The NCLs are relative latecomers to the arena of lysosomal storage diseases, because the underlying lysosomal enzyme defects have only recently been identified. In contrast to previously known lysosomal storage diseases (the gangliosidoses and mucopolysaccharidoses, for instance), the nature of the stored material did not provide clues as to the underlying biochemical defect. In retrospect, this may be explained by the heterogeneous nature of the undegraded material. For example, a deficiency of a lysosomal protease is known to cause LINCL, and one might expect that the resulting storage material might contain a complex mixture of undegraded peptides. In addition, it is possible that deficiencies in the NCL lysosomal proteins or enzymes cause a more general dysfunction of lysosomes in which secondary changes, only indirectly related to the primary defect, are more prominent. Previous studies on the biochemical nature of the storage material in the NCLs that were completed before the enzyme or protein defects were known are described in the next section. The relationship between these underlying defects and the observed biochemical storage is still unclear and remains an active area of investigation.
When examined by histochemistry, the storage material is resistant to lipid solvents and, in unstained sections, is colorless to slightly yellow. The granular deposits are acid-fast, strongly PAS-positive, and stain intensely with Sudan black B, indicating the presence of significant amounts of both carbohydrate and lipid. Strong acid phophatase activity is easily demonstrated, emphasizing the lysosomal nature of the storage bodies.45 Autofluorescent storage particles with preserved ultrastructure have been purified from INCL brains by homogenization and sucrose52 or cesium chloride53 density gradient ultracentrifugation. The storage material consists of 40 percent protein and 35 percent lipid by dry weight.53 Interestingly, the major identifiable proteins in the storage material are saposins A and D and fragments of glial fibrillary acidic protein.53 Saposins (sphingolipid activator proteins) accumulate in tissues of patients with several lysosomal storage diseases,54 and the significance of its striking accumulation in INCL is unknown. Mitochondrial ATPase subunit c accumulation is not a consistent feature, in contrast to late infantile and juvenile-onset NCL.55 Biochemical studies that have been performed on the brains of patients with INCL largely reflect the severe neuronal loss in the disease. Total lipids are reduced and the yield of myelin is less than 2 percent of normal.56,57 Long chain polyunsaturated fatty acids are reduced. A marked fall in lipid-bound neuraminic acid is seen.45 No directed biochemical analysis of INCL brain storage material has been reported in the short time since the discovery of the enzyme defect. Total lack of PPT activity was recorded from autopsy samples of INCL brain.1
Subunit c of mitochondrial ATP synthase is a major protein component of the storage material in ovine neuronal ceroid lipofuscinosis59 and in classical LINCL (and to a lesser extent, JNCL).55 A specific delay in the degradation of subunit c within late infantile cells has been shown.60,61 Furthermore, the slower degradation is accounted for by a defect in lysosomal proteolysis, as lysosomal extracts from late infantile cells degrade purified subunit c more slowly in vitro.62 It is not currently known whether subunit c is a substrate for the lysosomal protease defective in classical LINCL, nor whether the defect in this protease directly accounts for the subunit c accumulation. Other biochemical abnormalities in classical LINCL, including the accumulation of dolichol oligosaccharide and oligosaccharyl phosphates, are felt to be secondary to the lysosomal storage.63
The lipopigments accumulating in JNCL are very similar to the age-related pigments accumulating in normal tissues. The similarity of staining characteristics to those of lipofuscin and ceroid, confirmed by several authors, applies to all types and variants of NCL with insignificant variations. The lipopigments are insoluble in polar and nonpolar solvents, making their study possible in ordinary paraffin sections, where they can be stained for easier detection using PAS, Sudan black B, or acid fuchsin. The material also shows bright yellow autofluorescence under ultraviolet light. The lipopigment granules have an intense activity of acid phosphatase, suggesting an association with lysosomes. This impression has been confirmed by electron microscopic studies and, to some extent later, by the molecular defects identified.
