Missense Variation in TPP1 Gene causes Neuronal Ceroid Lipofuscinosis Type 2 in a Family from Jammu and Kashmir-India

We report diagnosis of Neuronal Ceroid Lipofuscinosis Type 2 (CLN2), a rare, hereditary neurodegenerative disease of childhood, in a four and a half year old girl, the first child of nonconsanguineous parents with no family history. Despite extensive efforts by the parents, her clinical condition remained undiagnosed and without management, until recently. Our published “Bottom-up Approach”, based on comprehensive and multidisciplinary clinical, pathological, radiographical and genetic evaluations, played key role in diagnosis of the disease. Detailed analyses involving Next Generation Sequencing confirmed a missense variation NC_00011.10:g.6616374C>T (NP_000382.3:p.Arg339Gln; rs765380155) in exon 8 of TPP1 gene. In silico analyses predicted it to be highly pathogenic. Further family screening (including her both unaffected parents and asymptomatic, one year old younger sister) of the identified variation through Sanger Sequencing, revealed a perfect autosomal recessive segregation in the family. This study is the first case report on classic CLN2 from Jammu and Kashmir-India. This study is also indicating the effectiveness of our “Bottom-up Approach” in understanding rare disorders in low resource regions and the importance of timely diagnosis. Like in the proband, had diagnosis been established a bit early, the family might have benefitted at least with reference to their second child through counselling programmes.


INTRODUCTION
Neuronal ceroid lipofuscinoses (NCLs), also referred to Batten Disease, are a diverse group of hereditary, progressive neurodegenerative diseases which predominantly affect the children [1,2]. The NCLs share marked clinical abnormalities that are associated with neurodegeneration and its consequences including a combination of progressive visual impairment (leading to blindness), cerebellar ataxia, telencephalic manifestations including epileptic seizures and progressive dementia resulting in motor and mental deterioration, and poor prognosis with severe cases leading to premature death of the patients [2,3]. Morphological hallmarks of NCLs include selective neuronal degeneration primarily in the cerebral and cerebellar cortices and retina, which is associated with progressive cerebellar and cortical degeneration and secondary fiber tract atrophy [2,3]. Other pathognomonic feature of NCLs includes the abnormal intra-lysosomal accumulation of NCL-specific lipopigment residual bodies of heterogeneous origin in neurons and extra-neuronal tissues [1]. That is why, these group of neurodegenerative disease are categorized as "Lysosomal Storage Diseases (LSDs) [2]. The characteristic heterogeneous lysosomal storage materials include auto-fluorescent ceroid-lipopigments and lipofuscin, "subunit c of mitochondrial ATP synthase (SCMAS)", or "sphingolipid activator proteins (SAPs)" namely saponins A and D [2,4].
NCLs usually have an onset in childhood, occasionally as early as first few months of life, or adulthood. Based on the age-of-onset of clinical symptoms, NCLs have been broadly categorised into congenital, infantile, late-infantile, juvenile, adult and late-adult forms [2]. About thirteen NCLassociated proteins are known so far, each differing in their biological functions and intracellular localization which happens predominantly in the lysosomes (CLN1, CLN2, CLN3, CLN5, CLN7, CLN10, CLN11, CLN12 and CLN13), endoplasmic reticulum (CLN6 and CLN8) and cytosol (CLN4 and CLN14) [2]. Except CLN9, the genetic basis of rest of the NCL types is delineated [5].
Furthermore, NCLs can be subdivided according to a number of disease categories as mentioned in Table 1. NCLs are genetically heterogeneous with most of the clinical types following an autosomal recessive mode of inheritance, except one adult form [1,5]. NCL-associated variations are listed online in the NCL Mutation database (http://www.ucl.ac.uk/ncl/). Cathepsin D (CTSD) CLN10 100,000 individuals with an estimated rate of incidence of 1 in 100,000 births [6]. However, there is a geographical variation in the prevalence of NCLs. In the US, NCL prevalence is estimated to be 1 in 12,500 births or 300 to 350 new cases per year [11]. In some regions including Finland, Newfoundland and others, certain forms of NCL are relatively frequent [3,12,13]. However, there have been no epidemiological data on NCLs from India. Only a few NCL case reports from India have been published [8,[14][15][16][17][18]. These studies were mainly based on enzymatic and clinicopathological studies and have indicated CLN2, also known as Late Infantile Neuronal Ceroid Lipofuscinosis (LINCL), as the most common NCL-type.
In this study, we report diagnosis of autosomal recessive, classic CLN2 (OMIM# 204500) in a four and a half year old girl from Jammu and Kashmir (J&K) -India, ascertained by employing our previously published "Bottom-up Approach" [19]. Using a targeted Next-generation Sequencing (NGS) approach, the proband was found to harbour a pathogenic homozygous missense variation NM_000391 Sequencing, the same variation was confirmed in the proband and both of her parents were found to be carriers. It is pertinent to mention that this is the first NCL case report from J&K region -India.
Despite the onset of early symptoms in the proband almost 2-3 years back between the age of 2.5-3 and its progression into a severe disease, the case had remained undiagnosed owing to limited resources until recently when investigated comprehensively using our multi-disciplinary Bottom-up Approach [19]. Due to lack of diagnosis and appropriate counselling programmes in the region, unfortunately, the family did not undergo genetic counselling earlier. It was disheartening to find out upon genetic screening of the family that the younger, year old asymptomatic kid is also harbouring the same variation in a homozygous manner. With new developments in place for the pharmacological treatment of symptoms presented in CLN2 paediatric cases as well as enzyme replacement therapy (ERT) and gene therapy [20][21][22][23], it is anticipated if interventions are provided at this juncture too, may help the family extensively and the affected kids (in particular) in leading a better life.

