Neuromuscular Notes: Genetic Testing for Neuromuscular Diseases

The human genome is composed of nuclear and mitochondrial DNA. Its reference sequence is available as part of the Human Genome Project. In 2021, the Telomere-to-Telomere Consortium assembled the first complete sequence of the human genome, which includes the remaining 8% of the genome (~2,000 genes) that had not yet been sequenced in the 2019 reference sequence (ie, Genome Reference Consortium Human Build 38 patch release 13 [GRCh38.p13]).^1-3

Historically, testing for suspicion of a genetic disorder was performed using single gene testing. Sanger sequencing (reading of 1 DNA base pair at a time), multiple ligation probe analysis to detect copy number variants (eg, Charcot-Marie-Tooth 1A), and Southern blot to determine the length of nucleotide repeats (eg, myotonic dystrophies) were the most commonly used technologies. Although these technologies are used frequently for the diagnosis of monogenic diseases, especially those caused by structural genetic variants (eg, duplications, deletions, copy number variations, repeat expansions or contractions), 2 developments have changed the usefulness of this approach. First, as genetic testing has become more ubiquitous, wide genotypic–phenotypic variability has been observed. Numerous clinical presentations have been associated with single genes (phenotypic variability) (see Table 1), and a clinical presentation can be observed in relationship to numerous genes (genotypic variability), as discussed in the following. Second, the development of next-generation sequencing (NGS) has dramatically altered clinical testing possibilities.⁴ NGS involves massively parallel gene sequencing of multiple short (100 to 200 base pairs) DNA fragments at once. This allows for time-efficient and cost-effective testing of multiple genes at once. As such, targeted gene panels using NGS are frequently used as the first diagnostic step in evaluation of neuromuscular disorders with broad genetic heterogeneity (eg, limb-girdle muscular dystrophies), even replacing the traditional need for muscle biopsy in certain circumstances.⁵ Whole exome sequencing and whole genome sequencing also use NGS methodology and are moving out of the research sphere and into clinical practice.⁵

Despite these momentous advances, the identification of a genetic cause in an individual with a highly suspected genetic neuromuscular disorder is still not always possible. Thus, emerging technologies such as long-read sequencing to capture structural variants that are frequently missed in short-read NGS approaches (eg, GGGGCC repeat expansion within intron 1 of C9orf72 in amyotrophic lateral sclerosis), transcriptomics, proteomics, and metabolomics are being used as complementary diagnostic tools.^5,6 This review discusses best practices for the clinical workup of suspected genetic neuromuscular diseases and provides an overview of how emerging technologies may further aid diagnosis in the future.

Nuclear and Mitochondrial DNA

Nuclear DNA (nDNA) contains ~3200 million nucleotides divided into chromosomes (1 pair of sex chromosomes and 22 pairs of somatic chromosomes). Most nDNA is noncoding; coding nDNA (ie, the exome) only constitutes ~1% to 2% of the total nuclear genome (~22,000 genes).⁴ Intronic disease-causing variants, such as the C9orf72 hexanucleotide repeat expansion in amyotrophic lateral sclerosis or the tetranucleotide (CCTG) repeat expansion within CNBP in myotonic dystrophy type 2, are missed with panel testing that only assesses the exome.

Mitochondrial DNA (mtDNA) is characterized by hundreds of copies of a 16.6-kilobase circular DNA molecule. Each copy contains 37 genes (13 protein-coding, 22 transfer, and 2 ribosomal RNA genes) and lacks introns. Heteroplasmy is defined by the presence of both normal and pathogenic mtDNA within a single cell. The proportion of pathogenic to normal mtDNA may vary in different tissues and determines the type and severity of the mitochondrial disease phenotype. High copy number, high sequence variation rate, and paternal leakage of mtDNA (which is mostly maternally inherited) have been proposed as contributors to heteroplasmy.⁷ Heteroplasmy tends to decline in nonpostmitotic tissues (eg, blood) as the individual ages, whereas pathogenic mtDNA copies tend to accumulate in postmitotic tissues over time. As a result, a pathogenic variant in mtDNA may be identified in an individual’s peripheral blood sample within the first 2 decades of life, but not later. For the same reason, mtDNA sequencing from an affected postmitotic tissue (eg, skeletal muscle) has a higher diagnostic yield than blood samples in adults with a suspected primary mitochondrial disorder.⁸