The storage material is already identifiable in the prenatal period, and the typical inclusion profiles have been reported in the first trimester.64– 66 This finding suggests that abnormal levels of accumulated material can be tolerated by cells and tissues to a certain extent. Two fundamental questions that remain unanswered are whether cell death is caused directly by the storage material or for other reasons, and why neuronal cells are most affected while storage material accumulates in multiple cell types.
The major component of the storage material in JNCL inclusions is subunit c of the mitochondrial ATP synthase complex.55,67 However, cells from JNCL patients show less intense staining for subunit c than cells from patients with LINCL.67,68 Subunit c is an essential membrane component of the proton channel of the large oligometric complex, ATP synthase, which generates ATP by oxidative phosphorylation. Another protein, subunit c of vacuolar ATPase, is also deposited to some degree in several of the naturally occurring animal models of NCL.69,70 Vacuolar subunit c (also known as ductin) is a membrane protein that is part of both the vacuolar H+ translocating ATPase (involved in acidification) and of gap junctions. Mitochondrial ATP synthase subunit c also accumulates in other lysosomal disorders including mucopolysaccharidosis I, II, IIIA, multiple sulfatase deficiency, mucolipidosis I, Niemann-Pick types A and C, and GM1 and GM2 gangliosidoses.68,71 However, in these cases, the accumulation is observed only in neurons, whereas in the NCL diseases, cells outside the nervous system also accumulate subunit c, suggesting that the subunit c storage is more specific for NCL diseases than for other lysosomal disorders.
Interestingly, the stored subunit c contains a trimethyllysine residue in the human NCLs72– 74 and in three animal models (mouse, dog, and sheep),72,75,76 although whether this modified amino acid residue is a direct cause of the accumulation, a result of its accumulation, or even a normal feature remains unknown.
Other minor storage components are two of the sphingolipid activator proteins or saposins.53 Saposins are small, heat-stable, lysosomal proteins, originating from a single precursor molecule that is cleaved to release four mature saposins.77 Saposins activate lysosomal hydrolases involved in glycosphingolipid degradation. Saposins A and D are the major accumulating saposins in all NCL diseases, including JNCL,78 as well as in animal models for NCL.70 Because both subunit c and saposins are extremely hydrophobic molecules with a tendency to self-aggregate, the underlying problem in the NCLs may be the processing and disposal of these hydrophobic molecules.
In addition to the accumulation of storage material, many other biochemical changes have been reported in tissues from NCL patients or affected animals. Based on the identified defects, various biochemical deficiencies have been proposed. These include a defect in proteases,79 abnormal lipid/fatty acid peroxidation,80 abnormal dolichol metabolism,81– 84 and abnormal carnitine biosynthesis.85 Although the gene defect is now known, the issue of the relationship between the gene defect and these biochemical findings remains unsolved until the cellular function of this membrane protein is clarified.
Enzyme and Protein Deficiencies in the NCLS
Palmitoyl-Protein Thioesterase (PPT; CLN1).
Virtually all cases of neuronal ceroid lipofuscinosis with predominantly granular osmiophilic deposits, or GROD (which would include all INCL), have an underlying deficiency in PPT activity as a result of mutations in the PPT gene.1,13,14 Palmitoyl-protein thioesterase [E.C. 188.8.131.52] is a recently described lysosomal enzyme that removes long chain fatty acids (usually palmitate) from cysteine residues in proteins.87 The covalent linkage involves a thioester bond between the carboxyl end of the fatty acid and the sulfhydryl group of the cysteine residue, hence the designation of the enzyme that hydrolyzes this reaction as a thioesterase (Fig. 154-3).
A “typical” palmitoylated protein is normally found in association with the inner surface of the plasma membrane. The thioester bond that is hydrolyzed by palmitoyl-protein thioesterase is shown. The natural substrates of lysosomal palmitoyl-protein thioesterase are palmitoylated proteins and peptides undergoing degradation in the lysosome.
