UBC Faculty Research and Publications

ALS-linked misfolded SOD1 species have divergent impacts on mitochondria Pickles, Sarah; Semmler, Sabrina; Broom, Helen R; Destroismaisons, Laurie; Legroux, Laurine; Arbour, Nathalie; Meiering, Elizabeth; Cashman, Neil R; Vande Velde, Christine Apr 27, 2016

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

Item Metadata


52383-40478_2016_Article_313.pdf [ 1.91MB ]
JSON: 52383-1.0368582.json
JSON-LD: 52383-1.0368582-ld.json
RDF/XML (Pretty): 52383-1.0368582-rdf.xml
RDF/JSON: 52383-1.0368582-rdf.json
Turtle: 52383-1.0368582-turtle.txt
N-Triples: 52383-1.0368582-rdf-ntriples.txt
Original Record: 52383-1.0368582-source.json
Full Text

Full Text

RESEARCH Open AccessALS-linked misfolded SOD1 species havedivergent impacts on mitochondriaSarah Pickles1,2, Sabrina Semmler1,4, Helen R. Broom5, Laurie Destroismaisons1, Laurine Legroux1,3,Nathalie Arbour1,3, Elizabeth Meiering5, Neil R. Cashman6 and Christine Vande Velde1,3*AbstractApproximately 20 % of familial Amyotrophic Lateral Sclerosis (ALS) is caused by mutations in superoxide dismutase(SOD1), which leads to misfolding of the SOD1 protein, resulting in a toxic gain of function. Several conformation-restricted antibodies have been generated that specifically recognize misfolded SOD1 protein, and have been usedas therapeutics in pre-clinical models. Misfolded SOD1 selectively associates with spinal cord mitochondria in SOD1rodent models. Using the SOD1G93A rat model, we find that SOD1 conformational specific antibodies AMF7-63 andDSE2-3H1 labeled a fibrillar network concentrated in the anterior horn; while A5C3, B8H10, C4F6 and D3H5 labeledmotor neurons as well as puncta in the neuropil. There is a time-dependent accumulation of misfolded SOD1 atthe surface of spinal cord mitochondria with AMF7-63-labeled mitochondria having increased volume in contrast toa mitochondrial subset labeled with B8H10. In spinal cord homogenates and isolated mitochondria, AMF7-63, DSE2-3H1 and B8H10 detect misfolded SOD1 aggregates. SOD1 that lacks its metal cofactors has an increased affinity fornaïve mitochondria and misfolded SOD1 antibodies B8H10 and DSE2-3H1 readily detect demetalated mutant andwild-type SOD1. Together, these data suggest that multiple non-native species of misfolded SOD1 may exist, someof which are associated with mitochondrial damage. Conformational antibodies are invaluable tools to identify andcharacterize the variation in misfolded SOD1 species with regards to biochemical characteristics and toxicity. Thisinformation is highly relevant to the further development of these reagents as therapeutics.Keywords: Amyotrophic Lateral Sclerosis, Mitochondria, Superoxide dismutase, Flow cytometryIntroductionThe defining feature of the neurodegenerative diseaseAmyotrophic Lateral Sclerosis (ALS) is the loss of motorneurons in the cortex, brain stem and spinal cord [1]. Lossof motor neurons leads to denervation resulting in muscleweakness, atrophy and eventual paralysis. Despite identifi-cation of the first gene linked to familial ALS (FALS),Superoxide Dismutase 1 (SOD1) [2] over twenty yearsago, and the discovery of many more ALS genes since, thecauses of motor neuron degeneration remain unknown.Mutations in SOD1 account for 15 to 20 % of all FALScases, and approximately 3 % of sporadic ALS (SALS) cases[3]. SOD1 mutations universally lead to conformationchanges within the native protein structure, resulting inthe acquisition of an elusive toxic function [4]. Several anti-bodies have been developed to specifically target thesealtered conformations, which are collectively referred to asmisfolded SOD1 (reviewed in [5, 6]). RecombinantSOD1G93A protein lacking its metals (apo), including thestructure stabilizing zinc cofactor, was used forimmunization which led to the generation of a heteroge-neous pool of antibodies with different affinities and re-activity to distinct epitopes located on one or more of theSOD1G93A protein conformers. These antibodies were sub-sequently clonally expanded to monoclonal antibodiesnamed as A5C3, B8H10, C4F6, and D3H5 [7, 8]. Otherantibodies, such as DSE2-3H1, SEDI, USOD, and a seriesof polyclonal antibodies produced by Forsberg and col-leagues, were produced via immunization with peptidescomprised of amino acids that are normally inaccessible inthe well folded protein [9–11]. All of these antibodiesrecognize epitopes that are exposed only when SOD1* Correspondence: c.vande.velde@umontreal.ca1Centre de recherche du Centre Hospitalier de l’Université de Montréal(CRCHUM) Université de Montréal, 900 rue Saint-Denis, Local R09.442,Montréal, QC H2X 0A9, Canada3Department of Neurosciences, Université de Montréal, Montréal, QC H2X0A9, CanadaFull list of author information is available at the end of the article© 2016 Pickles et al. 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.Pickles et al. Acta Neuropathologica Communications  (2016) 4:43 DOI 10.1186/s40478-016-0313-8adopts a non-native conformation induced either by muta-tion, loss of its zinc cofactor, and/or oxidation. While manyof these were developed with the intent to be potentialtherapeutics, these reagents have also become valuabletools with which to track the toxic forms of SOD1. Mis-folded SOD1 is detected predominantly within the motorneurons of ALS animal models [8, 11–14]. In humans,various antibodies report on misfolded SOD1 in neuronsof FALS patients as well as SALS patients, although thislatter finding remains controversial [7, 9, 15, 16]. In pre-clinical research using mutant SOD1 animals, it is nowappreciated that reducing misfolded SOD1 levels viaimmunization significantly increases survival [8, 17]. Thisprovides additional support that misfolded SOD1 lies atthe root of SOD1-mediated ALS [8, 17].Despite consensus in the field that misfolded SOD1 iscentral to disease pathogenesis, it remains unknownhow misfolded SOD1 causes motor neuron death. Mis-folded SOD1 has been implicated in the induction ofendoplasmic reticulum (ER) stress [12, 18], defectiveaxonal transport [7], alteration of motor neuron excitabil-ity [19], and mitochondrial dysfunction [11, 13, 14, 20] inSOD1-mediated ALS disease models. Multiple aspects ofmitochondrial physiology are disrupted in mutant SOD1cell culture and animal models including morphology[21–23], adenosine triphosphate (ATP) generation [24],calcium handling [25], axonal transport [26] and proteinimport [27]. Interestingly, misfolded SOD1 directly associ-ates with mitochondria derived from affected, but not un-affected tissues [11]. The selective association of misfoldedSOD1 to spinal cord mitochondria has just recently beenattributed to a lack of the putative chaperone macrophagemigration inhibitory factor (MIF) in this tissue, and morespecifically motor neurons [28].Recent evidence suggests that multiple non-native/misfolded SOD1 species may exist [29–31]. Consistentwith this concept, we have previously reported that theB8H10 antibody detects misfolded SOD1 in both cyto-solic and mitochondrial fractions prepared fromSOD1G93A spinal cords while the C4F6 antibody exclu-sively detects cytosolic misfolded SOD1 [13]. Otherwork in cultured cells made