Mutation update: Review of TPP1 gene variants associated with neuronal ceroid lipofuscinosis CLN2 disease

Abstract Neuronal ceroid lipofuscinosis type 2 (CLN2 disease) is an autosomal recessive condition caused by variants in the TPP1 gene, leading to deficient activity of the lysosomal enzyme tripeptidyl peptidase I (TPP1). We update on the spectrum of TPP1 variants associated with CLN2 disease, comprising 131 unique variants from 389 individuals (717 alleles) collected from the literature review, public databases, and laboratory communications. Previously unrecorded individuals were added to the UCL TPP1‐specific database. Two known pathogenic variants, c.509–1 G>C and c.622 C>T (p.(Arg208*)), collectively occur in 60% of affected individuals in the sample, and account for 50% of disease‐associated alleles. At least 86 variants (66%) are private to single families. Homozygosity occurs in 45% of individuals where both alleles are known (87% of reported individuals). Atypical CLN2 disease, TPP1 enzyme deficiency with disease onset and/or progression distinct from classic late‐infantile CLN2, represents 13% of individuals recorded with associated phenotype. NCBI ClinVar currently holds records for 37% of variants collected here. Effective CLN2 disease management requires early diagnosis; however, irreversible neurodegeneration occurs before a diagnosis is typically reached at age 5. Timely classification and public reporting of TPP1 variants is essential as molecular testing increases in use as a first‐line diagnostic test for pediatric‐onset neurological disease.

development, and followed by rapidly progressing dementia, psychomotor decline (loss of the ability to walk and talk), epilepsy, blindness, and death, typically between 6 years of age and the early teenage years (Mole et al., 2018;Mole, 2001;Nickel et al., 2016;Nickel et al., 2018;Steinfeld et al., 2002). Whereas classic late-infantile CLN2 disease has a very well defined natural history, there exists a phenotypic spectrum of TPP1 enzyme deficiency in small numbers of patients, some with later onset or protracted disease course (Kohan et al., 2013;Kousi, Lehesjoki, & Mole, 2012). One form of juvenile onset disease was initially described as spinocerebellar ataxia 7 (SCAR7; MIM# 609270) and was later attributed to a TPP1 enzyme deficiency . Other, variant forms of complex hereditary spastic paraplegia (Kara et al., 2016) and childhood-onset progressive ataxia (Dy, Sims, & Friedman, 2015) were described clinically before being linked to TPP1 enzyme deficiency. Occasional cases present before the age of 2 years (Nickel et al., 2018). With the knowledge of a shared molecular etiology, rather than being distinct entities, these diseases can be considered part of the same phenotypic spectrum that includes classic late-infantile CLN2 disease and forms of atypical CLN2 disease. Thus, NCL classification was revised to take into account such phenotypic variation (Williams & Mole, 2012).
Effective CLN2 disease management requires timely diagnosis; however, irreversible neurodegeneration often occurs before a diagnosis is typically reached at 5 years of age (Nickel et al., 2018).
Early diagnosis has become even more relevant as a recently approved intracerebroventricular enzyme replacement therapy has been shown to effectively slow the rapid decline in motor and language function in patients with CLN2 disease .
Aside from genetic testing, there are other medical procedures that may increase suspicion of CLN2 disease, for example, severe cerebellar atrophy is the principal sign seen at the time of diagnosis on magnetic resonance imaging (MRI) (Williams et al., 2006). Photosensitivity, as detected by electroencephalography, is also an early marker of CLN2 disease (Specchio et al., 2017).
The American College of Medical Genetics (ACMG) guidelines recommend that gene variants be reported in combination with their assessed pathogenicity (Richards et al., 2015). The purpose of this mutation update is to summarize the identified disease-related genetic variation in the TPP1 gene, with emphasis on clinical classification and genotype-phenotype correlations. There is a clear set of patients with atypical CLN2 disease which includes TPP1 deficiency, from SCAR7 and juvenile NCL. We collected and analyzed variant information from 389 individuals (131 different/independent variants) associated with CLN2 disease to uniformly summarize all Embase was searched using the following searches: OR 'e.c. 3.4.14.9' OR 'tripeptidyl peptidase 1′ OR 'tripeptidyl peptidase i' OR 'tripeptidyl peptide hydrolase i' OR 'tripeptidylpeptidase 1' OR 'tripeptidylpeptidase i' OR 'tripeptidylpeptide hydrolase i') AND ('mutation'/de OR 'gene alteration' OR 'genome mutation' OR 'mutation') AND ('human'/de) 2. 'neuronal ceroid lipofuscinosis'/de AND 'mutation' AND 'cln2' AND 'human'/de NOT (('tpp1 gene'/de OR 'cln2 gene'/de OR 'tripeptidyl peptidase i'/de OR 'e.c. 3.4.14.9' OR 'tripeptidyl All variants collected from the UCL TPP1-specific database and literature searches were assessed using ACMG standards and guidelines for interpretation of sequence variants using available information (Richards et al., 2015) (Table 1). It should be noted that the effect on TPP1 function has not been established for all disease-associated alleles. In some patients, further variant alleles were found, in addition to those presumed to be disease-associated. Four of these additional alleles are described in the UCL TPP1 Locus-specific Database.
