Introduction

Inherited peripheral neuropathy (IPN) constitutes a heterogeneous spectrum of hereditary neurological disorders primarily marked by dysfunction affecting the peripheral nerves. Clinically, IPN commonly presents with progressive muscle weakness and atrophy, often accompanied by sensory manifestations such as numbness, tingling sensations and pain. IPN mainly encompasses Charcot-Marie-Tooth disease (CMT), hereditary motor neuropathy (HMN), and hereditary sensory neuropathy. To date, more than 140 genes have been found associated with IPN.1 Nevertheless, more than 60% of IPN patients remain genetically undiagnosed.1–3

In recent findings, tandem repeat expansions within non-coding regions of genes have emerged as contributory factors to a spectrum of neurological disorders. Repeat expansions in the C9orf72 and ATXN2 genes have been identified as causative factors in amyotrophic lateral sclerosis (ALS).4 5 Since 2019, repeat expansions within multiple genes, including LRP12, GIPC1, NOTCH2NLC and RILPL1, have been linked to oculopharyngodistal myopathy (OPDM), designated as OPDM1–4, respectively.6–11 OPDM is characterised by distal muscle weakness, often accompanied by ptosis, ophthalmoplegia, facial muscle weakness and bulbar palsy.12 On the other hand, repeat expansions within genes such as RFC1 and NOTCH2NLC have been found to be associated with IPN,13–15 highlighting the significance of non-coding repeat diseases in the aetiological landscape of IPN. Given the heterogeneous clinical manifestations observed in NOTCH2NLC-related disorders, encompassing OPDM and IPN, our hypothesis suggests the potential contribution of other OPDM-related genes in the pathogenesis of an IPN phenotype.

Hence, in this study, we enrolled 1555 Japanese individuals clinically diagnosed with IPN but lacking established genetic aetiology, and conducted a comprehensive genetic screening focusing on CGG repeat expansions in the LRP12, GIPC1 and RILPL1 genes. Furthermore, we systematically examined and characterised the clinical, electrophysiological, radiological and pathological profiles of patients carrying these disease-associated repeat expansions.

Subjects/materials and methodsSample selection

We included 2424 unrelated Japanese patients clinically diagnosed with CMT or HMN between 2007 and 2021 (figure 1A). Patients diagnosed with hereditary sensory (and autonomic) neuropathy (HSN and HSAN) were specifically excluded from this study. All patients were examined and diagnosed by at least two experienced neurologists or paediatric neurologists, and their clinical data along with blood samples were transferred to our laboratory for genetic testing. All patients with demyelinating CMT were confirmed to be negative for PMP22 duplication/deletion using fluorescence in situ hybridisation or multiplex ligation probe amplification. To assess the normal range of repeat numbers in the LRP12 gene, we collected 208 cases with genetically confirmed spinocerebellar ataxia 6 or 31.

Figure 1

Flow chart and genetic findings in our Japanese case series of IPN. (A) Genetic analysis workflow conducted in 2424 patients with CMT/HMN (IPN case series). Following a series of prescreening studies, 1555 genetically unidentified cases were processed to CGG repeat screening in the genes of LRP12, GIPC1 and RILPL1 and repeat expansions were only detected in LRP12 from 44 cases. (B) LRP12 repeat expansion is the fourth most common causative reason in the present IPN case series in Japan, following MFN2, GJB1 and MPZ. (C) Representative results of RP-PCR, fluorescence AL-PCR and long read sequencing by GridION. AL-PCR, amplicon-length PCR; CMT, Charcot-Marie-Tooth disease; HMN, hereditary motor neuropathy; IPN, inherited peripheral neuropathy; NGS, next-generation sequencing; RP-PCR, repeat-primed PCR.