Diagnostic Approach to the Individual With Genetic Neuromuscular Disorder

The number of genetic neuromuscular diseases that benefit from disease-modifying treatments continues to grow. Early diagnosis may reduce morbidity and mortality rates in these individuals and provide an explanation for symptoms, preventing the need for further (sometimes invasive) investigation. Spinal muscular atrophy, transthyretin familial amyloid polyneuropathy, congenital myasthenic syndromes, dystrophinopathies, Pompe disease, and lipid storage myopathies, such as primary carnitine or multiple acyl-CoA dehydrogenase deficiencies, are examples of treatable genetic neuromuscular diseases.⁹

Defining an individual’s phenotype through a detailed history and complete examination is the first step before any testing, as this will help determine the best strategy for genetic testing. For example, a strong clinical suspicion for facioscapulohumeral muscular dystrophy (FSHD) would necessitate more targeted testing rather than broad panel testing. Moreover, a genetic underpinning is not always clinically obvious. Genetic testing would be easily advisable in an individual with early onset of symptoms, slow disease progression, examination findings out of proportion to symptoms, and a positive family history, but these features are not universal (Figure).⁵ For instance, the absence of family history does not rule out an underlying genetic defect. A de novo pathogenic variant, incomplete penetrance, mosaicism (presence of a pathogenic variant in some cell lines but not in all), or small families may all result in a lack of other affected family members.

Information from an individual’s history or specific examination findings may point to a specific genetic neuromuscular disorder; confirmatory genetic testing targeting suspected genes is the best next step in those cases. For example, the Beevor sign (ie, upward umbilicus deviation when an individual initiates an abdominal crunch while supine) prompts consideration of the possibility of FSHD; myotonia that worsens with exercise (“paradoxical myotonia”) points to paramyotonia congenita; and chin fasciculations and a deep furrow on the tongue favor spinobulbar muscular atrophy. The presence of affected relatives with similar or related symptoms (ie, phenotypic variability) (Table 1), certain ancestry, or history of consanguinity (see Table 2 which is available at www.practicalneurology.com) can also narrow down the genetic differential diagnosis. For example, a family history of early cataracts and sudden death in an individual with distal muscle weakness may be an example of intrafamilial phenotypic variability in myotonic dystrophy type 1. Likewise, the presence of French Canadian ancestry may favor an oculopharyngeal muscular dystrophy over a mitochondrial myopathy in an individual with ocular muscle weakness.

On the other hand, a specific and well-defined phenotype may be associated with a heterogenous genetic basis. For example, an Emery-Dreifuss muscular dystrophy phenotype can be caused by pathogenic variants in multiple genes (ie, LMNA, EMD, FHL1, SYNE1, SYNE2, TMEM43). A gene panel using NGS is the best next step for diagnostic confirmation in such cases.¹⁰

Even for the most experienced and astute clinicians who have witnessed the increasing complexity of phenotypic and genotypic variability in neuromuscular disorders, “deep phenotyping” and extensive genetic testing often are not enough to identify a well-established genotype–phenotype correlation in an individual. In such instances, it is important for clinicians to be open-minded and consider the possibility of an acquired disorder, a new phenotype that has not yet been associated with a genotype that individual carries (eg, exercise intolerance as the presenting symptom of a muscular dystrophy), or a new genetic defect for a well-known phenotype (eg, the identification of an intronic single nucleotide variant and a pentanucleotide expansion [compound heterozygosis] in RCF1 in an individual with cerebellar ataxia, neuropathy, vestibular areflexia syndrome (CANVAS) rather than the typical biallelic expansions associated with this disorder).¹¹ Despite a clinician’s best efforts, a definitive diagnosis cannot always be reached.

Single Gene vs Multigene Testing Approach

The decision to perform single gene or panel testing is not solely based on the information gathered from patient history, examination, and ancillary tests. The health system, insurance company, clinician administrative support, and patient income are key factors to consider when selecting a genetic test. The Figure proposes a diagnostic algorithm that can serve as a guide for the selection of genetic tests in individuals with neuromuscular disorders.

The ordering provider should be familiar with the limitations of the assay used for the genetic testing in question. False positive (FP) or false negative (FN) results may occur, which could lead to misdiagnosis. For example, the complex genetic basis of FSHD makes this muscular dystrophy susceptible to both FP and FN findings. Length of D4Z4 repeat contraction on chromosome arm 4q, haplotype (permissive A vs non-permissive B alleles), and methylation assays should be part of FSHD genetic testing. Identification of a D4Z4 repeat contraction may lead to an FSHD1 diagnosis; however, if this contraction resides in a B allele (instead of an A-permissive allele), it is not disease-causative. Conversely, mosaicism or recognition of a similar region on chromosome 10 by the Southern blot probe may miss a D4Z4 repeat contraction and the FSHD1 diagnosis. D4Z4 hypomethylation in the absence of a contraction in an A allele should prompt testing for FSHD2, which is clinically indistinguishable from FSHD1 and caused by other defective genes, such as SMCHD1 on chromosome arm 18p.¹² Table 3 lists neuromuscular disorders associated with abnormal repeat numbers.

