CLN7 disease is an autosomal recessive, childhood-onset neurodegenerative lysosomal storage disorder caused by the defective lysosomal membrane protein CLN7. We have disrupted the Cln7/Mfsd8 gene in mice by targeted deletion of exon 2 generating a novel knockout (KO) mouse model for CLN7 disease, which recapitulates key features of human CLN7 disease pathology. Cln7 KO mice showed increased mortality and a neurological phenotype including hind limb clasping and myoclonus. Lysosomal dysfunction in the brain of mutant mice was shown by the storage of autofluorescent lipofuscin-like lipopigments, subunit c of mitochondrial ATP synthase and saposin D and increased expression of lysosomal cathepsins B, D and Z. By immunohistochemical co-stainings, increased cathepsin Z expression restricted to Cln7-deficient microglia and neurons was found. Ultrastructural analyses revealed large storage bodies in Purkinje cells of Cln7 KO mice containing inclusions composed of irregular, curvilinear and rectilinear profiles as well as fingerprint profiles. Generalized astrogliosis and microgliosis in the brain preceded neurodegeneration in the olfactory bulb, cerebral cortex and cerebellum in Cln7 KO mice. Increased levels of LC3-II and the presence of neuronal p62- and ubiquitin-positive protein aggregates suggested that impaired autophagy represents a major pathomechanism in the brain of Cln7 KO mice. The data suggest that loss of the putative lysosomal transporter Cln7 in the brain leads to lysosomal dysfunction, impaired constitutive autophagy and neurodegeneration late in disease.
The neuronal ceroid lipofuscinoses (NCLs) are lysosomal storage disorders (LSDs) that are characterized by the selective damage and loss of neurons in the retina and the brain, neuroinflammation and accumulation of autofluorescent storage material, called lipofuscin-like lipopigments (1). The storage material consists of low molecular weight hydrophobic proteins, mainly subunit c of mitochondrial ATP synthase (SCMAS) or sphingolipid activator proteins (saposins) A and D with the remainder composed of carbohydrates, phospholipids, glycosphingolipids, bis(monoacylglycero)phosphate, dolichol and metals, predominantly iron (2,3). To date, defects in 13 different genes coding for soluble lysosomal proteins (palmitoyl protein thioesterase 1, tripeptidyl peptidase 1, CLN5, cathepsin D and cathepsin F), membrane proteins located in lysosomes (CLN3, CLN7 and ATP13A2) and the endoplasmic reticulum (ER; CLN6 and CLN8), respectively, cytosolic proteins associated to membranes (cysteine-string protein alpha and potassium channel tetramerization domain-containing protein 7) and secretory proteins (progranulin) have been identified in NCL patients (4). The CLN7/MFSD8 gene encodes the polytopic lysosomal membrane glycoprotein CLN7 of unknown function, which shares sequence homologies with the major facilitator superfamily (MFS) of transporter proteins (5). Localization of CLN7 protein in lysosomal membranes has been demonstrated by different independent approaches including proteomic analyses of purified human and rat tritosomes, immunoblotting of mouse liver tritosomes and co-localization studies in cultured hippocampal neurons (6–9). Potential substrates, the mode, directionality of transport and neuron-specific functions of CLN7 are unknown. In CLN7 disease (MIM no. 611 124), more than 30 missense, nonsense and frameshift mutations (listed in the web-based NCL Mutation Database, www.ucl.ac.uk/ncl/mutation) have been identified in the CLN7/MFSD8 gene which result with two exceptions in a relatively uniform clinical manifestation and variant late-infantile NCL (vLINCL) phenotype most likely due to a complete loss of CLN7 function (10). CLN7 disease, vLINCL phenotype, starts between 1.5 and 5 years of age with presenting symptoms (developmental regression or seizures) and rapidly progresses with loss of mental and motor functions, speech impairment, seizures, visual failure, ataxia and myoclonus leading to early death in the first or second decade of life (11). CLN7 patients carrying the CLN7 p.Ala157Pro mutation present with a juvenile phenotype and a protracted disease course (11). Recently, compound heterozygous adult patients carrying the mutations CLN7 p.Glu336Gln/CLN7 p.Glu381X and CLN7 p.Glu336Gln/CLN7 p.Lys333Lys fsX3 were described with non-syndromic autosomal recessive macular dystrophy, which lacked NCL symptoms (12).
Autopsy of human cases revealed progressive neuronal loss in layer V of the cerebral cortex, complete loss of the cerebellar granule cell layer, age-dependent progressive loss of Purkinje cells in the cerebellum and photoreceptor degeneration in the retina (1,8). We have recently generated a hypomorphic Cln7-lacZ gene trap mouse, which partly replicated the phenotype observed in CLN7 patients (9). In addition, a naturally occurring Chinese Crested dog model was identified presenting with autofluorescence in the brain, lysosomal storage, astrogliosis, blindness, anxiety and cognitive impairment (13).
In this study, we have generated a novel mouse model for CLN7 disease by deletion of exon 2 of the Cln7/Mfsd8 gene, which encodes the lysosomal membrane protein Cln7. Here, we show that the absence of Cln7 in mice leads to lysosomal dysfunction exemplified by the accumulation of autofluorescent lipofuscin-like lipopigments in lysosomes, storage of SCMAS and saposin D and increased levels of cathepsin B (CtsB), CtsD, CtsZ and β-hexosaminidase in the brain. Strong, progressive astrogliosis, subsequent microgliosis and neurodegeneration were observed in the brain of Cln7 knockout (KO) mice. Increased neuronal levels of the microtubule-associated protein 1 light chain 3-II (LC3-II) and the accumulation of p62- and ubiquitin-positive protein aggregates indicate an impairment of constitutive macroautophagy in the brain of Cln7 KO mice. The data suggest that loss of Cln7 in mice leads to defects in the lysosome/autophagy system in the brain and neurodegeneration at the end stage of disease.
