Novel homozygous CLN3 missense variant in isolated retinal dystrophy: A case report and electron microscopic findings

Abstract Background Biallelic CLN3 gene variants have been found in either juvenile‐onset neuronal ceroid lipofuscinosis (JNCL) or isolated retinal dystrophy. It has been reported that most JNCL patients carry a common 1.02‐kb deletion variant homozygously. Clinical characteristics of patients with biallelic CLN3 missense variants are not well elucidated. Methods We described a 26‐year‐old Japanese male patient with isolated retinal dystrophy. Whole‐exome sequencing (WES) and transmission electron microscopy (TEM) were performed. Results Whole‐exome sequencing identified a novel homozygous CLN3 missense variant [c.482C>T; p.(Ser161Leu)]. Ophthalmoscopy revealed retinal degeneration and macular atrophy, and later attenuated retinal vessels. Severely reduced responses were observed in both rod and cone electroretinograms. In TEM of the patient's lymphocytes, fingerprint profiles, which are specific findings in CLN3‐associated JNCL, were observed in 16/624 (2.56%) lymphocytes of the patient, who has never exhibited neurological signs during the 13‐year follow‐up period. Conclusion Our results indicated that this novel CLN3 missense variant is associated with teenage‐onset isolated retinal dystrophy. This is the first report of any patient with CLN3‐associated disorder in the Japanese population. Although fingerprint profiles have never been reported in CLN3‐associated isolated retinal dystrophy, these profiles were observed, albeit infrequently, suggesting that frequency of the fingerprint profiles might depend on variant types.


| INTRODUCTION
Juvenile-onset neuronal ceroid lipofuscinosis (JNCL, Batten disease, OMIM #204200) is a heterogeneous group of autosomal recessive neurodegenerative lysosomal storage disorders characterized by early-onset retinal degeneration, epilepsy, and progressive psychomotor deterioration. The CLN3 gene encodes a lysosomal/endosomal transmembrane protein of 438 amino acids and is ubiquitously expressed in human tissues (Mirza et al., 2019). Biallelic CLN3 (OMIM *607042) variants have been associated with not only JNCL but also early-or late-onset isolated retinal dystrophy without neurological signs (Chen et al., 2019;Ku et al., 2017;Lerner et al., 1995;Munroe et al., 1997;Wang et al., 2014). A previous study revealed that JNCL is associated with a common founder 1.02-kb deletion variant of CLN3, demonstrating homozygous states in 74% (139/188 cases), compound heterozygous states in 22% (41/188 cases), and other biallelic variants in only 4% (8/188 cases) of patients (Munroe et al., 1997). To date, CLN3-associated JNCL or isolated retinal dystrophy has never been reported in the Japanese population. Here, we described long-term ophthalmological observation of a Japanese patient with isolated retinal dystrophy, identification of a novel homozygous CLN3 variant, and characteristic electron microscopic findings.

| MATERIALS AND METHODS
The Institutional Review Boards of The Jikei University (approval no. 24-232 6997) and National Hospital Organization Tokyo Medical Center (R14-050) approved the study protocol. The protocol adhered to the tenets of the Declaration of Helsinki, and informed consent was obtained from participants.

| 1 Clinical assessment
A 26-year old-male patient (II-2, JU#0519, Figure 1a) from a Japanese family (JIKEI-010JIKEI) was ophthalmologically assessed at The Jikei University Hospital. We performed a comprehensive ophthalmic examination, including decimal best corrected visual acuity (BCVA), slitlamp examination, funduscopy, fluorescein angiography, and fundus autofluorescence imaging (FAF) using an Optos 200Tx, Ultra-Wide Field Retinal Imaging System (Optos), and spectral-domain Cirrus optical coherence tomography (OCT; Carl Zeiss Meditec AG). Horizontal B-scan OCT images through the fovea were obtained. Visual field testing using Goldmann perimetry (Haag Streit) was performed. Full-field electroretinography (ERG) using a Ganzfeld dome and Neuropack 2 (Nihon Kohden) with corneal contact lens electrodes was recorded by an EOG-ERG Ganzfeld stimulator (Electrophysiology system; LACE Elettronica) according to the protocol of the International Society for Clinical Electrophysiology of Vision (McCulloch et al., 2015). Details on the procedure and conditions have been previously reported (Katagiri et al., 2016;Kutsuma et al., 2019).

