Acta Neuropathol (2014) 127:845–860 DOI 10.1007/s00401-014-1262-6

Original Paper

Common pathobiochemical hallmarks of progranulin‑associated frontotemporal lobar degeneration and neuronal ceroid lipofuscinosis Julia K. Götzl · Kohji Mori · Markus Damme · Katrin Fellerer · Sabina Tahirovic · Gernot Kleinberger · Jonathan Janssens · Julie van der Zee · Christina M. Lang · Elisabeth Kremmer · Jean‑Jacques Martin · Sebastiaan Engelborghs · Hans A. Kretzschmar · Thomas Arzberger · Christine Van Broeckhoven · Christian Haass · Anja Capell  Received: 15 January 2014 / Revised: 13 February 2014 / Accepted: 13 February 2014 / Published online: 12 March 2014 © Springer-Verlag Berlin Heidelberg 2014

Abstract  Heterozygous loss-of-function mutations in the progranulin (GRN) gene and the resulting reduction of GRN levels is a common genetic cause for frontotemporal lobar degeneration (FTLD) with accumulation of TAR DNA-binding protein (TDP)-43. Recently, it has been shown that a complete GRN deficiency due to a homozygous GRN loss-of-function mutation causes neuronal ceroid lipofuscinosis (NCL), a lysosomal storage disorder. These findings suggest that lysosomal dysfunction may also contribute to some extent to FTLD. Indeed, Grn(−/−) mice recapitulate not only pathobiochemical features of Electronic supplementary material  The online version of this article (doi:10.1007/s00401-014-1262-6) contains supplementary material, which is available to authorized users. J. K. Götzl · K. Mori · K. Fellerer · G. Kleinberger · C. M. Lang · C. Haass (*) · A. Capell (*)  Adolf‑Butenandt Institute, Biochemistry, Ludwig-MaximiliansUniversity Munich, Schillerstrasse 44, 80336 Munich, Germany e-mail: [email protected] A. Capell e-mail: [email protected] J. K. Götzl  Institute of Neuroscience, Technical University Munich, 80802 Munich, Germany M. Damme  Department of Biochemistry, Christian-Albrechts-University Kiel, 20498 Kiel, Germany S. Tahirovic · E. Kremmer · T. Arzberger · C. Haass  German Center for Neurodegenerative Diseases (DZNE) Munich, 80336 Munich, Germany G. Kleinberger · J. Janssens · J. van der Zee · C. Van Broeckhoven  Department of Molecular Genetics, Neurodegenerative Brain Disease Group, VIB, 2610 Antwerp, Belgium

GRN-associated FTLD-TDP (FTLD-TDP/GRN), but also those which are characteristic for NCL and lysosomal impairment. In Grn(−/−) mice the lysosomal proteins cathepsin D (CTSD), LAMP (lysosomal-associated membrane protein) 1 and the NCL storage components saposin D and subunit c of mitochondrial ATP synthase (SCMAS) were all found to be elevated. Moreover, these mice display increased levels of transmembrane protein (TMEM) 106B, a lysosomal protein known as a risk factor for FTLD-TDP pathology. In line with a potential pathological overlap of FTLD and NCL, Ctsd(−/−) mice, a model for NCL, show elevated levels of the FTLD-associated proteins GRN and TMEM106B. In addition, pathologically phosphorylated TDP-43 occurs in Ctsd(−/−) mice to a similar extent as in

G. Kleinberger · J. Janssens · J. van der Zee · J.-J. Martin · S. Engelborghs · C. Van Broeckhoven  Institute Born‑Bunge, University of Antwerp, 2610 Antwerp, Belgium G. Kleinberger · C. Haass  Munich Cluster for Systems Neurology (SyNergy), 80336 Munich, Germany E. Kremmer  Institute of Molecular Immunology, Helmholtz Center Munich, German Research Center for Environmental Health (GmbH), 81377 Munich, Germany H. A. Kretzschmar · T. Arzberger  Center for Neuropathology and Prion Research, LudwigMaximilians-University Munich, 81377 Munich, Germany T. Arzberger  Department of Psychiatry and Psychotherapy, LudwigMaximilians-University Munich, 80336 Munich, Germany

