UBC Faculty Research and Publications

Novel antibodies reveal presynaptic localization of C9orf72 protein and reduced protein levels in C9orf72… Frick, Petra; Sellier, Chantal; Mackenzie, Ian R A; Cheng, Chieh-Yu; Tahraoui-Bories, Julie; Martinat, Cecile; Pasterkamp, R. J; Prudlo, Johannes; Edbauer, Dieter; Oulad-Abdelghani, Mustapha; Feederle, Regina; Charlet-Berguerand, Nicolas; Neumann, Manuela Aug 3, 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-40478_2018_Article_579.pdf [ 3.23MB ]
JSON: 52383-1.0370939.json
JSON-LD: 52383-1.0370939-ld.json
RDF/XML (Pretty): 52383-1.0370939-rdf.xml
RDF/JSON: 52383-1.0370939-rdf.json
Turtle: 52383-1.0370939-turtle.txt
N-Triples: 52383-1.0370939-rdf-ntriples.txt
Original Record: 52383-1.0370939-source.json
Full Text

Full Text

RESEARCH Open AccessNovel antibodies reveal presynapticlocalization of C9orf72 protein and reducedprotein levels in C9orf72 mutation carriersPetra Frick1, Chantal Sellier2, Ian R. A. Mackenzie3, Chieh-Yu Cheng1, Julie Tahraoui-Bories4, Cecile Martinat4,R. Jeroen Pasterkamp5, Johannes Prudlo6,7, Dieter Edbauer8,9,10, Mustapha Oulad-Abdelghani2, Regina Feederle8,9,11,Nicolas Charlet-Berguerand2 and Manuela Neumann1,12*AbstractHexanucleotide repeat expansion in C9orf72 is the most common genetic cause of frontotemporal dementia andamyotrophic lateral sclerosis, but the pathogenic mechanism of this mutation remains unresolved.Haploinsufficiency has been proposed as one potential mechanism. However, insights if and how reduced C9orf72proteins levels might contribute to disease pathogenesis are still limited because C9orf72 expression, localizationand functions in the central nervous system (CNS) are uncertain, in part due to the poor specificity of currentlyavailable C9orf72 antibodies.Here, we generated and characterized novel knock-out validated monoclonal rat and mouse antibodies againstC9orf72. We found that C9orf72 is a low abundant, cytoplasmic, highly soluble protein with the long 481 aminoacid isoform being the predominant, if not exclusively, expressed protein isoform in mouse tissues and humanbrain. As consequence of the C9orf72 repeat expansion, C9orf72 protein levels in the cerebellum were reduced to80% in our series of C9orf72 mutation carriers (n = 17) compared to controls (n = 26). However, no associationsbetween cerebellar protein levels and clinical phenotypes were seen. Finally, by utilizing complementaryimmunohistochemical and biochemical approaches including analysis of human iPSC derived motor neurons,we identified C9orf72, in addition to its association to lysosomes, to be localized to the presynapses and ableto interact with all members of the RAB3 protein family, suggestive of a role for C9orf72 in regulating synaptic vesiclefunctions by potentially acting as guanine nucleotide exchange factor for RAB3 proteins.In conclusion, our findings provide further evidence for haploinsufficiency as potential mechanism in C9orf72pathogenesis by demonstrating reduced protein levels in C9orf72 mutation carriers and important novel insights intothe physiological role of C9orf72 in the CNS. Moreover, the described novel monoclonal C9orf72 antibodies will beuseful tools to further dissect the cellular and molecular functions of C9orf72.Keywords: Frontotemporal dementia, Frontotemporal lobar degeneration, Amyotrophic lateral sclerosis, C9orf72, RAB3,Synaptic vesicles* Correspondence: Manuela.Neumann@dzne.de1German Center for Neurodegenerative Diseases (DZNE), Otfried-Müllerstr. 23,72076 Tübingen, Germany12Department of Neuropathology, University of Tübingen, Tübingen,GermanyFull list of author information is available at the end of the article© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Frick et al. Acta Neuropathologica Communications  (2018) 6:72 https://doi.org/10.1186/s40478-018-0579-0IntroductionIn 2011, abnormal expansion of a GGGGCC hexanucleo-tide repeat in a predicted non-coding region of the chromo-some 9 open reading frame 72 (C9orf72) gene wasidentified as the most common genetic cause of familialand sporadic forms of frontotemporal dementia (FTD) andamyotrophic lateral sclerosis (ALS) including families inwhich both conditions co-occur [13, 40]. How the repeatexpansion in C9orf72 contributes to neurodegeneration iscurrently unresolved. As for other repeat expansion muta-tions, three non-exclusive pathogenic mechanisms havebeen proposed: a toxic RNA gain of function by accumula-tion of transcripts with expanded repeats in RNA foci thatbind and sequester specific RNA binding proteins resultingin disruption of their function, a toxic protein gain of func-tion through aberrantly expressed proteins generated by re-peat associated non-ATG translation of transcripts withexpanded repeats, and haploinsufficiency as consequence ofexpanded repeats altering expression of their hosting gene.Based on studies using human tissues and model sys-tems there is evidence for all three mechanisms being inplay for C9orf72 repeat expansions. First, RNA foci com-posed of mutated sense and antisense C9orf72 tran-scripts are present in up to 50% of neuronal nuclei inkey anatomical regions [12, 13, 15, 32]. RNAs with theextended repeats are reported to fold into G quadruplexstructures able to bind several proteins [10, 19, 24, 34].Second, GGGGCC repeat expansions in sense and anti-sense transcripts have been shown to act as templatesfor the synthesis of five aberrantly expressed dipeptiderepeat (DPR) proteins by repeat-associated non-ATGtranslation. Neuronal inclusions composed of these DPRproteins are highly specific neuropathological hallmarkfeatures of C9orf72 mutation carriers [2, 15, 33, 35, 60]and have shown neurotoxic effects in various model sys-tems upon overexpression under artificial AUG startcodon [8, 29, 31, 37]. However, so far no consistent cor-relation of RNA foci and DPR protein pathology withthe regional pattern of neurodegeneration and/or pres-ence of TDP-43 pathology, the neuropathological hall-mark feature of ALS and FTD including cases withC9orf72 mutations [36], has emerged despite extensivequantitative analysis [11, 12, 26, 27]. Thus, the patho-genic relevancies of RNA foci and DPR proteins in dis-ease pathogenesis remain to be clarified. Finally, thepotential role of haploinsufficiency is supported by theconsistent demonstration of decreased C9orf72 mRNAlevels in different tissues including postmortem braintissue of C9orf72 mutation carriers [13, 17, 52], motordeficits reported in C9orf72 knockdown zebrafish [9] andC.elegans [50] models as well as contribution of reducedprotein levels to neurodegeneration in induced humanmotor neurons from C9orf72 mutation carriers [46]. How-ever, the absence of obvious neurologic phenotypes inC9orf72 knock-out mice argues against a sole role of hap-loinsufficiency in disease pathogenesis [3, 23, 38, 47], al-though it remains to be tested whether additional stressorsmight be required as second hits to induce neurodegenera-tion in these mouse models.Further investigations on the potential contribution ofC9orf72 haploinsufficiency in disease pathogenesis ofALS and FTD is hampered due to little knowledge aboutthe physiological functions of C9orf72. Based on bio-informatics analysis, C9orf72 is predicted to contain dif-ferentially expressed in normal and neoplastic cells(DENN) domains [25, 59], that are characteristics ofguanine nucleotide exchange factors (GEFs) for specificRab GTPases (Rabs) [55, 58]. Rabs are key determinantsof organelle identities that switch between two conform-ational states, an inactive form bound to GDP and an ac-tive form bound to GTP. The conversion of GDP-boundto GTP-bound Rabs is facilitated by GEFs and results inenrichment of activated Rabs at distinct target membraneswhere they recruit additional proteins in order to mediatepractically all membrane trafficking events in eukaryotes[39]. In agreement with the bioinformatics prediction, re-cent studies have shown that C9orf72 is part of a proteincomplex containing Smith-Magenis syndrome chromo-somal region candidate gene 8 (SMCR8) and WD repeatcontaining protein 41 (WDR41) [1, 21, 45, 48, 51, 54, 57],and that this complex possesses GEF activity for RAB8Aand RAB39B, two Rabs involved in autophagy [45, 57].Accordingly, C9orf72 has been found to modulate differ-ent steps of the autophagy and endo-lysosomal pathways[1, 21, 45, 48, 51, 54, 57].Beyond that, insights on functions of C9orf72 andwhether C9orf72 repeat expansions result in reducedC9orf72 protein levels are so far limited in part due to poorspecificity of currently available C9orf72 antibodies withconflicting and inconsistent results reported on its subcellu-lar distribution and expressed isoforms. Importantly, sincethe subcellular localization of a predicted GEF protein de-termines where in the cell specific Rab-GTPases might beactivated, it is essential to gain further knowledge on the ex-pression pattern of C9orf72 particularly in the central ner-vous system (CNS) in order to dissect its functions.Therefore, the aim of the present study was to gener-ate and characterize novel antibodies against C9orf72with the goal to investigate the subcellular distributionand function of C9orf72 in the CNS and to test for hap-loinsufficiency on the protein level in postmortem braintissue of ALS, ALS/FTD and FTD cases with C9orf72 re-peat expansions.Material and methodsGeneration of monoclonal antibodies against C9orf72Rat monoclonal antibodies (mAbs) were generated againstsynthesized peptides corresponding to amino acid residuesFrick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 2 of 17321–335 (peptide 1), 417–434 (peptide 2), 100–113 (pep-tide 3), 140–155 (peptide 4) and 184–203 (peptide 5) ofhuman C9orf72 (Protein ID: Q96LT7) coupled to bovineserum albumin (BSA) or ovalbumin (OVA) (Peptide Spe-cialty Laboratories GmbH, Heidelberg, Germany). Lou/crats were immunized subcutaneously and intraperitoneallywith a mixture of 50 μg OVA-coupled peptides, 5 nmolCpG 2006 oligonucleotide (Tib Molbiol, Berlin, Germany),and an equal volume of incomplete Freund’s adjuvant.After a 6 weeks interval, rats were boosted with 50 μgOVA-coupled peptides in PBS. Hyperimmune spleen cellswere fused with the mouse myeloma cell line P3X63Ag8.653using standard procedures. As primary screen supernatantswere screened in a solid-phase enzyme-linked immunosorb-ent assay (ELISA) with the BSA-coupled peptides. Positivesupernatants were tested by immunoblot analysis using pro-tein lysates from transiently transfected HEK293T cells. Cellsfrom identified and further characterized mAbs were sub-cloned at least twice by limiting dilution. Experiments wereperformed with cell culture supernatants from establishedclones 12E7 (rat IgG1) against peptide 4, 5F6 (rat IgG1) and12G10 (rat IgG1) against peptide 5, and 2H7 (rat IgG1) and15C5 (rat IgG2a) against peptide 2. For generation of mousemonoclonal antibodies, 100 μg of HIS-tagged C9orf72/SMCR8/ WDR41 complex purified from baculovirus in-fected insect SF2 cells and 200 μg of poly (I/C) as adjuvantwere injected intraperitoneally into 2 months old BALB/c fe-male mice. Five injections were performed at 2 week inter-vals. Spleen cells were fused with Sp2/0.Agl4 myeloma cells.Hybridoma culture supernatants were tested by ELISA forcross-reaction with GST-tagged C9orf72 purified fromBL21(RIL) pRARE E. coli. Positive supernatants were thentested by immunofluorescence and immunoblot on lysatesfrom HA-tagged C9orf72 transfected HEK293T cells. Spe-cific cultures were cloned twice on soft agar. One specificclone 1C1 was established (mouse IgG1k), and ascites fluidwas prepared by injection of 2 × 106 hybridoma cells intoFreund adjuvant-primed BALB/c mice. Epitope mapping re-vealed that 1C1 recognizes a region between amino acid170–200 of human C9orf72. All animal experimental proce-dures were performed according to the European authorityguidelines.Other antibodiesOther primary antibodies used in this study included:anti-C9orf72 (22637–1-AP, Proteintech Group; 66,140–1-Ig,Proteintech Group; GTX119776, GeneTex; sc-138,763,Santa Cruz; AP12928b, Abgent), anti-synaptoporin (102,002,Synaptic Systems), anti-synaptophysin (101,002, SynapticSystems; M0776, Dako), anti-GAPDH (MAB374, Millipore),anti-α-tubulin (T5168, Sigma), anti-RAB3 (107,003, SynapticSystems), anti-RAB3A (HPA003160, Atlas Antibodies),anti-PSD-95 (ab18258, Abcam), anti-LAMP1 (ab24170,Abcam), anti-LAMP2 (10397–1-AP, Proteintech Group),anti-SMCR8 (1D2, homemade), anti-RAB39B (1H1, home-made), anti-Histone H3 (ab21054, Abcam), anti-Cox IV(ab16056, Abcam), anti-myc tag (2276 and 2272, Cell Signal-ing), anti-FLAG tag (PA1-984B, Pierce), and anti-HA tag(26,183, Pierce or 3F10, Merck).Plasmids and constructspCMV6 containing C-terminally myc-DDK-tagged cDNAsof human C9orf72 variant 1 (NM_145005), human C9orf72variant 2 (NM_018325), mouse C9orf72 encoding for iso-form 1 (NM_001081343), mouse C9orf72 encoding for iso-form 2 (BC026738), C-terminally FLAG-tagged humancDNAs of SMCR8, and Rab proteins were purchased fromOriGene. cDNAs for C9orf72 were subcloned intopcDNA5/FRT/TO (Invitrogen) with or without myc-DDK-tag. Optimized cDNA for human N-terminally HA-taggedC9orf72 variant 1 cloned in pcDNA3 was described previ-ously [45] and is available through Addgene.Cell cultures and transfectionHEK293T cells were cultured in Dulbecco’s modified Ea-gle’s medium (DMEM) with Glutamax (Invitrogen) sup-plemented with 10% (vol/vol) fetal calf serum (FCS,Invitrogen) and penicillin/streptomycin (Invitrogen) at37 °C with 5% CO2. Transfection of cells for immuno-blot or immunofluorescence analysis was carried outwith Fugene HD (Promega) or calcium phosphatemethod and indicated plasmids. Medium was exchanged7 h after transfection and cells were examined 48 h aftertransfection. For immunoprecipitation experiments6.25 × 105 HEK293T cells were co-transfected for 24 hwith indicated plasmids using Fugene HD (Promega)and further processed as described below.Motor neurons were generated from human inducedpluripotent stem cells (iPSCs) as previously described[28]. Briefly, iPSCs (clone 56c2, Phenocell) were dissoci-ated with Accutase (Invitrogen) and resuspended in dif-ferentiation medium N2B27 (DMEM F12, Neurobasalvol:vol supplemented with N2 (Life Technologies), B27(Life Technologies), Pen-strep 1%, β-mercapthoethanol0.1% (Life Technologies), ascorbic acid (AA, 0.5 μM,Sigma Aldrich)) with Y-27632 (5 μM, STemGent),SB431542 (40 μM, Tocris), LDN 193189 (0.2 μM, Milte-nyi) and Chir-99,021 (3 M, Tocris) for 2 days.Chir-99,021 was maintained from day 0 to day 4 of dif-ferentiation and SB431542/LDN 193189 from day 0 today 7. Differentiation medium was supplemented withretinoic acid (RA, 100 nM, Sigma Aldrich) and smooth-ened agonist (SGA, 500 nM, Calbiochem) at day 2 of dif-ferentiation. At day 7, medium was changed for N2B27supplemented with RA (100 nM), SAG (500 nM) andBDNF (10 ng/ml). Two days later, DAPT (10 μM,Tocris) was added. Ten days after the beginning of thedifferentiation, cells were dissociated with Trypsin andFrick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 3 of 17plated on poly-ornithin/laminin coated glass coverslipsat a density of 33,000 cells per cm2 in N2B27 mediumcontaining RA (100 nM), SAG (500 nM), BDNF (10 ng/ml) and DAPT (10 μM). Seven days later, DAPT waswithdrawn and replaced by GNDF (10 ng/ml). Cellswere analyzed by immunocytochemistry after 30 days ofdifferentiation.Immunocytochemistry and immunofluorescenceHEK293T cells: Cells were fixed on coverslips for 15 minin 4% paraformaldehyde (PFA) in PBS, permeabilized for5 min in 0.25% Triton X-100 in PBS and subsequentlyincubated for 1 h in blocking buffer (2% fetal bovineserum and 0.1% Triton X-100 in PBS). Coverslips wereincubated overnight at 4 °C with the indicated C9orf72mAbs and anti-myc antibodies diluted in blocking buf-fer. After washing steps with PBS, coverslips were incu-bated for 1 h at room temperature with Alexa Fluor488- or 594-conjugated secondary antibodies (Invitro-gen), washed with PBS, incubated in Hoechst 33342 for5 min to stain nuclei, and mounted onto glass slidesusing fluorescence mounting medium (Dako Agilent).Human iPSC derived motor neurons: Glass coverslipswith plated cells were fixed in 4% PFA for 10 min andsubsequently washed three times with PBS. The cover-slips were incubated for 10 min in PBS plus 0.5% TritonX-100 and washed three times with PBS before incuba-tion for 1 h with indicated primary antibodies (1C1,12E7, SMCR8, LAMP1, LAMP2, RAB39B, RAB3, synap-tophysin). Coverslips were washed twice with PBS beforeincubation with a Cyanine 3-conjugated goat anti-rabbit,anti-mouse or anti-rat and Alexa 488-conjugated donkeyanti-rat or anti-mouse secondary antibodies for 1 h,washed with PBS, incubated for 2 min with DAPI (1/10000dilution in PBS) and rinsed twice with PBS before mount-ing onto glass slides using ProLong Mountant (MolecularProbes). For quantification, hundred C9orf72-positive vesi-cles per coverslip were counted and the mean and standarddeviation of co-localization was calculated from three inde-pendent experiments.Images were obtained with a fluorescence microscope(Leica) equipped with a CCD camera.ImmunoprecipitationCo-transfected HEK293T cells were scraped into radio-immunoprecipitation assay (RIPA) buffer (50 mM Tris–HCI pH 7.6, 150 mM NaCl, 1% NP-40) and centrifugedfor 15 min at 18,000 g at 4 °C; 20 μl of pre-washed HAmagnetic beads (Dynabeads) was added, and immuno-precipitation was carried out for 1 h at 4 °C with constantrotation. For immunoprecipitation of endogenous C9orf72from mouse brains, 100 μl of mAb 1C1 was incubated withmouse brain extract in RIPA buffer overnight at 4 °C.Pre-cleared A/G magnetic beads (Life Technologies) wereadded and immunoprecipitation was carried out for 3 h at4 °C with constant rotation.After three washes of beads with washing buffer (50 mMTris–HCI pH 7.6, 150 mM NaCl, 0.05% Tween), boundproteins were eluted by boiling in SDS–PAGE loading buf-fer at 95 °C for 5 min and analyzed by immunoblot usingthe indicated antibodies.Mouse tissueC57BL/6 N mice were bred in our facilities. Animalswere euthanized using CO2, organs were removed andfurther processed for histological and biochemical ana-lysis. Brain tissue from C9orf72 knock-out mice (C9−/−;n = 3) described previously [47] and non-transgenic lit-termates (C9+/+; n = 3) were kindly provided by the Pas-terkamp lab for knock-out validation experiments ofC9orf72 mAbs. All animal experiments were carried outin accordance with the institutional and European au-thority guidelines.Human postmortem tissueHuman post mortem tissues were obtained from thebrain banks affiliated with the University of Tübingenand the University of British Columbia. Consent for aut-opsy was obtained from probands or their legal repre-sentative in accordance with local institutional reviewboards.The study cohort consisted of 18 cases with a C9orf72repeat expansion mutation (C9+) covering the completeclinical spectrum presenting with ALS (n = 6), FTD/ALS(n = 5) or FTD (n = 7) and 33 control cases (C9-) con-sisting of neurologic disease controls with TDP-43 path-ology in the absence of a C9orf72 repeat expansionmutation clinically presenting with ALS (n = 16), ALS/FTD (n = 6) and FTD (n = 5), three neurologicallyhealthy controls, two Alzheimer’ disease cases and onecase with hypoxic encephalopathy. C9orf72 repeat ex-pansions have been identified by genetic testing usingrepeat-primed PCR or were inferred from the presenceof DPR protein pathology. Details for cases are providedin Additional file 1: Table S1.Immunohistochemistry and in situ hybridizationImmunohistochemistry on mouse and human CNS tis-sue was performed on 2–5 μm thick sections of formalinfixed, paraffin-embedded (FFPE) tissue using the Ven-tana BenchMark XT automated staining system with theiVIEW DAB detection kit (Ventana).To establish a protocol for the immunohistochemicaldetection of C9orf72, mAb 1C1 and 12E7 were first testedon sections from C9+/+ and C9−/− mouse brains (forma-lin fixation: 24 h) with different dilutions and differentantigen pretreatments (boiling for 60 min in CC1 or CC2buffer (Ventana), or no pretreatment). A specific, knock-outFrick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 4 of 17validated signal was obtained in mouse brains for mAb 1C1using 1:100 dilution and boiling for 60 min with CC1 bufferas pretreatment and this protocol was used for further ex-periments in this study.To test for the immunohistochemical detection ofC9orf72 in human postmortem FFPE brain tissue, hippo-campus, frontal cortex and cerebellum sections (n = 3 C9+and n = 3 C9- cases, formalin fixation times ranging from2 weeks to several months) were stained with the abovedescribed protocol for 1C1; however, no immunoreactivitywas observed. To test for the influence of formalin fixationtimes on C9orf72 immunoreactivity, mouse brains (n = 2)were fixed in formalin for 12, 24, 48, 72 or 96 h beforeparaffin embedding and sections stained using the abovedescribed 1C1 protocol. Best signals were obtained for 12and 24 h fixation time points, while a dramatic reductionin immunoreactivity was observed with increasing forma-lin fixation times. No improvement of 1C1 immunoreac-tivity signals could be achieved in the > 24 h formalinfixed mouse tissues by testing additional pretreatments(boiling for 90 min in CC1 or CC2 buffer (Ventana), en-zymatic digestion with Protease 1 and 2 (Ventana), or for-mic acid treatment for 5 min).No specific signal could be detected in mouse FFPEsections using mAb 12E7 with any tested pretreatmentin mouse tissue, indicating