CLINICAL INVESTIGATIONS
A four and a half year old girl (II-1; Figure 1    For participation in the study and collection of the blood samples, signed informed consents were obtained from the proband's parents. About 1-2 mL of whole peripheral blood samples were drawn and collected in ethylenediaminetetraacetic acid (EDTA) vacutainers from the proband (II-1), her father (I-1), mother (I-2) and her sister (II-2) for the molecular investigations.

Extraction of DNA
DNA was extracted from peripheral blood lymphocytes of the collected blood samples using Xpress DNA Blood Mini Kit (Cat. No.: MG17BI-S50/250; MagGenome Technologies, India) by following the manufacturer's protocol. QC-analysis of the extracted DNA samples was carried out through gel electrophoresis and spectrophotometric methods.

NGS Data Analysis
The raw data was requested and retrieved from the company through the family and after obtaining informed consent of the parents, detailed data re-analysis was performed. For the alignment of sequencing reads generated from the proband's genomic sample, the human reference genome hg38 was used. About 27,426,174 reads were mapped to the reference genome. Variants were filtered for their stringencies including evolutionary conservation, minor allele frequencies and pathogenicity predictions based on the public genomic databases such as the GERP++, gnomAD, 1000 Genomes Browser, and others. Filters were applied to the NGS data for the identification of variants with a minor allele frequency of ≤5%, variants classified as disease-causing in public databases, and variants with a minor allele frequency of ≤1% predicted to be loss-of-function variants. Genotype-phenotype correlations were then evaluated for each variant resulting from the filtering strategies. Only the variants with 30X or more coverage were considered for the clinical correlation. Non-synonymous and splice site variants found in the genes and relevant to the clinical symptoms were used for clinical interpretation. Silent variations that do not result in any change in amino acid in the coding region were not reported. Reanalyses of the NGS data for quality assurance was carried out by the pipeline developed by Key2Genes®, India and as previously described [24].

Validation of NGS findings using targeted PCR-based Sanger Sequencing
In order to validate the NGS findings and evaluate the inheritance pattern of identified variant in the recruited family, a targeted bi-directional PCR-based Sanger Sequencing (targeting the exonic region in which the variation was identified as well as its flanking exon-intron junctions) was performed twice for each participant (sequence of forward and reverse primers, and details of PCR reaction and amplicon purification are available upon request). Sanger sequencing of the amplicons were facilitated by Biodroid® Innovations, India. Sequence electrophoregrams were visualized with Sequence Scanner Software 2 (Applied Biosystems, US).