to overexpress mutantSOD1 indicates that the C4F6 antibody recognizes sol-uble mutant protein, whereas SEDI preferentially detectsmutant SOD1 within inclusions [31]. Additionally, aseries of polyclonal SOD1 peptide-specific antibodiesidentify two different forms of SOD1 aggregates (or“strains”) in mutant SOD1 mice based on epitope acces-sibility, with one such aggregate-type/strain correlatingwith an earlier age of onset [30]. Together these datasuggest that multiple forms of misfolded SOD1 arepossible.We hypothesized that if more than one form of mis-folded SOD1 exists, there may be conformer-specificdifferences in localization, potency and/or pathomecha-nistic consequences. To this end, we have employed apanel of misfolded SOD1-specific antibodies, to evaluatemisfolded SOD1 localization, ability to induce mitochon-drial toxicity and incorporation into aggregates. Herein,we report that the misfolded SOD1-specific antibodyDSE2-3H1 labels motor neurons and robustly detects fi-brils in the anterior horn of SOD1G93A spinal cords, afinding that is confirmed by a second independent anti-body raised against the same peptide immunogen(AMF7-63). Other misfolded SOD1-specific antibodies,A5C3, B8H10, C4F6 and D3H5 antibodies label predom-inantly to motor neurons and numerous neuropilpuncta. Despite their different labeling patterns withinthe spinal cord, both B8H10 and AMF7-63 antibodiesimmunolabel spinal cord mitochondria in a time-dependent manner. However, the presence of AMF7-63-reactive misfolded SOD1 at mitochondria correlates witha more severe dysregulation of mitochondrial volumecompared to mitochondria without associated misfoldedSOD1.Materials and methodsAnimalsSOD1G93A and SOD1WT transgenic rats have been previ-ously described [32, 33]. Non-transgenic littermates wereused in some experiments. Early symptomatic is definedas animals that have a noticeable gait defect, hopping orlimping, typically involving only one limb. Both maleand female rats were used. Animals were treated in strictadherence to approved protocols from the CRCHUM In-stitutional Committee for the Protection of Animals andthe Canadian Council on Animal Care (CCAC).AntibodiesRabbit anti-Cu/Zn SOD (Enzo Life Sciences), rabbit anti-SOD1 (Cell Signaling), mouse anti-VDAC1 (Calbiochem),mouse anti-Actin (MP Biomedicals), were used for immu-noblots. Anti-misfolded SOD1 mouse monoclonal anti-bodies D3H5 (1:250, generously provided by Dr. J-P Julien),A5C3 (1:50), B8H10 (1:250) and C4F6 (1:250) (Medimabs),DSE2-3H1 (1:1000), rabbit monoclonal antibody AMF7-63(1:1500) and rabbit polyclonal antibody SEDI (1:100, gener-ously provided by Dr. J. Robertson) were used for immuno-blotting, immunofluorescence and flow cytometry. Mouseand rabbit IgG (Jackson ImmunoResearch Labs) and mouseanti-IgG1 (BD Biosciences) were used as controls. Goatanti-mouse allophycocyanin-conjugated (BD Pharmingen),goat anti-rabbit PE (eBioscience) and goat anti-rabbit PE-Cy7-conjugated (Santa Cruz) secondary antibodies wereused for flow cytometry studies. For immunofluorescence,goat anti-ChAT (1:100; Millipore), mouse anti-SMI32(1:2000; Covance), mouse anti-SMI31 (1:2000; Covance)and mouse anti-MAP2 (1:500; Sigma) were used.Pickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 2 of 14Flow cytometry of isolated mitochondriaSpinal cord and liver mitochondria were isolated frommice and rats [11], and prepared for analysis by flow cy-tometry, as previously described [13, 34].Immunoprecipitation and immunoblottingIsolated mitochondria were solubilized and immunopre-cipitated as previously described [11]. Briefly, 50 μg ofmitochondria were incubated with 15 μL Protein G mag-netic beads (Invitrogen), overnight at 4 °C with rotation.Protein G beads were previously incubated with mis-folded SOD1-specific antibody. Immunoprecipitatedproteins were eluted from the beads in 2.5× Laemmlibuffer and electrophoresed on 15 % Tris-Glycine gels,and subsequently transferred to nitrocellulose.ImmunofluorescenceSections were labeled with anti-misfolded SOD1 anti-bodies, as previously described [13]. Briefly, sectionswere washed 10 min at room temperature in PBS, thenpermeabilized for 10 min at room temperature in PBSwith 0.4 % TX-100. Sections were blocked with 2 %normal donkey serum (Sigma), 2 % bovine serum albu-min (Sigma), in 0.4 % TX-100/PBS for 1 h at roomtemperature. Primary antibodies were incubated over-night at 4 °C in blocking solution. Appropriate second-ary antibodies were added in blocking solution for 1 h atroom temperature. Sections were mounted using Pro-Long antifade reagent (Invitrogen). Immunofluorescentimages were captured by confocal microscopy (LeicaSP5; 20× and 40× objective, 1.7 NA) and processed withLeica LAS AF software and/or PhotoshopCS4 (Adobe).Filter-trap assay20 μg of spinal cord homogenates or isolated spinal cordmitochondria in PBS were filtered through a 0.22 μmcellulose acetate membrane (GE Healthcare) using theBio-Dot Microfiltration Apparatus (Bio-Rad). Wells werewashed twice with PBS, the membrane was removedfrom the apparatus and then blocked 1 h at roomtemperature and immunoblotted with misfolded SOD1-specific antibodies. Mitochondria for these experimentswere prepared by floating upwards to their buoyantdensity so as to avoid possible co-sedimentation of ag-gregates, as previously described [11].Dot blot of recombinant SOD1 protein1 μg of recombinant SOD1 protein, produced as previ-ously described [35–37], in TBS (20 mM Tris, 500 mMNaCl, 1 mM EDTA pH 7.5) was spotted onto nitrocellu-lose membrane (BioRad) using the Bio-Dot Microfiltra-tion Apparatus (Bio-Rad). Wells were washed twice withTBS, and the membrane was removed from the appar-atus and blocked in TBS-T (as above plus 0.05 %Tween-20) with 1 % bovine serum albumin (BSA) for30 min at room temperature, and immunoblotted withmisfolded SOD1 antibodies. Primary and secondary anti-bodies were incubated in blocking buffer. For non-nativesamples, 5 % v/v BME, and 0.5 % v/v SDS was added,and samples were heat denatured by incubation for5 min at 95 °C.In vitro mitochondrial binding assay50 μg of isolated spinal cord mitochondria (2 μg/μL)from non-transgenic rats were incubated with 3 μMbaculovirus-produced SOD1WT and SOD1G93A recom-binant protein, purified as previously described [38], for20 min at 37 °C in HB Buffer (210 mM Mannitol,70 mM Sucrose, 10 mM Tris pH 7.5, 1 mM EDTA) [38].Mitochondria were washed once with HB buffer andthen re-suspended in HB and 4× Laemmli sample bufferand subjected to SDS-PAGE and immunoblotted withan antibody to human SOD1 (Cell Signaling). To deter-mine if modification of SOD1 structure would alter itsbinding to the mitochondrial surface, the protein was in-cubated with 5.5 mM EDTA or 10 mM hydrogen perox-ide in PBS overnight at 4 °C or room temperature,respectively, with protease inhibitors (Roche). EDTA andhydrogen peroxide were removed and replaced by PBSby dialysis with Slide-A-Lyzer Mini dialysis devices(Pierce). Untreated samples were treated equivalently.StatisticsTwo-way ANOVA was used to determine the interactionbetween groups and time for percentage of misfoldedSOD1+ mitochondrial subpopulations over time and differ-ences