Background

The neuronal ceroid lipofuscinoses (NCLs) are a group of inherited lysosomal storage disorder causing severe neurodegeneration due to neuronal loss (brain and retina), and accumulation of lipopigments in many cell types, including neurons. The NCLs incidence rate worldwide is 1 to 8 in 100,000 live births [1].

The NCLs share common clinical presentations like epilepsy, loss of motor and cognitive function, visual impairment, and premature death [2]. Though usually observed in childhood, the age of onset of the disease varies. Considering this, NCLs were initially classified into four groups – infantile (INCL-Haltia-Santavuori disease), late infantile (LINCL-Jansky-Bielschowsky disease), juvenile (Batten-Spielmeyer-Vogt disease) and adult (Kufs disease) [3]. Eventually, allelic heterogeneity in NCLs was identified due to an advancement in biochemical and genetic techniques and hence a new approach of molecular classification and diagnostic algorithms was designed [4, 5].

Until now, 14 types of NCL are identified (NCL1-NCL14) however the most commonly observed form are NCL1, NCL2, and NCL3 [6]. These NCLs subtypes are autosomal recessively inherited except for NCL4B, inherited in an autosomal dominant form [7]. An exception to the above inheritance pattern is a uniparental disomy case in NCL which occurred due to complete isodisomy of chromosome 8, leading to homozygosity of a maternally-inherited deletion in NCL8 [8]. Until May 2015, total 515 changes in 13 human genes have been reported in NCLs [9].

NCL1 (OMIM#256730) representing the early infantile disease, results due to a mutation in PPT1 (palmitoyl-protein thioesterase 1; OMIM*600722) gene located at 1p34.2. The gene codes for a lysosomal enzyme called palmitoyl-protein thioesterase 1 (PPT1) whose function is to remove fatty acids attached in thioester linkages to cysteine residues in the protein. The downstream effect of PPT1 deficiency involves deregulated cellular processes like vesicular trafficking, synaptic function, lipid metabolism, neural specification, and axon connectivity [10]. In addition, a study by Lyly et al. established that an alteration in cholesterol metabolism and ectopic F1-ATP synthase resulted due to PPT1 deficiency [11]. According to the NCL mutation and patients’ database 76 changes have been reported in PPT1 gene [9]. Amongst these, the mutation p.Arg122Trp has a founder effect in Finnish population with NCL1 [12]. This mutation causes a defect in the transport of the PPT1 from the endoplasmic reticulum to lysosomes [12, 13]. The mutation p.Thr75Pro and p.Leu10Ter have a founder effect in Scotland [14].

NCL2 (OMIM#204500), representing the late infantile disease, results due to a defect in TPP1 (tripeptidyl peptidase I; OMIM*607998) gene at the locus 11p15.4. This gene encodes the instruction for making the lysosomal enzyme called tripeptidyl peptidase 1 (TPP1). TPP1 deficiency results in accumulation of ceroid lipofuscin, an autofluorescent storage material, in cell’s lysosomes. Total 140 disease-causing mutations are in the TPP1 gene of the patients with NCL2 [9]. The most common TPP1 gene mutations are c.509-1G > C and p.Arg208Ter [14]. The mutation p.Gly284Val seems to be predominant in Canada suggesting a possible founder effect [14].

The genetics of NCL1 and NCL2 remain unknown in India. Hence, the aim of the present study is to identify the molecular spectrum and common molecular marker of these diseases in Indian patients. The study also aims to support the correlation between the null or reduced enzyme activity and the mutations causing disease and clinical phenotype in NCL1 and NCL2 patients.