S-acylation at cysteine residues is a common post-translational modification of proteins.88 The fatty acylated cysteine residue is most often located at the inner surface of the plasma membrane. Cycles of acylation and deacylation of proteins may regulate certain protein-protein or protein-membrane interactions that are important in cellular signal transduction.89 However, the function of lysosomal PPT appears to be the removal of covalently bound fatty acids from S-acylated proteins that arrive in the lysosome for degradation.90
PPT activity was first demonstrated in tissue extracts from bovine brain, a rich source of the enzyme.87 The substrate for the detection of enzyme activity was H-Ras, a protein that is both palmitoylated and farnesylated near the C-terminus. (Farnesylation is a lipid modification that occurs in a stable thioether linkage to cysteine residues in proteins.) Specific removal of the palmitate (but not the farnesyl group) from the protein could be demonstrated. The enzyme assay was used to guide the purification of PPT to homogeneity.87 Because the enzyme has a neutral pH optimum for acyl-protein substrates, its lysosomal nature was not suspected until the cloning of the cDNA91 and subsequent studies of its gene localization92 and involvement in neuronal ceroid lipofuscinosis, a storage disorder.1
cDNAs encoding human,92 bovine,91 rat,91 and mouse93 PPT have been characterized. A major 2.5-kb mRNA is detected in human, bovine, and rat tissues. (A second 1.4-kb mRNA is detected in the rat, a consequence of use of an alternative polyadenylation site present only in the rat gene.) The human PPT cDNA consists of a short 5′-untranslated region of 14 base pairs, an open reading frame of 918 base pairs, and a 1388-base pair 3′ untranslated region. The amino acid sequence deduced from the cDNA contains 306 amino acids, of which the first 25 amino acids constitute a leader peptide. Sequence motifs characteristic of thioesterases94 (glycine-X-serine-X-glycine and glycine-aspartic acid-histidine) are present at amino acid residues 113 to 117 and 287 to 289, respectively. Thioesterases contain the classical “catalytic triad” residues of serine, aspartic (or glutamic) acid, and histidine, as do many lipases and proteases.95,96 Three potential asparagine-linked glycosylation sites are found near the C-terminus of the protein at positions 197, 212, and 232. All of these sites are utilized to some degree in vivo.91,97
PPT mRNA is found ubiquitously in a variety of tissues from the human92 and rat,91 with fairly high and uniform levels in lung, spleen, brain, and testis, and relatively little mRNA in liver and skeletal muscle. The correlation between mRNA levels and levels of enzyme activity is imperfect, as much higher activity (and protein) levels are found in testis and brain as compared to other tissues.87 Interestingly, brain and testis are the two most severely affected organs histopathologically in INCL.44 The analysis of PPT transcript abundance by in situ hybridization of mouse brain has revealed that PPT transcripts are expressed widely but not homogeneously in the mouse brain and the expression follows a specific temporal pattern (J. Isosomppi, personal communication). The signal is most intensive in the cerebral cortex (layers II, IV, and V), hippocampal CA1 to CA3 pyramidal cells, dentate gyrus granule cells, and hypothalamus. Localization of PPT transcripts and immunoreactivity is spatially and temporally well correlated. Immunostaining is localized to axons and dendrites, especially in the pyramidal and granular cells of the hippocampus, correlating well, both spatially and temporally, with immunoreactivity of a presynaptic vesicle membrane protein, synaptophysin (J. Isosomppi, personal communication).