Prescreening analyses of patients with IPNs

We performed genetic screening of 2424 patients using DNA microarray (Affymetrix, Santa Clara, California, USA), next-generation sequencing on Illumina MiSeq (Illumina, San Diego, California, USA), or Ion Proton (Thermo Fisher Scientific, Waltham, Massachusetts, USA) by targeting disease-causing or disease-candidate gene panels of IPN, as previously described (online supplemental figure 1).16 Additionally, whole-exome sequencing was performed on 758 patients using the SureSelect v4+UTRs or v5+UTRs Kit (Agilent Technologies, California, USA) and sequenced using the Hiseq2000/Hiseq2500 platform (Illumina Inc., San Diego, CA, USA) or Ion Proton. Subsequently, repeat-primed polymerase chain reaction (RP-PCR) and amplicon-length PCR (AL-PCR) were utilized to identify repeat expansions in the RFC1 and NOTCH2NLC genes.15 17 Based on the aforementioned screening, 1555 patients with unidentified IPN were selected for further repeat analyses (figure 1A).

RP-PCR and fluorescence AL-PCR in LRP12, GIPC1 and RILPL1

Samples of 1555 patients were subjected to RP-PCR, which targeted repeat regions in LRP12, GIPC1 and RILPL1 genes. Samples that exhibited a saw-tooth curve were considered positive. Subsequently, fluorescence AL-PCR was performed to determine the repeat number. The RP-PCR and AL-PCR were performed using previously described procedures.6 8 10 The primers applied can be found in online supplemental figure 2A. All PCR products were subjected to capillary electrophoresis (ABI PRISM 3130xL Genetic Analyzer, Applied Biosystems, Foster City, California, USA). The resulting data were visualised using Peak Scanner Software (Applied Biosystems).

We conducted additional analysis of CGG repeat numbers within the LRP12 gene using a control cohort consisting of 208 patients.

Long-read sequencing of LRP12 by Oxford Nanopore

Among five patients, CGG repeat expansions were suspected using RP-PCR, but the repeat numbers could not be determined by fluorescence AL-PCR (figure 1C). We carried out long-read sequencing with adaptive sampling on the GridION platform (Oxford Nanopore Technologies, Oxford, UK)18 (online supplemental figure 2B). Briefly, the human genome reference GRCh38 was employed, and basecalling was performed using the Super Accuracy mode of Guppy software (V.6.5.7, https://community.nanoporetech.com). The generated FASTQ files underwent repeat detection analysis via tandem-genotypes (V.1.9.0). Alignment and consensus sequence generation for repeat expansions were conducted using lamassemble, with the determination of repeat numbers achieved through CLC Sequence Viewer V.8 (QIAGEN, Valencia, California, USA).

Comparative analysis

To comprehensively understand potential distinctions between our cases with LRP12 repeat expansion and LRP12-OPDM, we conducted a review of previously documented LRP12-OPDM cases.9 19–23 Our inclusion criteria encompassed cases with firmly established genetic diagnoses and readily accessible clinical data. Furthermore, we conducted a comparative analysis of the clinical data and nerve conduction study (NCS) findings between the cases involving LRP12 and NOTCH2NLC within this case series.

Statistical analysis

Statistical analyses were performed using GraphPad Prism V.9.3.1 (GraphPad Software, San Diego, California, USA). The Mann-Whitney U test and Pearson’s correlation test were used to perform comparisons between frequencies and numerical variables. Fisher’s exact test was applied for categorical data whenever available. We considered p<0.05 as significant in all statistical evaluations.

ResultsCGG repeat analysis in LRP12

Among 1555 patients with unidentified IPN, CGG repeat expansions were detected in the LRP12 gene (NM_013437.5) from 44 unrelated individuals. The LRP12 gene constituted 4.8% of 913 genetically diagnosed cases in our IPN case series, making it the fourth most common causative gene, following MFN2, GJB1 and MPZ (figure 1B).

