Article Text

Download PDFPDF

Original research
Long-read sequencing identified repeat expansions in the 5′UTR of the NOTCH2NLC gene from Chinese patients with neuronal intranuclear inclusion disease
  1. Jianwen Deng1,
  2. Muliang Gu1,
  3. Yu Miao2,
  4. Sheng Yao3,
  5. Min Zhu4,
  6. Pu Fang4,
  7. Xuefan Yu5,
  8. Pidong Li2,
  9. Yanan Su2,
  10. Jian Huang2,
  11. Jun Zhang6,
  12. Jiaxi Yu1,
  13. Fan Li1,
  14. Jing Bai1,
  15. Wei Sun1,
  16. Yining Huang1,
  17. Yun Yuan1,
  18. Daojun Hong4,6,
  19. Zhaoxia Wang1
  1. 1 Department of Neurology, Peking University First Hospital, Beijing, China
  2. 2 GrandOmics Biosciences, Beijing, China
  3. 3 Department of Neurology, Sixth Medical Center of PLA General Hospital, Beijing, China
  4. 4 Department of Neurology, First Affiliated Hospital of Nanchang University, Nanchang, China
  5. 5 Department of Neurology, Bethune First Affiliated Hospital of Jilin University, Nanchang, China
  6. 6 Department of Neurology, Peking University People's Hospital, Beijing, China
  1. Correspondence to Dr Daojun Hong, Department of Neurology, The First Affiliated Hospital of Nanchang University, Nanchang, China; hongdaojun{at}; Dr Zhaoxia Wang, Department of Neurology, Peking University First Hospital, Beijing, China; drwangzx{at}


Background Neuronal intranuclear inclusion disease (NIID) is a heterogenous neurodegenerative disorder named after its pathological features. It has long been considered a disease of genetic origin. Recently, the GGC repeated expansion in the 5′-untranslated region (5′UTR) of the NOTCH2NLC gene has been found in adult-onset NIID in Japanese individuals. This study was aimed to investigate the causative mutations of NIID in Chinese patients.

Methods Fifteen patients with NIID were identified from five academic neurological centres. Biopsied skin samples were analysed by histological staining, immunostaining and electron microscopic observation. Whole-genome sequencing (WGS) and long-read sequencing (LRS) were initially performed in three patients with NIID. Repeat-primed PCR was conducted to confirm the genetic variations in the three patients and the other 12 cases.

Results Our patients included 14 adult-onset patients and 1 juvenile-onset patient characterised by degeneration of multiple nervous systems. All patients were identified with intranuclear inclusions in the nuclei of fibroblasts, fat cells and ductal epithelial cells of sweat glands. The WGS failed to find any likely pathogenic variations for NIID. The LRS successfully identified that three patients with adult-onset NIID showed abnormalities of GGC expansion in 5′UTR of the NOTCH2NLC gene. The GGC repeated expansion was further confirmed by repeat-primed PCR in seven familial cases and eight sporadic cases.

Conclusion Our findings provided evidence that confirmed the GGC repeated expansion in the 5′UTR of the NOTCH2NLC gene is associated with the pathogenesis of NIID. Additionally, the GGC expansion was not only responsible for adult-onset patients, but also responsible for juvenile-onset patients.

  • genetics
  • neurology
  • NIID
  • GGC repeated expansion

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Neuronal intranuclear inclusion disease (NIID) is a neurodegenerative disorder pathologically characterised by the presence of eosinophilic hyaline intranuclear inclusions in the central and peripheral nervous systems, as well as in the multiple other organs.1 2 Based on the age of onset, NIID can be divided into infantile, juvenile and adult clinical subgroups. The neurological manifestations of adult-onset NIID show a great clinical heterogeneity, and can include dementia, encephalitic episode, autonomic dysfunctions, Parkinsonism, cerebellar ataxia, convulsions, pyramidal symptoms and sensory disturbances.3 4