that its epitope is masked inFFPE tissue. Subsequently, no staining was observed for12E7 in human postmortem FFPE tissue.Human C9orf72 specific antibodies 5F6 and 12G10were tested in parallel on mouse brain sections (used asnegative control) and human postmortem hippocampusand cerebellum sections with different dilutions and pre-treatments (boiling for 60 min in CC1 or CC2 buffer(Ventana), or no pretreatment). However, all conditionsresulted in similar staining profiles in human and mousetissue due to cross-reactivity with unrelated proteins(Additional file 1: Figure S1). Additional antibodies for im-munohistochemistry included polyclonal rabbit anti-synaptoporin, anti-synaptophysin and LAMP1.In situ hybridization was performed on FFPE mouse sec-tions using the RNAscope® 2.5 HD Reagent Kit-RED (Ad-vanced Cell Diagnostics) according to the user manuals322,452-USM and 322,360-USM. The target region of theprobe (Mm-3110043O21Rik) corresponds to nucleotides601–1539 of mouse C9orf72 mRNA (NM_001081343).RIPA lysatesMouse tissues and human postmortem brain tissueswere homogenized in RIPA buffer (50 mM Tris-HCl,150 mM NaCl, 1% (v/v) octylphenoxy poly(ethyleneox-y)ethanol (IPEGAL), 5 mM EDTA, 0.5% (w/v) sodiumdeoxycholate, 0.1% (w/v) sodium dodecyl sulfate (SDS),pH 8.0) at 1 g/2 ml ratio. Lysates were passed through18 and 21 Gauge needles and sonicated to shear nucleicacids. Cellular debris was removed by centrifugation for2 min at 3000×g at 4 °C and supernatant collected asRIPA lysate. Transfected cells were scraped into RIPAbuffer and subjected to brief sonication on ice. Cell deb-ris was removed by centrifugation for 5 min at 12,000×gat 4 °C. Protein concentration was determined using thePierce™ BCA Protein Assay Kit (Thermo Fisher Scien-tific) according to the manufacturer’s recommendations.Subcellular fractionationsCytoplasmic and nuclear proteins from mouse brainswere extracted as previously described [18]. Briefly, fore-brain tissue from wild-type mice was dounce homoge-nized in buffer containing 10 mM HEPES, 10 mM NaCl,1 mM KH2PO4, 5 mM NaHCO3, 5 mM EDTA, 1 mMCaCl2, 0.5 mM MgCl2 supplemented with protease in-hibitors (1 g/10 ml ratio). After 10 min incubation onice, sucrose was added to a final concentration of125 mM. An aliquot was collected representing whole lys-ate, and the remaining homogenate was centrifuged for10 min at 1000×g at 4 °C. The supernatant was collectedas cytoplasmic fraction. The pellet was washed in TSEbuffer (10 mM Tris, 300 mM sucrose, 1 mM EDTA, 0.1%IPEGAL supplemented with protease inhibitors) and cen-trifuged for 12 min at 6250×g at 4 °C. The supernatantwas removed and the pellet was washed twice in TSE buf-fer and centrifuged for 10 min at 4000×g at 4 °C. The finalpellet was re-suspended in RIPA buffer containing 2%SDS by sonication and collected as nuclear fraction.Enrichment of synaptosomal compartments frommouse forebrain was performed according to previouslydescribed protocols [7, 16, 20] with minor modifications(see schematic presentation Fig. 2e). Briefly, mouse fore-brains were homogenized in 5 volumes of ice-cold buff-ered sucrose (0.32 M sucrose, 4 mM HEPES-NaOH,pH 7.3) using a Dounce tissue grinder with tight-fittingglass pestle (Wheaton). The homogenate was centrifugedfor 5 min at 1000×g at 4 °C. The pellet (P1) containingcellular debris and nuclei was discarded, and the post-nuclear supernatant (S1) was centrifuged for 15 min at12,500×g. The supernatant (S2; synaptosome depletedfraction containing cytosol and microsomes) was re-moved, and the pellet (P2) representing the crude synap-tosomal fraction was washed in buffered sucrose andcentrifugation repeated as above. For preparation ofcrude synaptic vesicles, P2 was resuspended in bufferedsucrose, and subjected to osmotic lysis by adding 9 vol-umes of ice-cold water. After three Dounce strokes, 1 MTris-HCl pH 7.5 was added to a final concentration of10 mM. After 30 min incubation on ice, the suspensionwas centrifuged for 20 min at 25,000×g to yield a lysatepellet (LP1, enriched in presynaptic membrane proteins)and a lysate supernatant (LS1). LS1 was further centri-fuged for 2 h at 174000×g to yield LS2 containing theFrick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 5 of 17soluble synaptosomal content and LP2 representing thecrude synaptic vesicle (SV) fraction.For preparation of postsynaptic densities (PSDs) thewashed P2 fraction was suspended in solution B (0.32 Msucrose, 1 mM NaHCO3) by six Dounce strokes. Eightml of the resuspended P2 was loaded on a discontinuoussucrose gradient (10 ml each of 0.85, 1.0, and 1.2 M su-crose solutions containing 1 mM NaHCO3) and centri-fuged for 2 h at 82,500×g. The fraction between 1.0 and1.2 M sucrose containing synaptosomes was collectedand diluted with solution B. An equal volume of 1% Tri-ton X-100 in 0.32 M sucrose, 12 mM Tris-HCI pH 8.1was added, the suspension incubated for 15 min at 4 °Cunder continuous stirring and then centrifuged for20 min at 32,800×g. The pellet (P3) was resuspended insolution B, and was layered on a sucrose gradient (com-posed of 4 ml of 2 M sucrose, 3 ml of 1.5 M sucrose,and 3.0 ml of 1 M sucrose solutions containing 1 MNaHCO3) and centrifuged for 2 h at 201,800×g. Thefraction between 1.5 and 2.0 M sucrose was collectedand diluted to a final volume of 2 ml with solution B.An equal volume of 1% Triton X-100, 150 mM KCI wasadded and the suspension was centrifuged for 20 min at201,800×g. The resulting pellet P4 containing PSDs was re-suspended in solution B. For immunoblot analysis, 50 μgtotal protein per fraction were loaded for C9orf72 detec-tion and 20 μg for detection of other marker proteins.Biochemical analysis of human postmortem brain tissueThe sequential extraction of proteins with buffers of in-creasing stringency from fresh-frozen postmortem frontalcortex from C9orf72 mutation cases (n = 2) and controls(n = 4) was performed as described previously [36]. Briefly,gray matter was extracted at 5 ml/g (v/w) with low-salt(LS) buffer (10 mM Tris, pH 7.5, 2 mM EDTA, 1 mM di-thiothreitol (DTT), 10% sucrose, and a cocktail of proteaseinhibitors), high-salt (HS) buffer (50 mM Tris, pH 7.5,0.5 M NaCl, 2 mM EDTA, 1 mM DTT, 10% sucrose) with1% Triton X-100, myelin flotation buffer (HS buffer con-taining 30% sucrose), and Sarkosyl (SARK) buffer (HS buf-fer + 1% N-lauroyl-sarcosine). The detergent-insolublematerial was finally extracted in 0.25 ml/g of urea buffer(7 M urea, 2 M thiourea, 4% 3-[(3-cholamidopropyl)di-methylammonio]-1-propanesulfonate, 30 mM Tris, pH 8.5).Equal amounts of protein fractions per case (10 μl for LS,HS, SARK and UREA) were analyzed by immunoblot. Forquantification of C9orf72 levels, RIPA lysates were generatedfrom frozen postmortem cerebellum (n = 17 C9+, n = 26C9-) and frontal cortex (n = 10 C9-) and analyzed byimmunblot analysis as described below. To correlateC9orf72 expression in frontal cortex with levels of neurode-generation frontal cortex sections were assessed on H&Estains and neurodegeneration/cell death graded as absent(0), mild (1), moderate (2) or severe (3) based on the pres-ence of spongiosis, neuronal loss, and gliosis.Immunoblot analysisProteins were separated by SDS-polyacrylamide gel electro-phoresis (SDS-PAGE). With the exception of immunoblotsfor the quantification of C9orf72 levels in human lysates, im-munoblots were performed using enhanced chemilumines-cence detection. Therefore, proteins were transferred toeither polyvinylidene difluoride membranes (Millipore) ornitrocellulose membranes (GE Healthcare). Membraneswere blocked with Tris buffered saline containing 3–5%non-fat dry milk and incubated with indicated primary anti-bodies overnight at 4C. Bound antibodies were detectedwith horseradish peroxidase-conjugated anti-rat IgG (H+L), anti-mouse IgG (H+L), anti-mouse IgG (light chain spe-cific) or anti-rabbit IgG (H+ L) and signals were visualizedwith the chemiluminescence detection reagents LuminataForte (Millipore) or Amersham ECL Prime (GE Healthcare).Precision Plus Protein Dual Color Standards (Biorad) orPageRule Plus Prestained Protein Ladder, 10 to 250 kDa(ThermoFisher) were used as molecular weight size marker.Semi-quantitative immunoblot analysis of human tis-sues was performed using fluorescence detection andthe Odyssey® CLx Imaging System (LI-COR Biosciences).Proteins were transferred to nitrocellulose membranesand blocked with Odyssey blocking buffer (LI-COR Bio-sciences). Antibodies were detected with IRDye800CWor 680RD conjugated anti-rat or anti-mouse IgG(LI-COR). The linear range for C9orf72 detection wasdetermined by serial dilutions of RIPA lysate and conse-quently 50 μg of RIPA lysates per lane were loaded foranalysis. Total protein stains of membranes (BLOT-Fast-Stain, G-Biosciences) were used as loading controls fornormalization. This has been shown to be more suitableover measuring housekeeping genes such as GAPDHand is the recommended approach for normalization,particularly when loading of high amount of proteinsper lane is necessary as is the case for the detection ofC9orf72 [14]. Signal intensities for C9orf72 and totalprotein stains were analysed using the Image Studio™software (LI-COR).Statistical analysisStatistical analysis was performed with the GraphPadPr-ism software (version 7.01 for Windows). Student’s t test(two-tailed) was used for comparison of two groups andone-way ANOVA was used for comparison of multiplegroups followed by Tukey honestly significant difference(HSD) post hoc test. Associations between age at death,disease duration, postmortem delay and neurodegenera-tion with C9orf72 levels were analyzed by Spearman’srank correlation coefficient. Significance level was set atp < 0.05.Frick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 6 of 17ResultsCharacterization of highly specific novel monoclonalC9orf72 antibodiesNovel rat and mouse monoclonal antibodies (mAbs)against C9orf72 were generated and characterized recog-nizing 4 different epitopes of the human C9orf72 proteinsequence (Fig. 1a, Table 1). The specificity of identifiedmAbs was first demonstrated by immunoblot analysis ofprotein lysates from HEK293 cells transiently expressingeither untagged or myc-DDK-tagged human C9orf72short (C9-S) and long (C9-L) isoforms as well as murineC9orf72 isoform 1 (mC9–1) and isoform 2 (mC9–1)(Fig. 1b; Additional file 1: Figure S1a). In line with therespective epitopes recognized by the different mAbs,rat clone 12E7 detected C9-S, C9-L and mC9–1; mouseclone 1C1 detected C9-S, C9-L, mC9–1 and mC9–2; ratclones 5F6 and 12G10 specifically labeled human C9-Sand C9-L but not murine C9orf72; and rat clones 2H7and 15C5 labeled C9-L, mC9–1 and mC9–2 but notC9-S. However, clones 2H7 and 15C5 also revealed astrong unspecific band below 50 kDa of the size of un-tagged C9-L, limiting their usefulness