The average CGG repeat number observed was 76.41±17.97 (range: 50–152 repeats), predominantly below 100 repeats (41/44 cases). Detailed CGG repeat numbers are provided in online supplemental table 1. Pedigree analysis was conducted within five families exhibiting a positive family history, confirming the co-segregation of LRP12 repeat expansion with clinical phenotypes across family members, except for an asymptomatic sibling of patient 12 (online supplemental figure 3A). In two families, slight intergenerational changes in repeat number were identified, with counts decreasing from 61 repeats to 53 repeats and from 81 repeats to 63 repeats, respectively (online supplemental figure 3A). However, no notable differences were observed in the onset age or phenotypic diversity across generations. A~1.3 Mb core haplotype block was identified in 35 individuals (online supplemental figure 4). In contrast, among 208 control samples, the range of CGG repeats numbers in LRP12 varied from 2 to 28 (online supplemental figure 3B).

CGG repeat analysis in GIPC1 and RILPL1

No disease-associated repeat expansions were identified in the GIPC1 or RILPL1 genes.

Clinical characteristics

Among 44 patients identified with LRP12 repeat expansions, clinical data were collected from 37 cases (online supplemental table 1). Predominantly, these cases received clinical diagnoses of axonal CMT or HMN by neurologist. Their average onset age was 39.08 years, with a mean examination age of 53.78 years. 25/37 cases had a family history, indicating autosomal dominant inheritance, while 12/37 cases were sporadic. The initial symptom was variable, predominantly manifested with lower extremity muscle weakness (30/37 cases) and may also have presented with fatigue (1 case), right side limb weakness (1 case), handgrip weakness (1 case), lower extremity pain (1 case), muscle cramps in the lower extremities (1 case) and ptosis (1 case).

All 37 patients developed limb muscle weakness, with 23 cases showing distal predominance, 8 cases with proximal predominance and 6 cases with equal proximal and distal impairment. Muscle weakness was more pronounced in the lower limbs than the upper limbs. Muscle atrophy in the lower limbs was observed in all 36 cases for whom data were available, with the gastrocnemius muscle being the most frequently affected. On the 25 cases with lower limb muscle atrophy observed via visual inspection, atrophies were noted in the posterior leg compartment (21 cases), the thigh (2 cases) or showed diffuse atrophy (2 cases). Additionally, 5 cases presented with atrophy of the plantar muscles, resembling pes cavus. In the upper limbs, muscle atrophy was observed in 7 cases, with 2 cases showing proximal atrophy and 5 cases with distal atrophy. The majority of cases exhibited symmetrical muscle weakness, with 11/36 cases (30.6%) showing asymmetrical weakness between the left and right sides. Sensory disturbances were observed in 7/37 cases (18.9%), while ptosis (2.7%), ophthalmoplegia (0%), facial muscle weakness (11.1%) and bulbar palsy (14.7%) were uncommon. Cerebellar ataxia was detected in 2/34 cases (5.9%). No upper motor neuron signs, cognitive dysfunction, obstructive sleep apnoea or parkinsonism were observed. Mild cardiac disease was found in 5/32 cases (15.6%), consisting of 4 cases with arrhythmia and 1 case with ventricular wall thickening. 2/35 cases (5.7%) became non-ambulatory due to lower limb weakness. All 33 cases were capable of walking, with 27 cases (81.8%) able to walk independently, while 6 cases (18.2%) required canes or assistance. Among the 20 cases with comprehensive evaluation of lower limb motor functions, 15 cases (75.0%) were unable to stand on their tiptoes. None of these cases required ventilator support.

Additionally, among the five cases (P3, P4, P10, P15 and P42) with disease duration exceeding 30 years, four individuals did not exhibit typical symptoms commonly associated with LRP12-OPDM. All five patients maintained their walking ability without experiencing any cardiac or respiratory failure.

Serum creatine kinase (CK) levels were elevated (>200 IU/L) in 32/34 cases (94.1%), with mean CK level 715.9 IU/L. A weak negative correlation was observed between LRP12 CGG repeat number and CK level (online supplemental figure 5A).