Since skin biopsy was ascertained to assist the diagnosis of NIID, an increasing number of patients with adult-onset NIID have been reported in Japan,5 6 which indicated that the adult-onset NIID might have a greater susceptibility to affect individuals in Eastern Asia. However, very few cases have been reported in mainland China,7 8 although China has the largest population in the region. Recently, the GGC repeated expansion in the 5′-untranslated region (5′UTR) of the NOTCH2NLC gene has been found to be involved in the pathogenesis of NIID in the Japanese population.9 Our study aims to examine the genetic aetiology of NIID in Chinese patients. Over the past year, we have used next-generation whole-genome sequencing (WGS) to explore the causative genes associated with adult-onset NIID, but have failed to identify likely pathogenic mutations. More recently, we used long-read genome sequencing (LRS) on the Oxford Nanopore platform, and identified trinucleotide repeat expansion (GGC) in the NOTCH2NLC as the genetic cause of juvenile-onset and adult-onset NIID pathogenesis in China.

Materials and methods


Fifteen patients from 12 families with NIID were retrospectively assessed from five academic neurological centres between October 2017 and April 2019. The patients were examined by at least two experienced neurologists. The detailed medical history was obtained from the subjects and their relatives. Information regarding age of onset, progression of disease, family history and other clinical manifestations were collected. Mini-Mental State Examination and Frontal Assessment Battery were used to evaluate the executive and cognitive functions. A battery of laboratory tests was conducted to exclude metabolic, toxic and inflammatory causes. Motor and sensory nerve conduction velocity was performed in the nerves using a standard method with surface electrodes for stimulation and recording. Open biopsies from the skin in the distal part of the right leg (10 cm above the external malleolus) were performed in all patients. Standard pathological examinations included H&E staining, anti-p62 antibody (sc-28359, Santa Cruz Biotechnology) immunostaining and electron microscopic observation. All patients’ tissue samples were obtained after a written consent signed by each individual in compliance with the bioethics laws of China as well as the Declaration of Helsinki.

Whole-Genome sequencing

For short-read sequencing, a DNA sequencing library was established using MGIEasy DNA Library Prep Kit following the manufacturer’s instructions (BGI, Shenzhen, China) to generate DNA Nanoballs (DNB). Each DNB library was sequenced on the MGISEQ-2000 instrument, and 150 bp paired-end reads were generated. Burrows-Wheeler Alignment Maximal Exact Matches (BWA-MEM) were used for read mapping onto the hg38 human genome as a reference. Sequencing data passing quality control (QC) were subject to a computational pipeline for data processing and analysis, following the standard workflow. Calls with variant quality less than 20 were filtered out, and 95% of the targeted bases were covered sufficiently to pass our thresholds for calling SNPs and small insertions or deletions (indels). GATK was used for single nucleotide variants (SNVs)/insertions and deletions (InDels) calling, ANNOVAR12 was used for variant annotation and we manually examined all potential disease variants in Integrative Genomics Viewer.

Nanopore long-read sequencing

Since no disease-causing variant was found with conventional short-read sequencing, we hypothesised that the disease may be caused by complex structural variations (SV). Therefore, low-coverage sequencing was performed for three affected individuals on the Oxford Nanopore platform. Large insert-size libraries were created according to the manufacturer’s recommended protocols (Oxford Nanopore, Oxford, UK). Mapping-based methods were conducted for regular SV calling. At the time we were performing the nanopore LRS, Sone et al reported that a short tandem repeat (STR)-GGC repeated expansion in the 5′UTR of the NOTCH2NLC gene was responsible for adult-onset NIID in patients.9 Therefore, we directly used the repeatHMM algorithm for counting GGC motifs of this STR region of the NOTCH2NLC gene. Because the NOTCH2NLC gene is one of five homologous NOTCH2 genes, we used samtools10 to extract all reads aligned to 10 kb upstream and downstream of the STR region of the NOTCH2NLC and its homologous genes. We then aligned the extracted reads to the reference sequence of 10 kb upstream and downstream of the STR region of the NOTCH2NLC gene, and used the SNPs between the homologous sequences to distinguish these reads in order to screen for all reads located 10 kb upstream and downstream of the STR region of the NOTCH2NLC gene.