Deletion of Cln7/Mfsd8 exon 2 leads to complete loss of Cln7
To study Cln7/Mfsd8 gene function invivo, we generated a Cln7 KO mouse carrying the European Conditional Mouse Mutagenesis (EUCOMM) tm1d allele by Cre-mediated recombination of the floxed exon 2 of the murine Cln7/Mfsd8 gene (Fig. 1A, 14). To genotype homozygous Cln7 KO mice, a triplex polymerase chain reaction (PCR) was performed, which verified the deletion of exon 2 in genomic DNA prepared from mouse tail biopsies (Fig. 1B). Quantitative real-time PCR with primers annealing to exons 1 and 2 of the Cln7/Mfsd8 gene revealed the absence of exon 2 in the mutant Cln7 mRNA purified from brain of Cln7 KO mice (Fig. 1C). Sequence analysis of the cDNA reverse-transcribed from RNA purified from brain, kidney and liver of Cln7 KO mice revealed that the deletion of exon 2 of the murine Cln7/Mfsd8 gene leads to loss of nucleotides 67–115 in the Cln7 mRNA coding sequence (NM_028140, Supplementary Material, Fig. S1A). The predicted truncated 37 amino acid Cln7 protein is composed of 22 genuine amino acids and 15 novel amino acids (Cln7 p.Glu23Phe fsX16), lacking all 12 putative transmembrane domains and the C-terminal amino acids (Supplementary Material, Fig. S1B). Quantitative real-time PCR analyses in human embryonic kidney (HEK) 293 cells transfected with the 3xFLAG Cln7 c.67_115del cDNA revealed complete loss of exon 2 in the mutant Cln7 mRNA (Ex1/2 boundaries). In addition, decreased amounts (∼50%) of 3xFLAG Cln7 c.67_115del mRNA compared with wild-type Cln7 mRNA (Ex5/6 boundaries) were detected, suggesting that the mutant mRNA is unstable (Supplementary Material, Fig. S1C). In HEK293 cells transiently expressing wild-type 3xFLAG Cln7, a 55–70 kDa immunoreactive band was observed in total membrane fractions, which was absent in non-transfected and 3xFLAG Cln7 p.Glu23Phe fsX16 expressing cells indicating the lack of the truncated Cln7 protein in membranes (Supplementary Material, Fig. S1D).
Cln7 expression in the brain
To study expression of Cln7 in the brain and to verify the absence of Cln7 in Cln7 KO mice on the protein level, different brain regions from adult mice were prepared and total membrane fractions analyzed by Cln7 immunoblotting using a polyclonal antibody raised against a cytosolic loop of Cln7 (9). In immunoblots, a 40–50 kDa Cln7 immunoreactive band was detected in total membrane homogenates of cerebral cortex, cerebellum, hippocampus, medulla and olfactory bulb from wild-type mice, which was absent in homogenates of Cln7 KO mice (Fig. 1D). Highest Cln7 protein levels were found in the cerebral cortex and cerebellum, whereas moderate amounts were detected in the hippocampus and medulla. Lowest Cln7 expression was found in the olfactory bulb (Fig. 1D).
Increased mortality and neurological phenotype of Cln7 KO mice
Homozygous Cln7 KO mice were viable, phenotypically normal at birth and fertile. Kaplan–Meier survival analyses revealed increased mortality rates of mutant mice, compared with unaffected littermates (Fig. 2A). By 8 months of age, Cln7 KO mice began to manifest signs of neurological deterioration attested by clasping phenotype (Fig. 2B). In addition, hindleg paralysis, tremor and myoclonus epilepsies were observed in Cln7 KO mice. Cln7 KO mice surviving until the age of 10–11 months had to be euthanized for animal welfare reasons because of the inability to reach food and water and the presence of tremor and myoclonus epilepsies. In the brain of 10-month-old Cln7 KO mice, enzymatic activities of β-hexosaminidase were significantly elevated by 2.6-fold when compared with age-matched wild-type mice (Fig. 2C). Quantitative real-time PCR analyses revealed a 1.6-fold increase in the amounts of Hexb mRNA when compared with controls.
Cln7 deficiency in mice recapitulates key pathobiochemical features of human CLN7 disease
The presence of autofluorescent ceroid lipopigments in the brain and peripheral tissues is a key feature of CLN7 disease (8,9,13). Autofluorescence was analyzed in different brain regions of Cln7 KO mice by confocal laser scan microscopy of unstained brain sections (Fig. 3A). Autofluorescence was detected already by 3 months of age in several brain regions including hippocampus, cerebral cortex, cerebellum, thalamus and olfactory bulb, which was absent in age-matched wild-type mice. In 10-month-old Cln7 KO mice, generalized autofluorescence was detected in the hippocampus (CA1, CA2 and CA3 regions, dentate gyrus, Fig. 3A). In the cerebral cortex, autofluorescence was observed throughout layers I–VI. In the cerebellum, autofluorescence was most prominent in the Purkinje cell layer, but it was also strongly detected in the granule cell layer (Fig. 3A). In addition, strong autofluorescence was found in the thalamus and olfactory bulb. In CLN7 patients, SCMAS accumulation was found both in the brain and in peripheral organs (1). We analyzed SCMAS storage in 10-month-old Cln7 KO and in age-matched Cln6/nclf mice, a naturally occurring mouse model for CLN6 disease (15). By 10 months of age, storage of SCMAS was detected mainly in neurons of all brain areas of Cln7 KO mice, which was absent in age-matched wild-type mice (Fig. 3B). In the hippocampus, strong accumulation of SCMAS was observed in neurons of the CA1, CA2 and CA3 regions (Fig. 3B), whereas no SCMAS staining was detected in the granule cell layer of the dentate gyrus. Accumulation of SCMAS in the cerebellum was found in the Purkinje cell layer and moderately in the granule cell layer. Furthermore, strong SCMAS staining was observed in layers V and VI of the cerebral cortex and in the thalamus. Comparative analyses revealed clear differences in the amounts of stored SCMAS between Cln7 KO and Cln6/nclf mice (Fig. 3B). In Cln7 KO and Cln6/nclf mice, accumulation of SCMAS in neurons of the cerebral cortex and the thalamus was similar, whereas increased staining intensities were found in Purkinje and pyramidal cells of the hippocampus in Cln6/nclf mice compared with Cln7 KO mice. In peripheral tissues of Cln7 KO mice, strong and generalized SCMAS immunostaining in the cardiac muscle was detected (Supplementary Material, Fig. S2). In the kidney, SCMAS storage was not uniform and was detected in a number of renal tubules. In the liver, SCMAS staining was restricted to cells in the proximity to central veins (Supplementary Material, Fig. S2). To quantify increased levels of SCMAS in the brain, detergent-insoluble fractions of total brain homogenates of 10-month-old Cln7 KO mice were analyzed by immunoblotting. Accumulation of SCMAS was observed in brain homogenates of Cln7 KO mice, which was absent in homogenates of age-matched wild-type mice (Fig. 3C). Storage of SCMAS in Cln7 KO brains was already detected by 3 months of age (data not shown). In addition, increased levels of prosaposin (1.5-fold) and saposin D (6-fold) were detected in the brains of 9-month-old Cln7 KO mice, compared with age-matched wild-type mice (Fig. 3D). The data indicate that loss of Cln7 in mice results in lysosomal dysfunction and accumulation of storage material, containing both SCMAS and saposin D mainly in neuronal cells.