RESULTS
Detailed analyses of NGS data revealed that II-1 was harbouring a homozygous missense variation were found to be heterozygous carriers of the variation. This variation is already reported in limited CLN2 patients [32][33][34]. As the family had not known CLN2 as a genetic disease and did not undergo any genetic counselling programme, we proposed the parents for molecular screening of the variation in II-2. Unfortunately II-2, otherwise asymptomatic, was also found to be harbouring the same variation in a homozygous manner.
CLN2 is a rare progressive, fatal paediatric neurodegenerative disorder with typical onset of CNS and retinal manifestations between 2-4 years of age, progressing with worsening of clinical severity and leading to death by 6-15 years of life or later [37]. Clinical manifestation of CLN2 is usually heralded by partial, generalized tonic-clonic or secondarily generalized seizures, followed by ataxia, dementia, epilepsy, myoclonus, loss of vision (due to macular and retinal degeneration and optic atrophy) and speech, cerebellar, pyramidal and extra-pyramidal signs, and developmental regression, leading to death in the second decade of life [1,37,39]. The patients suffer loss of ambulation due to motor regression accompanied with progressive deterioration of speech and vision, resulting in inability to walk or sit unsupported and speak and blindness. The CLN2-associated neurohistopathological features include brain atrophy, neuronal loss and lysosomal accumulation of curvilinear bodies and auto-fluorescent storage material rich in SCMAS in neurons and other cell types in the CNS and peripheral tissues [40,41]. The accumulation of SCMAS in late-infantile and juvenile NCL has been described as an epiphenomenon and is not considered as disease-specific [42]. However, the underpinning biological mechanism for CLN2-associated neuronal and retinal degeneration has not yet determined. Despite CLN2-associated clinical heterogeneity/penetrance issues have not been reported in the literature, yet a few number of atypical CLN2 cases with late age-of-onset and prolonged disease course are reported worldwide [43,44]. Although CLN2 is an ultra-rare disease with an estimated worldwide incidence of 1:100,000 live births, it is the most common form of NCL among paediatric cases [3,45].
A precise diagnosis of CLN2 is quite challenging during early stages of its precipitation and is often delayed until its significant progression in severity, leading to complex diagnostic odyssey characterized by several misdiagnoses in a clinical set up [23]. Delayed diagnosis or misdiagnosis of a disease usually results in disjointed care and treatment delay for the patients. Although brain imaging provides relatively non-specific findings, marked neuroradiological signs of cerebellar atrophy are observed at early disease stage in CLN2 patients [9]. However, the characteristic neuroelectrophysiological findings in EEG include a triad consisting of early presentation of a typical photo-paraoxysmal response (PPR) known as "paraoxysmal spike-wave response" against low frequency intermittent occipital photic stimulation (using flash rates at 1 to 2 Hz), grossly enhanced SSEPs and a late presentation of grossly enhanced cortical VEPs and diminished EEG [39,46,47]. For the confirmation of a CLN2 case suspected through differential diagnosis, the demonstration of decreased TPP1 enzyme activity (in leukocytes, fibroblasts or dried blood spots) through enzyme activity assays and molecular detection of TPP1 (or CLN2) biallelic pathogenic variations have been regarded as the gold-standard diagnostic approaches [48].
The clinical management of CLN2 is complex, which primarily encompasses supportive and palliative care, pharmacological treatment of symptoms presented in CLN2 paediatric cases, enzyme replacement therapy (ERT) and gene therapy [20][21][22][23]. Antiepileptic drugs, including benzodiazepines (clobazam/clonazepam, lamotrigine, levetiracetum and valproate, are the common first-line therapeutics used for the management of epileptic as well as non-epileptic seizures and myoclonus [22]. A set of tailored inter-disciplinary therapies including physical, occupational, speech/orofacial myofunctional, nutritional, ophthalmologic, behavioural and complementary therapies are also sometimes sought for the management of CLN2-associated gait issues, pain, motor disturbances, retinal degeneration, sleep disorders, speech problems and behavioural issues in affected children as described in the literature [22]. For the treatment of symptoms in the proband, anti-epileptics including briviact (50 mg/day), rivotril (0.5 mg/day) and valparin oral solution (4 mL twice a day), and supplements including memocart oral solution (3 mL twice a day) and omilcal suspension (10mL/day) were clinically recommended. For the management of movement disorder, physiotherapy was recommended. Administration of a recently FDA-and EMA approved enzyme replacement therapy for CLN2, based on a recombinant proenzyme form of human TPP1 "cerliponase alfa", [20] was also recommended, an early adoption of which could essentially aid in the long-term management of the disease as well as prevention of further disease progression in II-1 and prevention of precipitation of CLN2-associated clinical manifestations in her younger asymptomatic sibling (II-2).
More than 150 distinct CLN2-associated TPP1 gene variations have been reported so far, information on which is available in the NCL Disease database (https://www.ucl.ac.uk/ncl-disease/mutation-and-patient-database/mutation-and-patient-datasheets-human-ncl-genes/cln2-tpp1). There is a marked allelic heterogeneity associated with CLN2. Different types of associated TPP1 genetic changes include several missense, non-sense and splice-junction variations, and frame-shift changes including single-nucleotide insertions and small deletions [49]. Many of these changes have been reported to cause either reduced activity or inactivation of hTPP1 due to its disrupted folding, intracellular processing, and trafficking [4,50,51]. It has been estimated that nearly half of the reported CLN2 patients harbour homozygous changes in TPP1 gene, while the remaining patients are compound heterozygous for different variations [34].