Repeat-primed PCR

The repeat-primed PCR (RP-PCR) protocol was performed according to the literature with small modifications.9 In brief, the PCR primer mix contained three primers: 0.3 µM of NOTCH2NLC-F: 5′-FAM-GGCATTTGCGCCTGTGCTTCGGACCGT-3′, 0.15 µM of M13-(GGC)4(GGA)2 R: 5′-CAGGAAACAGCTATGACCTCCTCCGCCGCCGCCGCC-3′ and 0.3 µM of M13-linker-R: 5′-CAGGAAACAGCTATGACC-3′. After incubation at 98°C for 10 min, the following cycling conditions were performed: 16 cycles of 98°C for 30 s, 66°C for 1 min with reduced 0.5°C per cycle and 68°C for 8 min; then, 32 cycles of 98°C for 30 s, 58°C for 1 min and 68°C for 8 min; then, a final elongation step of 68°C for 10 min. Electrophoresis was performed on a 3500xl Genetic analyser (Thermo Fisher Scientific, Waltham, Massachusetts, USA) and the data were analysed using GeneMapper software (Thermo Fisher Scientific). A saw-tooth tail pattern in the electropherogram was considered to be the disease-associated repeat expansion.


Clinical characteristics of patients with late-onset NIID

The clinical characteristics of the 15 patients are summarised in table 1. The cohort of cases included eight male and seven female patients. The mean age was 57.9±14.7 years. The median duration of disease was 4 (0.5, 29) years. The main clinical symptoms included cognitive impairment (14/15, 93.3%), episodic encephalopathy (10/15, 66.7%), bladder dysfunction (9/15, 60.0%), abnormal behaviour (8/15, 53.3%), convulsion (8/15, 53.3%), muscle weakness (8/15, 53.3%), visual abnormalities (8/15, 53.3%), sensory disturbance (7/15, 46.7%), dysarthria (7/15, 46.7%), miosis (6/15, 40.0%), tremor (6/15, 40.0%), rigidity (6/15, 40.0%) and ataxia (5/15, 33.3%). Among the patients with bladder dysfunction, five patients presented with the symptom prior to the development of other neurological symptoms; seven patients needed permanent cystostomy. Four patients reported similar cases in their families. Two families indicated an autosomal dominant inherited pattern, but two other families only showed siblings affected.

Table 1

The clinical features of the patients with NIID

Cerebral MRI showed diffuse high signals of cerebral white matter on T2-weighted images (figure 1A) and diffusion-weighted MRI high signals along the corticomedullary junction (figure 1B) in all cases. Linear enhanced lesions along the surface of the cortex were observed in two patients (figure 1C). Eosinophilic intranuclear inclusions were found in the nuclei of fibroblasts, fat cells and ductal epithelial cells of sweat glands through HE staining (figure 1D and E). These eosinophilic intranuclear inclusions were p62-positive (figure 1F and G) in all patients. Electron microscopy revealed a pile of round-halo filamentous materials (8–12 nm) without a limiting membrane in the centre of the nucleus (figure 1H and I).

Figure 1

Brain MRI findings and histopathological features of NIID patients. Brain MRI of patient 1 shows diffuse high signals of white matter on T2-weighted images (A), high signals along the corticomedullary junction on diffusion-weighted image (B) and linear enhanced lesions along the cortex (C, arrow). Eosinophilic intranuclear inclusions were observed in ductal epithelial cells of sweat gland (D and E magnification). Immunohistochemistry showed the intranuclear inclusions positive to p62 antibody in ductal epithelial cells (F) and fat cells (G). Electron microscopy indicates the intranuclear inclusions are a pile of filament materials without a limiting membrane (H and I magnification). NIID, neuronal intranuclear inclusion disease.

WGS screening

We initially performed WGS to determine potential causal SNVs, InDels, exonic rearrangements and SVs in the mapping region, and sequencing data passing QC were subject to a computational pipeline for data processing and analysis through the standard workflow. Since NIID is a rare disorder with a clear phenotype, there was a low likelihood that a causal mutation in patients was present in normal populations. We therefore filtered for novel variants by comparing our data to several databases including dbSNP build 132, the 1000 Genomes Project, Hapmap, YH project and the National Heart, Lung, and Blood Institute Exome Sequencing Project. However, comparing the genetic lists among the three patients, no causative variants were detected in the WGS screening.

LRS detecting the short tandem repeat variation

We performed WGS with ~15 × coverage on the Oxford Nanopore platform. The nanopore raw data were filtered to remove sequences with high base error rates and quality scores less than 7; the resulting qualified data were used for the next analysis. Mapping-based methods were conducted for calling SVs in a genome-wide scale. Candidate SVs were obtained for manual examination and further validation. However, no variant was identified or validated as a likely causative mutation.