for further studies.Immunoblot analyses were confirmed by double-labelimmunofluorescence of HEK293 cells transiently express-ing myc-DDK-tagged C9-S, C9-L or mC9–1. Completeco-localization of the diffuse cytoplasmic C9orf72 stainingwas seen between anti-myc and anti-C9orf72 mAbs 12E7and 1C1 for C9-S, C9-L, while mAbs 15C5 and 2H7 onlyrecognized C9-L but not C9-S and mAbs 5F6 and 12G10recognized specifically human but not murine C9orf72(Fig. 1c, data not shown). Further validation of the specifi-city of our antibodies to detect C9orf72 was performed byimmunoblot analysis of whole brain protein lysates ofwild-type mice and C9orf72 knock-out mice. A singleband around 50 kDa was obtained for rat mAb 12E7 andmouse mAb 1C1 in wild-type mice corresponding in sizeto the expected molecular weight of the murine C9orf72isoform 1 (Fig. 1d; Additional file 1: Figure S1b). Notably,no additional bands were observed at the expected mo-lecular weight size for the postulated murine isoforms 2and 3, although based on the recognized epitope of 1C1all murine isoforms should be recognized and absence ofisoform 2 was further demonstrated with the C-terminalmAb 15C5 (Additional file 1: Figure S1c). These results in-dicate that the 481 amino acid long murine isoform 1,which is the equivalent isoform to C9-L in humans,is the predominantly expressed C9orf72 isoform inthe mouse CNS. Importantly, this 50 kDa band wasabsent in lysates of C9orf72 knock-out mice, confirm-ing the specificity of our mAbs 1C1 and 12E7 to de-tect C9orf72.Of technical interest, none of the tested commerciallyavailable C9orf72 antibodies used in previous studies re-vealed a comparably high specificity for detecting C9orf72in our knock-out validation experiments using C9orf72−/− mouse brain (Additional file 1: Figure S2).Thus, we generated novel C9orf72 mAbs with knock-outvalidated specificity as valuable and powerful tools for fur-ther analysis of C9orf72 protein expression and localization.C9orf72 protein is enriched at the presynapse and co-localizes with a subset of synaptic vesicles in human iPSC-derived neurons in addition to its localization to lysosomesBy comparing different tissues of wild-type mice, C9orf72protein was found to be expressed as long isoform at high-est levels in the CNS (brain and spinal cord), at mediumlevels in tissue of the immune system (spleen) and at lowlevels in lung, heart, liver, kidney and skeletal muscle(Fig. 2a). These data are in agreement with transcriptomeprofiles reported in databases [30] and results in trans-genic mice with targeted LacZ insertion into the C9orf72locus [49]. Expression levels of different mouse brain re-gions did not reveal obvious regional differences (Fig. 2b).In nuclear-cytoplasmic fractionation experiments of mousebrain tissue, C9orf72 was exclusively found in the cytosolicprotein fraction (Fig. 2c). Interestingly, when we evaluatedthe expression levels of C9orf72 in the CNS over a timecourse from postnatal day 1 to 300 (n = 3), we noticed anincrease of C9orf72 levels within the first 2 postnatal weekswhile after that period no significant changes were observedin C9orf72 expression levels (Fig. 2d). Since this time periodwith increase of C9orf72 expression coincides with the on-set of synaptogenesis and synapse maturation, we specu-lated that C9orf72 might be localized at the synapse.To test this hypothesis, subcellular fractionations ac-cording to well-established protocols for the enrichmentof synaptosomal compartments [7, 16, 20] were per-formed with adult mouse brain as illustrated schematic-ally in Fig. 2e. Interestingly, a significant fraction ofC9orf72 was found to be present in the crude synapto-somal fraction P2 (Fig. 2f and g) and in the pure synap-tosomal fraction after sucrose gradient centrifugation(Fig. 2g). After hypotonic lysis of synaptosomal fractionP2 and centrifugation steps, C9orf72 was found to be re-leased into LS2, a fraction enriched for soluble cytoplas-mic contents of synaptosomes, while C9orf72 was notpresent in fraction LP2 enriched for synaptic vesicles(SVs) (Fig. 2f ) or in the Triton-extracted PSD fraction P4(Fig. 2g). Analysis of control proteins for specific frac-tions revealed expected distribution patterns, with en-richment of synaptophysin in the crude SV fraction LP2as expected for an integral membrane part of synapticvesicles, and enrichment of PSD-95 in pure PDS frac-tions P4 and in LP1 enriched for synaptosomal heavymembranes (Fig. 2f and g), thereby validating the qualityof our extractions.In line with our biochemical data, immunohistochem-istry for C9orf72 performed with mAb 1C1 in mouseFrick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 7 of 17brain sections revealed a fine punctate immunoreactivityin the neuropil consistent with a synaptic staining pat-tern throughout the CNS, although with variable signalintensities (Fig. 3). Specifically, C9orf72 expression wasmost pronounced in the entire hippocampal mossy fibersystem with strong labeling in the hilus, stratum lucidumand the infrapyramidal fiber bundles (Fig. 3a and b). Thisstrikingly resembles the staining pattern seen for otherpresynaptic marker proteins such as synaptophysin andsynaptoporin, a synaptic vesicle protein with an expressionFig. 1 Basic characterization of novel monoclonal antibodies against C9orf72. a Schematic representation of postulated human and murine C9orf72protein isoforms with epitopes recognized by novel monoclonal antibodies (mAbs) against C9orf72. In humans, two C9orf72 protein isoforms arepostulated with isoform 1 representing a 481 amino acid protein, also known as long isoform or C9-L (transcribed by transcript variant 2 with the GGGGCC repeat located in the promoter region and transcript variant 3 with the GGGGCC repeat located in the first intron); and isoform 2 representing a 222amino acid protein, also known as short isoform or C9-S (transcribed by transcript variant 1 with the GGGGCC repeat located in the first intron). In mice,three protein isoforms are postulated, with isoform 1 corresponding in size to human C9-L with 98% similarity on amino acid sequence. The red lines inthe murine isoforms illustrate two amino acid changes between the human and mouse C9orf72 sequence in the epitope recognized by mAbs 5F6 and12G10. b Immunoblot analysis of protein lysates of HEK293 cells expressing untagged or myc-DDK-tagged human C9-L and C9-S or myc-DDK-taggedmurine C9orf72 isoform 1 (mC9–1) with novel C9orf72 mAbs. Clones 12E7 and 1C1 recognize hC9-S and hC9-L as well as mC9–1. Clones 5F6 and12G10 specifically recognize human but not mouse C9orf72. Clones 2H7 and 15C5 specifically recognize an epitope in the C-terminus only present inhC9-L and mC9–1 but not hC9-S, however, both mAbs also recognize an unspecific band (asterisk). c Double label immunofluorescence for anti-myc(green) and anti-C9orf72 (red) of HEK293 cells transiently expressing myc-DDK-tagged hC9-L, hC9-S or mC9–1 confirms the specificity of the indicatedmAbs for specific C9orf72 isoforms or species. Hoechst 33342 staining of nuclei (blue) in the merged images. Scale bar: 20 μm. d Immunoblot analysisof total protein lysates from brains of wild-type (C9+/+) and C9orf72 knock-out (C9−/−) mice. Only a single band around 50 kDa corresponding in size tothe murine isoform 1 is detected with mAbs 12E7 and 1C1 in wild-type mice (arrowhead). Note, that this band is completely absent in C9−/− mice,validating the high specificity for C9orf72 of the mAbs 12E7 and 1C1. The weak band labeled with an asterisk seen in C9−/− with the mouse mAb 1C1represents mouse IgG heavy chain recognized by the anti-mouse IgG (H + L) detection antibody (see Additional file 1: Figure S1b for secondaryantibody control). GAPDH is shown as loading control. MW size marker: Precision Plus Protein Dual Color Standards (b and d)Frick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 8 of 17pattern mainly restricted to the mossy fiber system (Fig. 3k),a finding further confirmed by double-label immunofluores-cence for C9orf72 and synaptoporin (Additional file 1: FigureS3a). A strong neuropil staining was also seen in the globuspallidus (Fig. 3c and d), while weaker immunoreactivity waspresent in the caudate-putamen (Fig. 3c), throughout thecortex (Fig. 3e and f) and in the molecular layer of thecerebellum (Fig. 3g). In addition to the punctate neuropilstaining, particularly large motor neurons in the brain stemand spinal cord (Fig. 3h) demonstrated staining of smalldots in the cytoplasm consistent with staining of small ves-icles. These vesicles did not stain with antibodies againstsynaptophysin and LAMP1 (data not shown) and thustheir nature remains to be identified. No labeling of nucleiTable 1 Summary of basic characterization of novel monoclonal C9orf72 antibodiesClone (species)1C1 (mouse) 12E7 (rat) 5F6 (rat) 12G10 (rat) 2H7 (rat) 15C5 (rat)detection of recombinant proteinsC9-L yes yes yes yes yes yesC9-S yes yes yes yes no nomC9–1 yes yes no no yes yesmC9–2 yes no no no yes yesknock-out validated yes yes na na no# no#C9-L, long human C9orf72 isoform; C9-S, short human C9orf72 isoform; mC9–1, long murine C9orf72 isoform 1; mC9–2, murine C9orf72 isoform 2; na, notapplicable; #unspecific band at similar molecular size of endogenous long C9orf72 isoformFig. 