in AMF7-63+, B8H10+, and negative mitochondrialsubpopulations over time. Sidak’s multiple comparison testwas used to determine differences between misfoldedSOD1+ groups. One-way ANOVA was used to determinedifferences in AMF7-63+, AMF7-63+B8H10+ and B8H10+subpopulations. * P < 0.05, ** P < 0.01 *** P < 0.001,**** P < 0.0001. All analyses was done with GraphPadPrism software.ResultsMisfolded SOD1 specific antibodies DSE2-3H1 and AMF7-63 detect fibrils in the spinal cord of SOD1G93A ratsTo evaluate whether multiple antibodies targeted tonon-native/misfolded conformations of SOD1 yielded auniversal localization under similar conditions, lumbarsections of symptomatic SOD1G93A rat spinal cords werelabeled with a panel of misfolded SOD1 specific anti-bodies. In symptomatic SOD1G93A rats, all of the mis-folded SOD1-specific antibodies tested (A5C3, B8H10,C4F6, D3H5, DSE2-3H1, and AMF7-63) labeled motorneurons as marked by choline acetyltransferase (ChAT)(Fig. 1a). Antibodies DSE2-3H1 (mouse monoclonal)Pickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 3 of 14Fig. 1 (See legend on next page.)Pickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 4 of 14and AMF7-63 (rabbit monoclonal) were raised againstthe same epitope located in the electrostatic loop (aminoacids 125DDLGKGGNEESTKTGNAG142) of SOD1which is normally inaccessible in the well-folded SOD1structure [11]. In symptomatic animals, both antibodiesintensely labeled the anterior horn albeit with varying af-finities, with AMF7-63 demonstrating a more intense la-beling (Fig. 1a). This is consistent with a 103-foldenhanced affinity for the immunogenic peptide (N.Cashman, unpublished data). The labeling resembled anetwork of fibril-like structures, with some motorneuron soma being obviously labeled. In general, the fi-brillar pattern detected by DSE2-3H1 and AMF7-63 wasso robust, it was often difficult to discern individualmotor neurons. These two antibodies also revealed a fi-brillar network in pre-symptomatic animals (14 and15 weeks old, Fig. 1b). In contrast, fibrils were only occa-sionally detected in sections labeled with misfoldedSOD1 antibodies A5C3, B8H10 and C4F6, but notD3H5, in pre-symptomatic as well as symptomatic ani-mals (data not shown). Instead, we noted that A5C3,B8H10, C4F6 and D3H5 antibodies homogenously la-beled motor neuron somata, and this was accompaniedby numerous small puncta observed throughout theneuropil (Fig. 1a). As expected, none of the antibodiesyielded a signal in lumbar spinal cords of age-matchedrats expressing comparable levels of human wild typeSOD1 (SOD1WT), thereby confirming antibody specifi-city, nor was there non-specific labeling in IgG controlsor sections stained with secondary antibody alone(Additional file 1: Figure S1A, B). AMF7-63-reactivemisfolded SOD1 fibrils could be localized to differentmotor neuron subcompartments including the soma,dendrites (SMI32, MAP2) and axons (SMI31) (Fig. 1c).These initial studies indicate that misfolded SOD1-specific antibodies can be broadly considered as two dis-tinct groups: A5C3, B8H10, C4F6, and D3H5 whichlabel motor neurons and numerous puncta throughoutthe neuropil; and AMF7-63 and DSE2-3H1 which alsolabel motor neurons but also intensely reveal fibril-likestructures. Based on these spatial considerations, thesedata suggest that more than one type of misfoldedSOD1 species exists in vivo.To determine if these seemingly different misfoldedSOD1 conformations could co-exist within the samemotor neuron, spinal cord sections were co-labeled withAMF7-63 and B8H10. A partial co-localization of thesetwo antibodies within ChAT-positive motor neurons wasfrequently observed (Fig. 1d), suggesting that these anti-bodies recognize apparently distinct non-native SOD1species within the same neurons. In addition, we ob-served neurons that labeled with AMF7-63 uniquely (ie.void of B8H10), and vice versa.AMF7-63 antibody detects misfolded SOD1G93A at themitochondrial surfaceSeveral misfolded SOD1-specific antibodies (DSE2-3H1,A5C3, B8H10, SEDI) recognize misfolded SOD1 proteindeposited on the cytoplasmic face of the mitochondrialouter membrane [10, 11, 13, 14, 20]. However, this isnot a universally shared feature of misfolded SOD1 asconformers recognized by C4F6 are primarily cytosolicand exhibit little to no mitochondrial association [13].We sought to determine if AMF7-63-reactive misfoldedSOD1 also associates with mitochondria. MisfoldedSOD1-specific antibodies A5C3, B8H10, DSE2-3H1 andAMF7-63 were used to immunoprecipitate misfoldedSOD1 from spinal cord mitochondria isolated fromsymptomatic SOD1G93A animals (Fig. 2a). As previouslypublished, B8H10 and DSE2-3H1-reactive SOD1 wererobustly detected in mitochondrial fractions [11, 13, 28](Fig. 2a). AMF7-63 detected similar amounts of mis-folded SOD1 (Fig. 2a). A5C3, which we have previouslydemonstrated to label misfolded SOD1 on distal axonalmitochondria [11] also detected misfolded SOD1 withinmitochondrial fractions, but to a lesser extent than theother antibodies (Fig. 2a). While a non-specific bandmigrating just above SOD1 was detected when theAMF7-63 antibody was used for immunoprecipitation,the specificity of the AMF7-63 antibody for misfoldedSOD1 was confirmed by immunoprecipitation of spinalcord homogenates and isolated mitochondria fromsymptomatic SOD1G93A rats as well as age-matchedSOD1WT rats. As expected, AMF7-63 immunoprecipi-tated misfolded SOD1 exclusively from SOD1G93A ratspinal cord homogenates and mitochondria, but notsimilar preparations from livers or SOD1WT tissue(Additional file 2: Figure S2).We wondered if the presence of AMF7-63-reactive mis-folded SOD1 conformer negatively impacted mitochondrial(See figure on previous page.)Fig. 1 Misfolded SOD1-specific antibodies have distinct labeling patterns in SOD1G93A rat spinal cords. Immunohistochemistry for misfolded SOD1in SOD1G93A lumbar spinal cords. a Lumbar sections of a symptomatic SOD1G93A rat were stained with misfolded SOD1 specific antibody A5C3(green) and co-labeled with ChAT (blue). Additional representative images of B8H10, C4F6, D3H5, DSE2-3H1 and AMF7-63 are also shown. b TheAMF7-63 antibody detects fibrils in pre-symptomatic, 14 and 15 week SOD1G93A rat spinal cords. c Symptomatic SOD1G93A rat spinal cord waslabeled with AMF7-63 (green) and SMI32 (red), SMI31 (red), or MAP2 (red). d Lumbar sections labeled with misfolded SOD1 antibodies AMF7-63(blue), B8H10 (green), and co-labeled with ChAT (red). Two to three animals of each genotype were analyzed. Scale bar = 100 μm for (a, b and d)and 25 μm for (c)Pickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 5 of 14function. To this end, we employed surface-labeling withthe misfolded SOD1-specific antibodies AMF7-63 andB8H10 and subsequent flow cytometric detection of iso-lated spinal cord mitochondria [34]. Using early symptom-atic SOD1G93A rats, we established that AMF7-63preferentially detected misfolded SOD1 on isolated spinalcord mitochondria compared to liver, a tissue that is un-affected in ALS (Fig. 2b). Significantly more individualspinal cord mitochondria (as marked by the indicator dyeMitoTracker Green, MTG) from SOD1G93A rats labeled forFig. 