Electrophysiological findings

NCS data were available for 37 patients. The most frequent abnormality was reduced amplitude of the tibial nerve compound muscle action potential (CMAP) (normal range: >4.4 mV) in 15/36 cases (41.7%). Tibial motor nerve conduction velocity (MCV) (normal range: >41.7 m/s) was decreased in 8/36 cases (22.2%). Reduction in median nerve MCV (normal range: >49.6 m/s) was found in 5/35 cases (14.3%) and reduction in median nerve CMAP (normal range: >3.1 mV) was observed in 5/35 cases (14.3%). The median nerve MCV was higher than 38 m/s in all cases. Additionally, sensory nerve impairment was observed in seven cases (P5, P14, P15, P17, P23, P38 and P39), encompassing median nerve sensory conduction velocity (SCV; normal range: >47.2 m/s) in 2/35 cases, median nerve sensory nerve action potential (SNAP; normal range: >7.0 μV) amplitude in 1/33 cases, sural nerve SCV (normal range: >40.8 m/s) in 3/36 cases and sural nerve SNAP (normal range: >5.0 µV) in 5/34 cases. In one case (P23), the median nerve motor distal latency was found to be prolonged (4.5 ms; normal range: <4.5 ms), suggesting a concurrent diagnosis of carpal tunnel syndrome.

Among the 31 cases that underwent needle electromyography (nEMG), 30 (96.7%) exhibited neurogenic changes. In the subset of 29 cases with detailed findings, 27 exhibited high amplitude motor unit potentials (MUP), while 1 case showed low amplitude MUP. Neurogenic alterations were observed across multiple regions in 24 cases, including decreased interference pattern (IP; 19 cases), fibrillation potentials (13 cases) and fasciculation potentials (8 cases). Neither fibrillation nor fasciculation potentials were detected in the cranial nerve regions. Notably, P43 displayed a combination of neurogenic and myogenic changes, characterised by high and prolonged amplitude MUP and decreased IP observed in the anterior tibialis, while the quadriceps femoris showed low amplitude MUP. Myotonic discharges were detected in three cases (P16, P22 and P42). However, none of these cases exhibited percussion or grip myotonia, nor did they show repeat expansions in the DMPK or CNBP genes. Detailed NCS/nEMG information is summarised in online supplemental figure 6 and table 2.

Pathological findings in skeletal muscles and peripheral nerves

Skeletal muscle pathology, performed in 12 cases, consistently revealed neurogenic changes. Grouping atrophy was observed in 11 out of 12 cases (91.7%), and pyknotic nuclear clump was present in all 9 cases with records (100%). Rimmed vacuoles (RV) were identified in 2 out of 11 cases (18.2%), while p62 deposition was observed in 4 out of 8 cases (50.0%) (figure 2, online supplemental figure 5B). Notably, among these cases, those with less than 80 CGG repeats showed no RV, whereas two out of three cases (P32 and P40) with more than 81 CGG repeats displayed RV. Furthermore, within these two cases, p62 deposition was also identified. Peripheral nerve biopsy in two cases revealed no significant loss of large or small myelinated nerve fibres. Teased fibre analysis indicated segmental demyelination, paranodal demyelination and mild axonal impairment. p62 deposition was observed in one case (figure 3).

Figure 2

Representative pathological findings in skeletal muscle. H&E staining (A, E, F, I, J, M, N), ATP-ase (pH=10.5) (C), Gomori-Trichrome staining (B, K, O), NADH-TR staining (G), anti-p62 staining (D, H, L, P). (A–H) (Patients 15 and 17): fibre type grouping, small group atrophy and pyknotic nuclear clumps are observed, indicating neurogenic changes; rimmed vacuoles or p62-positive inclusions are not found. (I–P) (Patients 32 and 40): rimmed vacuoles, grouping atrophy, small angular fibres, pyknotic nuclear clumps and p62-positive deposition are observed. Black arrows indicate rimmed vacuole or p62-positive inclusions. Black triangles indicate pyknotic nuclear clumps. Black bar, 200 µm; red bar, 50 µm; white bar 20 µm.

Figure 3

Pathological findings of peripheral nerves from Patients 32 and 40. (A, D): Toluidine blue staining indicates that both large and small myelinated fibres are preserved. (C): Teased fibre analysis shows segmental and paranodal demyelination, as well as mild axonal impairment. (B, E): Anti-p62 staining shows deposition of p62 in Patient 40, but not in Patient 32. (F): Anti-p62 staining in sural nerve of a control sample. Black arrows indicate p62-positive inclusions. Black bar, 50 µm; white bar, 20 µm. NC, normal control.