13

846

Grn(−/−) mice. Consistent with these findings, some NCL patients accumulate pathologically phosphorylated TDP43 within their brains. Based on these observations, we searched for pathological marker proteins, which are characteristic for NCL or lysosomal impairment in brains of FTLD-TDP/GRN patients. Strikingly, saposin D, SCMAS as well as the lysosomal proteins CTSD and LAMP1/2 are all elevated in patients with FTLD-TDP/GRN. Thus, our findings suggest that lysosomal storage disorders and GRN-associated FTLD may share common features. Keywords  Frontotemporal lobar degeneration (FTLD) · Progranulin (GRN) · TDP-43 · Neuronal ceroid lipofuscinosis (NCL) · Cathepsin D · Lysosome · Neurodegeneration

Introduction Frontotemporal lobar degeneration (FTLD) is the second most frequent form of dementia in people under the age of 65 years with 30–50 % of the patients having a positive family history [28, 62, 64]. The pathological hallmarks of the major variant of FTLD are intracellular ubiquitin and TAR DNA-binding protein (TDP)-43 positive inclusions [3, 49]. This subtype of FTLD is therefore designated FTLDTDP [42] distinguishing it from other types of FTLD such as FTLD-tau, FTLD-FUS (fused in sarcoma) [43] or FTLD-DPR [45]. In FTLD-TDP patients, TDP-43, a DNAand RNA-binding protein involved in transcription and splicing, is hyperphosphorylated, proteolytically processed and frequently mislocalized to the cytoplasm [3, 31, 36, 49]. The majority of FTLD-TDP causing mutations were identified in the GRN gene [4, 16], which account for up to 20 % of familial FTLD-TDP cases [24, 27]. Of the 69 pathogenic mutations reported to date (http://www.molgen.vibua.be/FTDMutations/) [17], most are loss-of-function mutations leading to GRN haploinsufficiency [27], which result in a severe reduction of GRN levels in tissues and biological fluids of patients [20, 25, 67]. Additionally, missense mutations also lead to reduced functional GRN by impairing secretion or misfolding [46, 63, 77, 80]. While lysosomal dysfunction and impaired autophagy have been discussed to contribute to neurodegenerative diseases such as Alzheimer’s disease and Parkinson’s disease (reviewed in [51, 58]), less evidence exists for lysosomal malfunction in FTLD-TDP. However, recently it was reported that a complete loss of GRN surprisingly results in adult-onset neuronal ceroid lipofuscinosis (NCL), a neuronal lysosomal storage disease (LSD) [68]. In that study, two siblings were shown to carry a homozygous deletion of four base pairs in the GRN gene (c.813_816del), which leads to a frameshift and a premature termination

13

Acta Neuropathol (2014) 127:845–860

of translation. The families of the parents were distantly related and in both families some cases with dementia without clear diagnosis were reported [68]. The identical heterozygous mutation was previously shown to cause FTLD-TDP [5, 41, 87]. NCL, a fatal disorder with progressive neuronal loss, is the most common cause of neurodegeneration in children and young adults. NCL is a clinically and genetically heterogeneous disorder, with autosomal recessive inheritance, in most NCL variants (reviewed in [35]). Lipofuscin, an autofluorescent lipopigment, is commonly found in all forms of NCL. Depending on the NCL variant, the major protein components of lipofuscin are either the sphingolipid activator proteins (saposin) A and D or subunit c of mitochondrial ATP synthase (SCMAS) [19, 73]. Interestingly, lipofuscin accumulation [1, 54, 83] as well as accumulation of ubiquitin [1, 26, 54, 83, 85, 86] and p62 [83] positive protein aggregates have been detected in Grn knockout mice. These data provided some evidence that reduced GRN might be associated with a dysfunction of the autophagosomal–lysosomal system, even before the linkage of GRN to NCL was known. Besides the fact that a total loss of GRN results in NCL, rare FTLD-causing mutations in the charged multivesicular body protein 2B (CHMP2B) gene [66] or in the valosincontaining protein (VCP) gene [81, 82] may further suggest a lysosomal involvement in FTLD [72, 74]. In addition, TMEM106B, a recently identified risk factor for GRNassociated FTLD-TDP (FTLD-TDP/GRN), is localized in late endosomes and lysosomes [7, 13, 40]. It was suggested that TMEM106B variants confer a risk for FTLD-TDP by increased expression levels of TMEM106B [76]. Elevated TMEM106B mRNA and protein levels were confirmed in GRN mutation carriers [13, 21]. Furthermore, studies in cellular model systems have provided evidence that overexpression of TMEM106B leads to lysosomal impairment and might influence GRN expression [7, 13]. In addition, GRN levels are influenced by sortilin, a member of the vacuolar protein sorting (Vps) 10 receptor family. Most likely sortilin is responsible for uptake and clearance of extracellular GRN, by facilitating its transport to lysosomes [33]. In this study, we aimed to investigate potential similarities between NCL and FTLD-TDP. Our findings suggest that lysosomal storage disorders and FTLD may share common pathological mechanisms and could belong to two extreme ends of one disease spectrum.

