The evolving spectrum of complex inherited neuropathies

Purpose of review Inherited peripheral neuropathies can be divided into those diseases in which peripheral neuropathy is the sole or main feature of the disease (Charcot-Marie-Tooth disease) and those in which peripheral neuropathy is just one feature of a more complex syndrome. In recent years there has been a substantial expansion in the number of genes associated with complex neuropathy syndromes. Recent findings This review will focus on emerging themes in this group of diseases, namely the increasing number of diseases due to repeat expansions; the emergence of both recessive and dominant negative alleles in the same gene producing a common phenotype and diseases in which there is selective loss of the allele from haematopoietic stem cells making genetic diagnosis on blood derived DNA problematic. Summary In this review we provide a practical approach to investigating and diagnosing patients with peripheral neuropathy as part of a complex syndrome and provide an updated table of the genes associated with this group of diseases.


INTRODUCTION
Healthcare professionals involved in the diagnosis of individuals with inherited peripheral neuropathy have traditionally tested genes associated with diseases in which peripheral neuropathy is the sole or major component of the phenotype.Testing has therefore focused on those genes that have been described in association with Charcot-Marie-Tooth disease (CMT) and the closely related diseases, hereditary sensory neuropathy and distal hereditary motor neuropathy.With the widespread adoption of next generation sequencing into routine clinical practice, it has now become apparent that many complex genetic diseases associated with peripheral neuropathy such as the Hereditary Spastic Paraplegias (HSP) and Spinocerebellar Ataxias can present with peripheral neuropathy.This is relevant as many of the genes responsible for such diseases are absent from inherited neuropathy gene panels.In 2017, we published a clinical approach to these complex neuropathy syndromes [1].In this review, we revisit our approach and include an up-to-date table of more than 250 genes associated with complex neuropathy syndromes.

APPROACH TO GENETIC TESTING IN COMPLEX NEUROPATHY SYNDROMES
There are more than 250 genes associated with complex neuropathy phenotypes, many of which are exceptionally rare and reported in only a handful of families.It is therefore unrealistic to expect to have a working knowledge of the phenotypes of all the causative genes.We propose an approach in which the clinician defines the major feature of the nonneuropathy phenotype.In the majority of patients this will be a neurological phenotype of either ataxia, spasticity or global developmental delay (Table 1) although rarer neurological features including optic atrophy, ophthalmoplegia, deafness, myopathy and movement disorders may also be associated with peripheral neuropathy (Table 2).Important nonneurological phenotypes are diverse and include cardiomyopathy, dermatological disease and liver and renal failure (Table 2).Once the predominant phenotype has been defined the neuropathy should then be characterised based on the nerve conduction velocities (NCV) into those with normal or slow conduction.Neuropathies with normal nerve conduction velocities should then be further classified into those which are predominantly sensory, those predominantly motor and those that are mixed (see Fig. 1).Using this approach, one is able to sub classify the many genes associated with complex phenotypes into a workable format (see Tables 1 and 2).

RECENT ADVANCES
Since our last review in 2017, there has been a substantial increase in the number of genes associated with complex neuropathy syndromes.This includes an increased number of repeat expansion diseases, the discovery of dominant negative variants for known recessive complex neuropathy syndromes, a new treatable complex syndrome and the discovery of genes such as SMDA9l, which causes an ataxic neuropathy syndrome associated with pancytopenia in which the pathogenic variant is selected against in haematopoietic progenitor cells potentially leading to false negative testing [2].