Radiological findings

The muscle images of 18 cases were available (figure 4). The evaluation of lower limb muscle atrophy in 18 cases was conducted using the Mercuri score on CT or MRI. Muscle atrophy in the distal lower limbs was observed in all patients, particularly lower leg flexors muscle was predominant. Two cases (P15 and P38) had follow-up images available 8 or 10 years later, revealing a progressive atrophy over time. The highest average Mercuri score was obtained in the gastrocnemius (3.56), followed by the soleus (3.33). The mildest atrophy was observed in the gracilis, with a score of 0.53, followed by the adductor longus (0.73) and rectus femoris (0.87) (online supplemental figure 7).

Figure 4

CT or MRI images show lower limb muscle atrophy from 18 patients. Images of thighs (upper panel) and legs (lower panel). Muscle atrophy can be more prominently observed in the legs than in the thighs. In both regions, atrophy and fatty replacement are notably more pronounced on the posterior than anterior lower limbs. There is a tendency for more pronounced atrophy on the lateral sides compared with the medial sides in both regions. Atrophy of the posterior leg muscles, including the gastrocnemius and soleus muscles, is consistently noted across all cases. Follow-up imaging of patients 15 and 38 shows progressive atrophy over time. Patients 31, 40 and 42 did not undergo examination of their upper legs.

Brain MRI in 13 cases showed no abnormal signals in subcortical white matter on diffusion weighted image. Cerebellar atrophy was identified in three patients, one of whom exhibited mild cerebellar ataxia (online supplemental figure 8, table 1).

Comparative analysis

We conducted a retrospective review of 76 previously documented LRP12-OPDM cases, alongside our 44 cases (table 1).9 19–23 The CGG repeat numbers in our cases were significantly shorter than those in LRP12-OPDM (76.41 vs 162.33, p<0.0001). Our cases exhibited significantly lower frequencies of ptosis, ophthalmoplegia, facial muscle weakness and bulbar palsy compared with LRP12-OPDM (p<0.0001, respectively). Furthermore, skeletal muscle tissues obtained from our cases showed significantly lower frequencies of the presence of RV (18.2% vs 100%, p<0.0001) and p62 deposition (50.0% vs 100%, p=0.0015) compared with those with LRP12-OPDM. No statistical difference was observed in onset age (39.08 vs 40.73 years), limb weakness, limb muscle atrophy, sensory disturbance, cognitive disorder, cardiac disorder or respiratory failure. The incidence of loss of ambulation was significantly lower in our LRP12 cases (5.7% vs 24.2%, p=0.0273).

Table 1

Comparative analysis between our cases and LRP12-OPDM

Next, we compared the clinical data and nerve conduction study results for the 44 cases with LRP12 repeat expansions and 21 cases with NOTCH2NLC repeat expansions identified in our cases series. No significant differences were observed in onset age, CK levels, frequencies of ptosis, ophthalmoplegia, facial muscle weakness or bulbar palsy. Limb muscle weakness and atrophy were consistently present in all cases within the LRP12 group, whereas a subset of cases in the NOTCH2NLC group did not show these features. On the other hand, sensory disturbance was significantly more common in the NOTCH2NLC group (18.9% vs 50.0%, p=0.0315). Abnormal finding of head MRI and cognitive disorder were notably more prevalent in the NOTCH2NLC group, with one case showing subcortical white matter hyperdense resembling neuronal intranuclear inclusion disease (NIID). NCS results for the median, tibial and sural nerves indicated milder impairment in the LRP12 group compared with the NOTCH2NLC-IPN group, with further details provided in online supplemental table 3 and figure 9.

