Next-Generation Sequencing in a Cohort of Asian Indian Patients with the Duchenne Muscular Dystrophy Phenotype: Diagnostic Yield and Mutation Spectrum

Abstract

Multiplex ligation-dependent probe amplification (MLPA) detects exonic deletions and duplications in the DMD gene in around 65 to 70% of patients with the Duchenne muscular dystrophy (DMD) phenotype. This study looks at the diagnostic yield of next-generation sequencing (NGS) and the mutation spectrum in an Asian Indian cohort of MLPA-negative cases with the DMD phenotype. NGS-based sequencing of DMD gene was done in 28 MLPA-negative cases (25 male probands with the DMD phenotype and 3 obligate carrier mothers of deceased affected male patients) and disease-causing variants were identified in 19 (67.9%) of these cases. Further molecular testing in four of the remaining nine cases revealed gene variants associated with limb girdle muscular dystrophies. Thus, NGS-based multigene panel testing for muscular dystrophy-associated genes or clinical exome sequencing rather than targeted DMD gene sequencing appears to be a more cost-effective testing modality with better diagnostic yield, for MLPA-negative patients with the DMD phenotype.

Introduction

Duchenne muscular dystrophy (OMIM #310200) is one of the most common and severe muscular dystrophies, with an incidence of 1 in 3,500 live male births, which usually presents in boys within the first decade of life. ¹ It is caused by mutation in the DMD gene, which is one of the largest known human genes, spanning around 2.2 megabases with 79 exons, and codes for dystrophin, a muscle membrane protein.

A large proportion of DMD cases (around 65–70%) are caused by large exonic deletions/duplications in the DMD gene. ^{2 3} Multiplex ligation-dependent probe amplification (MLPA) of the DMD gene, which detects these exonic deletions and duplications, is the recommended first-line genetic test for male children presenting with the classic phenotype of DMD. However, MLPA cannot detect the single nucleotide variants (SNVs) and small indels that are present in the remaining 30 to 35% cases of DMD. To detect these sequence variants, sequencing of the DMD gene has to be done, which was expensive and time-consuming with the conventional Sanger sequencing technique, but has now become faster and more cost-effective with the availability of the high throughput next-generation sequencing (NGS) technology.

This study describes the diagnostic yield of NGS and the spectrum of sequence variants identified in an Asian Indian cohort of MLPA-negative patients presenting with the classic DMD phenotype.

Methods

This is a prospective observational study, done in a tertiary care center in southern India. Male patients presenting with the classic DMD phenotype to the Medical Genetics Department, from June 2017 to June 2019, for whom the MLPA test of the DMD gene was negative for exonic deletions/duplications, were recruited into the study. The clinical features based on which the diagnosis of DMD was considered included history of progressive, symmetrical, and proximal muscle weakness with onset in early childhood, positive Gowers' sign, calf muscle hypertrophy, raised serum creatine phosphokinase (CPK) to more than 10 times the normal reference range, and loss of independent ambulation by the age of 13 years (in patients presenting beyond 13 years of age), with or without a positive family history suggestive of X-linked inheritance. In addition, obligate carrier mothers of deceased affected male patients (i.e., mother of deceased affected male child with history suggestive of classic DMD phenotype and with family pedigree showing other affected male relatives consistent with X-linked recessive inheritance) who had negative MLPA results, were also included.

Targeted sequencing of the DMD gene using the NGS platform was performed in all 28 cases (25 probands and three obligate carrier mothers). DNA was extracted from the peripheral blood sample of each subject using the QIA symphony DNA Mini Kit (Qiagen, Hilden, Germany). Whole genome libraries were prepared using the Kapa Hyperprep kit (Kapa Biosystems, Roche Holding AG, Switzerland). The whole genome libraries were used to perform targeted gene capture using a custom capture kit (SeqCap EZ Target Enrichment Kit, Nimblegen, Roche Holding AG, Switzerland), through which selective capture of all 79 exons and flanking intron–exon junctions of the DMD gene was done. The libraries were sequenced to mean >80–100X coverage on Illumina sequencing platform. The sequences obtained were aligned to the human reference genome (GRCh37/hg19) using Sentieon aligner and analyzed using the Sentieon genomics tool. ⁴ Only nonsynonymous and splice site variants found in the DMD gene were considered. Polymorphic variants were filtered out based on allele frequency in 1000 genome ( https://www.internationalgenome.org/1000-genomes-browsers/ ), genome aggregation database (gnomeAD) ( https://gnomad.broadinstitute.org/ ), and the internal Indian population database maintained by the laboratory. Previously reported variants were annotated with the help of published literature and standard databases such as ClinVar ( https://www.ncbi.nlm.nih.gov/clinvar/ ), HGMD (Human Gene Mutation Database; http://www.hgmd.cf.ac.uk/ac/index.php ) and LOVD (Leiden Open Variation Database; https://www.lovd.nl/3.0/search ). Classification of the identified variants was done as per the guidelines outlined by the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP variant classification guidelines 2015), ⁵ with the help of the InterVar software ( http://wintervar.wglab.org/ ). The effect of nonsynonymous variants was calculated with the help of in silico prediction tools such as PolyPhen-2 ( http://genetics.bwh.harvard.edu/pph2/ ), SIFT (Sorting Intolerant from Tolerant; https://sift.bii.a-star.edu.sg/ ), MutationTaster2 ( http://www.mutationtaster.org/ ), and CADD (Combined Annotation Dependent Depletion; https://cadd.gs.washington.edu/ ). The effect of splice variants was checked with the help of MutationTaster2 and Human Splicing Finder ( http://www.umd.be/HSF/ ).