Accumulation of storage bodies with curvilinear and fingerprint profiles and neurodegeneration in Cln7 KO mice
The ultrastructure of the storage material in the brain of Cln7 KO mice was analyzed by electron microscopy (EM). In cerebellar Purkinje cells of 10-month-old Cln7 KO mice, large storage bodies were observed which were absent in age-matched wild-type mice. In the storage material, a mixture of irregular, curvilinear and rectilinear profiles as well as fingerprint profiles could be identified (Fig. 4A). Luxol fast blue staining was performed to detect lipofuscin in brain tissue sections (16). Storage of lipofuscin was observed in large neurons of the hippocampus and in the Purkinje cell layer of the cerebellum (Fig. 4B). To analyze neurodegeneration in Cln7 KO brains, magnetic resonance imaging (MRI) scanning of 9-month-old Cln7 KO and age-matched wild-type mice was performed (Fig. 4C). MRI showed bright signal due to enlarged cerebrospinal fluid space in the olfactory bulb, cerebral cortex and dilation of the fourth ventricle in Cln7 KO mice, suggesting brain atrophy in the olfactory bulb, cerebral cortex and cerebellum. Immunohistochemical stainings of brain sections from 10-month-old Cln7 KO mice revealed no obvious loss of Purkinje cells in the cerebellum (Supplementary Material, Fig. S3).
Neuroinflammation in the brain of Cln7 KO mice
Neuroinflammation in CLN7 patients (8) and early activation of astrocytes and subsequent activation of microglia in NCL mouse models have been reported (17). Glial fibrillary acidic protein (GFAP) immunohistochemistry revealed a strong astrogliosis in different brain regions of Cln7 KO mice, with most prominent GFAP staining observed in the hippocampus, cerebral cortex and thalamus (Fig. 5A). Activated microglia were observed in all brain regions of Cln7 KO mice, but microgliosis was not as pronounced as astrogliosis. To analyze the onset and progression of astrogliosis (GFAP) and microgliosis (Iba1), age-dependent analyses were performed in hippocampal regions of 1-, 2-, 5-, 7- and 10-month-old Cln7 KO mice and age-matched control mice. GFAP immunohistochemistry revealed the onset of astrogliosis by 5 months of age with a progressive increase of GFAP-positive astrocytes in 7- and 10-month-old Cln7 KO mice (Fig. 5B). Microgliosis was detected by 7 months of age in the hippocampus, with progressive increase in 10-month-old Cln7 KO mice. Comparative analyses using 10-month-old Cln7 KO and age-matched Cln6/nclf mice revealed a more pronounced astrogliosis and microgliosis in Cln7 KO mice in all brain regions analyzed (Supplementary Material, Fig. S4).
Increased levels of lysosomal cathepsins in the brain of Cln7 KO mice
In the brain of chloride channel 7 (ClC7)-deficient mice (18) and mouse models for CLN1 (19) and CLN11 disease (20), increased levels of cathepsins (Cts) have been reported. We therefore analyzed the levels of CtsB, CtsD and CtsZ in the brain of 10-month-old Cln7 KO and age-matched wild-type mice by immunoblotting. The levels of CtsB, CtsD and CtsZ were significantly elevated (2.5-, 3- and 2-fold, respectively) in total brain homogenates of Cln7 KO mice compared with age-matched control mice (Fig. 6A). Age-dependent analyses revealed a progressive increase of CtsB and CtsD expression in the brains of Cln7 KO mice compared with age-matched wild-type mice starting by 5 months of age (Fig. 6B). Quantitative real-time PCR revealed increased levels of Ctsd, Ctsz and Ctsl mRNAs (3.5-, 2- and 1.5-fold) in the brains of 10-month-old Cln7 KO mice compared with age-matched wild-type mice, suggesting that enhanced transcription contributes to higher amounts of the analyzed cathepsins in the brain (Fig. 6C). Immunohistochemistry confirmed increased levels of CtsZ in the hippocampus, cerebral cortex, cerebellum, thalamus and medulla of Cln7 KO mice (Fig. 6D and Supplementary Material, Fig. S5A). To analyze whether elevation of CtsZ in the brain was cell-type-specific, immunohistochemical co-stainings of CtsZ with antibodies against marker proteins of microglia (Iba1), neurons (NeuN) and astrocytes (GFAP; Fig. 6E and Supplementary Material, Fig. S5B) were performed. Increased levels of CtsZ were observed both in microglia and in neuronal cells. Astrocytes lacked increased CtsZ expression. The results suggest that both activation of microglia and increased expression in neurons contribute to increased levels of CtsZ in the brains of Cln7 KO mice.