Discussion

The identification of 5’ UTR CGG repeat expansions within the LRP12 gene in 44 out of 2424 Japanese patients with IPN underscores its significance as a notable causative factor. The prevalence of the LRP12 gene, accounting for 4.8% of genetically diagnosed cases in our IPN case series, positions it as the fourth most frequently implicated gene, as highlighted in figure 1B. Intriguingly, it emerges as the predominant gene associated with repeat diseases within our IPN case series.

LRP12 gene was initially identified as a causative gene for OPDM.6 OPDM represents a skeletal muscle disorder, characterised by distal limb muscle weakness and atrophy, ptosis, ophthalmoplegia, facial muscle weakness and bulbar palsy.12 19 24–26 Our cases also exhibited limb muscle weakness and atrophy, notably affecting the lower limb flexor muscles, such as the gastrocnemius and soleus muscles (figure 4, online supplemental figure 7), which resembled a typical finding often associated with OPDM.19 25 27 Conversely, the prevalence rates of specific symptoms in our cases were significantly lower when compared with LRP12-OPDM cases, comprising ptosis, ophthalmoplegia, facial muscle weakness and bulbar palsy. Furthermore, the five cases that demonstrated a long-term course of over 30 years were mild progression and did not exhibit typical symptoms of OPDM. Otherwise, only 5.7% prevalence of ambulation loss is significantly lower than that observed in the LRP12-OPDM (table 1). Taken together, our study demonstrates a distinct, milder and more slowly progressive disease course across our patients in comparison to patients with LRP12-OPDM.

Considering nEMG and skeletal muscle biopsy data, nearly all patients diagnosed with OPDM displayed myogenic changes, as documented in previous studies.19 22 Neurogenic changes, on the other hand, were infrequently observed, as reported in a limited number of cases.21 28 In our study, all 12 cases with muscle pathology data showed evident neurogenic changes, while nEMG findings revealed neurogenic changes in 30/31 cases, with 24 cases demonstrating the changes at multiple regions. These collective findings suggest the presence of neuropathy or neuronopathy in our cases. P15 showed low amplitudes in both upper and lower limbs on nEMG, alongside neurogenic changes indicated by NCS and muscle pathology, suggesting neuromyopathy. Furthermore, NCS highlighted peripheral nerve impairment in 22 out of 37 cases, with 7 cases presenting sensory nerve impairment and 15 cases showing isolated motor nerve impairment (online supplemental figure 6 and table 2). Additionally, 13 cases displayed no abnormal NCS findings but featuring limb muscle weakness and atrophy, along with neurogenic changes in nEMG or muscle pathology. These results suggest that these cases may be in an early stage without detectable NCS abnormalities, or may possibly indicate a neuronopathy. In summary, on comparing the available NCS data from 35 cases with nEMG or muscle pathology findings, 28 cases predominantly exhibiting motor nerve impairment were classified as HMN. Meanwhile, seven cases with abnormalities in both motor and sensory nerves were classified as CMT. Notably, in all cases, the median nerve MCV surpassed 38 m/s, further specifying the classification of CMT as CMT2 (axonal-CMT). Otherwise, although a novel LRP12-ALS phenotype was recently described,28 none of our patients developed pyramidal sign, therefore not supporting the diagnosis of ALS.

When comparing NOTCH2NLC-IPN and our LRP12 cases, it becomes evident that NOTCH2NLC-IPN cases commonly show sensory nerve impairment, resembling a CMT phenotype, notably demonstrating intermediate-CMT or demyelinating-CMT features. In contrast, our LRP12 cases predominantly present as HMN, although a subset of cases also display characteristics of axonal-CMT. Furthermore, NCS findings of LRP12 cases revealed milder abnormalities in both motor and sensory nerves compared with NOTCH2NLC-IPN (online supplemental figure 9). Interestingly, unlike NOTCH2NLC-IPN cases, none of the LRP12 cases exhibited neuroimaging features resembling NIID or cognitive impairment (online supplemental figure 8 and table 3). Despite both LRP12 and NOTCH2NLC are identified as causative genes for OPDM and IPN, they distinctly manifest different phenotypic expressions within the spectrum of IPN.