2 B8H10 and AMF7-63 reactive misfolded SOD1 is present in SOD1G93A spinal mitochondrial fractions. a Immunoprecipitation for misfoldedSOD1 in isolated SOD1G93A mitochondria with A5C3, B8H10, DSE2 3H1 and AMF7-63. Mouse (mIgG) and rabbit IgG (rbIgG) serve as controls. Inputis 10 μg isolated mitochondria. Upper bands and lower bands correspond to human (hSOD1) and rat (rSOD1) SOD1, respectively. Immunoprecipitationwith AMF7-63 resulted in a non-specific band (*) just above human SOD1, regardless of misfolded SOD1 status. Experiment shown is representative ofthree independent trials. b Immunolabeling of isolated spinal cord and liver mitochondria with misfolded SOD1-specific antibody AMF7-63 from symptomatic SOD1G93A rats and controls (aged matched SOD1WT and non-transgenic rats) by flow cytometry. Misfolded SOD1positive labeling is determined by comparing to isotype control (rabbit IgG, dotplots in first column) of SOD1G93A sample. Percentage ofmisfolded SOD1+ events is shown for each tissue and genotype in a representative sample. c Quantification of AMF7-63+ events in spinalcord (green circle) or liver (black square) of symptomatic SOD1G93A rats, and age- matched SOD1WT as well as non-transgenic rats. Dataare represented as percent of misfolded AMF7-63+ mitochondria, each dot represents one animal, n = 4 animals per genotype per tissue.d Comparison of spinal cord mitochondrial labeling positive for AMF7-63 (green, circle) or B8H10 (blue, square) in pre-symptomatic (10 and 14 weeks)and symptomatic SOD1G93A rats by flow cytometry. Data are represented as percentage of misfolded SOD1+ mitochondria, each dot represents oneanimal, (mean), n = 4–11 animals. ** P < 0.01. *** P < 0.001Pickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 6 of 14AMF7-63 (6.6 ± 1.9 %) compared to SOD1WT (0.1 ±0.03 %) and non-transgenic (0.2 ± 0.03 %) animals(SOD1G93A vs, SOD1WT, non-transgenic: P < 0.001) (Fig. 2b,c). As expected, AMF7-63-reactive misfolded SOD1 wasnot detected (ie. below 1 %) on liver mitochondria fromany group, confirming specificity of misfolded SOD1 for af-fected tissues (SOD1G93A: 0.1 ± 0.1 %; SOD1WT: 0.3 ±0.02 %; non-transgenic: 0.1 ± 0.2 %, n = 4 animals per geno-type) (Fig. 2c). Similarly, and consistent with our previouswork [13], B8H10+ mitochondria were robustly detected inmitochondrial preparations from symptomatic SOD1G93Aspinal cords (6.1 ± 0.4 %) but not SOD1WT (0.4 ± 0.2 %) ornon-transgenic cords (0.4 ± 0.1 %) (SOD1G93A vs, SOD1WT,non-transgenic: P < 0.0001, n = 3 animals per genotype)(Additional file 3: Figure S3A, B). There was no substantialB8H10 labeling of liver mitochondria in any animal modeltested (Additional file 3: Figure S3A, B).At an early symptomatic stage, surface labeling of iso-lated spinal cord mitochondria demonstrated that bothAMF7-63 and B8H10 antibodies detected misfoldedSOD1 at the cytoplasmic face of the mitochondrial outermembrane (Fig. 2d). Thus, we asked whether there wasa temporal difference in the accumulation of these twoconformers. Spinal cord mitochondria from 10 week old,14 week old and early symptomatic SOD1G93A rats wereprocessed for labeling with B8H10 and AMF7-63. Whileno mitochondrial signal for either antibody was detectedat 10 weeks, comparable labeling was detected in mito-chondria from 14 week animals (AMF7-63+: 1.5 ± 0.6 %;B8H10+: 2.1 ± 0.5 %). Higher proportions of mitochon-dria labeled for misfolded SOD1 at the early symptom-atic stage compared to the 10 and 14 week groups,demonstrating a significant age-dependent accumulationof each misfolded SOD1 conformer at the mitochondrialsurface (P < 0.0001). Comparison of the relative amountsof AMF7-63+ (5.2 ± 1.1 %) and B8H10+ (6.9 ± 0.6 %) sub-populations yielded no significant differences (Fig. 2d).Collectively, these data would suggest that there is nopreferential temporal accumulation of these two formsof non-native SOD1 conformers at the mitochondrialsurface.Given these latter data, one could argue that the anti-bodies are detecting the same conformer in vivo. Thus,to address this, we analyzed our data to determine if itwas possible to detect B8H10-labeled mitochondria voidof AMF7-63 labeling, and vice versa. By simultaneouslyimmunolabeling isolated mitochondria with bothAMF7-63 and B8H10, we were able to discern four dis-tinct mitochondrial subpopulations: i) double negative(AMF7-63−B8H10−); ii) AMF7-63 only (AMF7-63+B8H10−) 2.4 % in the illustrated example; iii) B8H10only (AMF7-63−B8H10+), 2.9 % in the illustrated ex-ample; and iv) double positive (AMF7-63+B8H10+),5.4 % in the illustrated example (Fig. 3a). Therefore, thetotal AMF7-63+ subpopulation (7.8 %) consists of bothB8H10+ (5.4 %) and B8H10− mitochondria (2.4 %). Thata proportion of mitochondria label positive with bothmisfolded SOD1 antibodies, while others label for onlyone conformer, suggests that AMF7-63 and B8H10recognize distinct misfolded SOD1 species.From 14 weeks to the early symptomatic stage, therewas a significant time dependent increase in the percent-age of misfolded SOD1 B8H10-labeled mitochondria(P < 0.01) (Fig. 3b). The relative amounts of the threesubpopulations with surface-bound misfolded SOD1 re-vealed no significant differences between them at 14 weeks.Interestingly, between 14 weeks and the early symptom-atic stage, the proportion of AMF7-63+B8H10+ andAMF7-63−B8H10+ mitochondrial subpopulations nearlydoubled, whereas the AMF7-63+ B8H10− subpopulationremained roughly constant (Fig. 3b). At the lattertime point, there is a significantly higher percentageof AMF7-63−B8H10+ than AMF7-63+B8H10− mito-chondria (P < 0.01) (Fig. 3b). These data suggest thereis either preferential removal of the AMF7-63 only sub-population or disturbed removal/enhanced accumulationof B8H10 only mitochondria.Volume dyshomeostasis and superoxide production isenhanced in mitochondria with surface-bound AMF7-63-reactive SOD1A distinct advantage of the flow cytometry-basedmethod to detect mitochondrially-associated misfoldedSOD1 is that it permits the simultaneous use of fluores-cent indicator dyes to report on mitochondrial functions.Since a difference in the timing of mitochondrial associ-ation between AMF7-63+ and B8H10+ mitochondria wasnot detected (Fig. 2d), we hypothesized that perhaps theymay exhibit variable toxicity towards mitochondria.Thus, we evaluated select aspects of mitochondrial func-tion in the AMF7-63+ subpopulation. Mitochondrialsize/volume was assessed by flow cytometry on the basisof the intensity of forward light scatter (FSC) of individ-ual mitochondria [39, 40]. Quantification of the FSCnormalized to the total population indicate that the totalmitochondrial subset bearing AMF7-63 reactive mis-folded SOD1 from 14 week and early symptomaticSOD1G93A animals were significantly larger than non-coated mitochondria (P < 0.001) (Fig. 4a). Intriguingly,the AMF7-63+ subpopulation was also significantly largerthan the B8H10+ subpopulation (regardless of AMF7-63status) when animals began exhibiting early symptoms(P < 0.0001) (Fig. 4a).Mitotracker Green (MTG) can be used not only toidentify mitochondria, but also to report on mitochon-drial volume. Specifically, dye uptake measured by