Biosynthetic and endocytic pathways to lysosomes are not impaired in Cln7-deficient mouse embryonic fibroblasts
To analyze whether increased cathepsin expression in CLN7 KO mice is caused by defective transport and/or proteolytic activation in lysosomes, pulse-chase analyses of wild-type and Cln7-deficient primary mouse embryonic fibroblasts (MEFs) were performed. These analyses revealed normal processing of the newly synthesized CtsD and CtsZ precursor forms to the intermediate CtsD and mature CtsZ forms, respectively (Fig. 7A). Furthermore, we did not observe increased amounts CtsD and CtsZ precursor forms in the media of Cln7-deficient MEFs, indicating correct sorting and processing of M6P-containing lysosomal enzymes in lysosomes. Of note, increased amounts of CtsZ were precipitated in Cln7-deficient MEFs (Fig. 7A). The data indicate that acidification of lysosomes was not impaired by the loss of Cln7. To analyze whether the uptake and processing of endocytosed [125I]-arylsulfatase B (ASB) were altered by Cln7 deficiency, internalization assays using Cln7-deficient primary MEFs were performed (Fig. 7B). Autoradiographs revealed that the amount of intracellular [125I]-ASB was not changed in Cln7-deficient MEFs, indicating normal M6P-dependent binding and endocytosis by the 300 kDa M6P receptor. During the 2 and 4 h chase periods, the ASB precursor form decreased in both genotypes, whereas the intermediate form increased. Furthermore, proteolytic processing to the intermediate 40 kDa ASB form after the 2 and 4 h chase periods was unchanged, indicating correct intracellular transport and processing of [125I]-ASB in lysosomes (Fig. 7B).
Impaired autophagy in the brain of Cln7 KO mice
Impaired autophagy has been shown to be a key feature of many lysosomal storage diseases (21). We observed a 2-fold increase in the amounts of the autophagy marker LC3-II in the brains of 10-month-old Cln7 KO mice compared with age-matched wild-type mice by immunoblotting (Fig. 8A). Amounts of beclin 1 were decreased, indicating that autophagy was not induced by the loss of Cln7. In detergent-insoluble fractions, 10-fold increased levels of the autophagy adaptor protein p62 were detected in the brains of 10-month-old Cln7 KO mice compared with age-matched wild-type mice by immunoblotting, indicating the accumulation of aggregated protein species (Fig. 8A). We found a 2-fold increase in the amounts of polyubiquitinated proteins in the brains of 10-month-old Cln7 KO mice compared with age-matched wild-type mice. Immunohistochemistry confirmed the presence of large protein aggregates positive for p62 and ubiquitin in different brain regions of Cln7 KO mice (Fig. 8B and C). These aggregates were detected mainly in neurons of the granule cell layer in the cerebellum, in deeper layers of the cerebral cortex and in hippocampal pyramidal neurons of the CA1, CA2 and CA3 regions.
In this study, we have generated a novel mouse model for CLN7 disease by deletion of exon 2 of the Cln7/Mfsd8 gene. Loss of exon 2 led to a truncated Cln7 p.Glu23Phe fsX16 protein that lacked all transmembrane domains and was not detectable in the brain of Cln7 KO mice. In CLN7 patients with a nonsense mutation in the N-terminal domain (CLN7 p.Arg35X), a severe disease course and variant late-infantile phenotype was reported (11). Overexpression studies in HEK293 cells confirmed the loss of truncated Cln7 p.Glu23Phe fsX16 in membrane fractions, suggesting the absence of mutant Cln7 in lysosomal membranes. Expression analyses in wild-type mice revealed the ubiquitous presence of Cln7 in the brain with high Cln7 expression in the cerebral cortex and in the cerebellum of mice. The CLN7 expression pattern correlated with the observed neurodegeneration in CLN7 patients (1) and in Cln7 KO mice (this study) in the cerebellum and cerebral cortex, suggesting an important function of CLN7 in these brain regions. Due to the rarity of the disease and the limited access to autopsy brain, the onset and age-dependent progression of biochemical and morphological changes in the brain and disease mechanisms contributing to the neuropathology in CLN7 patients remain unknown. By immunohistochemistry, immunoblotting and quantification of enzymatic activities, accumulation of autofluorescent lipofuscin-like lipopigments, storage of SCMAS and saposin D, increased expression of lysosomal proteins including CtsD, CtsB and CtsZ and elevated enzymatic activities of β-hexosaminidase were shown, indicating lysosomal dysfunction in the brain of Cln7 KO mice. Strong widespread SCMAS storage was observed in cerebellar Purkinje cells, in hippocampal neurons of the CA3 region, in the cerebral cortex and in the thalamus. Distribution of SCMAS in the brain of Cln7 KO mice was similar to another mouse model for vLINCL, the Cln6/nclf mouse (22). Comparative immunohistochemical analyses revealed that SCMAS amounts in the cerebral cortex and thalamus were comparable in both mouse models, whereas SCMAS storage in Purkinje cells and hippocampal neurons was more pronounced in Cln6/nclf mice. The ultrastructure of the storage material in neurons of Cln7 KO mice contained a mixture of irregular, curvilinear and rectilinear profiles as well as fingerprint profiles and resembled the storage material found in the brain of human CLN7 patients (1). The Cln7 KO mice also recapitulated pathological hallmarks detected in peripheral organs of CLN7 patients. In CLN7 patients, strong, generalized SCMAS storage was observed in the heart muscle and weaker SCMAS accumulation in the kidney and in the liver (1). In line with these findings, strong uniform SCMAS storage was observed in cardiac myofibers of the heart and to a lesser extent in kidney tubules and in the liver of Cln7 KO mice.
Here we first demonstrate in the brain of the CLN7 KO mice that both SCMAS and saposin D are protein components of the storage material that contained, as shown by EM, curvilinear, rectilinear and fingerprint profiles. Both prosaposin (1.5-fold) and saposin D (6-fold) protein levels were significantly increased in Cln7-deficient brains, suggesting that higher saposin D amounts are caused mainly by impaired degradation rather than by increased biosynthesis of the precursor protein. Saposins A and D were described as the major protein component of the storage material containing granular osmiophilic deposits in CLN1 disease (23), CLN4 disease (24) and CLN10 disease (3). In line with our findings from Cln7 KO brains, elevated saposin D levels were identified in the brains of CLN2, CLN3 and CLN11 patients and in the brain of a mouse model for CLN11 disease (20). In these NCLs, storage material containing curvilinear profiles (CLN2) and fingerprint profiles (CLN3 and CLN11) was found (1,25).