The initial LRP12-OPDM studies indicated that pathogenic LRP12 CGG repeat numbers exceeded 100 repeats (figure 5).6 19 22 28 Thereafter, CGG repeat lengths shorter than 100 have been found to be associated with patients manifesting LRP12-ALS and LRP12-PMA (progressive muscular atrophy) phenotypes.28 In our current cases, the observed repeat numbers were predominantly shorter than 100. Moreover, our study revealed a notable discrepancy in repeat numbers between our patients and documented cases with LRP12-OPDM. Our study, along with the existing literature, consistently suggests a threshold whereby the LRP12 repeat number in the control population should not exceed 50.6 28 29 The CGG repeat number ranging from 50 to 100 within the LRP12 gene may give rise to a spectrum of phenotypes, comprising PMA, ALS, HMN and CMT (figure 5). Although a negative correlation between repeat number and onset age has been reported previously,19 our study only revealed a statistically insignificant trend (online supplemental figure 5A).

Figure 5

Diagram showing correlation between CGG repeat numbers in LRP12 and phenotypes in current and previous studies. Each square displays the disease phenotype along with the number of CGG repeats in LRP12. CGG repeat numbers have been found shorter in our patients (50–152), patients with LRP12-ALS (64–100) and LRP12-PMA (61–87), in contrast to the longer repeat range observed in patients with LRP12-OPDM (76–630). All reports consistently show that the count of CGG repeats within the LRP12 gene among control individuals does not exceed 50. CMT, Charcot-Marie-Tooth disease; HMN, hereditary motor neuropathy; OPDM, oculopharyngodistal myopathy; PMA, progressive muscular atrophy.

The typical characteristic pathological feature of OPDM is the presence of RV and p62-positive inclusions in skeletal muscle.11 12 19 20 23 25 26 30 31 In our cases, however, RV was identified in only 2/11 cases, and p62 deposition was noted in 4/8 cases (online supplemental figure 5B). Moreover, the presence of p62 deposition observed in both muscle and nerve tissues was shared by cases with CGG repeats around 100, which was noted in our case of P40 (96 repeats) and a previously reported patient (106 repeats)21 (figures 2,3). Electrophysiological studies in P43 (120 repeats) revealed a combination of neurogenic and myogenic changes (online supplemental table 2). These findings imply a mixed phenotype characterised by manifestation from both IPN and OPDM, affecting both muscle and nerve tissues. Additionally, in the muscle pathology of our cases with CGG repeats below 80, none of the eight cases exhibited RVs, and only two out of five cases investigated showed p62-positive inclusions (online supplemental figure 5C). These findings suggest that a shorter number of CGG repeats is more indicative of a predominantly neurogenic phenotype. Conversely, as the CGG repeat count approaches 100, there emerges a potential for a mixed pathology encompassing both neurogenic and myogenic characteristics. In the two cases with sural nerve biopsy (P32 and P40), teased fibre analysis revealed segmental and paranodal demyelination in certain fibres, yet toluidine blue staining showed no reduction in nerve fibre density (figure 3). These observations, combined with normal NCS results for sensory nerves, including the sural nerve, consistently indicated an HMN phenotype for these cases.

Repeat expansions in non-coding regions may be implicated through three non-exclusive mechanisms, which include the sequestration of RNA-binding proteins as a result of RNA foci accumulation, polyglutamine aggregation via repeat-associated non-AUG (RAN) translation and splicing disruption due to hypermethylation.32–34 The clinical diversity imparted by the LRP12 gene may be attributed to a repeat length-dependent mechanism, similar to other CGG repeat expansion-related disorders, FXS/FXTAS and NIID/OPDM3. Longer CGG repeats in the FMR1 gene (≥200) can recruit proteins, leading to RNA aggregation, splicing issues causing the FXS phenotype.35 36 Shorter CGG repeats (55–200) induce the FXTAS phenotype as a consequence RAN translation.37–39

The pathogenicity of shorter CGG repeat expansions (50–100 repeats) in LRP12 is not yet fully understood at this time. However, the aggregation of MBNL1 and dysregulation of messenger RNA splicing have been observed in patients with LRP12-OPDM carrying >100 CGG repeats but not in patients with under 100 CGG repeats in LRP12.28 Similar to FMR1 and NOTCH2NLC, it is possible that mechanisms involving both loss of function and gain of function may vary based on the number of repeats, offering a potential explanation for the diversity in the broad spectrum of clinical manifestations in LRP12 gene.