Materials and methods Human brain tissue Frontal cortical brain tissue from six FTLD-TDP patients of a Belgian GRN-mutation founder family (GRN

Acta Neuropathol (2014) 127:845–860

IVS1+5G>C) [8, 16], six control cases without neurologic pathology and one case with familial adult NCL [44, 50] were provided from the Antwerp Brain Bank (Institute Born-Bunge at the University of Antwerp, Antwerp, Belgium). The London Neurodegenerative Disease Brain Bank and Brains for Dementia Research provided frontal cortical brain tissue from two juvenile NCL patients, one NCL type 2 (CLN2) patient caused by deficiency of the lysosomal enzyme tripeptidyl peptidase 1 (TPP-1), and four normal control cases. In addition, paraffin-embedded frontal cortical sections from three NCL cases and one control were provided. Frontal cortical brain tissue of three CLN2 cases and one normal control cases without neurologic pathology were obtained from the Human Brain and Spinal Fluid Resource Center, Los Angeles, and brain tissue of four CLN2 cases were distributed by the Batten Disease Registry, Genetic Services and Specialty Clinical Laboratories, NYS Institute for Basic Research in Developmental Disabilities. Detailed information on pathology and clinical representation are summarized in Supplementary Tables 1 and 2. Mouse tissue Mice were killed by CO2 inhalation according to animal handling laws. Brain tissues were dissected from the Grn knockout mouse strain with deletion of Grn exons 2–13 [37] and the Ctsd knockout mouse strain [60].

847

antibodies were horseradish peroxidase-conjugated donkey anti-goat IgG (Santa Cruz Biotechnology; 1:5,000), antimouse IgG (Promega; 1:10,000), anti-rabbit IgG (Promega; 1:10,000), goat anti-rat IgG + IgM (L+M) (Dianova; 1:5,000), and generated mouse anti-rat IgG2c (1:1,000). For immunohistochemistry the following antibodies were used: rat monoclonal anti-mLAMP1 antibody clone 1D4B (developed by J. Thomas August, distributed by Developmental Studies Hybridoma Bank, NICHD, maintained by the University of Iowa, Department of Biology; 1:200), mouse anti-hLAMP1 clone H4A3 (developed by J. Thomas August and James E.K. Hildreth, distributed by Developmental Studies Hybridoma Bank; 1:2,000), rabbit polyclonal anti-IBA1 antibody (Wako Chemicals; 1:200), rabbit anti-SCMAS antibody (1:5,000) [59], goat polyclonal anti-saposin D antibody (1:200,000) [38], goat polyclonal anti-cathepsin D antibody (Santa Cruz Biotechnology; 1:200,000), rabbit polyclonal anti-TMEM106B antibody clone N2077 (1:5,000) [13], rat monoclonal anti-phospho TDP-43 (Ser409/Ser410) antibody clone 1D3 (1:50) [48], anti-HLA-DP/DQ/DR (CR3/43) antibody (Dako; 1:100) and mouse anti-p62 antibody (BD Biosciences; 1:1,000). Secondary antibodies directed to rat or rabbit IgG were conjugated to Alexa Fluor 488 or Alexa Fluor 555 (Invitrogen Life Technologies; 1:200). The DCS Super Vision 2 HRP-Polymer-Kit for mouse/rabbit antibodies (DCS Innovative Diagnostik-Systeme) and rabbit anti-goat antibody (Dako; 1:1,000) as a bridging antibody were used for human sections.