A number of different disease mechanisms such as altered calcium homeostasis, oxidative stress, activation of inflammatory cells, altered lipid trafficking, impaired autophagy, ER stress and autoimmune responses have been shown to contribute to disease pathology in LSDs and NCLs (26). A common feature of mouse models for variant late-infantile NCL is the early and localized activation of astrocytes in the brain (17). The onset of astrogliosis by 5 months in Cln7 KO mice was similar to the onset in other vLINCL mouse models, like the Cln6/nclf (21 weeks, 22) and Cln8/mnd mice (20 weeks, 27). Comparative analyses of neuroinflammation in Cln7 KO and Cln6/nclf mice revealed that astrogliosis and microgliosis were more pronounced in Cln7 KO mice. MRI revealed brain atrophy in the olfactory bulb, cerebellum and cerebral cortex of Cln7 KO mice at the end stage of the disease. The Cln7 KO mice, therefore, recapitulate key neuropathological features of human CLN7 disease, in which neuroinflammation and neurodegeneration in the cerebellar and cerebral cortex were observed (1).
Increased levels of multiple cathepsins have been reported in neurodegenerative disorders such as Alzheimer disease (28) and several LSDs including mouse models for Niemann–Pick disease C1 (29), Gaucher disease (30), CLN1 (19) and CLN11 disease (20). By quantitative reverse transcriptase (RT)–PCR and immunoblot analyses, increased expression of CtsB, CtsD and CtsZ was detected in Cln7 KO brains, indicating transcriptional upregulation rather than impaired degradation as cause for higher cathepsin levels. Increased CtsZ expression was found in microglia and neurons, but not in astrocytes, suggesting a cell-type-specific transcriptional upregulation of lysosomal proteins by the transcription factor EB (TFEB) in Cln7 KO brains (31). In agreement with this hypothesis, lysosomal biogenesis in activated microglia was significantly increased in the brain of progranulin KO mice, a model for CLN11 disease, due to enhanced dephosphorylation and nuclear translocation of TFEB (32). It is unlikely that elevated cathepsin expression in Cln7 KO brains is due to increased numbers of microglia as we found higher CtsZ levels also in neurons and higher CtsZ amounts were observed prior to the onset of microglial activation (5 versus 7 months). A compensatory upregulation of CtsD due to impaired proteolytic processing and activation as observed in a mouse model for CLN1 disease can be ruled out as mature forms of CtsD were detected in Cln7 KO brains. Elevated expression of lysosomal hydrolases may result in the destabilization of lysosomes and increased lysosomal membrane permeability, followed by increased apoptosis and neurodegeneration (33,34).
Impaired trafficking and altered proteolytic processing of lysosomal enzymes in neuronal cell cultures and in the brain have been reported in mouse models for CLN1 and CLN3 disease, respectively (19,35). Pulse-chase analyses revealed unaltered expression/proteolytic maturation and no missorting of CtsD and CtsZ into the media of Cln7-deficient MEFs. These findings and the presence of mature CtsD forms in the brain of Cln7 KO mice indicate that biosynthetic sorting of CtsD and CtsZ to lysosomes and lysosomal acidification were not impaired by the loss of Cln7. Radioactive uptake assays using Cln7-deficient MEFs showed that the M6P-dependent binding, internalization, lysosomal transport and degradation of endocytosed ASB were not altered, compared with control cells. In contrast, increased uptake and impaired degradation of arylsulfatase A were observed in MEFs deficient in the polytopic ER membrane protein Cln6 (36). The data suggest that lysosomal dysfunction in Cln7-deficient MEFs is not caused by impaired M6P-dependent biosynthetic and endocytic transport to lysosomes or altered lysosomal acidification resulting in defective proteolytic maturation of lysosomal enzymes.
Impairment of autophagy was described to be a major pathological hallmark in LSDs (37) and mouse models for CLN2 (38), CLN3 (39), CLN6 (22) and CLN10 (40) disease. In addition, p62- and ubiquitin-positive protein aggregates have been reported in the brain of mouse models for CLN2 (38) and CLN6 (22) disease and in arylsulfatase G-deficient mice (41). In this study, we first present disturbed macroautophagy in the brain of Cln7 KO mice as shown by increased levels of the marker for autophagosomes LC3-II. Decreased beclin 1 levels suggested that macroautophagy was not upregulated. Impaired autophagic flux could be caused either by defective autophagosome–lysosome fusion or by reduced lysosomal degradation (42). Lysosomal dysfunction in CLN7 KO mice as shown by the accumulation of SCMAS and saposin D and increased expression of CtsB, CtsD and CtsZ suggests that impaired degradation in lysosomes is a major mechanism contributing to impaired autophagic flux. Immunoblot and immunohistochemical analyses revealed the presence of detergent-insoluble p62- and polyubiquitin-positive protein aggregates, mainly in large neurons of Cln7 KO brains. The p62/SQSTM1 protein targets polyubiquitinated proteins to autophagosomes where they are selectively degraded via the autophagic pathway (43). p62- and ubiquitin-positive aggregates were detected in brain regions with strong SCMAS accumulation, e.g. in neurons of the hippocampus, cerebral cortex and thalamus. In summary, our findings suggest impaired macroautophagy as a main pathomechanism in CLN7 disease and confirm, together with data from mouse models for CLN2 (38), CLN3 (39), CLN6 (22) and CLN10 disease (40), the importance of a functional autophagic pathway for the clearance of protein aggregates in neurons.