thedelta mean fluorescence intensity (ΔMFI) correlates withmitochondrial volume as the dye accumulates withinPickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 7 of 14mitochondria independent of the mitochondrial trans-membrane potential [41]. In agreement with our FSCdata, AMF7-63+ and B8H10+ subpopulations have a sig-nificantly higher ΔMFI compared to uncoated mito-chondria at both time points (AMF7-63+, P < 0.0001;B8H10+, P < 0.01), with AMF7-63+ mitochondria takingup more dye compared to B8H10+ mitochondria whenanimals are in the early symptomatic stage (P < 0.0001)(Fig. 4b). Together, these data indicate that the associationof AMF7-63-reactive misfolded SOD1 conformers withthe mitochondrial surface correlates with enlargedmitochondria.Superoxide is produced as a natural by-product of oxi-dative phosphorylation [42]. We evaluated the levels ofmitochondrial superoxide produced by the AMF7-63+and B8H10+ subpopulations using MitoSOX Red, amitochondria-specific superoxide indicator [43, 44].Following normalization for size differences, misfoldedSOD1+ mitochondrial subpopulations produced signifi-cantly higher levels of superoxide compared to thea0.5% 0%0.4%2.4% 5.4%2.9%B8H10Isotype b****Fig. 3 Four distinct mitochondrial subpopulations revealed by simultaneously immunolabeling with AMF7-63 and B8H10 misfolded SOD1-specificantibodies. a Flow cytometric analysis of isolated spinal cord mitochondria from symptomatic SOD1G93A animals labeled with AMF7-63 and B8H10.Analysis only includes events previously gated for MTG staining (MTG+). Percentage of misfolded SOD1+ events is shown for each subpopulation in arepresentative experiment. b Quantification of AMF7-63+B8H10− (green, circle), AMF7-63+B8H10+ (black, square) and B8H10+ AMF7-63− (blue, triangle)mitochondrial subpopulations from spinal cords of 14 week and early symptomatic SOD1G93A rats. Data are represented as percent of misfolded SOD1+mitochondria (mean), n= 4–11 animals per subpopulation. ** P < 0.0114 weeks early symp. MitoSOX+ mitos. rel. to Total14 weeks early symp. MFI TMRM rel. to total 14 weeks early symp.0123Delta MFI MTG rel. to total14 weeks early symp. rel. to totala b c d **** ** **** ns ** **** ns ** **** **** ** **** * Fig. 4 Presence of AMF7-63 reactive misfolded SOD1 at the mitochondrial surface coincides with increased mitochondrial volume. Flow cytometricanalysis of isolated spinal cord mitochondria from 14 week and early symptomatic SOD1G93A rats labeled with misfolded SOD1 specific antibodiesAMF7-63 and B8H10. Analysis includes only events previously gated for MTG staining (MTG+). a Quantification of the geometric mean of FSC of AMF7-63+ (green) B8H10+ (blue) and negative (white) mitochondrial subpopulations relative to total population in 14 week and symptomatic SOD1G93Arats. b Quantification of delta mean fluorescence intensity (ΔMFI) of MTG staining of mitochondrial sub-populations relative to total population.c Quantification of percentage of MitoSOX+ mitochondria from mitochondrial subpopulations relative to total population. Each subpopulation wasnormalized for MTG staining. d Quantification of ΔMFI of TMRM staining of mitochondrial subpopulations relative to total population and normalizedto size/MTG. n = 4–5 animals per time point. Not significant (ns), * P < 0.05, ** P < 0.01, **** P < 0.0001Pickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 8 of 14misfolded SOD1− subpopulation at both 14 weeks andwhen animals began to exhibit symptoms (P <0.05). Fur-ther comparison revealed that while the AMF7-63+ andB8H10+ subpopulations were not significantly differentfrom each other, the AMF7-63+ subpopulation was sig-nificantly increased compared to the unlabeled subpopu-lations at the early symptomatic stage (P < 0.05, Fig. 4c).These changes were independent of mitochondrial trans-membrane potential (ΔΨm) which was unchanged be-tween subpopulations (Fig. 4d). Taken together, thesedata suggest that there could be variable mitochondrialdamage associated with different conformers of mis-folded SOD1, given that AMF7-63-reactive misfoldedSOD1 is associated with more severe deregulation ofmitochondrial volume homeostasis, while superoxideproduction is equivalent to B8H10-coated mitochondria.Misfolded SOD1 conformers are aggregated onmitochondriaGiven that B8H10 and AMF7-63-reactive misfoldedSOD1 disturbed mitochondrial volume to varying de-grees, we speculated this might be attributed to differ-ences in biochemical properties. Certain misfoldedSOD1-specific antibodies are reported to detect cyto-plasmic aggregates/inclusions [30, 31], but whether mis-folded SOD1 aggregates at/on mitochondria remainsunknown. Therefore, we examined the accumulation ofmisfolded SOD1 into aggregates in isolated mitochon-drial fractions. A filter trap assay in which proteinaceousaggregates larger than 220 nm are retained on a celluloseacetate membrane [45], was performed on homogenatesand isolated mitochondria from spinal cords of pre-symptomatic (10 week) and early symptomaticSOD1G93A rats as well as age-matched SOD1WT animals.Given that this assay is performed in non-denaturingconditions, we reasoned that the misfolded SOD1conformational antibodies should retain their specificity.In agreement with this, misfolded SOD1 antibodiesB8H10, DSE2-3H1 and AMF7-63 preferentially labeledhomogenates of SOD1G93A spinal cords but not controls(Fig. 5a). Moreover, these three antibodies demonstratedmore intense immunoreactivity for mitochondrial sam-ples (which were isolated via buoyant density centrifuga-tion so as to avoid potential co-pelleting of cytoplasmicaggregates) (Fig. 5a). Furthermore, the formation ofaggregates was disease/age-dependent, with robust label-ing of homogenates and isolated mitochondria fromearly symptomatic animals, but little to no labeling at10 weeks. Note, the C4F6 antibody detected little to noaggregates in homogenates or isolated mitochondria atany age (Fig. 5a), consistent with reports by others thatC4F6 recognizes soluble misfolded SOD1 [31, 46].Western blots depict SOD1 expression between samples(Fig. 5b). These results suggest that misfolded SOD1conformers recognized by B8H10, DSE2-3H1, andAMF7-63, are components of protein aggregates in bothspinal cord homogenates and are enriched onmitochondria.Preferential recognition of demetallated and reducedrecombinant SOD1To determine if the misfolded SOD1-specific antibodieshave a particular affinity to certain gross perturbationsin SOD1 structure, demetallation, aggregation or reduc-tion of the intra-molecular disulfide bond, recombinantwild-type and various SOD1 mutants (G93A, G85R, andA4V) were spotted onto a nitrocellulose membrane andblotted with misfolded SOD1 antibodies B8H10 andAMF7-63 under native conditions. SOD1 proteins thatare properly folded in a native structure have both cop-per and zinc bound as well as an intact (oxidized) disul-fide bond between Cys57 and Cys146, and are referreda bHomogenate MitochondriaHomogenate Mitochondria  Blot: C4F6Blot: AMF7-63Blot: DSE2 Blot: B8H10 20 KDaActin50 KDa37 KDa10 wks G93A WT early sympHomogenate17 wks VDACSOD110 wks G93A WT early sympMitochondria17 wks hr10 wks G93A WT early symp 17 