2 C9orf72 is enriched in synaptosomes. a Immunoblot analysis of total protein lysates of different mouse tissues reveals widespread C9orf72 proteinexpression detected as single band around 50 kDa with highest expression levels in brain followed by spinal cord. GAPDH is shown as loading control. bNo obvious changes are observed between different brain regions by immunoblot analysis of protein lysates from cortex, hippocampus, striatum,cerebellum and spinal cord. GAPDH is shown as loading control. c Immunoblot analysis of nuclear and cytoplasmic protein fractionations extracted fromadult mouse brain reveals localization of C9orf72 to the cytoplasm. α-tubulin and Histone H3 are shown to demonstrate purity of the cytoplasmic andnuclear fractions, respectively. d Immunoblot analysis and quantification of C9orf72 expression levels over mouse brain development from P1 to P300showing increase of C9orf72 between P1 and P16. e Schematic of the purification protocols for synaptic vesicles and postsynaptic densities (PSDs).Mouse forebrains were homogenized and centrifuged to generate a crude synaptosomal fraction (P2). P2 fractions were fractionated into synaptosomalheavy membranes (LP1), synaptic vesicles (LP2), and synaptic cytosol (LS2) by hypotonic lysis and differential centrifugation. Alternatively, P2 fractionswere centrifuged through sucrose gradient to reveal a pure synaptosomal fraction which was further processed to isolate pure PSDs. f C9orf72 isdetectable in the synaptosomal fraction P2, and is released into the LS2 fraction containing the soluble cytoplasmic content of synaptosomes. g C9orf72is enriched in the pure synaptosomal fraction, but absent from PSD fractions. f and g The purity of fractions was confirmed with specific marker proteinsfor synaptic vesicles (synaptophysin), postsynaptic densities (PSD-95), mitochondria (Cox-IV), and lysosomes (Lamp1). MW size marker: PageRule PlusPrestained Protein Ladder (a); Precision Plus Protein Dual Color Standards (b-d, f, g)Frick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 9 of 17and, no immunoreactivity in the white matter and glial cellswas detectable. The expression pattern is in good agree-ment with our in situ hybridization experiments showingwidespread and predominant C9orf72 mRNA expression inneurons with strongest signals in the dentate granule cellsbut not in glial cells (Additional file 1: Figure S3b). Import-antly, the specificity of the observed immunoreactivity forC9orf72 with mAb 1C1 was validated by immunohisto-chemical analysis of C9orf72 knock-out mouse brain sec-tions showing absence of immunoreactivity (Fig. 3i, j;Additional file 1: Figure S2f). In contrast, all tested com-mercially available C9orf72 antibodies revealed similarstaining intensities and patterns in wild-type and C9orf72knock-out mice (Additional file 1: Figure S2f).Unfortunately, we failed to detect reliable immunore-activity in routinely sampled FFPE human postmortemCNS tissue using the established knock-out validatedprotocol for mAb 1C1 on mouse tissue. Since we ob-served that formalin fixation times > 24 h dramaticallydiminished C9orf72 immunoreactivity signals with 1C1also in mouse tissue, this is most likely due to the longformalin fixation times (weeks to months) of availablehuman postmortem tissue (for details see Material andMethods). Furthermore, a cross-reactivity of the humanspecific C9orf72 antibodies 5F6 and 12G10 with add-itional proteins as shown in brain lysates and tissue sec-tions (Additional file 1: Figure S1d and e) preventedtheir suitability for immunohistochemical analyses.Thus, to address the localization of C9orf72 in humanneurons, we analyzed motor neurons differentiated from hu-man iPSCs. Both knock-out validated C9orf72 antibodies(1C1 and 12E7) revealed presence of C9orf72 in cytoplasmicpuncta with complete overlap of signals for both antibodies(Additional file 1: Figure S4a). Double-label immunofluores-cence revealed co-localization of C9orf72 with SMCR8 in al-most all C9orf72 positive puncta (90 ± 9%) (Fig. 4),consistent with the tight association reported betweenthese two proteins in co-immunoprecipitation experimentsFig. 3 C9orf72 immunohistochemistry reveals synaptic staining pattern with enrichment in hippocampal mossy fiber terminals. Immunohistochemistrywith anti-C9orf72 mAb 1C1 (a-j); immunohistochemistry with anti-synaptoporin antibody (k). In the adult mouse brain (a-h), strong immunoreactivity forC9orf72 is seen in the hippocampal mossy fiber system (a) with labeling in the hilus (asterisk), stratum lucidum (arrow) and infrapyramidal mossy fiberbundles (arrowhead). (b) Higher magnification of punctate staining pattern of mossy fiber terminals in suprapyramidal (SPB) and infrapyramidal (IPB)mossy fiber bundles. Robust staining was also observed in the globus pallidus (GP) (c and d) while the caudate putamen (CPu) (c) and other gray matterregions showed weaker immunoreactivity of the neuropil as shown for frontal cortex (e and f) and cerebellum with predominant staining in themolecular and granular layer (g). No immunoreactivity is seen in the white matter and internal capsule (ic). In addition to punctate neuropil staining,neurons with large cytoplasm such as motor neurons in the spinal cord showed several cytoplasmic puncta (h). (i and j): Specificity of anti-C9orf72immunohistochemistry was validated by the complete absence of immunoreactivity in brain sections from C9orf72 knock-out mice as shown forhippocampus (i) and cerebellum (j). Note the strikingly similar staining pattern of the mossy fiber terminals in the hippocampus for C9orf72 (a) and forthe presynaptic marker protein synaptoporin (k). Scale bar: 533 μm (c); 400 μm (a, i, k); 267 μm (e); 80 μm (d, j, insert k); 40 μm (b, f, g); 20 μm (h);6,5 μm (insert h)Frick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 10 of 17[1, 21, 45, 48, 51, 54, 57]. Further analysis showedco-localization of C9orf72 with LAMP1 and LAMP2(Fig. 4) in a subset of C9orf72 positive puncta (41 ± 8%), in-dicating that a fraction of C9orf72 is present at lysosomes,consistent with previous observations [1, 46]. Finally, inagreement with our above described findings of a pre-synaptic localization of C9orf72 in mouse CNS, a fraction(11 ± 4%) of C9orf72 positive puncta co-labelled withsynaptophysin used as marker for SVs (Fig. 4) and a com-parable co-localization was seen between SMCR8 andsynaptophysin (Additional file 1: Figure S4b). However,only a subset (7 ± 3%) of synaptophysin positive SVs wasfound to co-label with C9orf72, consistent with a transientinteraction of C9orf72 with SVs as indicated by our bio-chemical fractionation experiments. Overall, our data im-plicate an association of C9orf72 with SVs in mouse andhuman neurons in addition to its association with lyso-somes and vesicles of yet unknown identity.C9orf72 interacts and co-localizes with all members of theRAB3 protein familyBased on our biochemical and immunohistochemicalanalyses, the subcellular distribution of C9orf72 is con-sistent with it being a presynaptic terminal associatedprotein with a transient and reversible interaction withSVs. Given the described function of C9orf72 as GEF forRAB8A and RAB39B, we speculated that C9orf72 mightbe able to interact with specific Rabs present at SVs.Thus, we performed immunoprecipitation experimentsin HEK293 cells co-expressing HA-tagged C9orf72 andSMCR8 and various FLAG-tagged Rabs (Fig. 5a). Excit-ingly, an interaction of the C9orf72/SMCR8 complexwas seen with all members of the RAB3 protein family(RAB3A, RAB3B, RAB3C, RAB3D), which are Rabsabundantly present at SVs (Fig. 2f ) and known to playkey roles in neurotransmitter release [6]. As controls, inter-actions of the C9orf72/SMCR8 complex with RAB39B andall members of the RAB8 subfamily (RAB8A, RAB10,RAB13 and RAB15) were identified, comparable to interac-tions previously described [45]. Importantly, immunopre-cipitation of C9orf72 in wild-type mouse brain lysateconfirmed the interaction between endogenous C9orf72and endogenous RAB3 (Fig. 5b). Consistently, double-labelimmunofluorescence of human iPSC derived motor neu-rons revealed a partial co-localization of C9orf72 withRAB3 and RAB39B, which was used as positive control(Fig. 5c and d). The interaction of C9orf72 with RAB3 pro-teins is likely to be transient as only a subset (8 ± 3%) ofC9orf72 positive puncta co-localized with RAB3 and con-versely, only a subset (5 ± 3%) of RAB3-positive vesiclesco-labeled with C9orf72 (Fig. 5d). Thus, our data suggest apotential novel function of C9orf72 by acting as GEF forRAB3 proteins.C9orf72 protein expression levels are reduced in thecerebellum as consequence of C9orf72 repeat expansionsTo investigate the effect of C9orf72 repeat expansionson protein expression, we decided to focus for ourFig. 4 C9orf72 co-localizes with synaptic vesicles in human iPSC derived motor neurons. a C9orf72 positive puncta (green) are seen in the axonsof 30 day old human iPSC derived motor neurons which consistently co-localize with SMCR8 (red, upper panel) and partially co-localize withLAMP2 as lysosomal marker (red, middle panel) or with the synaptic vesicle marker synaptophysin (red, lower panel). Nuclei are stained with DAPI(blue) in the merged images. C9orf72 labeled with 12E7 antibody in the upper panel and with 1C1 in the middle and lower panel. b Graphshowing the percentage of C9orf72 positive puncta co-localizing with SMCR8, LAMP2 or synaptophysin. Values are shown as mean ± SDFrick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 11 of 17quantitative immunoblot analysis on cerebellum as brainregion, since it is a region known to express high levelsof C9orf72 mRNA [40] and shows consistent and robustchanges in transcript levels between C9orf72 mutationcarriers and controls [52, 53]. Moreover and in line withour results showing a predominant neuronal C9orf72expression, we observed a strong negative correlation(rho = − 0.834, p = 0.004, Spearman rank correlation)between C9orf72 protein expression and the level ofneurodegeneration/cell death in a pilot experiment onfrontal cortex samples from selected cases with noC9orf72 mutation (Additional file 1: Figure S5), Theseresults highlight the potential bias that neuronal cell lossmay blur the interpretation of changes in C9orf72 proteinlevels. Therefore, we considered that the cerebellum, a re-gion not affected by overt neurodegeneration in ALS andFTD, is best suited for the analysis of C9orf72 mutationspecific consequences on its own protein levels by avoidingmisinterpretation of changes related to neuronal cell loss.The analyzed cohort consisted of n = 17 C9orf72 muta-tion carriers covering the complete clinical spectrumfrom pure ALS, mixed ALS/FTD and pure FTD and n =26 neurologic disease controls (ALS, ALS/FTD and FTDcases without C9orf72 mutation) with detailed informa-tion on each case given in Additional file 1: Table S1.There were no significant differences in the demograph-ics between both cohorts. Immunoblot analysis of totalRIPA protein lysates extracted from cerebellum revealedthat C9orf72 is a low abundant protein detectable as sin-gle band of ~ 50 kDa corresponding in size to C9-L inall samples for mAb 1C1 (Fig. 6a) and 12E7. No bandwas detectable corresponding to the molecular size ofthe predicted human C9-S isoform, although both anti-bodies are able to detect C9-L and C9-S isoformsexpressed in HEK293 cells (Fig. 1) with comparable sen-sitivity, implying that the 481 amino acid isoform (C9-L)is the main and predominant protein isoform expressedin the human CNS as in the mouse CNS. Importantly,subsequent quantitative analysis of C9-L levels normal-ized to total protein stains revealed a ~ 20% reduction ofC9-L levels in cases with C9orf72 repeat expansionscompared to controls (p = 0.001) (Fig. 6b). There wereno significant differences in C9orf72 protein levelswithin each cohort between cases presenting clinicallyFig. 