In this study, we show clear differences between the phenotypes of the hypomorphic Cln7-lacZ gene trap mice and the Cln7 KO mice (9). Pathological features including increased mortality, clasping phenotype, myoclonus, hindleg paralysis, generalized neuroinflammation and neurodegeneration and biochemical alterations such as increased expression of lysosomal enzymes and impaired macroautophagy were only observed in Cln7 KO mice (9), but were absent in Cln7-lacZ gene trap mice. Our data suggest that residual activity of wild-type Cln7 protein in the hypomorphic Cln7-lacZ gene trap mice most likely reduces the severity of the neuropathological phenotype and delays disease course. In line with these findings, compound heterozygous patients carrying missense mutations affecting residues located in cytosolic domains with milder or atypical CLN7 disease course were identified (11,12). Patients carrying the CLN7 p.Ala157Pro mutation present with a juvenile phenotype and compound heterozygous adult patients carrying a severe CLN7 nonsense or frameshift mutation (CLN7 p.Glu381X or CLN7 p.Lys333Lys fsX3) in combination with a mild missense mutation develop a non-syndromic autosomal recessive macular dystrophy (12).
In summary, we present a new mouse model for CLN7 disease, vLINCL phenotype, which recapitulates key clinical and neuropathological features of human disease. Our data suggest that deficiency of the lysosomal membrane protein Cln7 in the brain of Cln7 KO mice results in lysosomal dysfunction, impaired constitutive autophagy and neurodegeneration late in disease. The identification of the Cln7 substrate and the mode of transport (uniport, symport and antiport) will clarify whether impaired metabolite efflux or altered ionic homeostasis contributes to lysosomal dysfunction in Cln7-defective cells in the brain.
Materials and Methods
Antibodies and reagents
KAPA™ Mouse Genotyping Hot Start Kit and jetPei were from VWR (Darmstadt, Germany). Nitrocellulose membranes were purchased from GE Healthcare Life Sciences (Freiburg, Germany). Albumin standard, restriction enzymes, TRIzol® reagent, high capacity cDNA reverse transcriptase kit, RT-PCR TaqMan® gene expression assays, IODO-GEN® and enhanced chemiluminescence (ECL) reagent were from Thermo Fisher Scientific (Schwerte, Germany). Plasmid Mini, plasmid Midi filter, gel extraction and PCR purification kits were from Qiagen (Hilden, Germany). QuikChange® site-directed mutagenesis kit was from Agilent Technologies (Santa Clara, CA, USA). Primary antibodies were goat anti-mouse cathepsin Z (R&D Systems, Minneapolis, MN, USA), goat anti-mouse cathepsin B (Neuromics, Edina, MN, USA), goat anti-mouse cathepsin D (Santa Cruz, Dallas, TX, USA), polyclonal rabbit anti-glial fibrillary acidic protein (GFAP, DAKO, Glostrup, Denmark), polyclonal rabbit anti-Iba1 (WAKO, Osaka, Japan), monoclonal mouse anti-calbindin (Sigma-Aldrich, Deisenhofen, Germany), monoclonal mouse anti-Neu N (Merck Millipore, Darmstadt, Germany), monoclonal anti-α-tubulin (Sigma-Aldrich), polyclonal rabbit anti-p62 (MBL International, Woburn, MA, USA), polyclonal anti-LC3-B (Sigma-Aldrich), monoclonal mouse anti-ubiquitin (Cell Signaling, Danvers, MA, USA), polyclonal rabbit anti-glyceraldehyde-3-phosphate dehydrogenase (Gapdh) and rabbit anti-beclin 1 antibodies from Santa Cruz Biotechnology (Cruz, CA, USA) and monoclonal anti-FLAG (Sigma-Aldrich). The affinity-purified polyclonal rabbit anti-mouse Cln7 antibody directed against a cytosolic loop of Cln7 was described previously (9). The monoclonal rat anti-mouse Lamp-1 antibody (clone 1D4B) developed by J. Thomas August and the monoclonal anti-β-tubulin antibody (clone E7) developed by M. Klymkowsky were obtained from the Developmental Studies Hybridoma Bank, created by the NICHD of the NIH and maintained at the Department of Biology, University of Iowa, Iowa City, USA. The rabbit polyclonal antibody raised against the SCMAS was from Dr E. Neufeld (UCLA, Los Angeles, CA, USA) and the goat anti-saposin D antibody was provided by Dr K. Sandhoff (44). Secondary goat anti-mouse, goat anti-rabbit and rabbit anti-goat antibodies coupled to horseradish peroxidase (HRP) were from Dianova (Hamburg, Germany). 3,3′-Diaminobenzidine tablets and protease inhibitors were purchased from Sigma-Aldrich. Oligonucleotide primers used for PCR and DNA sequencing were purchased from MWG Biotech (Munich, Germany). Sequencing was performed by Seqlab (Göttingen, Germany).
Generation of Cln7 KO mice
By using embryonic stem cells from the EUCOMM project, Cln7/Mfsd8(tm1a/tm1a) mice were recently generated harboring a gene trap targeting cassette containing a lacZ gene and a neomycin marker gene flanked by Flp recombinase target sites (FRT), together with the critical exon 2 of the Cln7/Mfsd8 gene flanked by loxP sites (9). Homozygous Cln7/Mfsd8(tm1a/tm1a) mice were mated with C57BL/6J Flp-deleter mice to excise the gene-targeting cassette generating Cln7/Mfsd8(tm1c/+) mice (Fig. 1A). Cln7/Mfsd8 (tm1c/+) mice harboring the floxed exon 2 were mated with C57BL/6J Cre-deleter mice to remove the critical exon 2 through Cre-mediated recombination generating Cln7/Mfsd8(tm1d/+) mice. Heterozygous Cln7/Mfsd8(tm1d/+) mice were then mated to generate homozygous Cln7/Mfsd8(tm1d/tm1d) mice.