NOTCH2NLC
Neuronal intranuclear inclusion disease (NIID) is a multisystem neurodegenerative disease that may present with dementia, peripheral neuropathy, autonomic dysfunction, cerebellar ataxia, parkinsonism, seizures or stroke like episodes and fluctuating encephalopathy [3,4].The disease was described in 1968 and was defined by the presence of eosinophilic intranuclear inclusions on brain biopsy at post mortem that can also be seen in other organs including skin biopsy [5].In addition to a histopathological signature, brain MRI in NIID commonly reveals leukoencephalopathy on fluid-attenuated inversion recovery (FLAIR) sequences and symmetrical high signal on diffusion weighted imaging (DWI) in corticomedullary junctions [6,7].Whilst sporadic cases of NIID tend to present with dementia, familial cases cluster into two main phenotypes, a dementia dominant form and a limb weakness dominant form [8].
In 2019, three independent groups identified heterozygous CGG repeat expansions in cohorts of sporadic and familial cases of NIID from Japan and China in the 5'UTR of NOTCH2NLC [4,7,8].These original reports included familial cases presenting with motor and sensory neuropathy and autonomic disturbance.Since the discovery of the genetic cause of NIID, there have been several reports of similar expansions in NOTCH2NLC in Chinese and Japanese peripheral neuropathy cohorts [9,10,11 && ].These include one Taiwanese study in which NOTCH2NLC repeat expansions accounted for 10.6% of unassigned CMT2 cases.All patients developed a sensory predominant neuropathy with an average age of onset of 37.1 (range 21-55) [9].A separate study by Wang et al. in an unsolved peripheral neuropathy cohort, found that repeat expansions in NOTCH2NLC accounted for 3.52% of Chinese inherited neuropathy patients [12].
The largest study reporting on the prevalence and phenotype of NOTCH2NLC related peripheral neuropathy comes from a large case series of inherited peripheral neuropathy in Japan [11 && ].The cohort comprised 2692 Japanese patients with a diagnosis of CMT, 1783 of whom were without a genetic diagnosis.This cohort was screened for alternative causes with whole exome sequencing and RFC1 repeat analysis prior to undergoing NOTCH2NLC analysis by repeat prime PCR.Using this approach, NOTCH2NLC repeat expansions were identified in 26 cases from 22 unrelated families representing 1.2% of genetically undiagnosed Japanese patients with CMT and the seventh most common causative gene in their series excluding CMT1A.The phenotype was characterised by distal wasting and weakness (>90%) with sensory disturbance in 50% of patients.Autonomic dysfunction was common (44%) and characterised by neurogenic bladder, constipation, orthostatic hypotension and erectile dysfunction.Tremor was present in 29% of patients and myoclonus in 4%.Cognitive

KEY POINTS
Complex peripheral neuropathy syndromes such as NOTCHNLC2 NIID can present with peripheral neuropathy.Some autosomal recessive complex neuropathy syndromes may also be caused by heterozygous dominant negative variants.SMAD9L variants that cause neuropathy and pancytopenia may be negatively selected against in blood leading to reduced diagnostic sensitivity in leukocyte derived DNA.Whilst it is clear that repeat expansions in NOTCHNLC are an important cause of inherited peripheral neuropathy in far East populations, similar studies have not been reported in Northern European, North American and African populations.In Caucasian movement disorder patients in the UK, repeat expansions in NOTCH2NLC are rare suggesting a possible founder effect in East Asian populations [13].
Cerebellar ataxia, neuropathy, and vestibular areflexia syndrome and RFC1 Cerebellar ataxia, neuropathy, and vestibular areflexia syndrome (CANVAS) is one of the commonest causes of late onset inherited peripheral neuropathy.It is an autosomal neurodegenerative disease characterised by adult onset sensory neuropathy.Until recently, CANVAS syndrome was thought to be solely caused by biallelic repeat expansions (AAGGG) in the second intron of RFC1, which encodes for the large subunit of replication factor C, a 5 subunit DNA polymerase accessory protein [14].As widespread screening for the repeat expansion has been introduced into clinical practice, a small proportion of individuals with typical CANVAS have been identified that do not carry two (biallelic) repeat expansions.
In two unrelated studies, compound heterozygous variants in RFC1 comprising one repeat expansion in the second intron of RFC1 on one allele and a     ].Although the total number of cases in these two studies was small (n < 10), individuals with truncating variants seemed to display a more severe phenotype with two individuals requiring a wheelchair 50 years of age and one requiring gastrostomy feeding to severe dysphagia.In one patient the onset of the disease was in the 3 rd decade and characterised by spasticity [16 && ].