Further clinical exome sequencing (CES) was done in four patients (out of the 9 who tested negative for DMD gene sequence variants). CES was performed using a custom capture kit and the Illumina-sequencing platform.

Results

Twenty-eight MLPA-negative cases were included in the study. Of them, 25 were male patients with the DMD phenotype and 3 were obligate carrier mothers of deceased affected patients. Out of 25 probands, NGS detected known pathogenic variants in the DMD gene in 15 (60%) patients and a variant of uncertain significance (VOUS) in the DMD gene in one proband ( Table 1 ). In 9 out of the 25 MLPA-negative probands, no sequence variants were identified in the DMD gene, suggesting the possibility of either an alternative diagnosis (i.e., a DMD-mimicking muscular dystrophy) or of a deep intronic mutation in the DMD gene (as deep intronic regions were not covered in the sequencing). Further molecular genetic testing through CES could be done in only four of these nine cases, and all four were identified to have biallelic homozygous variants in genes associated with autosomal recessive limb-girdle muscular dystrophies (LGMDs; in SGCA , SGCB , SGCC , and FKRP in one case each). In five probands negative for DMD gene variants, further CES could not be done because of financial constraints. All three mothers of affected deceased patients tested were found to be carriers for pathogenic variants in the DMD gene. All the variants identified in our cohort of patients have been listed in Table 1 , along with the details of their in-silico analysis.

Table 1. List of variants identified in the study cohort.