Cln6-defective nclf (B6.Cg-Cln6nclf, 15) mice purchased from The Jackson Laboratory (Bar Harbor, ME, USA) were provided by Dr Thomas Braulke (UKE Hamburg) and CtsD(−/−) mice (45) by Dr Paul Saftig (University of Kiel). Animals were maintained under standard housing conditions on a 12 h light and 12 h dark schedule in a pathogen-free animal facility at the University Medical Center. All experiments were approved by local authorities (acceptance no. 73/10). Removal of tissues from mice was approved by the local animal welfare officer Dr A. Haemisch (no. ORG532). For preparation of tissues processed by histochemistry, immunofluorescence and immunohistochemistry, mice were deeply anesthetized and perfused from the left ventricle with phosphate-buffered saline (PBS), followed by 4% formaldehyde in PBS and post-fixation for 24 h in the same fixative. Tissues used for immunoblotting were either immediately frozen and kept at −80°C prior analysis or used directly for analytical experiments.
Genotyping of mice
For genotyping of mice, genomic DNA from tail tips was extracted using the KAPA Mouse Genotyping Hot Start Kit, according to the instructions of the manufacturer. For genotyping Mfsd8(tm1c/tm1c) and Mfsd8(tm1d/tm1d) mice, a multiplex PCR using primers Mfsd8-F (5′-TGGTGCATTAATACAGTCCTAGAATCCAGG-3′), Mfsd8-R (5′-CTAGGGAGGTTCAGATAGTAGAACCC-3′) and 15472-F (5′-TTCCACCTAGAGAATGGAGCGAGATAG-3′) was performed.
Cloning of 3xFLAG Cln7 expression constructs
The cDNA coding for mouse Cln7 (clone IRAMp995N0715Q) was purchased from Source BioScience GmbH (Berlin) and amplified with primers Cln7-F (5′-CGAAGCTTGCGAACCTGGGAAGTGAGGCC-3′) and Cln7-R (5′-CGTCTAGATTACTCCTGGATCCTCATATATCTG-3′). PCR products were separated and purified from agarose gels, cleaved with restriction enzymes HindIII and XbaI and cloned into the corresponding restriction sites of p3xFLAG-CMV®10 expression vector (Sigma-Aldrich) generating 3xFLAG Cln7. A QuikChange® site-directed mutagenesis reaction was performed to generate the mutant 3xFLAG Cln7p.Glu23Phe fsX16 cDNA construct using 3xFLAG Cln7 cDNA as a template and primers Ex2-F (5′-GGATCACCCGGAAGCAGGTTTTTCTATTGTGATA-3′) and Ex2-R (5′-TATCACAATAGAAAAACCTGCTTCCGGGTGATCC-3′).
Quantitative real-time PCR
Total RNA from brains of Cln7 KO and age-matched wild-type mice was prepared using TRIzol® reagent, following the manufacturer's instructions. About 1 µg of RNA was transcribed into cDNA using the high capacity cDNA reverse transcriptase kit. For RT–PCR, TaqMan gene expression assays including predesigned probes and primer sets for mouse Cln7 (Mm00511384_m1/exon boundary 1–2; Mm01294801_m1/exon boundary 5–6), Ctsb (Mm01310506_m1), Ctsd (Mm00515587_m1), Ctsz (Mm00517687_m1), Ctsl (Mm00515597_m1), Hexb (Mm00599880_m1) and Actb (Mm00607939_s1) were used. Actb was chosen as a housekeeping gene. Quantitative real-time PCR was performed as described previously (46) using the Mx3000P® QPCR system (Stratagene, The Netherlands). The relative expression of Cln7 mRNA in the same cDNA was normalized to the amount of Actb mRNA in the same cDNA using the comparative delta CT method (2 –ΔΔCT). For sequencing of Cln7 cDNA, mRNA from wild-type and Cln7 KO mice was transcribed into cDNA and amplified using primers Ex1-F (5′-ATGGCGAACCTGGGAAGTGAG-3′) and Ex4-R (5′-TGGCTGCCACCGAAATAGAGA-3′). PCR products were purified from agarose gels and sequenced in both directions.
Preparation and culture of primary MEFs
Preparation and culture of primary MEFs were performed, as described previously (47).
Cell cultures and cell transfection
MEFs and HEK293 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum and penicillin/streptomycin in a humidified atmosphere containing 5% CO2 at 37°C. Immortalization of MEFs was performed as described previously (47). HEK293 cells were transiently transfected using jetPei, according to the instructions of the manufacturer. Twenty-four hours after the start of transfection, cells were harvested and processed for western blotting.
Analysis of β-hexosaminidase enzymatic activities
The enzymatic activities of β-hexosaminidase were measured as described previously using 4-nitrophenyl-N-acetyl-β-d-glucosaminide as substrate (47).
Preparation of membrane extracts and detergent-soluble and -insoluble fractions from tissue homogenates
Preparation of total membrane extracts from mouse tissues was performed as described previously (9). For the preparation of detergent-soluble and -insoluble fractions, brains were homogenized in ice-cold 0.25 m sucrose buffer containing 50 mm Tris, pH 7.4, 1 mm EDTA and protease inhibitors, using a Glass teflon douncer and centrifuged at 500g to remove cell debris. After addition of Triton X-100 (final concentration 1%), the homogenates were incubated for 30 min and fractionated into Triton X-100-soluble and -insoluble fractions by centrifugation at 13 000g. Protein concentrations of the fractions were determined using the Bradford protein quantification kit. Detergent-insoluble fractions were incubated with 50 mm Tris, pH 7.5 containing 2% sodium dodecyl sulfate (SDS) and protease inhibitors. Detergent-soluble brain homogenates (100 µg protein) or detergent-insoluble fractions (corresponding to 100 µg total homogenate) were separated by SDS–polyacrylamide gel electrophoresis (PAGE) (12.5% acrylamide) and blotted onto nitrocellulose. Preparation of radioimmunoprecipitation assay buffer (RIPA) fractions from brain tissues for the detection of prosaposin and saposin D by immunoblotting was performed, as described previously (20). Prior to blocking, nitrocellulose membranes were boiled for 5 min in PBS. After blocking the membranes with phosphate-buffered saline containing 0.05% Tween (PBS-T) and 5% milk powder, membranes were incubated with primary antibodies, including cathepsin D (dilution: 1:1000), cathepsin Z (1:500), cathepsin B (1:1000), SCMAS (1:3000), saposin D (1:1000), p62 (1:1000), ubiquitin (1:1000), LC3 (1:3000), GAPDH (1:1000), α-tubulin (1:5000) and β-tubulin (1:500) for 16 h at 4°C followed by incubation with HRP-conjugated secondary antibodies (1:3000). Immunoreactive bands were visualized by ECL using a molecular imager (Model ChemiDoc XRS System, Bio-Rad, Munich, Germany). Bands were quantified using the software QuantityOne 4.5.0 (Bio-Rad). The experiments were performed in triplicate using three different Cln7 KO and age-matched wild-type mice.