5 C9orf72 complex interacts with members of the RAB3 protein family. a Immunoblot analysis of HA-immunoprecipitated proteins from lysates ofHEK293 cells co-expressing HA-tagged human C9orf72 and HA-tagged SMCR8 with various FLAG-tagged Rabs showing co-immunoprecipitation of allmembers of the RAB3 family with the C9orf72 complex. Other Rabs were used as positive (RAB8 subfamily, RAB39B) or negative (RAB1A, RAB7A,RAB5A) controls based on published reports. b Immunoblot against RAB3 of control (IgG alone) or endogenous C9orf72 immunoprecipitated proteinsfrom lysates of adult mouse brain. MW size marker: PageRule Plus Prestained Protein Ladder (a and b). c Double-label immunofluorescence of 30 dayold human iPSC-derived motor neurons showing co-localization of C9orf72 (green) with RAB39B (red, upper panel) or RAB3 (red, lower panel) in asubset of C9orf72-positive puncta. C9orf72 labeled with 12E7 in the upper panel and 1C1 in the lower panel. d Graph showing the percentage ofC9orf72 positive puncta co-localizing with RAB3 or RAB39B. Values are shown as mean ± SDFrick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 12 of 17with either ALS, FTD/ALS or FTD and there were noassociations between cerebellar C9orf72 protein levelsand disease duration, age at death or post-mortem delay.Finally, to further analyze for potential biochemical al-terations of C9orf72 in mutation carriers, proteins weresequentially extracted from frozen frontal cortex fromcases with or without a C9orf72 repeat expansion, usinga series of buffers with an increasing ability to solubilizeproteins. C9-L was found to be present in the proteinfractions enriched for soluble proteins (low salt and Tri-ton X-soluble fractions) but not in the fractions enrichedfor insoluble proteins (sarkosyl and urea soluble frac-tions). There were no changes in solubility betweenC9orf72 mutation carriers and controls (Fig. 6c). Overall,these results demonstrate that C9orf72 is a low abun-dant, cytoplasmic soluble protein in the human CNSwith C9-L being expressed as the predominant proteinisoform and with reduced protein levels as consequenceof C9orf72 repeats expansions.DiscussionAn abnormal hexanucleotide expansion in a non-codingregion of C9orf72 is the most common genetic cause ofALS and FTD [13, 40]. The mechanisms of how thismutation contributes to neurodegeneration are unclearwith haploinsufficiency of C9orf72 functions suggestedas one potential disease mechanism based on consist-ently observed reduced RNA transcript levels in tissuesof C9orf72 mutation carriers [13, 17, 52]. However, dueto little knowledge on C9orf72 expression, localizationand functions particularly in the CNS, further insights ifand how reduced C9orf72 proteins levels might contrib-ute to disease pathogenesis are currently limited.Here, we generated and characterized novel knock-outvalidated C9orf72 monoclonal rat and mouse antibodiesthat indicate that C9orf72 is predominantly, if not exclu-sively, expressed as the long 481 amino acid isoform inhuman and mouse tissues. Utilizing these antibodies wedetected reduced C9orf72 protein levels in the cerebel-lum as consequence of C9orf72 repeat expansions in ourstudied cohort of C9orf72 mutation carriers. Finally, weidentified C9orf72 to be localized to the presynapses andable to interact with members of the RAB3 protein fam-ily, suggestive of a role for C9orf72 in regulating SVfunctions by potentially acting as GEF for RAB3. Thesefindings provide important novel insights into theFig. 6 Reduced C9orf72 expression levels in the cerebellum of C9orf72 mutation carriers. a Immunoblot analysis of C9orf72 protein levels in RIPAlysates extracted from frozen cerebellar gray matter of C9orf72 mutation cases and neurologic controls reveals a single band ~ 50 kDa correspondingin size to the long 481 amino acid isoform of C9orf72 (C9-L). Total protein stains are shown as loading controls. The blot shown is representative ofthree independent experiments. b Quantification of C9orf72 protein levels in the cerebellum of n = 17 cases with C9orf72 repeat expansions (C9+) andn = 26 controls (C9-). Dot blot of normalized C9orf72 values with mean and standard deviation shown as line and error bars. Different colors representclinical phenotypes (green = FTD; red = ALS/FTD; blue = ALS). p = 0.001 by Student’s two-tailed, unpaired t test. c Proteins were sequentially extractedfrom frozen frontal cortex of C9orf72 mutation carriers (C9+) and controls (C9-) with a series of buffers of increasing stringency to receive low salt (LS),high-salt Triton-X-100 (TX), sarkosyl (SARK), and urea protein fractions for immunoblot analysis. Human C9orf72 (C9-L) is present in all cases in thefractions enriched for highly soluble proteins (LS and to lesser extent TX) with no changes observed in solubility between C9+ and C9- cases. MW sizemarker: PageRule Plus Prestained Protein Ladder (a and c)Frick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 13 of 17physiological role of C9orf72 in the CNS, with signifi-cant implications for future studies addressing the po-tential contribution of haploinsufficiency in C9orf72disease pathogenesis as well as therapeutic strategies.Based on C9orf72 RNA transcripts two human andthree murine C9orf72 protein isoforms have been pre-dicted. Previous biochemical studies of endogenousC9orf72 protein expression have provided inconsistent re-sults with reported presence of only C9-L [41, 53] or C9-Land C9-S [56] as well as presence of only murine variant 1[23, 38] or all three mouse variants [4]. This observedvariability is likely due to use of antibodies that lack suffi-cient specificity as illustrated by our results of commer-cially available C9orf72 antibodies on C9orf72 knock-outmouse brain tissue (Additional file 1: Figure S2), a tech-nical limitation also reported by others [42, 53]. Here, util-izing knock-out validated C9orf72 mAbs generatedagainst epitopes that allow detection of all predicted iso-forms in vitro, one band was observed in examined mouseand human tissues corresponding in size to human C9-Land the corresponding variant 1 of murine C9orf72. Whilethis does not exclude that additional C9orf72 protein iso-forms might be present at amounts below detection limitof our immunoblot assay, our data indicate that the long481 amino acid isoform is by far the most predominantlyexpressed C9orf72 protein isoform in mouse and human.Biochemical analyses of C9-L extracted from human post-mortem tissue revealed no changes in solubility of C9-Lbetween C9orf72 mutation carriers and controls, which isconsistent with a previous report [56]. However, we foundthat C9-L is mostly a soluble protein with presence in LSand TX fractions, in contrast to a previous study reportingsignificant amount of C9orf72 in TX-insoluble, urea-solubleprotein fractions [56]. Discrepancies between both studiesmight be explained by different antibody specificities anddifferences in extraction methods.Reduced C9-L levels in postmortem brain tissue ofC9orf72 mutation carriers have been reported for some cor-tical regions but surprisingly not for cerebellum [41, 53, 56],the region with highest C9orf72 RNA expression [40] andconsistently reported reduced transcript levels in C9orf72mutation carriers [17, 52, 53]. However, the overall inter-pretation of these data in providing evidence for haploinsuf-ficiency on the protein level has been complicated bynon-specific binding of antibodies, insufficient statisticalpower due to small sample size and the risk that reducedprotein levels observed in cortical regions might be influ-enced by neurodegenerative changes instead of C9orf72 mu-tation specific consequences. This latter limitation isunderpinned by our finding of a strong negative correlationbetween presence of neurodegeneration/cell death andC9orf72 levels in cortical regions. Therefore, we focused forour quantitative immunoblot analysis on the cerebellum, aregion without overt neurodegeneration in ALS and FTD.Utilizing our novel knock-out validated mAbs we foundcerebellar C9-L protein levels reduced to ~ 80% in our co-hort of C9orf72 mutation carriers (n= 17) compared to con-trols (n = 26). Notably, the observed degree of proteinreduction is in good agreement with the reported decreaseto 70% for RNA transcripts encoding for the long isoform inC9orf72 mutation carriers [13, 52]. No associations betweenclinical phenotypes (ALS, ALS/FTD or FTD), age at onsetand disease duration with cerebellar C9orf72 protein levelswere seen. This is consistent with the reported lack of asso-ciations between cerebellar transcript levels and clinical fea-tures [52]; however, the number of cases per clinicalsubgroup in our cohort might have been too small to detectsubtle associations with protein levels and this should befurther addressed in larger cohorts.While our data expand the evidence for reduced proteinexpression as a consequence of C9orf72 hexanucleotideexpansions, it remains to be established if and how re-duced proteins levels might contribute to disease patho-genesis. The fact that reduced protein levels are describedin unaffected and affected brain regions in C9orf72 muta-tion carriers, implies that reduced C9orf72 levels are notcausing neurodegeneration per se, an interpretation fur-ther supported by the absence of obvious neurologicalphenotypes in C9orf72 knock-out mice [3, 23, 38, 47].However, these data would be in line with a scenario thatdifferent cells and/or neuronal subpopulations might havedistinct vulnerabilities in tolerating reduced C9orf72 levelswhich might trigger neurodegeneration in combinationwith additional stressors [45], and/or as consequence of acooperative gain- and loss-of-function mechanism of re-peat expansions as recently proposed [46].One crucial prerequisite to further address the poten-tial role of reduced C9orf72 levels in disease pathogen-esis is to gain further insights on its physiologicalfunction in the CNS. Several recent studies have demon-strated that C9orf72 interacts with the SMCR8 proteinand regulates the endo-lysosomal and autophagy path-ways [1, 21, 45, 48, 51, 54, 57]. Our immunofluorescencedata of human iPSC derived human motor neurons areconsistent with these studies by demonstrating endogen-ous co-localization of almost all C9orf72-immunoreactivevesicles with SMCR8 and co-localization of ~ 40% ofC9orf72-positive vesicles with the lysosomal markerLAMP2. Most excitingly, our complementary histologicaland biochemical approaches provided strong evidence foran additional role of C9orf72 at presynaptic terminals byacting as putative GEF for members of the RAB3 proteinfamily. Immunohistochemically, we found a predomin-antly presynaptic staining pattern of C9orf72 in mousebrains most prominently in the synapse-rich hippocampalmossy fiber system, where it co-localized with SV markerproteins. No immunoreactivity was observed in glial cellsin mouse tissue in agreement with our in situ hybridizationFrick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 14 of 17data and with the predominant neuronal expression ofC9orf72 reported in transgenic mice with targeted LacZ in-sertion into the C9orf72 locus [49] under physiologicalconditions. However, in future studies