Regarding NGS methodology, sequence alignment and variant filtering are 2 steps susceptible to generating FP and FN results. An incorrect alignment of a short-read sequence to the reference genome sequence may yield an erroneous genetic variant (FP), whereas clinically relevant variants may be filtered out (FN). Furthermore, interpretation of each variant may differ among laboratories, leading to occasional discrepant results in their prediction of pathogenicity. Type of amino acid change, conservation of the involved sequence across species, and the absence or presence of the sequence in control databases are often used to classify variants as pathogenic or benign. Table 4 summarizes advantages and limitations of NGS approaches.⁴

Variants of Uncertain Significance

Not all identified genetic variants can be classified as pathogenic or benign, leading to the designation of a variant of uncertain significance (VUS).¹³ The likelihood of finding a VUS increases with longer genes and when a greater number of genes are tested. The largest human gene is DYS (dystrophin), and large coding regions are contained in TTN (titin; the largest human protein), NEB (nebulin), and RYR1 (ryanodine receptor), all of which are associated with neuromuscular disorders when defective.¹⁴

Detection of a VUS requires additional work on the part of the clinician. If a VUS is found in a gene that has not been associated with a neuromuscular disorder, it will be unlikely for that VUS to contribute to the individual’s neuromuscular phenotype. Conversely, if the VUS is identified within a gene that has been associated with a neuromuscular disorder, the clinician should further investigate that VUS because it may be associated with individual’s phenotype. Table 5 summarizes the main steps to investigate the clinical relevance of VUS in genes that have been previously associated with a neuromuscular disorder in individuals with a neuromuscular phenotype.

Novel and likely pathogenic VUS should be reported to https://www.ncbi.nlm.nih.gov/clinvar to help other clinicians in the diagnosis of other individuals with the same variant and phenotype and to support genotype–phenotype associations. A VUS eventually may be reclassified as pathogenic or benign based on additional data, so it is recommended to periodically check whether additional information for a VUS has been reported.

Transcriptomics, Proteomics, and Metabolomics⁶

In addition to bioinformatic tools to predict an abnormal transcriptome, unbiased RNA sequencing is emerging as a complementary tool to increase the diagnostic yield of genetic neuromuscular diseases. DNA variants that were originally overlooked or not detected may be identified as disease causative if associated with spliceopathy (ie, a variant causing an abnormal quantity of normally spliced or abnormally spliced forms) in affected tissues, rather than peripheral blood samples. For example, a synonymous genetic variant that was filtered out on whole exome sequencing may be reclassified as pathogenic if RNA sequencing shows an abnormal splicing pattern involving that gene transcript.

Untargeted proteomic assays in skeletal muscle may reveal higher or lower than expected levels of certain proteins that in turn will prompt revision of the variants within the genes from which those proteins were derived, or single gene testing to detect variants that might have been missed on NGS. Although accumulation or deficiency of metabolites is often used as a disease biomarker, their diagnostic role in neuromuscular (mainly metabolic) diseases is an area of investigation. One caveat is that factors other than genetics, such as aging, nutrition, medications, and comorbidities, may contribute to an individual’s metabolomic profile. Furthermore, new disease pathogenic mechanisms (ie, abnormal splicing, gain or loss of protein function, deficiency or accumulation of a metabolite) may be uncovered using these approaches.^5,6

Genetic Counseling

Counseling should include discussion about the pros and cons of genetic testing, potential implications for family members and family planning, and the possibility of results being nondiagnostic or revealing clinically relevant secondary findings unrelated to current symptoms, if applicable.^15,16 Clinicians should ensure that individuals have the necessary information before pursuing genetic testing, that informed consent is obtained, and that individual preferences are respected.

Genetic testing in asymptomatic family members needs to be considered in the event that a positive result will influence management (eg, monitoring for cardiomyopathy in an asymptomatic female dystrophinopathy carrier).¹⁷ Whereas the Genetic Information Act of 2008 prohibits discrimination in employment and health insurance based on genetic information (https://www.eeoc.gov/genetic-information-discrimination), it does not protect from discrimination in life or disability insurance, which should be discussed before pursuing genetic testing.

Conclusion

As the list of treatable genetic diseases and clinical trials targeting genetic defects continues to grow, the importance of timely, accurate genetic diagnosis increases. Careful clinical phenotyping and an awareness of genetic testing costs and limitations are key to designing a high-yield and feasible diagnostic approach for each individual. Emerging technologies investigating the transcriptome and proteomics may soon offer additional options in the clinician’s toolkit.