps were performed using our in-house software for variant annotation and segregation VarAFT (42). WES data from both patients were analyzed simultaneously and segregated using this tool. Variants were filtered based on an autosomal recessive inheritance pattern; considering the absence of consanguinity, we searched in priority for apparently compound heterozygous mutations shared by the two patients, although a search for shared homozygous variants was also performed. The analysis was performed in a two-phased approach: (i) analysis of a list of 88 genes known to be implicated in IPNs and (ii) analysis of the WES data. In order to refine the obtained lists of candidates, filtering was performed by removing all variants with a frequency above 1% on the gnomAD (http://gnomad.broadinstitute.org/). Additional filtering was then performed using several frequency datasets: the Greater Middle East Variome (http://igm.ucsd.edu/gme/) and an in-house exome database. Finally, manual filtering was performed by removing variants with frequencies above 1% in gnomAD subpopulations, or variants present at the homozygous state in gnomAD.

In order to predict the deleterious effect of the identified sequence variations, different bioinformatics tools were applied: MutationTaster (43), SIFT (44), PolyPhen-2 (45), UMD predictor (46) and CADD (47).

Segregation analysis by capillary Sanger sequencing

Candidate variants identified by WES sequencing in patients were tested by Sanger sequencing of PCR amplified fragments in family members, for whom DNA was available: the mother II.1, the two patients II.2 and II.3, and the two unaffected siblings II.1 and II.4 (Fig. 1).

Genomic and complementary DNAs (cDNA) sequences of the candidate gene (VRK1, Genbank accession number: NM_003384) were obtained from the UCSC Genome Browser, February 2009 human reference sequence (GRCh37 https://genome.ucsc.edu/). Primers used for PCR amplification were designed using Primer3 software (http://frodo.wi.mit.edu) to amplify the region surrounding the candidate DNA variations (available upon request). PCR products were purified by mixing with an equal volume (10 ul) of AMPure beads (Beckman Coulter, USA) according to the manufacturer’s instructions. Both strands were sequenced as described in Jobling et al. (48). Electrophoregrams were aligned to the reference sequence using Sequencher v5.4.6 (Genecodes, USA).

RT-PCR and qRT-PCR

Total RNA was extracted from cells using Pure Link RNA Minikit (Thermo Fisher Scientific, USA) and treated with DNAse I, on-column, according to the manufacturer's instructions. cDNA were synthesized from total RNA using the SuperScript III Reverse Transcriptase and random hexamers (Thermo Fisher Scientific, USA) according to the manufacturer’s protocol. For RT-PCR, specific primers were designed through the Primer3 software in order to detect the three possible isoforms of mouse vrk1 (available upon request). qRT-PCR experiments were performed on total RNA from patients and controls’ cells to measure VRK1 mRNA levels. Experiments were performed in triplicate on a Lightcycler 480 (Roche, Basel, Switzerland) using the Luminaris Color HiGreen qPCR Master Mix (Thermo Fisher Scientific, USA). 50 ng of RNAse H-treated cDNA were amplified in a 10 μl reaction mixture containing primers at a final concentration of 0.3 μM. For each sample, the expression levels of VRK1 were normalized to three different endogenous genes: GAPDH, GUS-B and B2M. Relative expression of VRK1 was calculated as fold-differences using the comparative C_T method (∆∆C_T). Two calibrator samples were used for each sample. Primer sequences are available upon request.

Cellular studies

Cell culture

Lymphoblastoid cells

Immortalized lymphoblastoid cell lines from patients and controls were prepared at the accredited Biological Resource Center (CRB TAC), Department of Medical Genetics, Timone Hospital of Marseille, from acid citrate dextrose blood samples transformed with Epstein–Barr virus using standard procedures. All used cell lines are stored at CRB TAC and belong to a biological sample collection declared to the French ministry of Health (declaration number DC-2008-429) whose use for research purposes was authorized by the French ministry of Health (authorization number AC-2011-1312 and AC-2017-2986). Human lymphoblastoid cells were grown as described in Jobling et al. (48).

Fibroblasts

Control fibroblast cell lines were purchased from the Coriell Cell Repository (Coriell Institute for Medical Research, Philadelphia, USA): AG08471 (male, foreskin fibroblasts) and AG13091 (40 years old male, foreskin fibroblasts). Patients’ fibroblasts II.2 and II.3 were prepared from a skin biopsy and stored by the CRB TAC, Department of Medical Genetics, Timone Hospital of Marseille according to the French regulation. Primary human fibroblasts were cultured as described in Jobling et al. (48).

hiPSCs maintenance and differentiation into spinal MNs

Reprogrammation of skin fibroblasts from patient II.2 (Fig. 1) into induced pluripotent stem cells (iPSC) was performed by our Cell Reprogramming and Differentiation Facility at U 1251/Marseille Medical Genetics (Marseille, France), using retroviral transduction of Oct4, Sox2, Klf4 and c-myc protocols (49). The control iPSC cell line was provided by the Cell Reprogramming and Differentiation Facility. All hiPSCs included in the study are declared in a collection authorized by the competent authorities in France (declaration number DC-2018-3207). For each cell line, we used classical quality control criteria, including expression of pluripotency markers, absence of expression of the reprogramming transgenes, phosphatase alkaline staining and karyotype for checking the absence of chromosomal abnormalities. Expression of several pluripotency markers was verified by flow cytometry: keratin sulfate antigens Tra1-60 and Tra1-81 and the glycolipid antigens SSEA3 and SSEA4 for the control hiPSC cell line, or intracellular pluripotency markers Oct3/4, Sox2 and Nanog for Patient II.2 hiPSC cell line. Pluripotency of hiPSC cell lines was also assessed by an in vitro embryoid body (EB) assay (50), where the formation of the three developmental germ layers (ectoderm, endoderm, and mesoderm) was tested by qRT-PCR analysis of germ layer-specific genes (FOXA2, AFP, TBXT, NCAM, COL2A1, ALDhA1) in the formed EBs at day 15.