it will be interestingto investigate whether cellular expression patterns mightchange under neuroinflammatory and neurodegenerativeconditions. In accordance with our immunohistochemicalfindings, C9orf72 was present in biochemical preparationsof synaptosomes and particularly in fractions enriched forsoluble cytosolic contents of synaptic vesicles, which issuggestive for a protein able to reversibly and transientlyinteract with SVs [22]. This interpretation is furthersupported by the co-localization of C9orf72 in a sub-set of synaptophysin-positive vesicles in human iPSCderived motor neurons. Like all other trafficking stepsin eukaryotes, SV cycle and presynaptic neurotrans-mitter release is governed by specific Rabs [6] withthe most abundant Rabs in neurons represented byhomologues of the RAB3 protein family specificallylocalizing to SVs and with well studied roles in regu-lating/modulating neurotransmitter release [43, 44].Our findings of an interaction of C9orf72 with RAB3family members by co-immunoprecipitation experi-ments and double-label immunofluorescence indicatethat C9orf72 might act as GEF for RAB3 therebymodulating the SV cycle. In support of this interpret-ation, it is noteworthy that subtle cognitive and im-aging alterations observed in a recent study ofpresymptomatic C9orf72 mutation carriers were pro-posed to represent an early and non-evolving pheno-type related to neurodevelopmental effects of C9orf72mutation [5]. However, the exact role of C9orf72 aspotential GEF in modulating neurotransmission andother steps of the SV cycle which also includes Rabsinvolved in endosomal and autophagosomal functions[6] will require further functional investigation.One limitation of our study is that the subcellular distri-bution of C9orf72 in postmortem human brain tissuescould not be investigated immunohistochemically due tothe lack of immunoreactivity in human FFPE tissue usingthe knock-out validated protocol successfully establishedin mouse FFPE tissue. This might be potentially explainedby protein degradation due to postmortem delay. How-ever, we observed no association between C9orf72 levelsand postmortem delay and mouse tissue with differentPM delay mimics in our biochemical analysis. An add-itional and perhaps more likely explanation seems to berelated to formalin fixation times. For mouse tissue we ob-served decreasing immunoreactivity signals for formalinfixation times > 24 h, while the available human postmor-tem tissue was routinely fixed for several weeks up tomonth. This issue needs to be addressed in future studiesusing differently processed autopsy and perhaps biopsytissues if available.ConclusionsIn summary, our data provide evidence for haploinsuffi-ciency at the protein level in C9orf72 mutation carriersand novel insights into the physiological role of C9orf72at the presynapse with a potential role as GEF for RAB3involved in SV exocytosis. These findings have signifi-cant implications for future studies aimed at addressingC9orf72 pathogenesis as well as therapeutic strategies.Furthermore, these novel mAbs against C9orf72 will beuseful tools to further dissect the cellular and molecularfunctions of C9orf72.Additional fileAdditional file 1: Table S1. Demographic, clinical and pathologicaldiagnosis of cases used in this study; Figure S1. Further characterizationof novel monoclonal C9orf72 antibodies; Figure S2. Commerciallyavailable C9orf72 antibodies tested on C9orf72 knock-out brain tissue;Figure S3. C9orf72 double-label immunofluorescence and C9orf72 in situhybridization; Figure S4. Immunofluorescence of human iPSC derivedmotor neurons; Figure S5. Immunoblot analysis of C9orf72 expressionlevels in frontal cortex. (PDF 14467 kb)AcknowledgementsWe would like to thank Manuel Gödan for excellent technical assistance.FundingThe study was supported by grants from the German Helmholtz-Association(W2/W3–036 and VHVI-510; to MN), the NOMIS foundation (to MN and DE), theFondation Thierry Latran (#57486; to NCB), the French Muscular DystrophyAssociation (#18605; to NCB), the European Research Council (ERC-2012-StG#310659, to NCB; ERC-2013-CoG #617198 to DE), ANR-10-LABX-0030-INRT(IGBMC); ANR-10-IDEX-0002-02 (IGBMC), the Canadian Institutes of HealthResearch (#74580; to IRM), the Canadian Consortium on Neurodegeneration inAging (#137794; to IRM), the Dutch ALS Foundation (TOTALS; to RJP), and theMunich Cluster of Systems Neurology (to DE and RF).Authors’ contributionsMN and NCB conceived and supervised the study and analyzed the data. PFperformed and analyzed experiments for the primary screen of supernatants,biochemical experiments of HEK293 cells, mouse and human tissue andimmunofluorescence of HEK293 cells. CS performed and analyzedimmunofluorescence experiments of iPSC derived motor neurons and co-immunoprecipitation experiments. JTB and CM generated iPSC derivedmotor neurons. CYC performed and analysed in situ hybridization experi-ments. RF and MOA generated monoclonal antibodies. DE performed experi-ments for the primary screen of supernatants. IRAM provided human tissue,RJP provided mouse tissue from C9orf72 knock-out mice. JP provided clinicaldata. MN performed and analyzed immunohistochemistry. MN, PF and NCBdrafted the manuscript with input and final approval from all co-authors.Competing interestsThe authors declare that they have no competing interestsPublisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1German Center for Neurodegenerative Diseases (DZNE), Otfried-Müllerstr. 23,72076 Tübingen, Germany. 2Institut de Génétique et de Biologie Moléculaireet Cellulaire (IGBMC), INSERM U964, CNRS UMR7104, Strasbourg University,67400 Illkirch, France. 3Department of Pathology, University of BritishColumbia and Vancouver General Hospital, Vancouver, Canada. 4INSERM/UEVE UMR 861, I-STEM, AFM, 91100 Corbeil-Essonnes, France. 5Department ofFrick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 15 of 17Translational Neuroscience, Brain Center Rudolf Magnus, University MedicalCenter Utrecht, Utrech University, 3584 CG Utrecht, The Netherlands.6German Center for Neurodegenerative Diseases (DZNE), Rostock, Germany.7Department of Neurology, University of Rostock, Rostock, Germany.8German Center for Neurodegenerative Diseases (DZNE), Munich, Germany.9Cluster of Systems Neurology (SyNergy), Munich, Germany.10Ludwig-Maximilians University Munich, Munich, Germany. 11Institute forDiabetes and Obesity, Monoclonal Antibody Core Facility and ResearchGroup, Helmholtz Zentrum München, Neuherberg, Germany. 12Departmentof Neuropathology, University of Tübingen, Tübingen, Germany.Received: 27 July 2018 Accepted: 28 July 2018References1. Amick J, Roczniak-Ferguson A, Ferguson SM (2016) C9orf72 binds SMCR8,localizes to lysosomes, and regulates mTORC1 signaling. Mol Biol Cell 27:3040–3051. https://doi.org/10.1091/mbc.E16-01-00032. Ash PE, Bieniek KF, Gendron TF, Caulfield T, Lin WL, Dejesus-Hernandez M etal (2013) Unconventional translation of C9ORF72 GGGGCC expansiongenerates insoluble polypeptides specific to c9FTD/ALS. Neuron 77:639–646.https://doi.org/10.1016/j.neuron.2013.02.0043. Atanasio A, Decman V, White D, Ramos M, Ikiz B, Lee HC et al (2016)C9orf72 ablation causes immune dysregulation characterized by leukocyteexpansion, autoantibody production, and glomerulonephropathy in mice.Sci Rep 6:23204. https://doi.org/10.1038/srep232044. Atkinson RA, Fernandez-Martos CM, Atkin JD, Vickers JC, King AE (2015)C9ORF72 expression and cellular localization over mouse development.Acta Neuropathol Commun 3:59. https://doi.org/10.1186/s40478-015-0238-75. Bertrand A, Wen J, Rinaldi D, Houot M, Sayah S, Camuzat A et al (2017) Earlycognitive, structural, and microstructural changes in PresymptomaticC9orf72 carriers younger than 40 years. JAMA Neurol. https://doi.org/10.1001/jamaneurol.2017.42666. Binotti B, Jahn R, Chua JJ (2016) Functions of Rab proteins at presynapticsites. Cells 5. https://doi.org/10.3390/cells50100077. Carlin RK, Grab DJ, Cohen RS, Siekevitz P (1980) Isolation and characterizationof postsynaptic densities from various brain regions: enrichment of differenttypes of postsynaptic densities. J Cell Biol 86:831–8458. Chew J, Gendron TF, Prudencio M, Sasaguri H, Zhang YJ, Castanedes-CaseyM et al (2015) C9ORF72 repeat expansions in mice cause TDP-43 pathology,neuronal loss, and behavioral deficits. Science 348:1151–1154. https://doi.org/10.1126/science.aaa93449. Ciura S, Lattante S, Le Ber I, Latouche M, Tostivint H, Brice A et al (2013)Loss of function of C9orf72 causes motor deficits in a zebrafish model ofamyotrophic lateral sclerosis. Ann Neurol 74:180–187. https://doi.org/10.1002/ana.2394610. Cooper-Knock J, Walsh MJ, Higginbottom A, Robin Highley J, Dickman MJ,Edbauer D et al (2014) Sequestration of multiple RNA recognition motif-containing proteins by C9orf72 repeat expansions. Brain 137:2040–2051.https://doi.org/10.1093/brain/awu12011. Davidson Y, Robinson AC, Liu X, Wu D, Troakes C, Rollinson S et al (2015)Neurodegeneration in frontotemporal lobar degeneration and motorNeurone disease associated with expansions in C9orf72 is linked to TDP-43pathology and not associated with aggregated forms of dipeptide repeatproteins. Neuropathol Appl Neurobiol. https://doi.org/10.1111/nan.1229212. DeJesus-Hernandez M, Finch NA, Wang X, Gendron TF, Bieniek KF, HeckmanMG et al (2017) In-depth clinico-pathological examination of RNA foci in alarge cohort of C9ORF72 expansion carriers. Acta Neuropathol 134:255–269.https://doi.org/10.1007/s00401-017-1725-713. DeJesus-Hernandez M, Mackenzie IR, Boeve BF, Boxer AL, Baker M,Rutherford NJ et al (2011) Expanded GGGGCC hexanucleotide repeat innoncoding region of C9ORF72 causes chromosome 9p-linked FTD and ALS.Neuron 72:245–256. https://doi.org/10.1016/j.neuron.2011.09.01114. Eaton SL, Roche SL, Llavero Hurtado M, Oldknow KJ, Farquharson C,Gillingwater TH et al (2013) Total protein analysis as a reliable loadingcontrol for quantitative fluorescent western blotting. PLoS One 8:e72457.https://doi.org/10.1371/journal.pone.007245715. Gendron TF, Bieniek KF, Zhang YJ, Jansen-West K, Ash PE, Caulfield T et al(2013) Antisense transcripts of the expanded C9ORF72 hexanucleotiderepeat form nuclear RNA foci and undergo repeat-associated non-ATGtranslation in c9FTD/ALS. Acta Neuropathol 126:829–844. https://doi.org/10.1007/s00401-013-1192-816. George JM, Jin H, Woods WS, Clayton DF (1995) Characterization of a novelprotein regulated during the critical period for song learning in the zebrafinch. Neuron 15:361–37217. Gijselinck I, Van Langenhove T, van der Zee J, Sleegers K, Philtjens S,Kleinberger G et al (2012) A C9orf72 promoter repeat expansion in aFlanders-Belgian cohort with disorders of the frontotemporal lobardegeneration-amyotrophic lateral sclerosis spectrum: a geneidentification study. Lancet