Methods

Patients

The patients in the present study were the clinical cases referred by pediatric neurologist and pediatricians from collaborating centres. The study is in accordance with the tenets of the Helsinki Declaration. The Ethics committee of the Foundation for Research in Genetics and Endocrinology (FRIGE) at the Institute of Human Genetics approved the study. As per the institutional ethics committee guidelines, a written informed consent for investigation and publication of the data was obtained from the parents/guardian of the patients. The 34 unrelated patients (22 males and 12 females) presented common clinical indications like epilepsy, cerebral atrophy, and cerebellar atrophy. They were in the age range of 4 months to 9 years at the time of investigations and were referred from different geographical/ethnic background in the time from 2015 to 2017 with a clinical suspicion of Batten disease (12 patients with NCL1, and 22 patients with NCL2). Total 16 patients (47.05%) had parental consanguinity. Table 1 provides the clinical details and the demographic profile of the patients.

Table 1 Clinical details and demographic profile of the patients with Batten disease (NCL1 and NCL2)
Full size table

Biochemical investigations

Leukocyte and genomic DNA (gDNA) isolation was carried out from six milliliters of blood, drawn from each patient in ethylenediaminetetraacetic acid (EDTA) vacutainer.

Palmitoyl protein thioesterase 1 (PPT1) enzyme activity

The patients suspected with the NCL1 disease were investigated for the lysosomal enzyme Palmitoyl protein thioesterase 1 (PPT1; EC 3.1.2.22) activity using a previously described protocol [15]. In brief, the samples were incubated with tritinX 100, β-glucosidase, dithiothreitol and the substrate 4-Methylumbelliferyl-6-sulpho-palmitoyl-β-D-glucoside at 370C for 1 h. The fluorescence was measured at 4mu.

Tripeptidy peptidase l (TPP1) enzyme activity

The NCL2 suspected patients were investigated for the Tripeptidy peptidase l (TPP1; EC 3.4.14.9) lysosomal enzyme activity using a previously described protocol [16]. In brief, the samples were mixed with EDTA/acetate/TritonX 100 in chilled condition. The substrate Ala-Ala-Phe-7-amido-4-methylcoumarin was added on limited intervals and incubated at 370C for 1 h. Chloroacetate/acetate was added following the incubation and the fluorescence was measured at 460 nm.

Molecular investigation

DNA extraction

DNA isolation followed the standard salting-out method and was quantified using a QIAxpert (Cat. No: 9002340) from Qiagen [17]. Sample purification was performed using The Genomic DNA Clean & Concentrator™-25 (DCC™) Kit, from Zymo Research, Irvine, California, U.S.A (Cat. No. D4064) and were stored at −20 °C until investigation.

Single gene sequencing (PPT1 and TPP1 gene)

The PPT1 and TPP1 genes were amplified using primer sets listed in the Additional file 1. The protocol followed 37 cycles consisting of initial denaturation (94 °C; 5 min), denaturation (94 °C; 30 s), annealing (58 °C-67 °C; 45 s), elongation (72 °C; 1 min), and final elongation (72 °C; 5 min). The polymerase chain reaction (PCR) products were run on the 2.5% agarose gel and visualized under ultraviolet transilluminator.

The capillary electrophoresis technology driven fluorescent dye-labeled genetic analysis system performed the Sanger sequencing on the Applied Biosystems™ SeqStudio™ Genetic Analyzer with SeqStudio™ Data Collection Software using a previously described protocol [18].

Multiplex ligation-dependent probe amplification (MLPA) analysis

The procedure followed the manufacturer’s recommendations of using gDNA (100 ng) and P470-A1 NCL probe mix (MRC-Holland, Amsterdam, the Netherlands). DNA was denatured (98 °C; 5 min) and hybridized (overnight; 60 °C) with the SALSA probe mix P470. The samples were then Ligase (54 °C; 15 min) and incubated at 98 °C for 5 min to stop the reaction. The PCR amplification was carried out with the specific SALSA FAM PCR primers. Amplified products were run on the ABI 3130 Genetic Analyzer (Applied Biosystems, USA) and the MLPA peak patterns were analyzed in control and test samples to detect the copy number differences of the exons.