For ultrastructural analyses, 100 μm thick vibratome sections from wild-type and Cln7 KO mice brains were cut with a vibratome (Leica VT 1000S). The sections were rinsed three times in 0.1 m sodium cacodylate buffer (pH 7.2–7.4) and osmicated using 1% osmium tetroxide in cacodylate buffer. Following osmication, the sections were dehydrated using ascending ethyl alcohol concentration steps, followed by two rinses in propylene oxide. Infiltration of the embedding medium was performed by immersing the pieces in a 1:1 mixture of propylene oxide and Epon and finally in neat Epon and hardened at 60°C. Semi-thin sections (0.5 μm) were prepared for light microscopy mounted on glass slides and stained for 1 min with 1% Toluidine blue. Ultrathin sections (60 nm) were examined using an EM902 (Zeiss, Germany). Images were taken with a MegaView III digital camera (A. Tröndle, Moorenweis, Germany).
Magnetic resonance imaging
MRI of mouse brains was performed at 36–37 weeks of age with a dedicated 7 T MR small animal imaging system (ClinScan, Bruker, Rheinstetten, Germany) with a four-element phased array mouse head receiver coil and a quadrature polarized whole body transmit coil. The imaging sequence was a three-dimensional (3D) constructive interference steady state (CISS) sequence with echo time (TE) = 3.87 ms, repetition time (TR) = 7.74 ms, flip angle (FA) = 50°, readout bandwidth = 200 Hz/pixel, number of excitations (NEX) = 4, matrix = 128 × 128 × 120, field of view (FOV) = 16 × 16 × 14.4 mm3, elliptical k-space sampling and 12.5 min scan time. Digital Imaging and Communications in Medicine (DICOM) image data were reviewed and multi-planar reconstructions were computed using the 3D viewer tool of the MRI system. ImageJ (http://imagej.nih.gov/ij) was used to review and adjust brightness and contrast of the data on the office PC.
Paraffin-embedded tissue samples were cut into 3 µm thick serial sections, mounted on glass slides and processed, following standard immunohistochemistry procedures using the Ventana Benchmark XT (Ventana, Tucson, AZ, USA). Briefly, deparaffinized sections were boiled for 30 min in 10 mm citrate buffer (10 mm sodium citrate, 0.05% Tween 20, pH 6.0) for antigen retrieval. Primary antibodies against SCMAS, GFAP, Iba1, cathepsin Z, p62 and calbindin were diluted in Tris-buffered saline (20 mm Tris, pH 7.3, 150 mm NaCl), containing 5% goat serum and 0.1% Triton X-100 and incubated for 1 h. Anti-rabbit or anti-mouse Histofine Simple Stain MAX Universal immunoperoxidase polymer kits (Nichirei Biosciences, Wedel, Germany) were used as secondary antibodies and detected with DAB solution. Sections were covered (Sakura Finetek, Staufen, Germany) and dried and pictures were taken using a Leica DMD108 (Wetzlar, Germany). For histochemistry, formaldehyde fixed tissues were embedded in paraffin, processed and stained, according to standard procedures.
Pulse-chase analyses of metabolically labeled MEFs
Wild-type and Cln7-deficient MEFs were starved for 1 h in medium lacking methionine followed by metabolic labeling in medium containing [35S] -methionine (100 µCi/ml) for 1 h. Cells were either harvested or chased in full medium supplemented with 5 mm mannose 6-phosphate (M6P) and unlabeled methionine (0.25 mg/ml) for 5 h. After cell lysis, CtsD and CtsZ were immunoprecipitated from cells and media and precipitates analyzed by fluorography, as described previously (47).
Uptake of radiolabeled [125I] -ASB
ASB was iodinated with sodium [125I] (75 TBq/mmol) and IODO-GEN, as described previously (48). Wild-type and Cln7-deficient MEFs grown in 35 mm dishes were pre-incubated with DMEM containing 0.1% bovine serum albumin (DMEM/BSA) for 30 min followed by incubation with [125I] -labeled ASB for 20 min in the absence (−) or presence (+) of 10 mm M6P in the medium. After removal of the medium and washing, the cells were either harvested or chased for 2 and 4 h, respectively. Total protein homogenates were separated by SDS–PAGE, followed by fluorography and autoradiography.
This work was supported by grants from the Deutsche Forschungsgemeinschaft (DFG grant number STO 761/3-1), the DFG Research Training Group (GRK1459) and the Batten Disease Support and Research Association (BDSRA).
We thank Dr Irm Hermans-Borgmeyer [University Medical Center Hamburg—Eppendorf (UKE), Germany] for providing the C57BL/6J Flp- and Cre-deleter mouse strains and Dr Kerstin Cornils (UKE) for the immortalization of primary MEFs. Dr Elizabeth F. Neufeld (UCLA, CA, USA) is acknowledged for the generous gift of the polyclonal anti-SCMAS antibody. We thank Dr Konrad Sandhoff (University of Bonn, Germany) for the anti-saposin D antibody. Dr Paul Saftig (University of Kiel, Germany) and Dr Thomas Braulke (UKE) are acknowledged for providing the CtsD (−/−) and Cln6/nclf mice, respectively. We thank Kirstin Hartmann from the Core Facility Mouse Pathology (UKE) for immunohistochemical stainings of mouse brains. We thank Dr Sandra Pohl (UKE) for critically reading the manuscript. Sarina Päsler is acknowledged for excellent technical assistance.
Conflict of Interest statement. None declared.