Mild cardiac disease was observed in 5/32 cases (15.6%) in this study, exhibiting arrhythmia or ventricular wall thickening. This cardiac involvement, though infrequent, aligns with observation from prior studies involving patients with LRP12-OPDM.19 40 Further studies have confirmed LRP12 RNA expression within cardiac muscle,41 42 and autopsy revealed p62-positive intranuclear inclusions in cardiac tissue.20 Collectively, these findings suggest a potential link between CGG repeat expansions in LRP12 and cardiac disorders.

There are specific limitations that require addressing within this study. The clinical data were obtained from various hospitals, which introduces considerable bias due to operational diversity in conducting the NCS analysis. For instance, there is variability in the specific finger used for recording the median nerve SNAP and the choice between antidromic and orthodromic techniques for sensory nerves. Furthermore, pathological analysis of peripheral nerves, crucial for confirming peripheral nervous system involvement and aiding in phenotype determination, was conducted in only two patients. Consequently, this assessment was constrained, and a larger sample size would provide more robust evidence regarding the involvement of peripheral nervous system. To elucidate the underlying mechanism and genotype-phenotype correlations of LRP12-related disorders, further investigations are warranted, such as functional assessments using patient-derived induced pluripotent stem cells, pathological analyses of other nervous system tissues and studies examining RNA accumulation and splicing levels.

In conclusion, our study disproves the existence of GIPC1 and RILPL1 repeat expansions in our IPN case series, but reveals an unexpected high prevalence of CMT/HMN patients linked to CGG repeat expansions in the LRP12 gene. This genetic abnormality emerges as the fourth most prevalent disorder within the Japanese IPN case series, establishing it as the most common disorder caused by repeat expansions. The observed repeat numbers predominantly fall within the range of 50–100, shorter than the typical lengths detected in LRP12-OPDM. Given the high diagnostic yield, we strongly recommend for the inclusion of LRP12 repeat expansion screening in IPN cases. Our comprehensive analysis of clinical, electrophysiological and pathological data suggests the existence of a potential subgroup of patients with LRP12-related disorders, exhibiting either axonal-CMT or HMN phenotypes. Additionally, we have postulated the presence of a mixed phenotype involving HMN and OPDM. Our investigation significantly broadens the observed phenotypic spectrum associated with the LRP12 gene.

Acknowledgments

The authors thank R Kato, S Ohya, M Beppu, T Kudo, K Nishoka, A Horiuchi, R Tokimura, H Sekiya, M Morita, T Abe, H Kato, M Matsuo, K Saito, M Hamaguchi, Y Adachi, S Mitsuma, M Okabe, T Ohkubo, M Shigehisa, H Ishikawa, M Yoshikawa, Y Aoshima, S Enomoto, M Eriguchi, N Kokubun, M Nakagawa, T Matsushita, Y Sano, K Yabuuchi, T Hayashi, H Shibata, N Iwasa, H Nanaura, N Iguchi, R Tohge, N Kimura, K Ogaki, J Kaneko, A Sugiyama, Y Ono, T Matsui, K Kanai, M Togo, H Suehiro, T Kato, K Shibuya, T Ishiguro, F Morii, H Fujimura, D Matsuse, M Mikekado and D Taniguchi for collecting gene sample and evaluating patients’ clinical findings. The authors appreciate I Nishino at the National Center of Neurology and Psychiatry for providing histopathological findings in muscles. The authors also thank T Ohnishi, M Minami and N Hirata at Kagoshima University for their vital technical assistance. The authors extend their appreciation to the Division of Gene Research, Research Support Centre, Kagoshima University, for the use of their facilities. Finally, the authors thank ENAGO (www.enago.jp) for English language editing.