In silico analysis

Prediction of the functional effect of the variants

The variants identified were looked up in public databases like The Human Gene Mutation Database (http://www.hgmd.cf.ac.uk), SNP database (http://www.ncbi.nlm.nih.gov/SNP/index.html) and UCL-London’s Global university database (http://www.ucl.ac.uk/ncl). The in silico tools like MutationTaster2, SIFT, FATHMM, PolyPhen2, PROVEAN, and MutationAssessor predicted the pathogenicity of the coding and non-coding DNA variants and amino acid substitution.

Homology modeling, structure validation and protein stability of the novel variants

The wild-type template crystallographic structures of the PPT1 (PDB ID: 1EI9) and TPP1 (PDB ID: 1EDY) were used to calculate the Root Mean Square Deviation (RMSD) of the novel mutant protein structures [19].

Orthologous conservation of the residues harboring the novel variant

Clastal Omega (an online multiple sequence alignment program) aligned the PPT1 (NP_000301) and TPP1 (NP_000382) protein sequence of Homo sapiens with different species to check the conservation of the residues incorporating novel variants [19].

Results

In the present study the clinical assessment, biochemical and molecular investigation confirmed the diagnosis of 34 patients with Batten disease (12 with NCL1 and 22 with NCL2). The most common clinical indications observed in these patients were epilepsy (91.2%). The neuroimaging study including Computed Tomography (CT scan) and/or brain Magnetic Resonance Imaging (MRI) from all the patients also revealed cerebral atrophy (64.7%), and cerebellar atrophy (79.41%).

Biochemical investigations

A significant deficiency of PPT1 and TPP1 enzyme activity was observed in the leukocytes of twelve patients with NCL1 and twenty-two patients with NCL2 respectively. The patients’ enzyme activity reduced to 0 to 2.8% compared to the control value (Tables 2 and 3).

Table 2 Biochemical and molecular alalysis of patients with NCL1 disease
Full size table
Table 3 Biochemical and molecular analysis of patients with NCL2 disease
Full size table

Molecular analysis

Bi-directional sanger sequencing (PPT1 gene)

The biochemical investigation confirmed 12 patients with NCL1. Sanger sequencing identified five homozygous missense variants, one frameshift variant, and one nonsense variant from total nine patients (Table 2).

Known mutation detected in the PPT1 gene

In total three known variants were revealed by Sanger sequencing. Patient P1 and P2 harbor the pathogenic homozygous mutations p.Val181Met in exon 6 and p.Asn110Ser in exon 3 respectively [20, 21]. Patient P3 carried compound heterozygous mutation p.Trp186Ter (known pathogenic) in exon 6 and a novel variant p.E178Nfs*13 in exon 5 [22].

Novel variants detected in PPT1 gene

Overall, four novel variants amongst nine patients responsible for NCL1 were detected (Fig. 1). Patient P4 to P7 were homozygous for the variant p.Pro238Leu in exon 7. Patient P8 and P9 presented homozygous missense variants p.Cys45Arg in exon 2 and p.V236G in exon 7 respectively.

Fig. 1
figure 1

Identification of novel variants in PPT1 gene. a Illustrative representation of the distributions of the novel variants identified in Indian NCL1 patients investigated in this study. b Sanger sequencing discovered one missense variant (p.Cys45Arg) in exon 2, two missense variants (p.Pro238Leu and p.Val236Gly) in exon 7, and one frameshift variation (p.Glu178Asnfs*13) in exon 5 of PPT1 gene. The common variant p.Pro238Leu was identified in 44% of the patients. The point of variation is indicated by an arrow. c The multiple alignment of the protein sequence surrounding the novel variants against various orthologous sequence revealed the conservative status of the wildtype residues (marked red).

Full size image

Three patients (P10, P11, and P12) did not carry any variation in the exon or exon-intron boundaries of PPT1 gene but their common clinical presentations like epilepsy, cerebral atrophy, and cerebellar atrophy indicated NCL. Their biochemical analysis depicting four times decrease in the PPT1 enzyme activity confirmed the NCL1 diagnosis. Since Sanger sequencing was uninformative, these patients were analyzed through MLPA. However, no large deletion/duplication was discovered in the PPT1 gene. In such cases, the possibility of deep intronic variations cannot be ruled out.