Variants identified in DMD gene (transcript id: ENST00000357033)
Test done in proband
Sl. no.	Variant identified				Known/Novel variant	Classification as per ACMG/AMP 2015 guidelines (using InterVar)	Database/publication with previous report of the variant	In silico prediction for variants classified as VOUS
	Exon/intron	Nucleotide change	Protein change	Type of variant
1	Exon 17	c.2047G > T	p.Glu683Ter	Nonsense	Known	Pathogenic	HGMD, LOVD	–
2	Exon 22	c.2833C > T	p.Gln945Ter	Nonsense	Known	Pathogenic	HGMD, LOVD	–
3	Exon 23	c.3154A > T	p.Lys1052Ter	Nonsense	Known	Pathogenic	LOVD	–
4	Exon 23	c.3037G > T	p.Glu1013Ter	Nonsense	Known	Likely Pathogenic	LOVD	–
5	Exon 43	c.6283C > T	p.Arg2095Ter	Nonsense	Known	Pathogenic	ClinVar, HGMD, LOVD	–
6	Exon 44	c.6423C > G	p.Tyr2141Ter	Nonsense	Known	Pathogenic	ClinVar, HGMD, LOVD	–
7	Exon 59	c.8692C > T	p.Gly2898Ter	Nonsense	Known	Pathogenic	HGMD	–
8	Exon 63	c.9268G > T	p.Glu3090Ter	Nonsense	Novel	Pathogenic	–	–
9	Exon 74	c.10546G > T	p.Glu3516Ter	Nonsense	Known	Pathogenic	ClinVar, HGMD	–
10	Exon 24	c.3257dupA	p.Gln1087AlafsTer11	Small indel (duplication)	Known	Pathogenic	LOVD	–
11	Exon 39	c.5478_5490del	p.Leu1827GlufsTer17	Small indel (deletion)	Known	Pathogenic	LOVD	–
12	Exon 74	c.10454del	p.Leu3485ArgfsTer11	Small indel (deletion)	Known	Pathogenic	Roberts et al 6	–
13	Intron 17 Exon 17	c.2168–2168 + 1delGCinsCT	–	Small indel (deletion) and 5′ splice site variant	Known	Pathogenic	LOVD	–
14	Intron 19	c.2381–2A > G	–	3′ splice site variant	Known	Pathogenic	LOVD	–
15	Intron 69	c.10086 + 2dup	–	5′ splice site variant	Known	Pathogenic	Santos et al 7	–
16	Intron 46	c.6762 + 3A > C	–	5′ splice site variant	Novel	VOUS	–	Damaging by MutationTaster and human splicing finder
Test done in mother of deceased affected individual
1	Exon 19	c.2302C > T	p.Arg768Ter	Nonsense	Known	Pathogenic	HGMD, LOVD	–
2	Exon 58	c.8608C > T	p.Arg2870Ter	Nonsense	Known	Pathogenic	ClinVar, HGMD	–
3	Exon 39	c.5478_5490del ATTGTTGCAAAGA	p.Leu1827GlufsTer17	Small deletion	Known	Pathogenic	LOVD	–
Variants identified in LGMD-associated gene
Test done in proband
Sl. no.	Variant identified				Known/novel variant	Classification	Database/publication with previous report of the variant	In silico prediction for variants classified as VOUS
	Exon/intron	Nucleotide change	Protein change	Mutation
1	Exon 5	Homozygous c.496C > T in SGCG gene (ENST00000218867.3) associated with LGMDR5	p.Arg166Ter	Nonsense	Known	Pathogenic	LOVD	–
2	Exon 3	Homozygous c.197T > A in SGCA gene (ENST00000262018) associated with LGMDR3	p.Leu66His	Missense	Known	VOUS	ClinVar (VOUS)	Probably damaging by PolyPhen-2; damaging by SIFT; CADD score-16
3	Exon 4	Homozygous c.1343C > T in FKRP gene (ENST00000318584) associated with LGMD dystroglycanopathy type C5	p.Pro448Leu	Missense	Known	VOUS	ClinVar (Likely pathogenic);HGMD (known mutation)	Probably damaging by PolyPhen-2; damaging by SIFT and mutation taster; CADD score-31
4	Exon 6	Homozygous c.799C > T in SGCB gene (ENST00000381431) associated with LGMDR4	p.Arg267Cys	Missense	Known	VOUS	ClinVar (VOUS)	Probably damaging by PolyPhen-2. Damaging by SIFT and mutation taster; CADD score-26.2

Abbreviations: ACMG, American College of Medical Genetics and Genomics; AMP, Association for Molecular Pathology; CADD, combined annotation dependent depletion ( https://cadd.gs.washington.edu/ ); HGMD, Human Gene Mutation Database ( http://www.hgmd.cf.ac.uk/ac/index.php ); LGMDR, limb-girdle muscular dystrophy, autosomal recessive; LOVD, Leiden Open Variation Database ( https://www.lovd.nl/3.0/search ); SIFT, Sorting Intolerant From Tolerant ( https://sift.bii.a-star.edu.sg/ ); VOUS, variant of uncertain significance.

Thus, the diagnostic yield of NGS-based DMD gene sequencing in our cohort was 67.9%. Out of the19 DMD gene variants identified (16 in probands and 3 in obligate carrier mothers), 11 were nonsense variants (61.1%), 4 were small deletions/insertions involving a few bases (22.2%), and 4 were splice variants (16.7%) as listed in Table 1 .

Discussion

The MLPA technique detects large/exonic deletions and duplications in the DMD gene but does not detect SNVs and small indels. Therefore, though it is recommended as the first-line test for the molecular diagnosis of male patients presenting with the classic DMD phenotype, it has an overall diagnostic yield of around 65 to 70%. ² In a previous study, published by our group, in a cohort of 510 male patients with the DMD phenotype from our center, the diagnostic yield of MLPA was found to be 72.9%. ³

For the MLPA-negative patients, further analysis is required for confirmation of the diagnosis. Though muscle biopsy with histopathology and immunohistochemistry (IHC) can be done for confirmation, it is an invasive procedure and does not identify the gene mutation. Identification of the exact disease-causing mutation not only helps in molecular confirmation of the diagnosis and, thereby in accurate prognostication and management of the disease, but also helps in offering definitive prenatal testing and carrier screening for other family members. Confirmation of the carrier status of other at-risk female relatives through targeted mutation analysis helps in appropriate genetic counseling, prevention of recurrence in their offspring through prenatal genetic testing, and planning surveillance for cardiomyopathy. ⁸