The variant c.299 A>G (p.(Gln100Arg)) has a frequency below 5% (National Center for Biotechnology Information, 2018) and is predicted benign. It occurred as an additional allele in three unrelated patients with classic late-infantile CLN2 disease (Sleat et al., 1999;Tessa, Simonati, Tavoni, Bertini, & Santorelli, 2000), but it also occurred as the disease-associated allele in trans with c.1266 + 5 G>A in a patient from Canada (disease phenotype unknown) (Kousi et al., 2012). The underlying sequence change for variant p.(Val426Val) is not available, therefore this variant cannot be assessed for potential alteration to splicing. In our data set, this variant occurred in two unrelated patients from Argentina, together with another variant of uncertain significance (c.89 + 4 A>G; Kohan et al., 2013;Noher de Halac et al., 2005). This latter variant potentially causes alteration of splicing and has been described as disease-associated in one patient from Argentina (disease phenotype unknown). Finally, c.1501 G>T (p.(Gly501Cys)) is predicted as probably damaging and occurs in one patient from Turkey (disease phenotype unknown) (Kousi et al., 2012). It occurs (phase unknown) with c.622 C>T (p.(Arg208*)) and c.1343 C>T (p.Ala448Val)), which is also predicted as probably damaging. For the latter three additional alleles, there is no information on population frequency. Thus, for these four cases, the assignment of disease-association is equivocal. Other additional alleles excluded from the analyses were listed in ClinVar as benign and/or have a prevalence in the population > 5%. To date, of the 131 variants reported in the UCL database as disease-associated, only 39 (30%) are recorded in ClinVar with an associated clinical classification.
Mutations identified in TPP1 are distributed over the whole protein structure (Figure 2) and the majority are likely linked to loss of enzyme activity, though very few will have been confirmed to do this biochemically. Loss of TPP1 activity leads to neuropeptide degradation failure and significant accumulation of subunit c of ATP synthase. However, accumulation of subunit c has been identified in most forms of NCL and other lysosomal storage disorders, suggesting that this may not be the primary metabolic error in TPP1 deficiency (Palmer et al., 2013;Ryazantsev, Yu, Zhao, Neufeld, & Ohmi, 2007 (Haltia, 2006;Palmer et al., 2013).

| Clinical and diagnostic relevance
Diagnosis of CLN2 disease may be reached through a mixture of clinical findings, TPP1 enzyme deficiency, and/or molecular findings in TPP1 (Fietz et al., 2016). Historically, diagnoses of NCL subtypes have relied on histopathological techniques, such as an electron microscope evaluation of autofluorescent storage material morphology, together with a clinical review of disease onset and symptoms (Williams et al., 2006). Assaying of white blood cell TPP1 activity is now the mainstay of diagnosis for TPP1-related diseases (Fietz et al., 2016). Whereas this provides a direct test for CLN2 disease, it requires a specific suspicion of CLN2 or other NCL. By that point, there will have been significant disease progression and neurodegeneration (Nickel et al., 2016).
Alongside the demonstration of deficient TPP1 enzyme activity, detection of two pathogenic mutations in trans is considered the gold standard for CLN2 disease diagnosis (Fietz et al., 2016). Unlike biochemical testing, molecular genetic testing can be used to test multiple etiologies, and potentially lead to a patients phenotype. This means that no specific suspicion of an etiology is required, positioning these broad next-generation sequencing (NGS)-based tests as a tool for earlier diagnosis of genetic diseases. NGS techniques such as whole exome sequencing (WES) have emerged in recent years as useful tools for enhancing NCL subtype classification, particularly when mutations in different genes F I G U R E 3 Most common alleles listed in the TPP1 locus-specific database by region of origin. The number of times an allele was encountered is shown in parentheses. North America includes Newfoundland. Note: The emphasis now is on collecting new variants; frequency of the most common variants is, therefore, underrepresented here as new reports for these are no longer included in the UCL TPP1 locus-specific database.
Timely diagnosis facilitates the early initiation of appropriate disease-specific care and enables families to make informed decisions about treatment goals (Williams et al., 2017). Unexplained seizures, especially if preceded by early language developmental delay, can be an early symptom of CLN2 disease. In children without a specific CLN2 indication who present with delayed language skills, experts recommend investigating pediatric-onset seizures using an epilepsy gene panel (Fietz et al., 2016;Lemke et al., 2012), as an approach to decrease time to the differential diagnosis of CLN2 disease.