The presence of the c.656G>T (p.Arg219Ile) and c.761G>T (p.Trp254Leu) mutations in VRK1 has been checked, in the patient II.2 hiPSC clone, after reprogrammation, by fluorescent Sanger sequencing.

iPSC clones were maintained and expanded on matrigel (BD Biosciences)-coated dishes in mTeSR1 medium (STEM CELL Technologies 05851) under standard procedures.

For differentiation of iPSCs into spinal MNs, we used an established 30 days differentiation protocol based on early activation of the Wnt signaling, coupled to activation of the Hedgehog pathway and inhibition of Notch signaling (51). Using this protocol, we obtained mature MNs, which express acetylcholine transferase and can fire Action Potentials (unpublished data). Differentiation efficiency was assessed by HB9 and Islet1 immunostaining as described in Maury et al. (51). Both iPS and differentiated MNs were tested for mycoplasma contamination every other week.

Protein extraction and immunoblotting

Protein extraction and immunoblotting were performed from fibroblasts or lymphocytes pellets as mentioned in Jobling et al. (48). Primary antibodies were rabbit polyclonal antibody to VRK1 (SIGMA-ALDRICH, #HPA017929) and GAPDH (Santa Cruz Biotechnology, #sc-48167), diluted at 1:200 and 1:1000, respectively. GAPDH served as a loading control. Secondary antibodies were donkey anti-rabbit IRDye 800 and donkey anti-goat IRDye 680, (Li-COR Biosciences), diluted at 1/10000.

Immunostaining and microscopy

Conditions for immunostaining have been described in Jobling et al. (48). VRK1 was detected with a rabbit polyclonal antibody (1:200) (SIGMA-ALDRICH, #HPA017929), coilin with mouse Pdelta monoclonal antibody (1:50) (Santa Cruz Biotechnology, #sc-56298), Neurofilament M with chicken polyclonal antibody (1:1000) (Covance, # PCK-593P). The following secondary antibodies were used: Goat Anti-Rabbit IgG H&L (Alexa Fluor®488) ab150077 (abcam, UK) at 1:400 dilution, donkey Anti-Mouse IgG H&L (DyLight® 550) ab96876 (abcam, UK) at 1:400 dilution, Donkey Anti-Rabbit IgG H&L (DyLight® 550) ab96892 (abcam, UK) ) at 1:400 dilution, Goat Anti-Chicken IgY H&L (Alexa Fluor® 488) ab150169 (abcam, UK) at 1: 1000 dilution and Donkey anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 647, A-31571 (Thermo Fisher Scientific, USA) at 1:1000 dilution.

Proteasome blockade

Fibroblasts from patients and controls were cultured with culture medium containing MG132 (474790, Merck Chemical LTD) at a final concentration of 1.5 μM, or the same volume of vehicle (DMSO, 0.025% v/v) for 1 h at 37°C and 5% CO2.

Constructs and production of AAV vectors

AAV2/6 vectors were produced by the Bertarelli platform for Gene Therapy at EPFL (Lausanne, Switzerland). Production and titration of AAV2/6-CMV-GFP was performed as described by Dirren et al. (52).

Quantification of neurite length and branching in GFP+ hiPSC-derived infected MNs

Quantification of neurite length and branching was performed in GFP positive-hiPSC-derived MNs after infection with AAV2/6-CMV-GFP (2 TU/cell). After fixation of the cells, GFP signal was amplified by immunohistochemistry using a goat anti-GFP antibody (Abcam, ab5449) following incubation with an anti-goat secondary antibody (Abcam, ab96931). Fluorescence was observed under a Zeiss ApoTome.2 fluorescence microscope with a 10× objective. Neuronal processes were then analyzed using the plugin NeuronJ on ImageJ software. We measured total neurite length and branching: (i) neurite length corresponds to the sum of the length of all neuritic branches per cell; (ii) a branch is defined as a neurite segment starting at a branching point or at the cell body, and ending at the next branching point or at the extremity of the neurite processes; (iii) branching points were defined when a secondary or tertiary neurite branch diverges from the primary path.

Statistical analyses

Statistical analyses were performed using unpaired two-tailed Student’s t-test, a one-way or a two-way analysis of variance (ANOVA) test for parametric data, and a Mann–Whitney test for non-parametric data, as appropriate. Details of the used test are described in the legends of the figures. Data are reported as mean ±SEM. Significance was accepted as the level of P < 0.05 (^*P < 0.05, ^**P < 0.01 and ^***P < 0.001).

Acknowledgements

We would like to thank the patients and families for their kind cooperation and their participation in this study. We thank the Bertarelli platform for Gene Therapy at EPFL (Lausanne, Switzerland) for viral vector production of AAV-CMV-GFP. We thank Claire El Yazidi and Morgane Thomas, from the Cell Reprogramming and Differentiation Facility at U 1251/Marseille Medical Genetics (Marseille, France) for reprogrammation and characterization of patient’s II.2 fibroblasts into iPSC cell line and for providing control iPSC lines. We thank Dr Antoine Salloum for performing skin biopsies.

Conflicts of Interest statement. None declared.

Funding

French Association against Myopathies `Association Française contre les Myopathies'; the `Agence Universitaire de la Francophonie' (AUF); A.MIDEX Foundation; Fellowship from the HERMES Programme of the European Union's Seventh Framework Programme for research, technological development and demonstration (to L.E.-B.); French Association against Myopathies `Association Française contre les Myopathies' (AFM to L.E.-B.); `Association ADN' (to K.R.).