Background Infantile neuroaxonal dystrophy (INAD) is a recessive disease that results in total neurological degeneration and death in childhood. PLA2G6 mutation is the underlying genetic defect, but rare genetic heterogeneity has been demonstrated. One of the five families we studied did not link to PLA2G6 locus, and in the family one of the two affected siblings additionally had atypical features including facial dysmorphism, pectus carinatum, scoliosis, pes varus, zygodactyly and bilateral cryptorchidism as well as cerebellar atrophy, as previously reported.
Methods Sural biopsy was investigated by electron microscopy. PLA2G6 was screened for mutations by Sanger sequencing. In the mutation-free family, candidate disease loci were found via linkage analysis using data from single nucleotide polymorphism genome scans. Exome sequencing was applied to find the variants at the loci.
Results PLA2G6 mutations were identified in four families including the one with an unusually severe phenotype that led to death within the first 2 years of life. In the remaining family, seven candidate loci totalling 15.2 Mb were found and a homozygous truncating mutation p.Q642X was identified in NALCN at 13q32.3. The patients are around 20-years-old.
Conclusions NALCN is the gene responsible for INAD with facial dysmorphism. The patients have lived to adulthood despite severe growth and neuromotor retardation. NALCN forms a voltage-independent ion channel with a role in the regulation of neuronal excitability. Our findings broaden the spectrum of genes associated with neuroaxonal dystrophy. Testing infants with idiopathic severe growth retardation and neurodegeneration for NALCN mutations could benefit families.
Statistics from Altmetric.com
Infantile neuroaxonal dystrophy (INAD) is characterised by progressive motor, mental and visual deterioration that begin in infancy. Onset is before the age of 3 years and death typically ensues before the age of 10 years.1 Early clinical manifestations include bilateral pyramidal tract signs, truncal hypotonia, cognitive decline and optic atrophy. Distal axonal swelling and spheroid bodies are found in tissue biopsies in most cases.2 ,3 PLA2G6 has been identified as the gene responsible for this recessive disorder.3 ,4 Studies on other patients indicated that the INAD clinical phenotype for PLA2G6 mutation is not uniform.5 ,6 Moreover, PLA2G6 mutation was found to lead to other recessive disorders, namely, parkinsonism (PARK14)7–9 and Karak syndrome.3 Linkage analysis in 12 families indicated locus heterogeneity for INAD, supporting the existence of one additional locus.3
Five INAD families were investigated in this study. PLA2G6 mutations were identified in four families. We searched for the disease gene in the remaining family; the two affected siblings were reported previously as atypical INAD.10 We performed linkage analysis to identify the candidate gene loci. Using exome sequencing, a truncating NALCN mutation was identified.
Materials and methods
We studied five INAD families. In the family with atypical clinical findings (family 5), the parents were first cousins and had two affected siblings and a healthy sibling (see online supplementary figure S1). The affected siblings were referred to the Department of Medical Genetics, Cerrahpasa Medical School, University of Istanbul, for multiple congenital anomalies, facial dysmorphism and neurodevelopmental delay. Their phenotypes were published previously.10
Informed consent was obtained from/for all participants. The Boğaziçi University Institutional Review Board for Research with Human Participants approved the study protocol.
Genotyping and statistical analysis
DNA was extracted from peripheral blood samples. A genome scan was performed for all five members of the family using Illumina Human 610-Quad Beadchip that included 610 000 single nucleotide polymorphism (SNP) markers. To find the candidate loci possibly harbouring the disease gene, both linkage analysis and a search for identical by descent (IBD) homozygous genotypes in the patients were performed. Multipoint logarithm of odds (LOD) scores were calculated using GeneHunter V.2.1r5, using markers at 0.07 cM spacing and in sets of 100. Autosomal recessive inheritance with full penetrance was assumed, and disease frequency was set to 0.0001. Regions exibiting loss of heterozygositiy (LOH) thus possibly homozygous for >50 SNP markers in both affected siblings were detected via LOH Detector plug-in on Illumina GenomeStudio V.2011.1 software. At those regions, genotypes were analysed using MS Excel. At the loci where the homozygous genotypes were shared between the patients and not their unaffected sibling, LOD score calculations were repeated with the same parameters but using all markers to assess the final scores, and copy number variation (CNV) analysis was performed via cnvPartition v3.1.6 to further assess homozygosity. Finally, haplotypes were constructed via Allegro v1.2c, and haplotype segregation was investigated via HaploPainter 029.5.