We next focused on the possible STR expansion in the NOTCH2NLC gene. However, the NOTCH2NLC gene is found to have five homologous genes including NOTCH2, NOTCH2NLA, NOTCH2NLB, NOTCH2NLC and NOTCH2NLR in the reference hg38 genome (figure 2A). Therefore, we initially used the reference sequence of 10 kb upstream and downstream of the STR region of these genes; we compared the reference sequence from 10 kb upstream and downstream of the STR to the NOTCH2, NOTCH2NLA, NOTCH2NLB, NOTCH2NLR and NOTCH2NLC sequences to find different SNPs (figure 2B). We then aligned the filtered reads to the hg38 reference genome by NGLMLR to distinguish the STR region of the NOTCH2NLC gene using the SNPs list (online supplementary table 1). The RepeatHMM algorithm was used to count GGC motifs of the STR region of the NOTCH2NLC gene (online supplementary table 2). In patient 1, repeatHMM revealed that 23 reads of GGC repeats were located in the STR of NOTCH2NLC, including 15 reads of repeats between 20 and 25, and 8 reads of repeats between 139 and 152 (figure 2C). In patient 12, repeatHMM identified that the STR of NOTCH2NLC had 17 reads of GGC repeats, including 8 reads of repeats between 16 and 20, and 9 reads of repeats between 120 and 376 (figure 2D). In patient 15, repeatHMM found 5 reads of GGC repeats between 120 and 130 existed in the STR of NOTCH2NLC (figure 2E).

Supplemental material

Supplemental material

Figure 2

The GGC repeats expansion detected by long-read sequencing in NIID patients. (A) Gene structure of Notch2NLC homologous genes. (B) Alignment of Notch2NLC homologous genes with SNPs highlighted by red box. (C–E) The GGC repeats of Notch2NLC genes in NIID patients were estimated by repeatHMM algorithm. NIID, neuronal intranuclear inclusion disease.

Variant verification

RP-PCR was used to verify the repeat expansion in the STR of NOTCH2NLC. Initially, repeat expansion was demonstrated by RP-PCR as a saw-tooth pattern in the above three patients (figure 3). Subsequently, the repeat expansions were identified in affected family individuals but not in unaffected familial members, indicating a family co-segregation. Finally, we explored the repeat expansion in seven sporadic NIID patients confirmed by pathological diagnosis of skin biopsy. All the sporadic patients had the expansion (figure 3). Taken together, (GGC)n repeat expansion in NOTCH2NLC was consistently found in both familial NIID patients and sporadic NIID patients, further suggesting a causative role for the repeat expansions in NIID patients.

Figure 3

Electropherogram of repeat-primed PCR. Control panel showed no repeat expansion in unaffected individual. Patient 1 to Patient 15 panels revealed a saw-tooth tail pattern of the repeat expansion in affected individuals.


Since the first case of NIID was reported by Lindenberg et al in 1968, the clinical symptoms of NIID have shown high heterogeneity.11 12 Based on the age of onset, the multisystem degenerative disease can be divided into infantile, juvenile and adult subgroups. Therefore, it has long been considered that NIID was closely associated with genetic causes. Initially, familial NIID cases were identified in two male and female siblings,13 as well as a set of identical twins,14 but their parents were unaffected, which was suggestive of autosomal recessive inheritance. However, several cases with adult-onset NIID were subsequently reported in autosomal dominant inherited families, all of which made the genetic cause of NIID more obscured.

The similar hereditary pattern was also observed in our cohort NIID patients. In this study, we described 15 NIID cases including five patients from two autosomal dominant inherited families, two patients from two families with multiple single-generation patients and eight sporadic individuals (online supplementary figure 1). In addition, no clinical anticipation was observed in our patients or other reported cases. Clinically, they showed a great heterogeneity that was mainly focused on cognitive impairment, episodic encephalopathy and bladder dysfunction, which were consistent with the typical clinical features of NIID reported in Japanese individuals.15 Sone et al have identified that the eosinophilic p62-positive intranuclear inclusion of skin biopsy can be used to pathologically diagnose NIID.5 Similarly, our NIID patients were pathologically confirmed to have a pile of round-halo filament materials without a limiting membrane in the nucleus of fibroblasts, fat cells and ductal epithelial cells of sweat gland.