These variations are particularly common in CLN2 patients of North American and European descent [53]. The second most common TPP1 variation in North America is c.851G>T (p.G284V) [53,54].
In order to deduce the functional nature of NC_00011.10:g.6616374C>T (p.R339Q) variation, MDS for wild-type R339 and variant Q339 hTPP1 protein models were performed. On subjecting the MDS trajectories of these models to RMSD, Rg, RMSF and SASA analyses, the overall results indicated slight conformational changes in Q339 model rendering it relatively less compact in structure and more flexible in comparison to the wild-type R339. It is to be noted that conformational changes induced in a protein molecule by an amino acid change can usually impact its overall molecular dynamicity, function, its folding into a functional secondary structure, and functional interaction between its amino acid compositions and the surrounding environmental factors or molecules. All these can be simulated in silico through MDS indicating the RMSD, RMSF, Rg and SASA profiles of a protein model. SASA analysis of a protein molecule can be highly essential for the functional annotation of a disease-associated variant, in the sense that pathogenicity is highly associated with the buried property of a protein molecule and should always be considered while performing MDS for a potentially pathogenic protein variant [57]. Since in our case study, SASA analysis revealed hTPP1 Q339 variant to be slightly buried than R339, the variation can be considered as potentially pathogenic in nature. Intermolecular hydrogen bond analysis depicted more hydrogen bonds in the catalytic site of Q339 variant in comparison to the wild-type R339 hTPP1 protein model. Studies have shown that Ser 475 amino acid residue is the active site nucleophile, which along with Asp 360 and Asp 517 residues are essential for the catalytic activity of TPP1 enzyme [58]. Besides, Glu 272 , Asp 276 and Ser 495 residues in human TPP1 are also crucial to its catalytic activity [4]. The additional hydrogen bond formation in the catalytic site between D 276 (Asp 276 ) and S 475 (Ser 475 ) residues of the variant Q339 can be considered as contributory to its compromised biological function. Furthermore, the findings of our studies suggest that the Arg 339 residue is also crucial for the enzymatic activity as well as overall molecular dynamics of the hTPP1 protein.
Studies have shown that newly synthesized misfolded proteins including lysosomal hydrolases or variants resulting in protein-misfolding undergo oligomerization in endoplasmic reticulum (ER) and are later sorted for ER-associated degradation (ERAD) which includes retro-translocation from ER to the cytoplasm for their ubiquitin/proteasomal degradation, resulting in their reduced half-life [51,59,60]. Overall MDS analyses of hTPP1 R339 and Q339 protein models in our study have revealed slight conformational discrepancies in Q339 which can likely cause a slight misfolding of the hTPP1 variant Q339, rendering it to be either functionally compromised or resulting in a faulty cellular processing and, subsequently, contributing pathogenically towards the disease manifestation in this case. However, further in vivo studies are required to prove the potential functional aspects (protein misfolding, low half-life, residual or no activity, excessive secretion to the extra-cellular space) of hTPP1 Q339 variant, which is beyond the current context of this study.
In general, CLN2 gene encodes tripeptidyl-peptidase 1 (TPP1, EC 3.4.14.9), a biologically crucial and ubiquitously expressed lysosomal endopeptidase that belongs to a family of S53-type of serinecarboxyl proteinases ubiquitously expressed in mammalian tissues (including of human, rat and mouse) [41,[61][62][63]. It has been demonstrated that the expression of TPP1 in human cerebral cortex usually increases with development and reaches the optimal levels after the age of 2 [64]. However, studies have demonstrated that fibroblast of CLN2 patients harbouring pathogenic TPP1 variations have less than 5% of TPP1 activity [65]. Similar findings were also observed through fluorimetric assay for TPP1 enzymatic activity in the proband. Some TPP1 variations that result in atypical disease form probably do not completely abolish the function of the encoded TPP1 protein.
TPP1 is synthesized as a 563 amino acid containing inactive proenzyme or precursor known as "pro-TPP1" that when exposed to in vivo low pH conditions in the lysosome, gets auto-proteolytically processed into an active 368 C-terminal amino acids containing TPP1 [65][66][67][68]. During this maturation process, the initial signal sequence (19-amino acids) and further propeptide (176 amino acids) are removed from the pro-TPP1. Studies have shown that the autocatalytic processing of pro-TPP1 into active TPP1 is a Ca 2+ -dependent process [61]. The pro-TPP1 has an apparent Mr of 67 kDa, whereas that of the mature TPP1 is 46 kDa [67]. The amino-terminus of mature TPP1 begins at residue L196 [67]. The mature TPP1 contains a catalytic domain and five N-glycosylation sites [69]. The function of the active TPP1 protease is to sequentially cleave off tripeptides from the free unsubstituted aminotermini of oligopeptides during lysosomal protein catabolism [35,61,65]. Several in vitro studies have demonstrated that TPP1 degrades certain hormones and neurotransmitters such as Angiotension II and III, β-amyloid, a Bcl2-interacting protein named Bid, cholecystokinin, glucagons, Neuromedin B, and Substance P, as well as SCMAS (that gets accumulated in most forms of NCL) [41,[70][71][72]. However, all the in vivo substrates of active TPP1 are still unknown. Experimental studies have demonstrated that 3,4-dichloroisocoumarin and diisopropyl fluorophosphate inhibit the activity of TPP1 [67].