Exome sequencing analysis
Exome sequencing was performed on the DNA sample from patient 2 (402), and bioinformatics analyses were done as described previously.11 Briefly, exome capture probes targeted a total of 62 Mb genomic regions aiming to cover 95% of the CCDS database. Sequencing and exome capture libraries were prepared using the TrueSeq DNA Sample Preparation Kit and TrueSeq Exome Enrichment Kit (Illumina), respectively. The enriched library was sequenced on the Illumina Hiseq2000 platform. We aligned the reads to a human genome reference sequence using Burrows-Wheeler Alignment (BWA-0.5.9) software, allowing approximately 2% base pair (bp) mismatch in read alignments. Single nucleotide variations (SNVs) and indels were detected by SAMTOOLS-0.1.14. The 62 519 variants listed in the output were annotated for novelty after comparing them with databases dbSNP (build 131) and 1000 Genomes using ANNOVAR software. Variants with total read depths <5 were filtered out. The resulting coverage for targeted exons was computed using BEDTools to detect possible exonic deletions within the candidate regions. The human GRCh37/hg19 assembly was used as reference for all genetic locations.
Among the variants in the exome sequencing results, the novel and rare (frequency ≤0.01) homozygous exonic and splicing variants within the candidate loci were considered. The reason for a cut-off frequency of ≤0.01 was that an allele frequency of 0.01 is expected to result in a homozygote frequency of 0.0001, a value above the expected frequency for the novel trait in the family. The variants were prioritised according to the predicted severity and gene function.
Coding sequences of PLA2G6 were analysed for mutations by Sanger sequencing. Mutation in NALCN identified by exome sequencing was verified by Sanger sequencing in all the family members. To investigate whether this mutation was a common variant in the Turkish population, 110 subjects randomly selected from the population were screened via the high-resolution melt (HRM) assay on a LightCycler 480 (Roche).
Quantification of transcript isoforms
To investigate the abundance of NALCN transcripts, commercial total RNA samples from adult human tissues (Clontech Laboratories, Inc) were used as templates to synthesise cDNA using a RevertAid First Strand cDNA Synthesis Kit (Fermentas). A real-time PCR was performed using intron-spanning primer pairs that were specific to either NALCN or housekeeping gene β-tubulin (TUBB)12 transcripts and a SYBR Green I Master Kit on a LightCycler 480 (Roche). Each assay was performed in triplicate, and NALCN transcript levels were normalised to the endogenous reference gene TUBB using the E-method in the LightCycler480 Relative Quantification Software (Roche). PCR efficiencies were calculated using relative standard curves.
The two affected siblings in family 5 have been investigated for multiple congenital anomalies and neurodevelopmental delay since infancy. Psychomotor retardation began in infancy, after which the patients showed little reaction to people around, had no eye-to-eye contact and never gained speech. Their clinical phenotypes in childhood were published previously.10 Both patients were born at term, with low birth weight and without any neonatal abnormalities. Growth was slow during the first year of life but became severely retarded later. At the age of 5.5 years, their anthropometric measures were below the 3rd percentile. At the last visit, patient 1 (a girl) was 21 years of age and patient 2 (a boy) was 18 years of age. Their weights were 9.7 kg (<<-5 SD) and 9.2 kg (<<-5 SD), heights 112.5 cm (<<-5 SD) and 110.5 cm (<<-5 SD), and head circumferences were likewise very small, 45.8 cm (<-4 SD) and 45.3 cm (<-4 SD), respectively. A picture of patient 2 is presented in figure 1. Progression was dramatic involving spastic tetraplegia, complete psychomotor regression and total loss of audiovisual functions, leading to the terminal stage of the disease.
