Abstract
Background:
GNE myopathy is an adult onset recessive genetic disorder that affects distal muscles sparing the quadriceps. GNE gene mutations have been identified in GNE myopathy patients all over the world. Homozygosity is a common feature in GNE myopathy patients worldwide.
Objectives:
The major objective of this study was to investigate the mutation spectrum of GNE myopathy in India in relation to the population diversity in the country.
Materials and methods:
We have collated GNE mutation data of Indian GNE myopathy patients from published literature and from recently identified patients. We also used data of people of Indian subcontinent from 1000 genomes database, South Asian Genome database and Strand Life Science database to determine frequency of GNE mutations in the general population.
Results:
A total of 67 GNE myopathy patients were studied, of whom 21% were homozygous for GNE variants, while the rest were compound heterozygous. Thirty-five different mutations in the GNE gene were recorded, of which 5 have not been reported earlier. The most frequent mutation was p.Val727Met (65%) found mainly in the heterozygous form. Another mutation, p.Ile618Thr was also common (16%) but was found mainly in patients from Rajasthan, while p.Val727Met was more widely distributed. The latter was also seen at a high frequency in general population of Indian subcontinent in all the databases. It was also present in Thailand but was absent in general population elsewhere in the world.
Conclusion:
p.Val727Met is likely to be a founder mutation of Indian subcontinent.
INTRODUCTION
GNE myopathy is an adult onset autosomal recessive genetic disorder that leads to progressive skeletal muscle weakness starting from distal muscles. The unique feature of this disease is that quadriceps are relatively spared, which are normally severely affected in other myopathies. It was initially identified in the Iranian Jewish population in Israel, and in Japan [1]. It was known by various names like Quadriceps Sparing Myopathy, Hereditary Inclusion Body Myopathy (due to the presence of inclusion bodies in affected muscle cells), and Nonaka Myopathy. With the identification of the disease-causing gene, GNE, it became evident that these various disorders were in fact the same disease, which was named GNE myopathy [2, 3]. Although the disease was initially thought to be localized to a few ethnicities, knowledge of the causative gene, and improved DNA diagnostic facilities have revealed its presence all over the world.
The GNE gene encodes a bifunctional enzyme UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase that catalyses the first two rate limiting steps in the sialic acid biosynthetic pathway. Homozygous and compound heterozygous mutations in both the epimerase and kinase domains of the GNE gene have been identified in patients of GNE myopathy, and the majority of these are missense mutations. About 70% of GNE myopathy patients display compound heterozygous mutations. In addition, a few indels, intronic variants, rearrangements and copy number variations with recombination hotspots have been reported [4, 5]. GNE variants presumably lead to reduction in the ability of the enzyme to synthesize sialic acid, and consequent hypo sialylation of glycoconjugates in muscle tissue [6]. A large number of GNE myopathy-causing GNE mutations have been recorded so far [for a list of variations http://www.dmd.nl/nmdb2/home.php?select_db=GNE, 4]. A few founder mutations have also been noted in some ethnic populations, such as p.Met743Thr (originally p.Met712Thr in old hGNE1 nomenclature) in patients of Middle Eastern origin [7, 8] and p.Val603Leu (originally reported as p.Val572Leu in hGNE1) are common in Japanese patients [9]. Reports of GNE myopathy in Indian patients have appeared relatively recently. GNE mutations in 9 GNE myopathy patients belonging to 8 Indian families were previously reported [10]. Recently Khadilkar and colleagues reported 26 genetically confirmed patients with GNE myopathy. In this study a number of patients from Rajasthan, India displayed the c.1853T>C;p.Ile618Thr variant [11]. Interestingly, this variant has also been seen in the Roma (Gypsy) population of Europe (a founder mutation) and shows extensive homozygosity. The mutation c.2179G>A;p.Val727Met was also found in many Indian GNE myopathy patients, and was reported in patients from Thailand, making it a common pathogenic variant [4].
We have used the currently available mutation information of GNE myopathy from published data and newly diagnosed patients from India to identify the mutations unique to the Indian subcontinent, and to estimate the disease frequency in India. Many Indian ethnic groups are small and isolated, and consanguineous marriages continue to be practiced. Both of these factors are conducive to an increase in prevalence of a recessive disease compared to a randomly-mating population in Hardy-Weinberg equilibrium.