Bi-directional sanger sequencing (TPP1 gene)

As confirmed by biochemical investigation, 22 patients affected with NCL2 were analyzed for pathogenic variants in TPP1 gene. Overall, 15 variants, comprising seven missense variants, four frameshift variants, three nonsense variants, and one intronic variant were identified amongst 19 patients (Table 3).

Known mutations detected in TPP1 gene

The TPP1 gene analysis detected seven known mutations distributed amongst 11 patients. Patient P13 was identified with a heterozygous copy of the known mutation p.Met345Leu in exon 8 [23]. However, the second variant responsible for the disease was unidentified may be due to its presence in the deep intronic region or due to possible large deletion/duplication. The patient has reduced TPP1 enzyme activity and clinical phenotypes like epilepsy, regression of mental and motor milestone, choreoathetosis, cerebral atrophy, and cerebellar atrophy were in line with the diagnosis. Patient P14 and P15 carried the homozygous missense mutation p.Arg339Trp and p.Arg339Gln respectively in exon 8 [2, 24]. Patient P16 was detected with a nonsense mutation p.Phe516Ter in exon 12 [25]. Patient P17 to P21 harbor a homozygous missense mutation p.Arg206Cys in exon 6 [26]. Patient P22 suffered NCL2 due to a homozygous nonsense mutation p.Arg208Ter in exon 6 [27].

Novel variants detected in TPP1 gene

Sanger sequencing revealed eight novel variants amongst nine patients (patients P23 to P31). This includes four frameshift variants, two missense variants, one intronic insertion, and one nonsense variant (Fig. 2). Patient P23 was found with a compound heterozygous variants p.Asn286Ser (known pathogenic) and p.Ser62Argfs*19 in exon 7 and exon 3 respectively [28]. In case of the patient P24, a homozygous missense variant p.Arg152Ser was detected in exon 5. In the patient P25 and P26, the variant p.Tyr459Ser in exon 11 resulted in NCL2. A frameshift termination p.Ser153Profs*19 in exon 5 was identified in the patient P27. Sanger sequencing detected the patient P28 with a homozygous nonsense termination p.Tyr157Ter in exon 5. Patient P29 was compound heterozygous for the variants p.F230Sfs*28 in and p.Ile484Aspfs*7 in were detected in exon 7 and exon 12 respectively. An intronic variant g.2023_2024insT intron 4 was identified in the Patient P30 and P31.

Fig. 2
figure 2

Identification of novel and most common variants in TPP1 gene. a Illustrative representation of the distributions of the novel and most common variants identified in Indian NCL2 patients investigated in this study. b Sanger sequencing discovered one intronic variant (In4:g.2023_2024insT), three missense variants (Ex5:p.Arg152Ser, Ex11:p.Tyr459Ser and Ex6:p.Arg206Cys), four frameshift variants (Ex3:p.Ser62Argfs*19, Ex5:p.Ser153Profs*19, Ex7:p.Phe230Serfs*28, and Ex12:p.Ile484Aspfs*7), and one nonsense variant (Ex5:p.Tyr157*) in TPP1 gene. The common variant Ex6:p.Arg206Cys occurred in 26% of the patients. The point of variation is indicated by an arrow. c The multiple alignment of the protein sequence surrounding the novel variants against various orthologous sequence revealed the conservative status of the wildtype residues (marked red). However, the residue isoleucine at the position 484 was found to be conserved in four out of nine species.

Full size image

However, no variation in the exonic or exon-intronic boundaries of TPP1 gene was detected in three patients (P32, P33, and P34). These patients presented common clinical indications like epilepsy, cerebral atrophy, cerebellar atrophy, and visual impairment as indicated in NCL patients. Their biochemical analysis resulting in a significant decrease in the TPP1 enzyme activity confirmed the NCL2 diagnosis. Despite the uninformative Sanger sequencing results, these patients could not be analyzed through MLPA due to unavailability of enough samples. In such cases, the possibility of deep intronic variations cannot be rule out.