wks 10 wks G93A WT early symp 17 wks Fig. 5 AMF7-63 and B8H10 antibodies detect aggregated misfolded SOD1 in spinal cords. a Size exclusion filter-trap assay of homogenates orisolated mitochondria from spinal cords of pre-symptomatic (10 week) and early symptomatic SOD1G93A rats or age-matched SOD1WT controls,blotted for misfolded SOD1-specific antibodies B8H10, DSE2-3H1, AMF7-63 and C4F6. Data is representative of three independent trials. b Homogenatesand isolated mitochondria immunoblotted for SOD1 to verify expression levels. Actin and VDAC serve as loading controlsPickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 9 of 14to as holo SOD1. Extended incubation of this protein inambient conditions can generate misfolded or low levelsof aggregated holo protein [37]. Note, as recombinantholo SOD1WT aggregates poorly, buffer was applied tothe membrane at that position (dashed box). Proteinlacking both metal cofactors is referred to as apo SOD1.A reduced apo form of the protein also lacks the crucialCys57-Cys146 disulfide bond. For all mutations, bothAMF7-63 and B8H10 had an increased preference forapo and apo reduced proteins (Fig. 6a). Fully denaturedprotein served as a positive control.To determine if apo SOD1 mutants had a preferentialassociation with isolated mitochondria, we performed anin vitro mitochondrial binding assay. Briefly, using sili-conized tubes, recombinant human SOD1 proteins wereincubated with non-transgenic spinal cord mitochondria,and after washing away unbound protein, mitochondriawere recovered and analyzed by western blot for thepresence of human SOD1. Recombinant SOD1G93A proteinshowed an increased binding to mitochondria compared toSOD1WT protein. Treatment with ethylenediamine tetraa-cetic acid (EDTA), to chelate the metal cofactors of SOD1,resulted in significantly increased binding of SOD1G93A(Fig. 6b, c). SOD1WT displayed a trend toward increasedbinding to mitochondria following treatment with EDTA(Fig. 6b, c). Treatment with hydrogen peroxide, previ-ously published to oxidize SOD1 [7], did not signifi-cantly affect the ability of either recombinant wild-type ormutant SOD1 to associate with mitochondria (Fig. 6b, c).Taken together, misfolded SOD1 antibodies B8H10 andAMF7-63 preferentially detect apo and apo/reduced mis-folded SOD1, and this form of mutant SOD1 has an in-creased association with mitochondria in vitro.DiscussionMisfolded SOD1 specific antibodies recognize distinctnon-native SOD1 conformersIn the literature, there are numerous reports of con-formational antibodies detecting misfolded SOD1 invarious models, tissues and via different methodologiesyielding sometimes contradictory conclusions and/orgeneralizations. We hypothesized that these disparate re-sults could be attributed to differences in the selectivityof these reagents for misfolded SOD1, especially if oneconsiders that “misfolded SOD1” is comprised of morethan one species. Thus, we performed a comprehensivecomparison of six different antibodies in a single geneticrodent model of ALS using multiple approaches. Wefind that misfolded SOD1-specific antibodies partitioninto distinct patterns with A5C3, B8H10, C4F6 andaWT G93A InputEDTAH2O2--+- - +++-- - -VDACSOD1b***Untreated(Holo)EDTA(Apo)H2O2cG93A WT A4V G85R G93A WT native  denaturedHoloHolo IncApo Reduced Apo Blot: AMF7-63 Blot: B8H10 G93A WT A4V G85R G93A WT native  denaturedFig. 6 Misfolded SOD1 specific antibodies show preferential reactivity for demetallated (apo) SOD1. a Recombinant SOD1 proteins (WT, G93A,G85R and A4V) were produced (i) with its full complement of metals (holo); (ii) with its full complement of metals and incubated so as toproduce low levels of aggregated protein (holo inc); (iii) lacking metals (apo); and (iv) lacking metals and a reduced Cys57-Cys146 disulfide bond(apo reduced) were spotted onto nitrocellulose and probed for misfolded SOD1 with AMF7-63 (left) and B8H10 (right). b In vitro mitochondrialbinding assay. Recombinant SOD1WT and SOD1G93A were incubated with spinal cord mitochondria from a non-transgenic rat, washed and subjectedto analysis by western blot. Recombinant SOD1 was either left untreated or incubated with EDTA or H2O2 before addition to mitochondria.c Quantification of b normalized to SOD1G93A binding SOD1WT (white) and SOD1G93A (black). * P < 0.05, ** P < 0.01, n = 5Pickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 10 of 14D3H5 antibodies predominantly labeling misfoldedSOD1 in motor neurons and numerous puncta withinthe neuropil. In contrast, the DSE2-3H1 and AMF7-63antibodies labeled an extensive fibrillar network localizedto motor neuron cell bodies, axons, and dendrites.Fibrils are a subset of aggregates composed of β-sheetsobserved in many neurodegenerative diseases [47, 48].Whether SOD1 forms fibrils in SOD1-mediated FALScases, remains controversial [16, 49]. However, inclu-sions found in the spinal cords of mutant SOD1 animalmodels contain fibrils that stain positive for ThioflavinT, a molecule that fluoresces upon binding to β-sheets[50, 51]. Interestingly, fibrils have the propensity to seedaggregation in vitro [47], and apo reduced wild-type andmutant SOD1 readily form fibrils in vitro [51]. More-over, injection of spinal cord homogenates from miceoverexpressing wild-type or mutant SOD1 into naïveanimal heterozygous YFP-SOD1G85R led to transmissionof motor neuron disease and interestingly, differentabundances and localizations of SOD1 inclusions and fi-brils. These findings suggest that different SOD1 mu-tants or non-native species may differ both in theirability to “seed” further SOD1 aggregates and the prop-erties of such aggregates [29].Misfolded SOD1 conformation-specific antibodies maybe especially useful at detecting distinct non-nativeforms of SOD1 and aid in dissecting which species con-tribute to pathology and potentially help to define themechanisms implicated. Our work finds that A5C3,AMF7-63 and B8H10-misfolded SOD1 localize to mito-chondria whereas as C4F6 does not. Interestingly, al-though C4F6 and B8H10 were raised against the sameimmunogen (full length apo SOD1G93A protein), the lo-cations of the epitopes are distinct. The C4F6 epitope iscentralized around the G93A mutation (encoded in exon4) [52], while the B8H10 epitope has been grosslymapped to the loop region encoded by exon 3 [5]. It isnoteworthy that these two epitopes are located on op-posite sides (~180°) of the three-dimensional structureof the SOD1 protein [5]. That only a subset of neuronscarried both epitopes recognized by B8H10 and AMF7-63 whereas other neurons were labeled with only one ofthese antibodies within the same animals strongly sup-ports that there are indeed multiple non-native mis-folded SOD1 conformers in vivo. Moreover, we clearlydemonstrate that currently available antibodies representpowerful tools differentiating these conformers thatcould be used to address the impact of distinct mis-folded SOD1 conformers on neuronal properties.AMF7-63-reactive misfolded SOD1 correlates withmitochondrial dysfunctionThat several misfolded SOD1-reactive conformers con-verge at the mitochondria highlights mitochondrialdysfunction as an important disease