Patients with CLN2 may encounter diagnostic delay due to the inexperience of their treating physicians and lack of awareness of NCL disorders. This may be a particular challenge in countries with an abundance of variant phenotypes due to diverse ethnic populations (Kohan et al., 2009). In an era where broad molecular tests are being used In addition, if a second variant is not identified, but enzyme activity is deficient, this can be used as evidence to classify any other potentially deleterious variants in the patient as well as provide a laboratory-based diagnosis of CLN2 disease (Fietz et al., 2016;Richards et al., 2015).
Zebrafish with a TPP1 deficiency die prematurely and show ubiquitous storage material containing ATP synthase subunit c, with it being more evident in the CNS and muscles (Mahmood et al., 2013). The early stop codon in exon 3 (which is also described in humans but at a different amino acid position) leads to an early-onset neurodegenerative phenotype and functional motor impairment preceded by a phase of hyperactivity that could be consistent with seizures. The zebrafish model also shows significant apoptotic cell death and aberrant proliferation in the optic tectum, cerebellum, and retina. As mouse models for CLN2 disease do not seem to suffer from visual problems or retinal degeneration (see below), the study of this aspect of the disease could utilize these findings from the zebrafish model.
The first TPP1 deficient mouse model (mixed background [C57BL/6:129S6]) was generated by knock-in of the CLN2-specific p.(Arg447His) missense mutation into the Tpp1 gene in combination with a large intronic insertion (Sleat et al., 2004). The lack of activity of TPP1 protein does not affect the initial stages of development but evolves with signs of progressive neurological deficits with aging. The lifespan is drastically reduced (median survival 138d) and the mice display early motor deficits, seizures, spontaneous tremors, and ataxia (Sleat et al., 2004). The neurological impairment is visible in the brain, spinal cord, and peripheral sensory neurons, with an accumulation of autofluorescent material in the lysosomes. The severe loss of neurons in the cerebral cortex that is observed in the human late-infantile CLN2 disease is not that obvious, but there is a clear loss of Purkinje cells which could be linked to the cerebellar ataxia. Studies on the histology of the retina do not show any loss of photoreceptors or any reduction in cell layers (Sleat et al., 2004).
More recently, a mouse model encoding the most common nonsense mutation found in humans, an early stop codon instead of arginine in the 208 positions (p.Arg207* in mice), has been generated and described (Geraets et al., 2017). The resulting transcript reduction leads to reduced enzymatic activity in different organs, such as the liver, spleen, or cerebellum. Consequently, mice show a reduced lifespan, with most dying between 3 and 6 months of age. As observed in the previous mouse model (Sleat et al., 2004), impaired motor behavior is observed and characterized by tremors, seizures, hyperactivity, and strength deficits. The visual phenotype was not studied. Histological evaluation of the brain displayed an accumulation of the mitochondrial ATP synthase subunit c in superficial or deep cortical layers. Showing a similar phenotype to the pre-existing mouse model, this new transgenic mouse could be used for the preclinical evaluation of all therapeutic approaches including mutation-guided therapies.
To date there is evidence of NCL in over 20 canine breeds and mixedbreed dogs (Katz, Rustad et al., 2017). The canine Tpp1 gene sequence (GenBank AF114167) includes all 13 exons that are present in the human TPP1 gene, and exonic sequences are highly conserved between both species (Drögemüller, Wöhlke, & Distl, 2005).
The first report of NCL in Dachshund dogs described a neurodegenerative disease starting with hind-leg weakness at 3 years of age (Cummings & de Lahunta, 1977 (Katz et al., 2014Vuillemenot et al., 2015). If the therapeutic approaches to treat the CNS succeed, animal models could provide valuable insight into further challenges affecting the life expectancy and the quality of life of the patients.

| Future prospects
The newly approved intracerebroventricular enzyme replacement therapy has rendered CLN2 disease from an untreatable to treatable disease, especially if treatment is started early before significant neurodegeneration has already taken place . The ultimate effort to improve early diagnosis of a now treatable disease is newborn screening (NBS). Experts suggest that assaying TPP1 activity with enzyme substrates compatible with tandem mass spectrometry detection could support future large-scale NBS programs (Barcenas et al., 2014;Fietz et al., 2016). The adoption of successful NBS programs for CLN2 also relies heavily on the clarity of genotype-phenotype correlations. There must be a concerted effort to ascertain the disease liability of TPP1 variants to facilitate interpretation of variants detected through population-based screening and diagnostic molecular genetic testing. The algorithm for detection should maximize specificity, achieve a high positive predictive value, and have a low false-positive rate (Pitt, 2010).
There is also the more distant possibility of whole exome or whole genomic sequencing on all newborns, which could be followed by specific testing for predicted enzyme deficiencies. In addition, a