In silico analysis of the novel variants

The in silico tools described above established the functional effects of the variants identified [see Additional file 2]. The novel variants were found to be disease causing. These predicting tools suggest the probably damaging and deleterious effect of the novel variants on protein function. These variants were found neither in the 1000 Genomes database nor in the Exome Aggregation Consortium (ExAc). The protein sequence alignment of Homo sapiens along with other species using Clastal Omega-an online multiple sequence alignment program suggests that these variations occurred at highly evolutionarily conserved and functionally active residual domain in the protein (Figs. 1 and 2).

The protein homology modeling of the missense point variants in the PPT1 gene (p.Cys45Arg, p.Val236Glu, and p.Pro238Leu) and TPP1 gene (p.Arg152Ser and p.Tyr459Ser) suggest their damaging effect at highly conserved residues. The variant p.Cys45Arg is in close proximity to Met41 and could affect the active site of PPT1 either by decreasing the oxyanion stabilization, altering the binding pocket or disrupting the active site by perturbing the position of Met41. The variant p.Val236Glu and p.Pro238Leu are very close to catalytic site Asp233. This could affect the N-linked glycosylation process and cause conformational changes in the protein (Fig. 3).

Fig. 3
figure 3

Homology modeling of novel missense variants identified in PPT1 gene. The native structure (blue) and mutant structure (brown) are superimposed. a The model of p.Cys45Arg depicting the amino acid change from polar to basic at the codon number 45 (TGT-CGT). b The model of p.Val236Gly depicting the amino acid change from non-polar to acidic at the codon number 236 (GTG-GAG). c The model of p.Pro238Leu depicting the amino acide change from non-polar to hydrophobic at the codon number 238 (CCT-CTT). All the models reveal the conformational changes in the PPT1 protein structure.

Full size image

The amino acid substitution in the variant p.Arg152Ser causes disruption in β-strand conformation, which might disturb processing of TPP1. While the variant p.Tyr459Ser probably compromise the active center and destabilizes hydrophobic pocket (Fig. 4).

Fig. 4
figure 4

Homology modeling of novel missense variants identified in TPP1 gene. The native structure (blue) and mutant structure (brown) are superimposed. a The model of p.Arg152Ser depicting the amino acid change from basic to polar at the codon number 152 (AGG-AGC). b The model of p.Tyr459Ser depicting the amino acid change from cyclic to non-cyclic at the codon number 459 (TAC-TCC). Both the models reveal the conformational changes in the TPP1 protein structure.

Full size image

The novel variants of PPT1 and TPP1 genes were screened in 100 control individuals, however, none carried the given variants. The most common mutation p.Pro238Leu observed in 44% of patients with NCL1 and p.Arg206Cys observed in 26% of patients with NCL2 were screened in 100 unrelated healthy subjects and were found to have normal allele. The novel variants identified in the present study are submitted to the NCBI ClinVar repository [see Additional file 3].

Discussion

Data presented here is the first study from India demonstrating the mutation spectrum of Batten disease (NCL1 and NCL2) in a large cohort. The given study reveals 34 cases of NCLs (12 with NCL1, and 22 with NCL2) with maximum NCL2 cases. Similarly, in a study by Santorelli et al., the highest numbers of cases confirmed the NCL2 (24%) [29].