| 2 Molecular genetic study
Genomic DNA was extracted from venous blood samples of all participants (Figure 1a). Whole-exome sequencing (WES) was performed as previously described (Katagiri et al., 2018). Whole-genome sequencing (WGS) was also performed by Macrogen Japan. Briefly, paired-end sequence libraries were constructed using the TruSeq DNA PCR Free Library Kit (Illumina). The libraries were subjected to 150-bp paired-end sequencing on the HiSeq X Ten system (Illumina). Reads from the fastq files were mapped/aligned to the human reference genome GRCh37/hg19 using Burrows-Wheeler Aligner Software. Duplicate reads were removed using Picard Mark Duplicates. The mapped reads around insertions/deletions (Indels) were realigned by the Genome Analysis Toolkit (GATK) Version 3.0. Variant calling was performed using GATK HaplotypeCaller. The called single-nucleotide variants (SNVs) and Indels were annotated using the SnpEff software 4.3. We focused on nonsynonymous variants and splice site variants, which are within 10 bp of exon-intron boundaries, and excluded synonymous and noncoding variants. We removed variants with allele frequency greater than 0.01 for recessive variants or 0.001 for dominant variants in any of the ethnic subgroups found in the following databases: 1,000 Genomes database (https://www.inter natio nalge nome.org/), Genome Aggregation Database (https://gnomad.broad insti tute.org/), and Human Genetic Variation Database (http:// www.hgvd.genome.med.kyoto -u.ac.jp/).
In all available family members, we employed targeted next-generation sequencing (NGS) and performed hybridization capture (xGen Predesigned Gene Capture Pools, Integrated DNA Technologies) experiments using CLN3 exon-specific oligonucleotide probes (Table S1), and amplicon sequencing using NextSeq 500 (Illumina), with an average depth of coverage of 700× for all targeted amplicons. The experiments were performed basically based on previously described methods (Fujiki et al., 2018) at Kazusa DNA Research Institute, Chiba, Japan.
A potential pathogenic variant in the CLN3 gene was confirmed by Sanger sequencing for cosegregation in all available family members. The following primer set for exon 7 of the CLN3 gene was used: forward primer 5′-GTCTGATAACTGGGTGGATG-3′ and reverse primer 5′-AAGGAGAACACAGGAACATTC-3′. The region around the identified CLN3 variant was visualized using Integrative | 3 of 9 MIZOBUCHI et al.

| 3 Transmission electron microscopy analysis
Sample preparation from the patient's peripheral lymphocytes for transmission electron microscopy (TEM) was performed based on previously described methods Kuniyoshi et al., 2020). Finally, ultrathin sections of about 70 nm were stained with uranyl acetate and lead citrate and observed on a JEM1400 Plus electron microscope (JEOL) at 100 kV.