The patients had severe psychomotor retardation, optic atrophy, muscular atrophy and truncal hypotonia. The electron microscopic analysis of the sural nerve biopsy performed in patient 2 at age 5 years showed complete unmyelinated axons, spheroid formations and generalised oedema, compatible with INAD (figure 2). The patients had epileptic seizures starting at age 13 years in patient 1 and at age 8 years in patient 2. They gradually developed contractures of extremities. In addition to these INAD features, in both siblings facial dysmorphism was present including prominent forehead, large and low-set ears, small nose, hypoplastic mandible, micrognathia, and strabismus. For patient 1 (401), the results of electroencephalograghy (EEG) at 14 months of age, brain MRI at 3 years of age and electromyogram (EMG) at 5 years of age were all normal.10 However, for patient 2, EEG at 7 months of age revealed diffuse bioelectric delay and EMG at 5 years of age indicated motor conduction velocity and response at the lower limits of the normal ranges whereas MRI at 5 years of age showed cerebellar atrophy.10 Miscellaneous features such as pectus carinatum, scoliosis, pes varus, zygodactyly and bilateral cryptorchidism were present in patient 2.
We analysed PLA2G6 in all the families and identified mutations in four of them (see online supplementary figure S2A). All mutations were novel and in the homozygous state in the patients (table 1). We undertook linkage analysis in the remaining family. Analysis using markers at 0.07 cM spacing yielded the highest multipoint LOD score of 1.92 at four loci, namely, 9p24.1, 13q32.3–33.2, 18q12.2 and 20p12.1 (see online supplementary figure S3). Eight other loci yielded multipoint LOD scores >1, namely, 1p34.1, 4p15.33, 8q24.23, 11p15.2, 14q21.1, 15q25.3, 16q24.1 and 20p12.3. All of the 12 loci were investigated further by haplotype inspection in the family members. At 4p15.33, 9p24.1, 13q32.3–33.2 and 18q12.2, the affected siblings shared homozygosity whereas the unaffected sib was heterozygous, indicating that they could possibly be the gene locus. At the remaining eight loci the affected siblings were not homozygous, and thus the loci were eliminated.
We extended homozygosity investigation to whole genome and detected nine loci including the abovementioned 4p15.33, 9p24.1, 13q32.3–33.2 and 18q12.2 (see online supplementary table S1) where the affected siblings shared homozygosity whereas the unaffected sib was heterozygous. Linkage analysis was repeated at those loci using all markers, and all except 1q32.1 yielded the highest LOD score of 1.92. Haplotypes of affected siblings at 1q32.1 did not support IBD from a recent ancestor, excluding this locus. Finally, non-identical CNV genotypes in the patients excluded 6p21.31. We hypothesise that only the remaining seven loci shown in online supplementary table S1 could possibly harbour the disease gene. Shared homozygosity in the affected siblings at the largest locus 13q32.3–33.2 encompassed a 10.7-Mb region between rs1412934 and rs7334038 (94 893 706 and 105 630 414 bp). The unaffected sib also shared the homozygous genotype in part, between rs1412934 and rs998733, narrowing down the candidate locus to approximately 6.5 Mb, between markers rs998733 and rs7334038 (99 089 156 and 105 630 414 bp). The next largest candidate locus was 9p24.1, a 4.1-Mb region between rs4740801 and rs2196066 (4 790 165 and 8 894 910 bp). 1p34.1–1p33 was <1.5 Mb, between markers rs12022162 and rs785475 (44 934 432 and 46 382 323 bp). The remaining four candidate loci were much smaller, between 0.69 and 0.85 Mb.