CONCLUSIONS
A four and a half year old girl presenting signs of neuroregression and an epileptiform disorder was diagnosed with CLN2 using a combinatorial approach named "Bottom-up Approach" [19]. Through a genotype-phenotype correlation based approach followed by sequencing data analyses and in silico predictions, the proband was diagnosed with CLN2 associated with a potential loss-of-function variation NC_00011.10:g.6616374C>T (rs765380155) in exon 8 of TPP1 gene. The relevance of this study in the current context is that it would add information to the clinical literature and mutational landscape of CLN2-associated TPP1 gene, particularly that of the population of J&K -India.
The authors highly recommend a comprehensive evaluation of paediatric cases suspected with CLN2 presenting with global development delay, seizures, loss of ambulation, ataxia and neuronal manifestations like brain atrophy and inclusion of CLN2-associated clinical guidelines to clinical specialists in their routine practice [73]. A combined clinical, pathological and molecular approach like "Bottom-up Approach" [19] for the precise delineation of rare diseases such as clinically suspected, paediatric NCL cases is highly recommended. Altogether this would ensure an early precise diagnosis of the disease (as depicted through this case report), thereby, facilitating an earlier clinical management or treatment to the symptomatic as well as even pre-symptomatic individuals using available interventional approaches. An early diagnosis would allow timely genetic counselling to be offered to the families of the patients, thereby, enabling genetic screening of the suspected families and their close-relatives as well as pre-implantation genetic diagnosis of the suspected couples or prenatal diagnosis through enzyme assay, identification of typical NCL inclusions through microscopy and genetic screening. In the current study, an immediate genetic counselling of the proband's family had prompted genetic screening of her parents and younger sibling which has lead to the detection of familial segregation of the identified variation. It is to be mentioned that genetic screening has also lead to the delineation of CLN2 in proband's younger sister, even before the

Conflict of Interest
The authors declare no conflict of interest associated with the manuscript. For declaration purposes, SS is the founder and chief scientific advisor of Biodroid® Innovations Private Limited, India.

Ethics and Reporting
The study was approved by the IERB, SMVDU, J&K, India. Authors declare at manuscript submission that all relevant ethical guidelines have been followed, all necessary IRB and/or ethics committee approvals have been obtained, all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived.

Funding Statement
The authors have nothing to declare.