MATERIALS AND METHODS
Population and patient data
We have restricted our analysis to the data pertaining to people of Indian subcontinent (India, Srilanka, Bangladesh, Pakistan, Nepal). The databases used for the analysis are 1000 genomes (http://www.internationalgenome.org/, South Asian Genomes and Exomes (http://clingen.igib.res.in/sage/), Exome Variant Server (http://evs.gs.washington.edu/EVS/) and Strand’s internal patient pooled database (PPDB). In addition, data from different publications pertaining to patients of Indian subcontinent have been curated and the sources have been cited at the relevant places. Data for 51 patients were obtained from publications, and that for 16 patients were generated as part of this study. We searched the PubMed database for all GNEM-related papers using a number of different queries (including GNE myopathy, HIBM, DMRV, Nonaka myopathy, and other relevant search terms). We have included all papers relevant to GNE gene mutations in Indian GNEM patients in the current report.
The Indian GNE myopathy patients reported here were mainly from two major clinics (Bengaluru, and Mumbai) with very few patients diagnosed in clinics from other parts of India. More patients in our study were from Southern and Western Indian States which is probably due to the location of clinics in these regions. States with the majority of recorded patients include Karnataka, Maharashtra, Gujarat, Rajasthan and Punjab. The age of diagnosis of Indian GNE myopathy patients varied from 20–50 years and about 60% of patients were females.
Patients’ diagnostic evaluation methods have previously been described [10, 11]. Briefly, patients were evaluated clinically with detailed history and examination of muscle power as per MRC (Medical Research Council) Grading. Serum creatine kinase, heart electrophysiology tests and diagnostic muscle biopsies were performed. Tibialis anterior or biceps brachii muscles were chosen for biopsy and stained with haematoxylin and eosin and modified Gomori Trichrome stains for histology evaluation and detection of rimmed vacuoles. Blood samples were subjected to genetic evaluation, using next generation and Sanger sequencing as previously described. GNE mutation information of Indian GNE myopathy patients came from a number of collaborating Centres, including National Institute of Neuroscience (Tokyo, Japan), Strand Life Sciences (Bengaluru, India), Eurofins Genomics (India), Perkin Elmer (India) and Northern Molecular Genetics Service (Newcastle Upon Tyne, UK). The majority of GNE gene sequencing for this report was performed at Strand Life Sciences using next generation sequencing, described below.
Next generation sequencing
Genomic DNA was extracted from peripheral blood lymphocytes of individuals and sequenced using TruSight One Sequencing Panel (Illumina, San Diego, CA, USA), containing >4500 genes associated with known inherited diseases, including GNE. An analytical validation of this panel has shown sensitivity of >96.5%, specificity of 99.95% and reproducibility of ∼97%. This panel has a coverage of >99% (≥10 reads) of exonic and flanking intronic regions of the GNE gene (NM_001128227.2).
DNA from each person was used for the individual library preparation for Next Generation Sequencing (NGS) as previously described [12], followed by enrichment which involved two successive hybridization steps with target specific biotinylated probes that are targeted at the exons and exon-intron junctions of genes in the panel. The target library was amplified using limited cycles PCR (ABI9700, Life Technologies), followed by sequencing using NextSeq platform (Illumina) according to the manufacturer’s instructions.
Data analysis and interpretation
NextSeq Reporter from Illumina was used to generate the trimmed fastq files and the reads were aligned against the whole genome build hg19 using Strand NGS v2.6 (Strand Life Sciences) as described [12] and Strand NGS variant caller was used to detect variants in the target region. Variants were then imported into StrandOmicsv3.4 (a proprietary clinical genomics interpretation and reporting platform from Strand Life Sciences), that includes publicly available data sources such as Uniprot, OMIM, HGMD, ClinVar, dbSNP, 1000 Genomes, Exome Variant Server and Exome Aggregation Consortium (ExAC) as well as bioinformatics prediction tools, such as SIFT, PolyPhen HVAR/HDIV, Mutation Taster, Mutation Assessor, FATHMM and LRT integrated in order to assess the pathogenicity of the variants. The identified GNE variants were labelled according to the ACMG recommended standards for reporting of sequence variations [13].
Analysis of genome databases and prediction of the effect of the mutations
In order to find the population distribution/allele frequency of c.2179G>A and other alleles, different databases were queried. The potential effect of the mutations on GNE protein function was estimated from SIFT and Polyphen analysis using the tool “variant effect predictor” available at 1000 genome portal.