Antibodies Biochemical analysis and immunoblotting The following antibodies were used for immunoblotting: mouse monoclonal anti-β-actin antibody (Sigma-Aldrich; 1:10,000–20,000), rabbit polyclonal anti-calnexin antibody (StressGen; 1:2,000), mouse monoclonal anti-α-tubulin antibody (Sigma-Aldrich; 1:5,000–1:40,000), rabbit polyclonal anti-TDP43 antibody (ProteinTech Group; 1:1,000), rat monoclonal anti-phospho TDP-43 (Ser409/Ser410) antibody clone 1D3 (1:10) [48], rat monoclonal anti-C-terminal TDP-43 antibody clone 2H4 (1:50) generated against amino acid residues 404–413 of human TDP-43, rat monoclonal anti-TMEM106B antibody clone 6F2 (1:50) [40], goat polyclonal anti-cathepsin D antibody (Santa Cruz Biotechnology; 1:500), rat monoclonal anti-GRN antibody clone 8H10 (1:50) raised against the C-terminus of mouse GRN (amino acid residues 580–597), mouse monoclonal anti-hLAMP1 antibody clone H4A3 (Santa Cruz Biotechnology; 1:500), mouse monoclonal anti-hLAMP2 antibody clone H4B4 (Santa Cruz Biotechnology; 1:500), goat polyclonal anti-saposin D antibody (1:1,000) [38], rabbit anti-GFAP (glial fibrillary acidic protein) antibody (Dako; 1:5,000) and mouse monoclonal anti-GAPDH antibody (Invitrogen Life Technologies; 1:10,000). Secondary

Aliquots of crunched frozen brain tissue were used for sequential biochemical protein extraction. All buffers were freshly supplemented with protease inhibitor cocktail (Sigma-Aldrich) and phosphatase inhibitor (Roche Applied Science). First, the brain material was homogenized in high salt (HS) buffer (0.5 M NaCl, 10 mM Tris–HCl pH 7.5, 5 mM EDTA, 1 mM DTT, 10 % sucrose) to extract soluble, non-transmembrane proteins, such as GRN, CTSD, GFAP, prosaposin and soluble TDP-43. HS-soluble proteins were separated by centrifugation at 15,000×g, 4 °C for 30 min. To extract transmembrane proteins (Lamp1/2, TMEM106B, calnexin) and less soluble non-transmembrane proteins (prosaposin, saposin D, TDP-43, pTDP-43), the HS-insoluble protein pellet was further extracted with RIPA buffer (150 mM NaCl, 20 mM Tris–HCl pH 7.4, 1 % NP40, 0.05 % Triton X-100, 0.5 % sodium-desoxycholate, 2.5 mM ETDA), followed by centrifugation at 150,000×g, 4 °C for 45 min. Finally, insoluble proteins (TDP-43, pTDP-43) were further extracted with urea buffer (30 mM Tris–HCl pH 8.5, 7 M urea, 2 M thiourea, 4 % CHAPS) and centrifuged at 150,000×g, 4 °C for 45 min. The protein

13

848

concentrations of HS and RIPA fractions were determined by BCA protein assay (Pierce, Thermo Scientific). The urea fractions were adjusted according to protein concentration of the corresponding RIPA fractions. Equal amounts of protein were separated by SDS-PAGE and transferred onto polyvinylidene difluoride membranes (Immobilon-P, Merck Millipore) if not stated otherwise. For TMEM106B detection, sample preparation and separation by urea SDSPAGE was performed as previously described [40]. For detection of saposin D, separated protein samples were transferred onto nitrocellulose blotting membrane (Protran BA85, GE Healthcare Lifesciences) followed by boiling the membrane in phosphate buffered saline (PBS). For protein detection, membranes were probed with the indicated primary antibodies followed by horseradish peroxidaseconjugated secondary antibody. Bound antibodies were detected with ECL (Amersham Western Blotting Detection Reagent, GE Healthcare Lifesciences) or ECL Plus (Pierce ECL Plus Western Blotting Substrate, Thermo Scientific). Images were acquired with the LAS-4,000 image reader (Fujifilm Life Science) and quantitatively analyzed using the Multi-Gauge V3.0 software (Fujifilm Life Science).

Acta Neuropathol (2014) 127:845–860

was used as bridging antibody for 1 h. After brief rinsing with 0.02 % Brij35 in PBS, antibody binding was detected and visualized with DCS Super Vision 2 HRP-Polymer-Kit for mouse/rabbit antibodies (DCS Innovative DiagnostikSysteme) using DAB as chromogen. For phosphorylated TDP-43 (clone 1D3) and HLA-DP/DQ/DR (clone CR3/43) staining, automated staining with the Benchmark system (Roche Applied Science) was performed according to manufacturer’s protocol. Counterstaining for cellular structures was performed with haemalum. Microscopic images were obtained using a BX41 microscope with a 40×/0.65 objective (Olympus) and cellSens Ver 1.6 software (Olympus). Statistical analysis For statistical analysis the unpaired, two-tailed student t test or the Mann–Whitney test was performed. All statistical analyses were performed using the GraphPad Prism 5.04 program (GraphPad Software) and statistical significance was set at *P