mechanism in ALS.To date, misfolded SOD1 antibodies SEDI [10], DSE2-3H1 [11, 20], A5C3 [11] B8H10 [13] and AMF7-63 (thisreport) detect misfolded SOD1 at the surface of spinalcord mitochondria. Importantly, in the same spinal cord,AMF7-63- and B8H10-reactive misfolded SOD1 con-formers were detected both separately and together ondistinct mitochondrial subpopulations again supportingpotentially distinct impacts of different SOD1 con-formers on mitochondria.AMF7-63+ mitochondria have increased size/volumecompared to B8H10+ mitochondria, and exhibit a trendtoward elevated superoxide production. However, separ-ation into discrete subpopulations, AMF7-63+B8H10−,AMF7-63+B8H10+, or B8H10+AMF7-63− mitochondriayielded no significant differences between the groups interms of mitochondrial size/volume, although AMF7-63+and AMF7-63+B8H10+ showed a trend toward in-creased volume (data not shown). That the AMF7-63+and B8H10+ mitochondrial subpopulations demon-strate differences in mitochondrial size/volume sug-gest that these antibodies recognize distinct misfoldedspecies that potentially elicit disparate degrees ofdamage, with AMF7-63 reactive misfolded SOD1 hav-ing increased potency. The misfolded SOD1 antibodyDSE2-3H1 detects misfolded SOD1 interacting withVoltage-dependent anion channel 1 (VDAC1) [20], amitochondrial outer membrane protein important forion homeostasis [53]. It is reported that recombinantmutant SOD1 inhibits VDAC1 conductance in a reconsti-tuted lipid bilayer [20]. Another group, focused on mutantbut not misfolded SOD1, reports that the interaction ofmutant SOD1 with B-cell lymphoma 2 (Bcl-2) and corre-sponding exposure of the pro-apoptotic BH3 domain isnecessary for Bcl-2 to alter VDAC1 permeability [54]. Ourdata does not address whether misfolded SOD1 (DSE2-3H1 or B8H10-reactive) interacts with Bcl-2. However,B8H10-reactive misfolded SOD1 and the pro-apoptoticform of Bcl-2 preferentially accumulate on the same mito-chondria [13], but this is not indicative of a direct inter-action. Furthermore, a portion of B8H10+ mitochondriaalso contain AMF7-63-reactive SOD1 on their surface.Therefore, DSE2-3H1-reactive SOD1 could have an in-creased association with the pro-apoptotic Bcl-2/VDAC1complex, resulting in altered mitochondrial ion homeosta-sis. Future knowledge of the interactome of each mis-folded SOD1 conformer may provide insight into thepossible differences in toxicity elicited by AMF7-63 andB8H10-reactive misfolded SOD1.We speculated that DSE2-3H1/AMF7-63-reactive mis-folded SOD1 may be prone to aggregation, as fibrils arecomposed of insoluble, ordered oligomeric chains [55].However, both B8H10 and AMF7-63 (and DSE2-3H1)labeled aggregates in spinal cord homogenates andPickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 11 of 14isolated mitochondria. Therefore, the increases in mito-chondrial size/volume elicited by AMF7-63-reactive mis-folded SOD1 cannot be due solely to its participation inaggregate formation at the mitochondrial surface. We can-not exclude the possibility that AMF7-63-reactive mis-folded SOD1 is included in aggregates of differing size/properties compared to the B8H10-reactive conformer orthat the solubility of these two forms of misfolded SOD1may differ so as to account for the increased toxicity. C4F6-reactive misfolded SOD1 is not detected in aggregates bythis assay, consistent with reports that this antibody recog-nizes a soluble form of misfolded SOD1 [31, 56].There is considerable debate over whether SOD1monomers [17], oligomers [57] or large aggregates [58]mediate toxicity. A caveat to these studies is they havefocused on cytosolic SOD1. Mitochondria are vulnerableto proteotoxic stress [59], particularly aggregated pro-teins [60] and thus, have developed multiple layers ofquality control mechanisms to combat this form ofstress [61]. Mutant SOD1 has been reported to form ag-gregates in the matrix of brain mitochondria from ALSanimal models [62] and at the surface of mitochondriaof cells over-expressing mutant SOD1 [63]. Whetherthese internal- or surface-localized aggregates containmisfolded SOD1 or cause mitochondrial dysfunction wasnot determined. However, several recent studies suggestthat aggregated SOD1 can perturb mitochondrial mem-brane integrity in vitro [64, 65]. Our results highlightthat multiple misfolded SOD1 conformational antibodiesdetect misfolded protein, some of which is found in anaggregated form, at the surface of mitochondria. Fur-thermore, the presence of misfolded SOD1 coincideswith disruptions in mitochondrial volume and super-oxide production, reinforcing that mitochondria are abona fide target of SOD1 toxicity.Demetallated SOD1 is preferentially detected bymisfolded SOD1-specific antibodies AMF7-63 and B8H10Although broadly considered as a cytosolic protein, asmall portion of SOD1 is localized to the mitochondrialintermembrane space (IMS) in normal physiologicalconditions [66]. In order for SOD1 to be imported intomitochondria, it must be in its apo reduced form [67].Given this, a pool of apo SOD1 at the mitochondrialsurface is expected. Interestingly, in our in vitro mito-chondrial binding assay, apo SOD1 readily associatedwith the outer mitochondrial membrane. Import ofmitochondrial substrates is slowed in spinal cord mito-chondria from SOD1G93A [27], and the regulation ofmutant SOD1s import into mitochondria is altered [68],therefore apo mutant SOD1 en route to the IMS may beaccumulating at the outer mitochondrial membrane anddisturbing normal mitochondrial physiology. BothAMF7-63 and B8H10 detected recombinant apo andapo reduced SOD1 more readily than recombinant holoSOD1.ConclusionsConformational antibodies targeted to misfolded SOD1show promise not only as therapeutics for ALS, but alsoas valuable tools with which to probe the mechanisms ofmisfolded SOD1 toxicity. These antibodies have revealedthat multiple non-native species of misfolded SOD1 existto contribute to motor neuron degeneration, possibly viadistinct mechanisms [31, 69]. Our study further supportsthis premise and highlights that variable potency/toxicityof different SOD1 species is possible even when onlyone SOD1 mutation is present (Fig. 7). Furthermore, weidentify the mitochondria as a target of several of thesemisfolded SOD1 conformers. This finding may haveB8H10D3H5 C4F6 AMF7-63A5C31 153aaExon 1 Exon 2 Exon 3 Exon 5Exon 4DSE2-3H1Neurons FibrilsAggregatesAssociation Aggregates FunctionVolumeSuperoxide levels  ++n.dn.dn.dn.d+++n.d+n.dn.d++---n.d+++++++n.d+++++++++Spinal cord Mitochondria ++++++ +Fig. 7 Summary of misfolded SOD1 antibody characteristics. Attributes of various misfolded SOD1 antibodies in spinal cords (presence inneurons, fibrils and aggregates) and isolated spinal cord mitochondria (outer mitochondrial membrane association, presence in aggregates,correlation with damage). Epitopes to misfolded SOD1 antibodies used in this study are grossly mapped to the encoding regions. +, positivefinding; −, negative finding; n.d, not determinedPickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 12 of 14profound implications for therapeutics aimed at neutral-izing misfolded SOD1.Additional filesAdditional file 1: Figure