RESULTS AND DISCUSSION
Mutations in Indian GNE myopathy patients
We obtained GNE mutation data for 67 confirmed patients (by clinical findings and muscle biopsy) of Indian origin. The data were compiled from various sources including published data and newly diagnosed patients (Table 1). Fifty-four patients had confirmed bi-allelic GNE mutations, and for 13 patients only one GNE variant was detected. Of the 54 GNE myopathy cases with bi-allelic GNE variants, 41 were compound heterozygotes, and in 38 of those c.2179G>A;p.Val727Met was one of the variant alleles, while the other allele was either unique or present in other cases. A total of 35 different pathologic GNE variants were found. As noted above, p.Val727Met was the most frequent (44/67 patients or 46/121 alleles). Another variant, c.1853T>C;p.Ile618Thr was also common (11/67 cases, 18/121 alleles) and appeared to occur mainly in patients from Rajasthan (11 patients), and was also more frequently found as homozygous (7/11 patients), which could result from marriage restrictions outside the ethnic boundary. Of the 35 mutations in this study, five mutations were not previously reported. Of these the c.52-1G>T mutation is interesting as it is located at the exon-intron boundary of exon 1. The resultant splicing defect might lead to retention of exon 1 which is normally spliced to exon 3 or 4 in different isoforms. This exon is skipped in hGNE1 and is not translated. In the isoform hGNE2, 17 amino acids of exon 1 are spliced to exon 3 resulting in 31 additional amino acids. This form has highly reduced epimerase activity, although the kinase activity is normal [14].
GNE mutations in Indian GNE myopathy patients
TH This study. $Also found in Japan [21]. #Novel GNE mutations. α Uncertain homozygous. – Not reported. &Stop codon.
Homozygosity is a common feature in GNE myopathy patients worldwide. For example, amongst Jews of Persian origin all patients were homozygous for p.Met743Thr [8, 15]. In Japanese patients the most common mutation is p.Val603Leu and a large number of patients with this mutation were homozygous. However, another common mutation amongst Japanese patients, p.Asp207Val is rarely homozygous. Our data show that amongst Indian patients homozygosity is infrequent. Of 67 patients only 14 (21%) were homozygous. The most common mutation p.Val727Met was homozygous only in one patient, and in another patient this allele was present on both chromosomes in combination with a second mutation, making it compound heterozygous. A single mutation, p.Ile618Thr accounted for 7 out of 14 homozygous patients, and as mentioned above these patients belonged to a particular ethnicity. The near absence of homozygous p.Val727Met patients probably suggests that disease may not develop in such individuals. The alternate possibility that homozygosity may lead to embryonic lethality is unlikely to be true as two homozygous patients have been seen.
In order to estimate the probable effect of the mutations on GNE function we carried out Polyphen analysis. A number of mutations (including p.Val727Met) displayed a high Polyphen score, suggesting that these mutations are likely to affect protein function (Table 2). A number of missense variants are predicted unlikely to be functionally different from wild type protein. It appears that many of these variants map to the surface of the protein, and could affect multimerisation or other properties, resulting in loss of protein function [16].
Frequency of pathogenic GNE variants in Indian patients and general population&
&Only alleles with frequency greater than1 are listed. #Novel GNE pathogenic variant. $Allele frequency is given within brackets. *Analysis was done for SAS population. **See Materials and Methods for details. ***http://clingen.igib.res.in/sage/
In addition, a truncated GNE protein is likely to occur in 6 patients (with 5 different nonsense mutations) due to presence of stop codons introduced by nonsense mutations and due to frameshift variants (2 reported). All seven truncating mutations were heterozygous, all with p.Val727Met occurring on the other allele. Generally, stop gain mutations are never seen on both alleles in GNE myopathy as this likely results in a null phenotype leading to lethality [1].
Population distribution of GNE variants
Overall allele frequency of p.Val727Met was 38% in patients, with 65% of all reported Indian GNE myopathy having at least one p.Val727Met allele (Table 2). The observed frequency may be a little lower due to inability to get information about the second allele in a few patients. Interestingly, this allele is rarely seen in patients from other parts of the world, except Thailand and appears to be common only in patients of Indian subcontinent. The spread of Indian culture and its religions (Hinduism and Buddhism) to Southeast Asia took place before the beginning of Christian era up till the thirteenth century. This historical migration and mixing of people of Indian origin with those of Southeast Asia could explain the appearance of the p.Val727Met allele in Thailand [17]. This allele has been reported in one middle-east patient [18] and one Chinese patient [19]. From our data it appears that the majority of Indian patients carry the p.Val727Met allele along with another pathogenic allele that varies from patient to patient. Since we have reported the available mutation data of all Indian patients there is no obvious regional bias. Yet, given the size of the country, it is possible that as more country-wide data from Indian patients is obtained, other pathogenic alleles of high frequency get discovered. The p.Ile618Thr allele was found commonly in patients from Rajasthan. Unlike the p.Val727Met allele, this variant was frequently homozygous [11]. It appears to be a common allele among Roma Gypsy population in Europe with a carrier frequency of 1.67% [20]. This supports the suggestion that Rajasthani people have historical link with Roma population [11].