The patients’ clinical appearance like seizures, myoclonic jerk, visual impairment, and neuroimaging examination showing cerebral atrophy and cerebellar atrophy were in concordance with the previously established phenotypes in NCL patients [3]. The mutations identified in this study resulted in a broader spectrum of clinical presentation and hence hampered the genotype-phenotype correlation in NCL patients. Such clinical presentations due to PPT1 and TPP1 gene mutations are also observed in the orthologous species. For instance, a study by Sanders et al. identified a homozygous mutation c.736_737insC in exon 8 of PPT1 gene in a canine presenting NCL-like signs including, visual impairment, disorientation, behavioral changes, lack of PPT1 activity in the brain, and accumulation of autofluorescent lysosomal inclusions with the granular osmiophilic deposit in neurons [30]. Also, a study by Mahmood et al. established that a homozygous TPP1 gene mutation in a zebrafish results in the progressive early onset of neurodegenerative phenotypes, small retina, accumulation of subunit c of mitochondrial ATP-synthase, and localized apoptotic cells death in the retina, optic tectum, and cerebellum [31].

Several country-specific mutations are reported in NCL1 and NCL2. The most common NCL1 mutation identified in Finland is p.Arg122Trp in PPT1 gene, which accounts for 98% Finish variants [12]. A study by Das et al. revealed two common mutations, p.Arg151Ter and p.Thr75Pro, in PPT1 gene of American NCL1 patients [32]. The absence of these common mutations in the present study of Indian patients suggests the molecular heterogeneity of NCL1 in India. In this study, a novel variant c.713C > T (p.Pro238Leu) was identified in the PPT1 gene of four unrelated NCL1 positive families (44%) from the southern part of India. This suggests its possible founder effect in the Indian origin settlers. However, a detailed study in larger cohorts is essential.

In case of NCL2, two common mutations p.Arg208Ter and c.509-1G > C (as per old nomenclature T523-1G > C) accounting for approximately 60% of all identified TPP1 mutant alleles worldwide and at least one of these mutation can be identified in more than 75% of patients [33, 34]. In the present study, one patient with NCL2 was identified with a homozygous p.Arg208Ter mutation but the variant c.509-1G > C was not observed in our cohort which indicate its uncommon occurrence in Indian NCL2 patients. The NCL2 country-specific mutation includes p.Gly284Val in Canada and p.Asp276Val in Argentina [14, 35]. In the present study, a known pathogenic mutation p.Arg206Cys was observed most commonly in the unrelated NCL2 patients (26%) suggesting its possible founder effect.

However, in the given study, the genetic diagnosis of about 17% of patients remained ambiguous. A similar percentage was also observed in previously published data were around 10% of patients were without any genetic identification [29]. This suggests that the deep intronic variants, large deletion or duplication in the NCL genes might also play a role in disease occurrence. In addition, as suggested by Santorelli et al., studying large informative families might identify new NCL genes and help in understanding NCLs molecular pathology [29].

A study by Das et al. established that a reduction in PPT1 and TPP1 enzyme activity ranges from 0 to 2.5% [33]. In the given study also, the reduction in these enzymes activity was from 0 to 2.8%. Based on these biochemical observations, the therapeutic approaches are tested to regain the enzyme activity. For instance, a study in a canine model with TPP1-deficiency revealed that the administration of a recombinant adeno-associated virus (rAAV) expressing canine TPP1 in the ependyma resulted in elevation of TPP1 expression leading to delay in clinical presentation and extension of life span [36]. In addition, studies to diminish the clinical phenotypes of NCL have been directed. Tracy et al. reported an alternative approach of using stem cell based delivery of therapeutic components to the retina, as the systemic administration would be ineffective [37]. This study reported the inhibition of the retinal degeneration in the canine model after a single intravitreal administration of autologous bone marrow-derived stem cells transduced with a TPP1 expression construct [37].

Conclusions

The given study contributes four novel variants in PPT1 gene and eight novel variants in TPP1 gene mutation spectrum. Our results with remarkable heterogeneity provide new insight into the molecular pathology of NCL1 and NCL2. In addition, it was observed that the novel variant p.Pro238Leu was common in Indian NCL1 patients while a known pathogenic mutation p.Arg206Cys was commonly observed in Indian NCL2 patients. This can give a new insight into the molecular pathology of NCL patients with Indian origin.