Childhood-onset autosomal recessive limb girdle muscular dystrophies, which mimic the DMD phenotype constitute a significant proportion of MLPA-negative patients, and IHC for these conditions is available in only a limited number of centers. ^{3 9} Until recently, DMD gene sequencing of the MLPA-negative cases was expensive, tedious and time-consuming, as polymerase chain reaction (PCR)-amplification and Sanger sequencing of all 79 exons had to be done. ⁹ Further, in the cases where no sequence variants were identified in the DMD gene, performing mutation analysis for each of the DMD-mimicking LGMDs, individually through Sanger sequencing, was practically feasible in only a small number of cases. But the availability of the NGS technology has enabled targeted DMD gene sequencing, as well as multigene panel testing for muscular dystrophy-associated genes and focused exome sequencing, to become cost effective and faster. ^{9 10} NGS-based testing helps to identify sequence variants in the DMD gene, as well as in other genes associated with autosomal recessive LGMDs and the other DMD-mimicking phenotypes. In addition to SNVs and small indels, large exonic deletions and duplications can also be detected from the NGS data, using tools, such as ExomeDepth, which are based on comparison of the read depths of the test data with the matched aggregate reference dataset. ^{1 11} However, the deletions/duplications detected thus, have to be validated through MLPA.

We found NGS-based DMD gene sequencing to have a diagnostic yield of 67.9% in our cohort of 28 MLPA-negative cases. In 9 out of the 25 MLPA-negative probands, no sequence variants were identified in the DMD gene, suggesting the possibility of either an alternative diagnosis (i.e., a DMD-mimicking muscular dystrophy; confirmed in four of these nine cases) or of a complex rearrangement or deep intronic mutation in the DMD gene (as deep intronic regions were not covered in the sequencing). One of the limitations of NGS-based sequencing of only the coding regions of the DMD gene is that it cannot detect deep intronic variants and complex rearrangements that constitute around 2% of the mutations in the DMD gene. Though it is possible to include the entire intronic regions of the DMD gene in the sequencing coverage, it is not practically feasible to do so in a targeted DMD gene sequencing platform because the intronic regions of the DMD gene are very large; out of approximately 2.2 Mb length of the gene, only around 14 kb constitute the coding portion. Whole-genome sequencing approach and/or transcriptional analysis of the mRNA can help to identify these variants, if the muscle biopsy and IHC is suggestive of dystrophinopathy. ¹²

The present study has helped to expand the spectrum of sequence variants in the DMD gene in Asian Indian patients with DMD. There are only a few reports of sequence variants in MLPA-negative Asian Indian patients available in published literature. The studies by Singh et al in 2017 and Aravind et al in 2019, each reported 5 DMD gene sequence variants and in the study published by our group, we reported 10 DMD gene variants, some of which are included in this study. ^{2 13 14} A recently published study in 2020 by Kohli et al has reported the sequence variants identified through targeted NGS of the DMD gene, in 68 patients of MLPA-negative dystrophinopathy patients of Asian Indian origin. In this cohort, they identified nonsense variants in 50%, missense variants in 5.9%, small deletions in 27.9%, small insertions in 8.8%, and slice site variants in 7.4% of the patients. ¹⁵ In this study, we identified a greater proportion of nonsense variants (61.1%), followed by small indels (22.2%) and splice site variants (16.7%).

Identification of the exact disease-causing variants also helps in identifying patients amenable to the various mutation-specific therapies. ⁸ Nine probands in our cohort were found to have nonsense mutations and these patients may potentially benefit with stop codon read-through therapy. In the phase-3 trial reported by the Ataluren Confirmatory Trial (ACT) DMD study group, the change in the 6-minute walk distance test (6MWD) between ataluren-treated and placebo-treated patients was not significant, but a significant effect of ataluren was noted in the group of patients with a baseline 6MWD of 300 to 400 m. ¹⁶ Trials are ongoing or planned to test efficacy of read-through therapy in providing long-term benefit in ambulatory patients and in retaining pulmonary function in nonambulatory patients. ¹⁷ Gene editing using CRISPR/Cas9 is another mutation-specific therapeutic modality that has been found to be efficacious in DMD patient-derived induced pluripotent cells and mouse models and clinical trials in DMD patients are likely to be initiated in the near future. ¹⁸

Conclusion

NGS of the DMD gene is an effective noninvasive method for confirming the diagnosis of DMD in MLPA-negative patients. However, as a significant proportion of such cases may be DMD-mimicking LGMDs, NGS-based multigene panel testing for muscular dystrophy-associated genes (which includes the LGMD-associated genes), or CES, would be a more cost-effective testing modality with a better diagnostic yield, rather than targeted DMD gene sequencing.