Of the other mutations in Indian patients, 5 were also reported in Japanese patients (p.Arg160*, p.Arg277Gln, p.Arg308Cys, p.Arg451*, p.Ala661Thr) [21]. Interestingly we have also noticed different stop-gain mutations. In these cases, truncated GNE protein with severely impaired function are likely to be made.
We have combined the data available from indicated portals to compute allele frequency of GNE variants in the general population of Indian subcontinent. p.Val727Met was observed at a frequency of 0.79%, 1.4%, or 2% depending upon the dataset used (Table 3). Since different datasets have data from combination of different ethnicities, it may explain the variations seen in allele frequencies. For example, IGIB dataset also has data from people belonging to other regions of Indian sub-continent, which may explain the lower frequency seen in this dataset. This allele was seen at a very high frequency in Gujarati community (1000 genome data). About 13 out of 90 individuals carried one p.Val727Met allele. This also corroborates the relatively higher number of patients currently seen from the same community. p.Arg277Gln was also seen at a frequency of 1% in the population, but only in the 1000genome dataset. None of the other variants was seen in the general population, and a much larger sample size would be needed to estimate their frequency. We also investigated occurrence of p.Val727Met variant in other populations (Table 4). We did not find this variant in normal American and European population data available at the NHLBI Exome Sequencing Project (http://evs.gs.washington.edu/EVS/). It is clear that this allele is present only among people from Indian subcontinent and is not found in other ethnicities. In a recent estimate of GNE myopathy prevalence based on allele frequency derived from some of the publicly available datasets, the number arrived at was 6 in 1,000,000 [4]. In this estimation, a population under Hardy-Weinberg equilibrium was assumed and predicted about 40,000 patients world-wide, which far exceeds the currently known diagnosed patients. We estimate the prevalence of the p.Val727Met allele in the general Indian population from the three data bases to be 1.4%. From this, using the method of Celeste et al. [4] the prevalence is calculated to be 49 in 1,000,000. A better estimate of the number of GNE myopathy patients in India will be obtained once more sequence data become available for larger numbers of individuals of Indian origin, which would also provide allele frequency of the less frequent pathogenic variants.
Predicted effect of novel GNE variants on GNE protein function
#Polyphen prediction (score).
Allele frequency of c.2179G>A in different populations
Frequency (total alleles) was obtained from indicated databases. 1Databases used were 1000genome, Strand and South Asian genome. 2Database used is 1000genome. 3Database used is 1000genome. 4Database used is 1000genome. 5Database used is 1000genome.
As mentioned before there is no approved treatment for GNE myopathy. Currently a clinical trial is being conducted by National Institutes of Health, USA to test the efficacy of N-acetyl mannosamine supplement in the treatment of this disease (https://www.genengnews.com/gen-news-highlights/ultragenyx-halts-ace-er-development-after-phase-iii-failure/81254845). The patient data presented here will help to carry out trials in India based on future potential therapy. Moreover, this information will also be useful in ongoing attempts to develop a rare disease registry in India. All these efforts pave the way for developing future therapies towards rare diseases (http://irdr.icmr.org.in/irdr/index.php/diseases-included).
CONCLUSION
In this study we have collated the largest number of GNE myopathy patients (67 cases) from the Indian subcontinent to date. However, given the size of the Indian population, this is still a minority of the total number of Indian GNE myopathy patients. Under diagnoses of GNE myopathy in India, as in the rest of the world, is due to non-specific clinical symptoms at adult onset of the disease, unfamiliarity of neurologists with this rare disease, practical difficulty of muscle biopsy diagnosis, unavailable genetic testing and other factors. It is clear that p.Val727Met is a major pathogenic allele of GNE myopathy in India. Its presence at a high frequency in the normal population of Indian subcontinent, particularly among Gujaratis, and its absence in most other parts of the world suggests that it is likely to be a founder mutation. The high frequency of this allele further indicates the likelihood of many undiagnosed patients of GNE myopathy in this country. Therefore, we suggest to survey the presence of at least the GNE p.Val727Met allele (if complete GNE gene testing is unavailable) to be carried out routinely in Indian patients with unexplained adult onset- neurological symptoms of upper and lower extremities (foot drop, muscle atrophy and weakness, rimmed vacuoles on muscle biopsy).
CONFLICT OF INTEREST
There is no conflict of interest of any of the authors in this study.
Footnotes
ACKNOWLEDGMENT
We thank Department of Science & Technology for JC Bose Fellowship (AB, SB and PPM) and BNP-Paribas (AB) for financial support. We also thank all GNE myopathy patients to whom this work is dedicated.