Neurol 11:54–65. https://doi.org/10.1016/S1474-4422(11)70261-718. Guillemin I, Becker M, Ociepka K, Friauf E, Nothwang HG (2005) A subcellularprefractionation protocol for minute amounts of mammalian cell culturesand tissue. Proteomics 5:35–45. https://doi.org/10.1002/pmic.20040089219. Haeusler AR, Donnelly CJ, Periz G, Simko EA, Shaw PG, Kim MS et al (2014)C9orf72 nucleotide repeat structures initiate molecular cascades of disease.Nature 507:195–200. https://doi.org/10.1038/nature1312420. Huttner WB, Schiebler W, Greengard P, De Camilli P (1983) Synapsin I(protein I), a nerve terminal-specific phosphoprotein. III Its association withsynaptic vesicles studied in a highly purified synaptic vesicle preparation. JCell Biol 96:1374–138821. Jung J, Nayak A, Schaeffer V, Starzetz T, Kirsch AK, Muller S et al (2017) Multipleximage-based autophagy RNAi screening identifies SMCR8 as ULK1 kinase activityand gene expression regulator. Elife 6. https://doi.org/10.7554/eLife.2306322. Kahle PJ, Neumann M, Ozmen L, Muller V, Jacobsen H, Schindzielorz A et al(2000) Subcellular localization of wild-type and Parkinson's disease-associated mutant alpha -synuclein in human and transgenic mouse brain. JNeurosci 20:6365–637323. Koppers M, Blokhuis AM, Westeneng HJ, Terpstra ML, Zundel CA, Vieira deSa R et al (2015) C9orf72 ablation in mice does not cause motor neurondegeneration or motor deficits. Ann Neurol 78:426–438. https://doi.org/10.1002/ana.2445324. Lee YB, Chen HJ, Peres JN, Gomez-Deza J, Attig J, Stalekar M et al (2013)Hexanucleotide repeats in ALS/FTD form length-dependent RNA foci,sequester RNA binding proteins, and are neurotoxic. Cell Rep 5:1178–1186.https://doi.org/10.1016/j.celrep.2013.10.04925. Levine TP, Daniels RD, Gatta AT, Wong LH, Hayes MJ (2013) The product ofC9orf72, a gene strongly implicated in neurodegeneration, is structurallyrelated to DENN Rab-GEFs. Bioinformatics 29:499–503. https://doi.org/10.1093/bioinformatics/bts72526. Mackenzie IR, Arzberger T, Kremmer E, Troost D, Lorenzl S, Mori K et al(2013) Dipeptide repeat protein pathology in C9ORF72 mutation cases:clinico-pathological correlations. Acta Neuropathol 126:859–879. https://doi.org/10.1007/s00401-013-1181-y27. Mackenzie IR, Frick P, Grasser FA, Gendron TF, Petrucelli L, Cashman NR et al(2015) Quantitative analysis and clinico-pathological correlations of differentdipeptide repeat protein pathologies in C9ORF72 mutation carriers. ActaNeuropathol 130:845–861. https://doi.org/10.1007/s00401-015-1476-228. Maury Y, Come J, Piskorowski RA, Salah-Mohellibi N, Chevaleyre V,Peschanski M et al (2015) Combinatorial analysis of developmental cuesefficiently converts human pluripotent stem cells into multiple neuronalsubtypes. Nat Biotechnol 33:89–96. https://doi.org/10.1038/nbt.304929. May S, Hornburg D, Schludi MH, Arzberger T, Rentzsch K, Schwenk BM et al(2014) C9orf72 FTLD/ALS-associated Gly-ala dipeptide repeat proteins causeneuronal toxicity and Unc119 sequestration. Acta Neuropathol 128:485–503.https://doi.org/10.1007/s00401-014-1329-430. Mele M, Ferreira PG, Reverter F, DeLuca DS, Monlong J, Sammeth M et al(2015) Human genomics. The human transcriptome across tissues andindividuals. Science 348:660–665. https://doi.org/10.1126/science.aaa035531. Mizielinska S, Gronke S, Niccoli T, Ridler CE, Clayton EL, Devoy A et al (2014)C9orf72 repeat expansions cause neurodegeneration in Drosophila througharginine-rich proteins. Science 345:1192–1194. https://doi.org/10.1126/science.125680032. Mizielinska S, Lashley T, Norona FE, Clayton EL, Ridler CE, Fratta P et al(2013) C9orf72 frontotemporal lobar degeneration is characterised byfrequent neuronal sense and antisense RNA foci. Acta Neuropathol 126:845–857. https://doi.org/10.1007/s00401-013-1200-z33. Mori K, Arzberger T, Grasser FA, Gijselinck I, May S, Rentzsch K et al (2013)Bidirectional transcripts of the expanded C9orf72 hexanucleotide repeat aretranslated into aggregating dipeptide repeat proteins. Acta Neuropathol126:881–893. https://doi.org/10.1007/s00401-013-1189-3Frick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 16 of 1734. Mori K, Lammich S, Mackenzie IR, Forne I, Zilow S, Kretzschmar H et al (2013)hnRNP A3 binds to GGGGCC repeats and is a constituent of p62-positive/TDP43-negative inclusions in the hippocampus of patients with C9orf72 mutations. ActaNeuropathol 125:413–423. https://doi.org/10.1007/s00401-013-1088-735. Mori K, Weng SM, Arzberger T, May S, Rentzsch K, Kremmer E et al (2013)The C9orf72 GGGGCC repeat is translated into aggregating dipeptide-repeatproteins in FTLD/ALS. Science 339:1335–1338. https://doi.org/10.1126/science.123292736. Neumann M, Sampathu DM, Kwong LK, Truax AC, Micsenyi MC, Chou TT etal (2006) Ubiquitinated TDP-43 in frontotemporal lobar degeneration andamyotrophic lateral sclerosis. Science 314:130–133. https://doi.org/10.1126/science.113410837. Ohki Y, Wenninger-Weinzierl A, Hruscha A, Asakawa K, Kawakami K, Haass Cet al (2017) Glycine-alanine dipeptide repeat protein contributes to toxicityin a zebrafish model of C9orf72 associated neurodegeneration. MolNeurodegener 12:6. https://doi.org/10.1186/s13024-016-0146-838. O'Rourke JG, Bogdanik L, Muhammad AK, Gendron TF, Kim KJ, Austin A et al(2015) C9orf72 BAC transgenic mice display typical pathologic features ofALS/FTD. Neuron 88:892–901. https://doi.org/10.1016/j.neuron.2015.10.02739. Pfeffer S, Aivazian D (2004) Targeting Rab GTPases to distinct membranecompartments. Nat Rev Mol Cell Biol 5:886–896. https://doi.org/10.1038/nrm150040. Renton AE, Majounie E, Waite A, Simon-Sanchez J, Rollinson S, Gibbs JR et al(2011) A hexanucleotide repeat expansion in C9ORF72 is the cause ofchromosome 9p21-linked ALS-FTD. Neuron 72:257–268. https://doi.org/10.1016/j.neuron.2011.09.01041. Saberi S, Stauffer JE, Jiang J, Garcia SD, Taylor AE, Schulte D et al (2017)Sense-encoded poly-GR dipeptide repeat proteins correlate toneurodegeneration and uniquely co-localize with TDP-43 in dendrites ofrepeat-expanded C9orf72 amyotrophic lateral sclerosis. Acta Neuropathol.https://doi.org/10.1007/s00401-017-1793-842. Satoh J, Tabunoki H, Ishida T, Saito Y, Arima K (2012) Dystrophic neuritesexpress C9orf72 in Alzheimer’s disease brains. Alzheimers Res Ther 4:33.https://doi.org/10.1186/alzrt13643. Schluter OM, Basu J, Sudhof TC, Rosenmund C (2006) Rab3 superprimessynaptic vesicles for release: implications for short-term synaptic plasticity. JNeurosci 26:1239–1246. https://doi.org/10.1523/JNEUROSCI.3553-05.200644. Schluter OM, Schmitz F, Jahn R, Rosenmund C, Sudhof TC (2004) Acomplete genetic analysis of neuronal Rab3 function. J Neurosci 24:6629–6637. https://doi.org/10.1523/JNEUROSCI.1610-04.200445. Sellier C, Campanari ML, Julie Corbier C, Gaucherot A, Kolb-Cheynel I, Oulad-Abdelghani M et al (2016) Loss of C9ORF72 impairs autophagy andsynergizes with polyQ Ataxin-2 to induce motor neuron dysfunction andcell death. EMBO J 35:1276–1297. https://doi.org/10.15252/embj.20159335046. Shi Y, Lin S, Staats KA, Li Y, Chang WH, Hung ST et al (2018)Haploinsufficiency leads to neurodegeneration in C9ORF72 ALS/FTD humaninduced motor neurons. Nat Med. https://doi.org/10.1038/nm.449047. Sudria-Lopez E, Koppers M, de Wit M, van der Meer C, Westeneng HJ, ZundelCA et al (2016) Full ablation of C9orf72 in mice causes immune system-relatedpathology and neoplastic events but no motor neuron defects. ActaNeuropathol 132:145–147. https://doi.org/10.1007/s00401-016-1581-x48. Sullivan PM, Zhou X, Robins AM, Paushter DH, Kim D, Smolka MB et al(2016) The ALS/FTLD associated protein C9orf72 associates with SMCR8 andWDR41 to regulate the autophagy-lysosome pathway. Acta NeuropatholCommun 4:51. https://doi.org/10.1186/s40478-016-0324-549. Suzuki N, Maroof AM, Merkle FT, Koszka K, Intoh A, Armstrong I et al (2013)The mouse C9ORF72 ortholog is enriched in neurons known to degeneratein ALS and FTD. Nat Neurosci 16:1725–1727. https://doi.org/10.1038/nn.356650. Therrien M, Rouleau GA, Dion PA, Parker JA (2013) Deletion of C9ORF72results in motor neuron degeneration and stress sensitivity in C. Elegans.PLoS One 8:e83450. https://doi.org/10.1371/journal.pone.008345051. Ugolino J, Ji YJ, Conchina K, Chu J, Nirujogi RS, Pandey A et al (2016) Loss ofC9orf72 enhances Autophagic activity via deregulated mTOR and TFEBsignaling. PLoS Genet 12:e1006443. https://doi.org/10.1371/journal.pgen.100644352. van Blitterswijk M, Gendron TF, Baker MC, DeJesus-Hernandez M, Finch NA,Brown PH et al (2015) Novel clinical associations with specific C9ORF72transcripts in patients with repeat expansions in C9ORF72. ActaNeuropathol 130:863–876. https://doi.org/10.1007/s00401-015-1480-653. Waite AJ, Baumer D, East S, Neal J, Morris HR, Ansorge O et al (2014)Reduced C9orf72 protein levels in frontal cortex of amyotrophic lateralsclerosis and frontotemporal degeneration brain with the C9ORF72hexanucleotide repeat expansion. Neurobiol Aging 35:1779 e1775–1779e1713. https://doi.org/10.1016/j.neurobiolaging.2014.01.01654. Webster CP, Smith EF, Bauer CS, Moller A, Hautbergue GM, Ferraiuolo L et al(2016) The C9orf72 protein interacts with Rab1a and the ULK1 complex toregulate initiation of autophagy. EMBO J 35:1656–1676. https://doi.org/10.15252/embj.20169440155. Wu X, Bradley MJ, Cai Y, Kummel D, De La Cruz EM, Barr FA et al (2011)Insights regarding guanine nucleotide exchange from the structure of aDENN-domain protein complexed with its Rab GTPase substrate. Proc NatlAcad Sci U S A 108:18672–18677. https://doi.org/10.1073/pnas.111041510856. Xiao S, MacNair L, McGoldrick P, McKeever PM, McLean JR, Zhang M et al(2015) Isoform-specific antibodies reveal distinct subcellular localizations ofC9orf72 in amyotrophic lateral sclerosis. Ann Neurol 78:568–583. https://doi.org/10.1002/ana.2446957. Yang M, Liang C, Swaminathan K, Herrlinger S, Lai F, Shiekhattar R et al (2016)A C9ORF72/SMCR8-containing complex regulates ULK1 and plays a dual rolein autophagy. Sci Adv 2:e1601167. https://doi.org/10.1126/sciadv.160116758. Yoshimura S, Gerondopoulos A, Linford A, Rigden DJ, Barr FA (2010) Family-wide characterization of the DENN domain Rab GDP-GTP exchange factors.J Cell Biol 191:367–381. https://doi.org/10.1083/jcb.20100805159. Zhang D, Iyer LM, He F, Aravind L (2012) Discovery of novel DENN proteins:implications for the evolution of eukaryotic intracellular membranestructures and human disease. Front Genet 3:283. https://doi.org/10.3389/fgene.2012.0028360. Zu T, Liu Y, Banez-Coronel M, Reid T, Pletnikova O, Lewis J et al (2013) RANproteins and RNA foci from antisense transcripts in C9ORF72 ALS andfrontotemporal dementia. Proc Natl Acad Sci U S A 110:E4968–E4977.https://doi.org/10.1073/pnas.1315438110Frick et al. Acta Neuropathologica Communications  (2018) 6:72 Page 17 of 17


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items