The age of single-gene neurological disorders is not dead

This edition of Brain sees the first of a series of articles describing new genes responsible for inherited neurological disease, thus expanding the clinical neurogenetics portfolio. Sebahattin Cirak from the Institute of Child Health in London led a transcontinental consortium of clinical and basic science researchers who map the locus for a gene responsible for an autosomal dominant distal myopathy (Cirak et al., 2010). Their work casts new light on the disease mechanisms of these poorly understood disorders; but at the same time, they broaden concepts on the potential mechanisms involved. Distal myopathies provide a classical example of the way that molecular genetics has revolutionized classifications in clinical neurology (Table 1). First described in the early part of the last century (Gowers, 1902), they were the focus of that old neurological chestnut ‘is the distal weakness due to muscle or nerve dysfunction?’ and were latterly subdivided into discrete entities based on astute clinical examination (reviewed in Barohn et al., 1998). Early gene mapping studies identified several loci, highlighted the likely aetiological heterogeneity and led to the further delineation of different clinical groups (Udd, 2009). A major subgroup was associated with pathological myofibrils (myfibrillar myopathies) (Udd, 2009), overlapping with inherited limb girdle syndromes. Positional cloning and candidate gene analysis led to the rounding-up of some obvious culprits [MYO7, the myosin heavy chain 7 gene (Meredith et al., 2004) and fast IIa myosin as recently published in Brain (Tajsharghi et al., 2010)] and provided a satisfying explanation for the muscle group specificity of some distal myopathies [including the ophthalmoplegia in inclusion body myopathy-3, due to mutations in the myosin heavy chain IIa gene (Martinsson et al., 2000)], but the real power of these ‘reverse genetic studies’ was in revealing new proteins required for intact muscle structure and function, including the largest protein known to man (Titin, TTN) (Lange et al., 2005). Rather than simply disrupting the muscle cell structure or the contractile apparatus, some of these disease genes were ubiquitously expressed and affected apparently essential general cellular functions. Why these genes should cause a pure skeletal myopathy and why the mutations should preferentially affect specific groups of distal muscles remain enigmatic. The advent of muscle magnetic resonance imaging has elegantly confirmed these restricted distributions at a higher resolution, with adjacent embryologically and physiologically related muscles being affected at different stages in different distal myopathies (Birchall et al., 2005; Fischer et al., 2005). Distal myopathiesa a Adapted from Griggs et al., 2007. Inh = inheritance pattern; AD = autosomal dominant; AR = autosomal recessive. Distal myopathiesa a Adapted from Griggs et al., 2007. Inh = inheritance pattern; AD = autosomal dominant; AR = autosomal recessive. The paper by Cirak et al. (2010) is an important contribution to this growing body of knowledge. By studying four generations of a large German family, they identified mutations in KLHL9, a gene encoding a new disease protein called kelch-like homolog-9. A heterozygous point mutation segregated with the phenotype in this family, causing a myopathy presenting between 8 and 16 years of age and leading to bilateral foot drop due to symmetric ankle dorsiflexion weakness. Despite the early age of onset (for a distal myopathy, see Table 1), the disorder remains mild and progresses very slowly, with hand involvement in the later stages and a curious sensory disturbance in all four extremities despite normal nerve conduction studies. Careful examination showed that ankle dorsiflexion and inversion are more severely affected than plantar flexion and eversion, and further evidence of muscle selectivity is noted on magnetic resonance imaging where the pattern of subclinical involvement is almost the direct opposite of that seen in some other distal myopathies (Birchall et al., 2005; Fischer et al., 2008), despite similar features on clinical examination. Ankle contractures were noted in some subjects, which is an unusual feature of distal myopathies, and despite the insidious clinical course, the serum creatine kinase is ∼1400 IU/l in some subjects consistent with the destructive muscle pathology and replacement of muscle fibres with fat and fibrous connective tissue. For most distal myopathies, the creatine kinase remains normal [although there are some notable exceptions (Bashir et al., 1998)], and the pathological features are often subtle and include the presence of blue rimmed vacuoles and eosinophilic inclusions that were notably absent from the muscle biopsies taken from patients in the German kindred. By studying 10 affected subjects and 3 unaffected subjects, Cirak et al. (2010) have sufficient statistical power to map the chromosomal region segregating with the phenotype [typically each meiosis/individual in a dominant disorder contributes on average ∼0.3 to the logarithm of the odds score, which is the standard statistical test of genetic linkage]. By defining the markers flanking the block of DNA shared by all affected individuals (fine mapping), they narrowed the interval down to 22.59 cM (approximately 23 million base pairs of DNA), which contains 81 genes. Using bioinformatic approaches and some ‘common sense’ they listed the genes in order of priority, based on their likely involvement in muscle disease, and KLHL9 was quickly identified. What does not come out in a scientific report is the component of luck. Unlike familial epilepsy, where most of the monogenic causes involve ion channel genes, ranking candidate genes is not always a precise science; and Cirak et al. (2010) had good fortune on their side. KLHL9 was only the fifth gene they sequenced. Having defined a genetic variant present in every affected individual in the family, they showed that the normal KLHL9 gene product is preferentially expressed in skeletal muscle at high levels, modelled the likely consequences of the mutation on protein structure and went on to perform functional studies in vitro. The mutation present in the German family was predicted to alter the 95th amino acid, changing it from a leucine to a phenylalanine (p.L95F). This residue is close to the Cul3 binding site, which is essential for the E3 ubiquitin ligase function of the protein. By expressing both the mutant and wild-type protein in a cell line and analysing the partners interacting with the proteins, they showed that the p.L95F mutation inhibits association with Cul3 without affecting the total amount of the protein. This supports the idea that the pathogenic mutation in this family affects E3 ubiquitin ligase activity in skeletal muscle, impairing the ubiquitin ‘tagging’ of abnormally folded proteins and thus preventing their incorporation into the proteasome, leading to accumulation within skeletal muscle fibres. These studies provide an explanation for the muscle-specific phenotype of this disorder and show that the segregating genetic variant has a functional effect—but they do not prove the underlying mechanism, and as the authors point out in their discussion, there are other possibilities, such as the misregulation of the Aurora B kinase. The work of Cirak et al. (2010) illustrates how much more straightforward it is now to find new disease genes in families with a distinct phenotype, provided there are enough clearly phenotyped subjects. Although they used microsatellite markers, many investigators now employ single-nucleotide polymorphisms that are technically easy to analyse at great depth across the whole genome using commercially available arrays. Indeed, many of us now outsource these analyses to one of several companies worldwide, who offer the service at ∼£300 per sample. Finding the causal variants within an often large chromosomal interval does still pose a challenge. Bioinformatics can help, but all too often there is a long ‘gene hunt’ that may take many months, if not years. However, both approaches can now be short-circuited by whole-exome capture next generation sequencing, where every segment of coding DNA is directly sequenced many times from the same subject. By comparing the sequence variants to perhaps only one other person in the same family with the disease, it is possible to identify the causal variant directly, provided that the substitution is not a large-scale rearrangement nor is within large intronic gene regulatory regions (Choi et al., 2009; Ng et al., 2009, 2010). Comparing these data to the emerging 1000 genomes project data (http://genome.wellcome.ac.uk/doc_WTX047611.html) will dramatically speed up the process of mutation identification. Cirak et al. (2010) take an important step by showing that the p.L95F substitution alters protein–protein interactions, but it is still not clear how this causes a muscle disease that targets some muscles and not others. Despite being first raised over 50 years ago, the fundamental issue of highly specific muscle selectivity in different inherited muscle diseases remains largely unanswered. However, Cirak et al. (2010) demonstrate the continued importance of carrying out detailed studies of families with rare presumed monogenic disorders. The reverse genetic approach, using the phenotype to lead to the gene, remains hugely powerful. Curiously, in the field of distal myopathies, there does not, as yet, appear to be a single unifying theme. Rare families are telling us something new that is of general importance, and here it is the role of the kelch-like homolog-9 protein. The age of single-gene neurological disorders is not dead, and families like these continue to help us dissect out cellular processes that are crucial in heath and disease. There has been a trend for birth notices to be given of new discoveries in genetics followed by a detailed report of the phenotype and putative mechanisms. The last 12 months has seen a different style of publishing, at least in Brain, with several new disease genes being reported in full. This, in my view, leads to a much more satisfying and well-rounded paper, where equal attention is given to the molecular methods, gene mapping and identification and the clinical phenotype. After all, the ultimate aim is to disseminate this information to advance clinical neurology for the good of our patients, and by publishing the ‘full story’ in one manuscript, Cirak et al. have done just that. Theirs is a model we encourage others to follow. Medical Research Council (UK); UK Parkinson’s Disease Society; Association Française contre les Myopathies; UK NIHR Biomedical Research Centre for Ageing and Age-related disease award (to the Newcastle upon Tyne Foundation Hospitals NHS Trust). P.F.C. is a Wellcome Trust Senior Fellow in Clinical Science and a UK NIHR Senior Investigator.