The lysosomal targeting and intracellular localization of PPT have been demonstrated through studies of mannose 6-phosphate-mediated cellular uptake98– 100 and immunofluorescence studies.99 The synthesis and intracellular targeting of PPT are typical for a lysosomal enzyme.91,98,99 In contrast to a number of other lysosomal enzymes, activation by internal cleavage does not appear to be a feature of the thioesterase. The mature enzyme is a doublet at 37 kDa/35 kDa, due to heterogeneous glycosylation.91,98,99 Two amino-terminal amino acids (histidine-leucine) that are predicted to be present from analysis of the cDNA, and which are present in the recombinant enzyme, are not present in the mature enzyme purified from brain. This minor processing is probably of little significance, as both unprocessed and amino-terminally processed forms are fully active.91
Studies with the purified, recombinant enzyme have shown a preference for long chain fatty acids of 14 to 18 carbons, with virtually no activity using substrates of 6 carbons or less.91 The enzyme will degrade noncysteine-containing thioesters such as palmitoyl CoA. Sensitivity to diethylpyrocarbonate indicates that an essential histidine residue may be involved in catalysis, and insensitivity to phenylmethanesulfonyl fluoride distinguishes this enzyme from other cellular esterases and was an aid to its identification and purification.87 The pH optimum is broad and varies with different substrates.
Metabolic labeling studies have provided some insight into the metabolic defect in INCL.101 When cells from PPT-deficient subjects are incubated with [35S]cysteine and [3H]palmitate, small lipid thioesters containing these labels accumulate in a time-dependent manner (Fig. 154-4). Furthermore, the lipid cysteine thioesters appear to be derived from protein, because cycloheximide blocks their formation. This observation implies that the cysteine must first be incorporated into protein in order to appear in the abnormal products. The fatty acid cysteine thioesters are substrates for the thioesterase in vitro. In addition, their accumulation can be reversed by the addition of PPT to the cell culture medium in cross-correction experiments,101 especially if the enzyme is modified with mannose 6-phosphate and is therefore competent for uptake through the mannose 6-phosphate receptor pathway.
Accumulation of [35S]cysteine-labeled lipids in normal and INCL lymphoblasts and correction of the metabolic defect in INCL lymphoblasts by recombinant PPT. Panel A, cells from normal control (lanes 1 to 4) or INCL patients (lanes 5 to 8) labeled with [35S]cysteine for varying times up to 6 h. Panel B, cells from INCL patients labeled for 3 h, then incubated with an excess of unlabeled cysteine in the absence (lanes 1 to 4) or presence (lanes 5 to 8) of recombinant human PPT (500 ng/ml) supplied from transfected COS cell- conditioned medium. (Reprinted with permission from Lu et al.101 Copyright 1996 National Academy of Sciences, U. S. A.)
The mechanism whereby deficiency of PPT leads to severe and selective neurodegeneration is unknown. One possibility is that the lipid-cysteine thioesters consist of small peptides that are resistant to degradation by virtue of their fatty acid modification. This “indigestible” proteolipid material may be particularly harmful to neurons. By analogy, other neurodegenerative disorders, such as Alzheimer disease, the prion diseases, and the polyglutamine disorders, are characterized by the accumulation of unmetabolized peptides. Unfortunately, in none of these diseases is the link between the storage and selective neuronal toxicity well understood.
Classical LINCL is caused by a deficiency in a novel pepstatin-insensitive lysosomal protease, pepinase. The deficiency in this enzyme was demonstrated by a technique that may be generally applicable to other lysosomal storage diseases.102 The mannose 6-phosphate modification of newly synthesized lysosomal enzymes was used as an affinity marker to identify a protein that was absent in brain tissue from a patient with the disorder (Fig. 154-5). (This approach was successful because brain tissue contains relatively high levels of mannose 6-phosphate-modified lysosomal enzymes, whereas in peripheral tissues, the mannose 6-phosphate is normally removed in the lysosome.102) Amino acid analysis of this novel 46-kDa protein revealed strong similarity to bacterial proteases, and a deficiency of pepstatin-insensitive protease activity was identified in cells from classical LINCL patients. Furthermore, the gene encoding the enzyme was found to reside on human chromosome 11p15, the site of the CLN2 gene determined by linkage studies in a large number of classical LINCL cases.27 Mutations in the coding region of the pepinase cDNA were described.102 Classical LINCL is only the second recognized human lysosomal storage disease to result from the deficiency of a lysosomal protease.