| 1 Case presentation
The 26-year-old patient was referred to The Jikei University hospital for the assessment of poor visual acuity. He manifested night blindness in his early teens. His parents, who were nonconsanguineous and unaffected, were from the same prefecture in Japan. His decimal BCVA was 0.7 (Snellen equivalent 20/28, spherical − 4.50 diopter) in the right eye and 0.2 (Snellen equivalent 20/100, spherical − 4.25 D) in the left eye. No remarkable finding was observed in the anterior segments or media in both eyes (BE). Fundus photographs showed retinal degeneration with pigmentation and mild macular involvement in BE (Figure 2a). Late-phase fluorescein angiography 10 min after injection showed hyperfluorescent patterns due to transmission defects in the maculae and outside vascular arcades in BE (Figure 2b), indicating degenerative changes in the retina and retinal pigment epithelium. GP showed ring scotoma patterns with decreased central sensitivities in BE (Figure 3a). Full-field ERG findings revealed severely decreased responses in rod (dark-adapted [DA] 0.01) ERG, in combined/rod-plus-cone (DA 3.0) ERG, cone (light-adapted 3.0) ERG, and 30-Hz flicker ERG (Figure 2c). Follow-up observation revealed progressive deterioration of visual function. At 38 years of age, BCVA was 0.2 (Snellen equivalent 20/100) in BE. GP revealed that the visual fields were progressively constricted in BE (Figure 3b). Ultrawide-field fundus photographs and fundus autofluorescence showed severe retinal degeneration with attenuated retinal vessels in BE (Figure 2d,e). OCT showed loss of the central macular outer layers in BE (Figure 2f). During the 13-year follow-up period, the patient has never exhibited neurological signs seen in protracted JNCL (Munroe et al., 1997). In protracted JNCL patients, the first progressive loss of vision occurs, with delayed neurologic onset, usually in the third or fourth decade.

| 2 Molecular genetic findings
After filtering WES data, 11 variants were investigated based on the patient's phenotype and inheritance pattern. Eventually, we identified a novel homozygous variant [c.482C>T; p.(Ser161Leu) in exon 7 (GRCh38chr16: 28486629)] of the CLN3 gene (Figure 1b). The heterozygous CLN3 variant was found in his unaffected mother and brothers (Figure 1a,b). Amino acid sequence alignments showed that the Ser161, located on the second luminal loop (Mirza et al., 2019), is well conserved among vertebrates (Figure 1c), suggesting functional importance for the residue. The variant was not found in the Genome Aggregation Database (https://gnomad. broad insti tute.org/) or the Human Gene Mutation Database (http://www.hgmd.cf.ac.uk/ac/index.php). According to the American College of Medical Genetics guideline (Richards et al., 2015) using the VarSome program (https://varso me.com/), p.(Ser161Leu) was classified as likely pathogenic based on the following lines: PM1 (hot spot of length 61 base-pairs has 5 non-VUS coding variants (5 pathogenic and 0 benign), pathogenicity = 100.0%), PM2 (the variant was not found in GnomAD exomes), PP2 (24 of 41 missense variants in CLN3 were pathogenic), and PP3 (pathogenic computational verdict because nine pathogenic predictions from DANN, DEOGEN2, EIGEN, FATHMM-MKL, M-CAP, MVP, MutationAssessor, MutationTaster, and SIFT vs. two benign predictions from PrimateAI and REVEL). Unexpectedly, the p.(Ser161Leu) variant was present within the region normally deleted in the common 1.02-kb deletion. A question was raised whether his deceased father might have carried the common 1.02-kb deletion heterozygously. To resolve this issue, we performed WGS and targeted NGS to exclude presence of the heterozygous 1.02-kb deletion in the patient. The coverage track on IGV revealed no decreased read depth around the p.(Ser161Leu) variant (Figure 1d), indicating no evidence of the heterozygous 1.02-kb deletion. Sequence read alignment data on IGV also denied the possibility of presence of the 1.02-kb deletion (Figure 1d). Homozygosity mapping using WGS showed a high rate of homozygosity (>80%) of single nucleotide variants around the region including the p.(Ser161Leu) variant ( Figure S1). In addition, targeted NGS showed to be negative for any copy number variation including heterozygous deletion in each CLN3 exon in the patient, as well as other family members and a control ( Figure S2). These results confirmed that the patient carried the CLN3 variant [p.(Ser161Leu)] homozygously.

| 3 Transmission electron microscopy
Transmission electron microscopy was next used to evaluate the patient's lymphocytes. Generally, each lysosome is surrounded by a limiting membrane and contains various electron-dense materials. Fingerprint profiles, which indicate lysosomal storage material, were observed in the patient (Figure 4). The fingerprint profiles, which are specific findings seen in CLN3-associated JNCL, were observed in 16/624 lymphocytes (2.56%) analyzed.