We investigated possibly deleterious homozygous gene variants in the exome sequencing results at the seven candidate loci. Three novel and one rare variants were found (table 2), all at either 9p24.1 or 13q32.3–33.2. All but one variant were either synonymous changes or non-synonymous changes that were predicted to be benign and tolerated by online tools PolyPhen-2 and SIFT, respectively. The remaining variant was a novel C to T conversion in NALCN exon 16 (c.1924C>T), which created a premature translational termination signal at codon 642 (p.Q642X) (see online supplementary figure S2). The mutation segregated with the disease in the family and was not found in the population control subjects. Additionally, we analysed the total exome sequence results for homozygous possibly damaging mutations and found 20 such variants in total, but again only NALCN p.Q642X was located in a region of shared homozygosity.
Quantitative PCR analysis showed that NALCN was expressed in all brain regions but not in the non-neural tissues investigated, indicating neural specificity (figure 3). The brain tissues assayed were the cerebellum, corpus callosum, frontal cortex, occipital cortex, parietal cortex, brain stem, striatum, substantia nigra, putamen, pons and spinal cord; the non-neural tissues assayed were blood, liver, skeletal muscle, bone marrow and adipose tissue.
The two affected siblings with NALCN mutation had been diagnosed with INAD based on physical and neurological examinations, nerve biopsy studies and brain MRI, which showed cerebellar atrophy in one sib only.10 The features not common to INAD included facial dysmorphism and skeletal anomalies. Facial dysmorphism was reported in two other INAD siblings;13 however, the additional atypical findings such as early onset peripheral gangrene and a rapidly lethal course were not found in our patients. The older of our patients is 21-years-old at present, surviving much longer than INAD patients with PLA2G6 mutation. However, they have severe neuromotor and developmental retardation, with lengths corresponding to 5.5 and 5 years for patients 1 and 2, weights to 16 and 10 months, and head circumferences to 16 and 10 months, respectively.
Exome sequencing revealed novel NALCN p.Q642X as the only damaging gene variant. The mutation is deduced to lead to the truncation of the protein after residue 641, resulting in the deletion of 1097 amino acids (63%) of the native protein with 1738 amino acids (see online supplementary figure S2C). However, the mutant protein is likely not synthesised at all, as the mutant mRNA is expected to undergo nonsense-mediated decay, the premature stop codon being located more than 50 nucleotides upstream of the 3′-most exon–exon junction.14 The mutation was not found in the 110 population samples tested, showing with >80% power that the mutation was not a normal sequence variant in the Turkish population.15
NALCN is expressed mainly in the central nervous system. Mice without functional Nalcn showed neonatal lethality, indicating that the human analogue is also most probably an essential protein.16 The protein is fully conserved between human and chimpanzee and highly conserved among other mammals, with only 24 of the total 1738 residues differing between humans and rats (HomoloGene). In the light of all this evidence, we concluded that NALCN was responsible for the disease in our patients.
NALCN (sodium leak channel, non-selective) was cloned in 199917 and encodes a voltage-independent cation channel permeable to sodium, potassium and calcium that is composed of four homologous domains, each with six transmembrane segments (see online supplementary figure S2C). In contrast, the 20 other members of the human gene family encode voltage-gated sodium-selective or calcium-selective channels.16 No other channel with the same function as NALCN is known in mammals, and the deficit of such a channel with a unique function would be expected to lead to a severe disease. Another gene family responsible for background currents is the two-pore domain potassium channels with 15 members.18 Knock-out animal models were created for some of those channel genes, and the resulting moderate phenotypes were attributed to possible compensation by other channel proteins.