Identification of a protein deficient in LINCL. Detergent-solubilized extracts of gray matter from normal (top) or LINCL (bottom) brain autopsy specimens were fractionated in two dimensions by isoelectric focusing and SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose, and the mannose 6-phosphorylated glycoproteins were detected with 25I-labeled mannose 6-phosphate receptor. The mannose 6-phosphate glycoprotein (pepinase) that is absent in LINCL extracts is indicated with an arrow. (Reprinted with permission from Sleat et al.102 Copyright 1997 American Association for the Advancement of Science.)
Analysis of cDNA clones encoding pepinase revealed two major transcripts at 3.7 and 2.7 kb, which are the products of alternative use of a polyadenylation signal in the 3′ untranslated region of the message. The mRNA was detected in all tissues examined with highest levels in heart and placenta. The transcripts contain one long, open, reading encoding a 563-residue protein, predicted to contain a 16-residue signal peptide. Amino acid sequencing by Edman degradation of the mature enzyme indicates that a proteolytic cleavage occurs between amino acid residues 194 and 195 during maturation. It was not possible to determine whether the enzyme has a large propiece that is removed, or whether the native enzyme consists of an amino-terminal light chain and a C-terminal heavy chain.
The enzyme has both endoprotease and carboxypeptidase activity102 (P. Lobel, personal communication). Little is yet known about the substrate specificity of the enzyme, and it will be interesting to see whether its deficiency directly, or indirectly, accounts for the accumulation of subunit c of mitochondrial ATPase or other storage material in this disease.
The CLN3 gene encodes a 438-amino acid protein with a deduced molecular weight of about 48 kDa. The predicted amino acid sequence of the polypeptide shows no homology with any known protein, but computer-based structural analysis indicates the presence of several hydrophobic regions, suggesting that it is an integral transmembrane protein.103,104 The function of the CLN3 protein remains unknown, but the high degree of evolutionary conservation demonstrated by orthologous genes in Saccharomyces cerevisiae (BNT1), Caenorhabditis elegans, mouse, and dog, suggests an important role for this protein in eukaryotic cells.105,106
Recent in vitro studies in HeLa cells have suggested that the recombinant, expressed CLN3 cDNA encodes a 43-kDa single-chain polypeptide. Analyses of the intracellular synthesis and maturation of the CLN3 polypeptide demonstrate that it is translocated into the ER membrane and is glycosylated on asparagine residues. Based on immunofluorescence studies using confocal microscopy, the polypeptide appears to be localized to lysosomal membranes as it colocalizes with Lamp1, an integral lysosomal membrane protein.107 However, a mitochondrial location has also been suggested.72
Yeast, a simple model organism, has been used to analyze the function of the CLN3 protein. The homologous gene in yeast, BTN1, was deleted in Saccharomyces cerevisiae 105 but no obvious disease phenotype was observed. However, mutant strains were resistant to the chemical aminonitrophenyl propanediol (ANP),108 particularly when grown at low pH, and this resistance could be complemented by the human CLN3 gene, indicating that CLN3 protein functions in yeast. Moreover, the degree of ANP resistance resulting from specific missense mutations was related to the severity of the human disease caused by the same mutations. Interestingly, the gene ANP1, when mutated, makes yeast more sensitive to ANP. The ANP1 gene product is located as part of a complex in the cis Golgi where it is implicated in the process of glycosylation.109– 111 Identification of a CLN3 phenotype in yeast will facilitate the identification of genes that will suppress the phenotype, and thereby provide clues to the function of the CLN3 protein.
CLN5 Protein (Finnish vLINCL).
The causative gene112 in Finnish variant LINCL is predicted to encode a novel membrane protein of 407 amino acids with a molecular weight of 46 kDa and a calculated pI of 8.41. No homologous proteins exist in protein or DNA databases. It remains to be seen whether the CLN5 protein is a lysosomal membrane protein, as is the CLN3 protein.107 The tissue expression pattern of the CLN 5 gene is ubiquitous on northern analysis, and a relatively low steady state transcript level is observed. The highest transcript levels have been found in aorta, kidney, lung, and pancreas.112