| DISCUSSION
In this study, we reported a patient with onset of night blindness in his teens and decreased visual acuity that progressively deteriorated over 13 years of follow-up, who was found to carry a novel homozygous variant in the CLN3 gene.
To date, 38 CLN3 variants, which are associated with JNCL or isolated retinal dystrophy, have been reported (Chen et al., 2019;Kousi, Lehesjoki, & Mole, 2012;Ku et al., 2017;Lerner et al., 1995;Munroe et al., 1997;Sarpong et al., 2009;Wang et al., 2014;Wisniewski et al., 1998). As for genotype-phenotype correlations, the homozygous 1.02-kb deletion has been found only in the most severe CLN3-associated phenotype, JNCL (Lerner et al., 1995;Munroe et al., 1997). In contrast, to date, 16 patients with biallelic CLN3 variants and isolated retinal dystrophy have been reported (Chen et al., 2019;Ku et al., 2017;Wang et al., 2014). The genetic details of these 16 patients are as follows (Table 1): seven patients (43. 8%,Cases 2,3,5,6,7,8,9) had the 1.02-kb deletion and missense or splice site variants, three patients (18.8%, Cases 12, 13, 14) had nonsense and missense variants, one patient (6.2%, Case 1) had splice site and missense variants, and five patients (31.2%,Cases 4,11,15,16,17) had biallelic missense variants. Only 11 of these 16 patients were ophthalmologically described in detail previously (Chen et al., 2019;Ku et al., 2017), which were summarized in Table S2. Among the 11 patients, 2 patients (Cases 16 and 17) carried a homozygous missense variant [p.(Arg405Trp)] and exhibited a milder phenotype (late-onset isolated retinal dystrophy), including onset in their 30s, retinal degeneration without macular atrophy, preserved foveal ellipsoid zone in OCT, and no central scotoma (Ku et al., 2017), compared with our patient, who exhibited onset in his teens, progressive macular atrophy and central scotoma in GP, and severely decreased ERG responses (Figure 2c). Fingerprint profiles in TEM, which indicate lysosomal material storage, are considered a specific finding of JNCL (Santavuori, 1988). Moreover, previous studies have revealed that fingerprint profiles were not observed in three patients with isolated retinal dystrophy (Ku et al., 2017). However, our patient exhibited fingerprint profiles in lymphocytes, albeit infrequently (2.56% of analyzed lymphocytes) (Figure 4), which might explain why the patient did not manifest neurological signs. In the previous study (Ku et al., 2017), if more lymphocytes had been examined by TEM, the fingerprint profiles might have been seen in the three patients. It is speculated that low frequency of the fingerprint profiles may be associated with isolated retinal dystrophy without neurological signs. In CLN3-associated isolated retinal dystrophy, further TEM analysis is needed for elucidating frequency of fingerprint profiles. The above-mentioned consequences suggest that the frequency of fingerprint profiles might be positively correlated with disease severity.
As for the phenotype, although our patient did not exhibit any neurologic signs during follow-up, two scenarios are considered in future. One possibility is that the patient has lifelong isolated retinal dystrophy, the other possibility is that the patient would present a protracted phenotype, manifesting first retinal dystrophy long before the manifestations of NCL become apparent.
In conclusions, we described clinical and genetic characteristics of a Japanese patient with teenage-onset isolated retinal dystrophy in whom a novel homozygous CLN3 missense variant was identified. This is the first report of a CLN3associated genetic disorder in the Japanese population. TEM F I G U R E 4 Transmission electron microscopic findings of four different lymphocytes. Images on the right panel are magnified images of circled areas in images on the left panel. Fingerprint profiles in each lysosome surrounded by a limiting membrane are present and depicted by arrows findings of the patient's lymphocytes revealed characteristic fingerprint profiles, albeit infrequently. In consideration of genotype-phenotype correlations, severities of CLN3associated phenotypes might depend on variant types.