NALCN is expressed the highest in brain, moderately in heart and weakly in pancreas.16 ,17 ,19 The channel regulates neuronal excitability by setting the neuronal membrane potential. Subsequent studies increased our understanding of its roles in various cellular mechanisms. It is activated by substance P and neurotensin in mouse neurones and action potentials are fired.20 A study on mice strains with Nalcn variants unravelled the role of the protein in osmoregulation.21 Several studies have shown that NALCN regulates rhythmic behaviours in animals.16 ,22–25 Nalcn mutant mice had a severely disrupted respiratory rhythm that led to death within 24 h of birth.16 Mutations in the NALCN homologue genes in Drosophila melanogaster and Caenorhabditis elegans were not lethal but resulted in altered locomotor rhythms; flies had inverted light–dark locomotor activities,22 whereas worms lost the ability to switch from crawling to swimming.23 No disruption in respiratory rhythm was observed in our patients, and the mutation was not lethal at least until the third decade. It is not uncommon that deficiencies in mouse and human orthologues result in different phenotypes, although it is generally the human phenotype that is the more severe. With referral to the high expression of the gene in heart and the possible role of the gene product in pace making,25 ,26 our patients had no history of altered heart rates or arrhythmia as assessed by electrocardiography (EKG). PLA2G6, the gene responsible for INAD, encodes calcium-independent phospholipase A2 group IV, an enzyme that does not seem to have a function related to NALCN. Thus, it is difficult to speculate how NALCN deficiency could possibly lead to pathology similar to PLA2G6 deficiency. Nevertheless, the two proteins have some similar or related functions, although not related to neuroaxonal dystrophy. Both stimulate insulin secretion in pancreatic cells,27 ,28 and PLA2G6 has a role in β-cell apoptosis, whereas NALCN is implicated in pancreatic cancer.29 ,30 NALCN is implicated also in lung cancer.31
We investigated the abundance of NALCN transcripts in various tissues. The gene was variably expressed in all regions of the brain tested but none of the non-neural tissues. Our finding that NALCN is largely neurone specific is consistent with previous studies.16 ,17 ,19
With regard to the PLA2G6 mutations that we identified, two were missense and the other two were truncating (table 1). The substitution of polar thereonine with non-polar methionine in missense mutation T661M is predicted to adversely affect protein function. Moreover, the residue at position 661 is within a stretch of 47 amino acids that is fully conserved in mammals. Both PolyPhen-2 and SIFT predicted as damaging with high confidence. PolyPhen-2 predicted the other missense mutation D484G as benign and SIFT predicted it as damaging with low confidence (score 0.4). However, the substitution of acidic aspartic acid with small, neutral glycine is expected to have an adverse effect on protein function. Residue 484 is within a stretch of 103 amino acids that is fully conserved among human, chimpanzee and macaque. All four mutations are deduced to affect both of the enzymatically active isoforms of the protein, as were all other mutations reported to date.3 ,6 ,32 The phenotypes of the patients were typical of INAD, except for the two siblings with homozygous c.402delC, who had very severe phenotypes and died in hospital at about 2 years of age.
All the evidence supports the existence of genes other than PLA2G6 responsible for neuroaxonal dystrophy in mammals: genetic heterogeneity in INAD was demonstrated in humans by linkage analysis,3 a mouse model for INAD did not link to Pla2g6 locus33 and a homozygous MFN2 mutation caused fetal-onset neuroaxonal dystrophy in a colony of laboratory dogs.34 Therefore, the identification of a second gene responsible for neuroaxonal dystrophy in humans was not a surprise.
Our findings expand the spectrum of genes responsible for neuroaxonal dystrophy and the associated phenotypes. The two siblings presented herein did not share all the clinical features; therefore, the general clinical manifestations of the disease need to be clarified further. It is hoped that future studies will facilitate the identification of other patients with NALCN mutation. NALCN is highly conserved across species, and its large size should be predisposing it to more frequent mutations. We propose that families with infants having growth and neuromotor retardation could benefit from NALCN mutation screening.
We thank the family for participating in this study.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
Contributors ÇK generated and analysed genetic data and co-wrote the manuscript. MS performed the clinical investigations and co-wrote the manuscript. AT supervised the genetic studies, obtained funding and co-wrote the manuscript.
Funding This study was supported by the Scientific and Technological Research Council of Turkey, grant number 110T252 and Boğaziçi University Research Fund, grant number BAP 5708.
Competing interests None.
Patient consent Obtained.
Ethics approval Boğaziçi University Institutional Review Board for Research with Human Participants approved the study protocol.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.