FGF14
Autosomal dominant missense variants in FGF14 were first described as a cause of SCA27, a rare, early onset ataxic syndrome accompanied by tremor [17].More recently, a GAA repeat expansion in intron 1 of the same gene, FGF14 was described as a common cause of autosomal dominant late onset cerebellar ataxia in a cohort of 128 patients [18].The mean age of onset was in the 6th decade and the phenotype was characterised by gait ataxia, down beat and gaze evoked nystagmus, diplopia, cerebellar dysarthria and cerebellar atrophy on MRI.Peripheral neuropathy was rarely seen in this cohort.
In a study of 45 patients with a combination of cerebellar ataxia and either peripheral neuropathy or bilateral vestibular failure or both (and negative for the RFC1 repeat expansion), screening for the FGF14 GAA repeat expansion was positive in 43% of patients with ataxia and vestibular failure, 38% with ataxia and neuropathy and 27% with ataxia, neuropathy and vestibular areflexia [19 & ].This has led some to claim that heterozygous FGF14 repeat expansions are a phenocopy of CANVAS syndrome.Whilst this may be technically true, it is important to recognise that in this study the neuropathy was extremely mild and was a mixed motor and sensory neuropathy whereas in CANVAS syndrome due to RFC1 biallelic repeat expansions, the neuropathy is an early and prominent feature and usually purely sensory.From the available data, it does not appear that repeat expansions in FGF14 present with peripheral neuropathy unlike CANVAS syndrome.

Dominant negative variants in recessive complex diseases
Interpreting the pathogenicity of novel variants is dependent on the pattern of inheritance.For example, single heterozygous variants are usually not relevant to a recessive disease.The exception to this is the increasing discovery of heterozygous, often de novo dominant variants in genes such as SLC12A6, UCHL1 and NARS1 traditionally associated with autosomal recessive inheritance, which are thought to act through a dominant negative mechanism.

SLC12A6
Potassium chloride co-transporters (KCC) are encoded by the SLC12A family of solute carriers, of which SLC12A6 encodes for the co-transporter protein KCC3 [20].Recessive variants in SLC12A6 cause Aldermann's syndrome, a severe early onset motor predominant peripheral neuropathy associated with variable degrees of agenesis of the corpus callosum and developmental delay [21].In 2016, a child with a severe and progressive peripheral neuropathy was reported in association with a de novo heterozygous variant in SLC12A6 [22].In the 4 years, there have been several series and case reports describing patients with heterozygous variants in SLC12A6 and a broad phenotype ranging from severe cases with a childhood onset motor predominant axonal neuropathy and developmental delay similar to Alderman's syndrome through a late onset (5th decade) predominant peripheral neuropathy with no central nervous system involvement [23][24][25][26].
Ubiquitin C-terminal hydrolase L1 Autosomal recessive missense variants in Ubiquitin C-terminal hydrolase L1 (UCHL1) are a recognised cause of autosomal recessive spastic paraplegia type 79 (SPG79) which is characterised by early onset cerebellar ataxia, spastic paraplegia and optic atrophy [27][28][29].In two independent case-control gene burden analyses performed in German and UK cohorts (the latter within the 100 000 genomes project), loss of function heterozygous variants in UCHL1 were found to be enriched in patients with hereditary ataxia and spastic paraplegia [30].In the German cohort this was in addition to rare variants in SPAST, KIF5A but also POLR3A, the latter also having only previously been associated with autosomal recessive disease.The core clinical phenotype included ataxia (present in 83%), spasticity in 77% and sensory or sensory motor axonal neuropathy in 52% although only 61% underwent neurophysiology testing.Global developmental delay was a less common feature in dominant UCHL1 disease and the age of onset ranged from childhood to 70 years of age.
Asparaginyl-tRNA synthetase Asparaginyl-tRNA synthetase (NARS1) is a member of the amino acyl tRNA synthetases, a group of enzymes that attach amino acids to their respective tRNA.Recessive loss of function variants have been reported in several AARS genes as a cause of complex hereditary spastic paraplegia including AARS, GARS1 and DARS2 [31].This is in contrast to toxic gain of function heterozygous variants that cause a peripheral neuropathy and fall within the CMT spectrum [32].Recently, both recessive and de novo dominant variants in NARS1 have been described to cause a complex phenotype characterised by severe global (and motor) developmental delay, microcephaly, epilepsy, ataxia and a peripheral neuropathy with slow conduction in 25% [33].Both recessive and denovo variants showed a similarly severe phenotype and were associated with loss of amino acetylation activity.
More recently, two variants in two families were reported in association with a pure axonal neuropathy, however a mouse model showed no phenotype in the heterozygous state and it remains uncertain as to whether dominant NARS1 variants can cause a pure neuropathy [34].A diagnostic approach for patients with complex inherited neuropathy syndromes.