Supplemental material

The long history of searching for the causative mutation of NIID suggested a significant complexity of disease mechanisms. In our previous work, we failed to identify any causative mutations in three NIID patients using WGS. This indicated that next-generation sequencing (NGS) was insufficient to resolve the challenge of identifying causative genes in NIID. Adult-onset NIID clinically and pathologically showed a similarity with fragile X-associated tremor/ataxia syndrome that is caused by expansion of a (GGC)n repeat to 55–200 copies in the 5′UTR of the FMR1 gene.16 Therefore, these clinical findings suggested that the expansion of STR should also be considered in NIID. As an alternative strategy, we used whole-genome low-coverage LRS to search for potential causative SVs. At the time when we began to analyse the big data, Sone et al reported that repeat expansions in the 5′UTR of the NOTCH2NLC gene were the genetic cause for NIID.9 The report of causative 5′UTR repeat expansion in NOTCH2NLC helped us focus on STRs in the candidate region with more detailed examination. The NOTCH2NLC gene has five homologous genes including NOTCH2, NOTCH2NLA, NOTCH2NLB, NOTCH2NLC and NOTCH2NLR, in the reference hg38 genome.17 In this study, we developed an SNPs list to efficiently distinguish NOTCH2NLC from the other four genes. Intriguingly, this repeat expansion was also identified in a juvenile-onset case, indicating that the GGC repeats in the 5′UTR of NOTCH2NLC might be the common pathogenesis of different age-onset in NIID.

The genetic diagnostic rate of Mendelian disorders is still relatively low, even with recent improvements in NGS technologies. It is believed that this is partially due to the presence of many long repetitive elements, copy number alterations and SVs that are relevant to the disease, undetected by conventional sequencing approaches.18 Although differentiating many of these complex elements was beyond the capacity of short-read sequencing (SRS)-related technologies, the complementary strengths make LRS a promising approach.19 The successes of identifying the causative variation for NIID indicated that LRS might be a powerful tool for genetic diagnosis of human diseases when conventional NGS fails to yield a positive diagnosis.

This study had some limitations that need to be explicitly acknowledged. First, the precise numbers of GGC repeats were not calculated in our RP-PCR, similar with the study in the Japanese patients, which caused an inaccessibility to analyse the relationship between the repeat numbers and phenotypic severity. Second, due to funding limitations, we did not perform the LRS genome screening in all of the family individuals. Third, most parents of sporadic patients were unavailable to be screened for GGC repeats, so the origin of variation was uncertain; nonetheless, the parents of juvenile-onset NIID (patient 6) were healthy but the mother carried a certain amount of GGC repeats as detected by RP-PCR although the number of repeats did not reach the pathogenic level (online supplementary figure 2), indicating a possibility of genetic anticipation in NIID.

Supplemental material

In summary, our study identified that repeat expansions in the 5′UTR of the NOTCH2NLC gene were responsible for juvenile-onset and adult-onset NIID in Chinese patients. In the preparation process of this manuscript, a newly published study by another Chinese research group also identified that GGC repeat expansions were the genetic cause of adult-onset NIID.20 Our study suggested that LRS was an effective tool for molecular diagnosis of genetic disorders, especially for neurological diseases that cannot be diagnosed by conventional clinical NGS technologies.


We appreciated the cooperation of the patients and their families.



  • JD and MG contributed equally.

  • Contributors DJH and ZXW conceived the research, designed studies and supervised the project; JWD, MLG and JXY designed and carried out experiments and analysed data; SY, MZ, PF, XFY, JZ, FL, JB, WS, YNH, YY, DJH and ZXW contributed to the clinical diagnosis and biopsy of NIID patients; YM, PDL, YNS and JH analysed the LRS data; JWD, DJH and ZXW wrote and edited the manuscript. All authors read and approved the final manuscript.

  • Funding This work was supported by the National Natural Science Foundation of China under Grant No. 81460199, 81870996 and 81571219.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval The research was approved by the ethics committee of Peking University First Hospital and Peking University People’s Hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.