Peripheral nerve and neuro-muscular junction disease
Treatable complex diseases SLC5A6 encodes a nonselective transporter for biotin (B7), pantothenic acid (B5) and lipoic acid.Recessive variants in the gene were originally described in 2017 [35] in individuals presenting in the first few years of life with failure to thrive, developmental delay or regression, seizures, diarrhoea and vomiting, immunodeficiency and osteopenia.As clinical testing has become more widespread, recessive variants have been reported in families with mixed sensory and motor axonal neuropathy with variable conduction velocity slowing associated with optic atrophy and recurrent episodes of pseudo obstruction [36].In one study, recessive variants in SLC5A6 were described in three families with onset in the second decade characterised by a motor predominant neuropathy and with subjective improvement or stabilisation with biotin, pantothenic and lipoic acid supplementation [37].
Ataxia pancytopenia syndrome and loss of pathogenic germ line variants in SAMD9L from leukocyte derived DNA Ataxia-pancytopenia syndrome (ATXPC) due to heterozygous variants in SAMD9L was first reported in 2016 [38].The initial clinical descriptions were of a syndrome of cerebellar ataxia and atrophy with mild pyramidal signs and white matter disease in addition to haematological abnormalities characterised by cytopenia, immunodeficiency and myelodysplasia.More recently a length dependent peripheral neuropathy with slow conduction velocity and pes-cavus has been described and in some families this can be the initial presenting feature [2,39].
Clinical and research genetic testing in the vast majority of cases of peripheral neuropathy is conducted on DNA extracted from white blood cells.This can be problematic for detecting pathogenic SAMD9L variants in affected individuals for the reasons I will outline below.
SAMD9L is a tumour suppressor gene that inhibits unregulated proliferation of haematological progenitor cells in the bone marrow.Heterozygous pathogenic variants in SAMD9L causing ATXPC exhibit a toxic gain of this function resulting in reduced cell proliferation and cytopenia.This leads to a genetic pressure whereby spontaneous somatic variants including deletions in SAMD9L that disrupt or render the inherited pathogenic SAMD9L variant null are preferentially selected for that may result in a low level of the mutant allele in blood below the limit of variant calling for many algorithms.This can be overcome by extracting DNA from nonhaematological cells or tissues such as skin.A patient with cytopenia and a peripheral neuropathy, especially with slow NCV in whom gene panel testing on blood has been negative may therefore require further testing on nonhaematological derived DNA.

CONCLUSION
As more patients are screened for an increasing number of genes, the phenotypes and presentations associated with each gene continue to expand.In the case of inherited peripheral neuropathies, it has become apparent that repeat expansions are emerging as a major cause of complex neuropathy syndromes often in association with cerebellar ataxia.As many of the 250 genes associated with a complex neuropathy syndrome can present with peripheral neuropathy, routine clinical genetic testing in CMT may need to expand to include many of these genes

FIGURE 1 .
FIGURE 1.A diagnostic approach for patients with complex inherited neuropathy syndromes.

Table 1 .
A summary of the complex inherited neuropathy syndromes with one of the three major core clinical phenotypes of ataxia, spasticity or global neurodevelopmental impairment AD PRKCGUsually adult onset isolated cerebellar ataxia.Missense mutation in catalytic domain of exon 11 associated with complex syndrome including cerebellar ataxia, sensory motor axonal neuropathy, parkinsonism, dystonia, myoclonus and pyramidal syndrome.

Table 1 (
Continued)The evolving spectrum of complex inherited neuropathies Rossor et al.

Table 2 .
A summary of the complex inherited neuropathy syndromes with one of the minor 10 clinical phenotypes associated with neuropathy

Table 2 (
Continued)The evolving spectrum of complex inherited neuropathies Rossor et al.

Table 2 (
Continued)The evolving spectrum of complex inherited neuropathies Rossor et al.

Table 2 (
Continued)Disorder of mitochondrial beta oxidation of fatty acids.Severe neonatal, infantile and late adolescent onset described, the latter characterised by a progressive myopathy with recurrent rhabdomyloysis and a sensory-motor axonal neuropathy.Abnormal urine organic acids.The evolving spectrum of complex inherited neuropathies Rossor et al.