S1. Misfolded SOD1 specific antibodies donot label SOD1WT. A) Lumbar spinal cord sections of a symptomaticSOD1G93A rat and age-matched SOD1WT were labeled with misfoldedSOD1 specific antibodies A5C3, B8H10, C4F6, D3H5, DSE2-3H1, AMF7-63and SEDI (green). B) No non-specific labeling as determined by IgGcontrols (mouse and rabbit), or incubation with secondary antibodyalone, was detected. (PPTX 2948 kb)Additional file 2: Figure S2. Misfolded SOD1 antibody AMF7-63specifically identifies mutant SOD1 in spinal cord but not liver fromSOD1G93A rats. The capacity for AMF7-63 to detect misfolded SOD1 inhomogenates or isolated mitochondria from spinal cords and livers wasassayed by immunoprecipitation. Rabbit IgG (IgG) serves as control. Inputis 10 μg of homogenate or isolated mitochondria. From top to bottombands correspond to non-specific (ns), human (hSOD1) and rat (rSOD1)SOD1. (PPTX 562 kb)Additional file 3: Figure S3. Misfolded SOD1 antibody B8H10specifically identifies misfolded SOD1 on the surface of isolatedmitochondria from spinal cord but not liver of SOD1G93A rats. A)Immunolabeling of isolated spinal cord and liver mitochondria withmisfolded SOD1 antibody B8H10 from symptomatic SOD1G93A rats andcontrols (age-matched SOD1WT and non-transgenic rats) by flow cytometry.Misfolded SOD1 positive labeling is determined by comparing to isotypecontrol (mouse IgG1) of SOD1G93A sample. Percentage of misfolded SOD1+events is shown for each tissue and genotype in a representative sample. C)Quantification of B8H10+ events in spinal cord (blue circles) or liver (blacksquares) of symptomatic SOD1G93A rats, and age- matched SOD1WT andnon-transgenic rats. Data is represented as percent of misfolded B8H10+mitochondria (mean ± SEM). *** P < 0.001, n = 3 animals per genotype pertissue. (PPTX 253 kb)AbbreviationsALS: Amyotrophic Lateral Sclerosis; Bcl-2: B-cell lymphoma 2; ChAT: cholineacetyltransferase; EDTA: ethylenediamine tetraacetic acid; FALS: FamilialAmyotrophic Lateral Sclerosis; FSC: forward side scatter; IMS: intermembranespace; MAP2: microtubule associated protein 2; MFI: mean fluorescenceintensity; MIF: macrophage inhibitory factor; MTG: Mitotracker Green;SALS: Sporadic Amyotrophic Lateral Sclerosis; SOD1: superoxide dismutase 1;VDAC1: voltage-dependent anion channel.Competing interestsThe authors declare that they have no competing interests. NRC is ChiefScientific Officer of ProMIS Neurosciences.Authors’ contributionsSP, SS, HRB, LL and LD performed research; SP, NA, EM and CVV analyzed thedata; EM and NRC supplied reagents; SP and CVV designed the research andwrote the manuscript. All authors approved the final manuscript.AcknowledgementsWe thank L. Hayward, J.P. Julien, and J. Robertson for sharing of reagents,the CRCHUM cytometry and cell imaging core facilities, M. O’Neill and S.Boillée for helpful comments, S.L. Peyrard for help with animal husbandry,and G.A. Rouleau for contributing to baculovirus protein production. Thiswork was supported by the Canadian Foundation for Innovation, MuscularDystrophy Association, ALS Society of Canada, Brain Canada, and the FrickFoundation for ALS Research (CVV). CVV and NA are Canadian Institutes ofHealth Research New Investigators. SP was partially supported by the TimNoël Studentship from the ALS Society of Canada. LL holds a studentshipfrom the Multiple Sclerosis Society of Canada. Funding bodies had no inputin the design of study, nor collection, analysis or interpretation of the data.Author details1Centre de recherche du Centre Hospitalier de l’Université de Montréal(CRCHUM) Université de Montréal, 900 rue Saint-Denis, Local R09.442,Montréal, QC H2X 0A9, Canada. 2Department of Biochemistry, Université deMontréal, Montréal, QC H2X 0A9, Canada. 3Department of Neurosciences,Université de Montréal, Montréal, QC H2X 0A9, Canada. 4Integrated Programin Neuroscience, McGill University, Montréal, QC H3A 2B4, Canada.5Department of Biochemistry, University of Waterloo, Waterloo, ON N2L 3G1,Canada. 6Department of Medicine (Neurology), University of British Columbiaand Vancouver Coastal Health Research Institute, Brain Research Centre,Vancouver, BC V6T 2B5, Canada.Received: 23 February 2016 Accepted: 13 April 2016References1. Harms MB, Baloh RH. Clinical neurogenetics: amyotrophic lateral sclerosis.Neurol Clin. 2013;31(4):929–50.2. Rosen DR, et al. Mutations in Cu/Zn superoxide dismutase gene are associatedwith familial amyotrophic lateral sclerosis. Nature. 1993;362(6415):59–62.3. Renton AE, Chio A, Traynor BJ. State of play in amyotrophic lateral sclerosisgenetics. Nat Neurosci. 2014;17(1):17–23.4. Gurney ME, et al. Motor neuron degeneration in mice that expressa human Cu, Zn superoxide dismutase mutation. Science.1994;264(5166):1772–5.5. Pickles S, Vande Velde C. Misfolded SOD1 and ALS: zeroing in onmitochondria. Amyotroph Lateral Scler. 2012;13(4):333–40.6. Rotunno MS, Bosco DA. An emerging role for misfolded wild-type SOD1 insporadic ALS pathogenesis. Front Cell Neurosci. 2013;7:253.7. Bosco DA, et al. Wild-type and mutant SOD1 share an aberrantconformation and a common pathogenic pathway in ALS. Nat Neurosci.2010;13(11):1396–403.8. Gros-Louis F, et al. Intracerebroventricular infusion of monoclonal antibodyor its derived Fab fragment against misfolded forms of SOD1 mutant delaysmortality in a mouse model of ALS. J Neurochem. 2010;113(5):1188–99.9. Forsberg K, et al. Novel antibodies reveal inclusions containing non-nativeSOD1 in sporadic ALS patients. PLoS One. 2010;5(7), e11552.10. Rakhit R, et al. An immunological epitope selective for pathologicalmonomer-misfolded SOD1 in ALS. Nat Med. 2007;13(6):754–9.11. Vande Velde C, et al. Selective association of misfolded ALS-linked mutantSOD1 with the cytoplasmic face of mitochondria. Proc Natl Acad Sci U S A.2008;105(10):4022–7.12. Saxena S, et al. Neuroprotection through excitability and mTORrequired in ALS motoneurons to delay disease and extend survival.Neuron. 2013;80(1):80–96.13. Pickles S, et al. Mitochondrial damage revealed by immunoselection forALS-linked misfolded SOD1. Hum Mol Genet. 2013;22(19):3947–59.14. Vande Velde C, et al. Misfolded SOD1 associated with motor neuronmitochondria alters mitochondrial shape and distribution prior to clinicalonset. PLoS One. 2011;6(7):11.15. Ayers JI, et al. Conformational specificity of the C4F6 SOD1 antibody; lowfrequency of reactivity in sporadic ALS cases. Acta Neuropathol Commun.2014;2:55.16. Kerman A, et al. Amyotrophic lateral sclerosis is a non-amyloid disease inwhich extensive misfolding of SOD1 is unique to the familial form. ActaNeuropathol. 2010;119(3):335–44.17. Liu HN, et al. Targeting of monomer/misfolded SOD1 as a therapeuticstrategy for amyotrophic lateral sclerosis. J Neurosci. 2012;32(26):8791–9.18. Fujisawa T, et al. A novel monoclonal antibody reveals a conformationalalteration shared by amyotrophic lateral sclerosis-linked SOD1 mutants. AnnNeurol. 2012;72(5):739–49.19. Ruegsegger C, et al. Aberrant association of misfolded SOD1 with Na/KATPase-alpha3 impairs its activity and contributes to motor neuronvulnerability in ALS. Acta Neuropathol. 2016;131(3):427–51.20. Israelson A, et al. Misfolded mutant SOD1 directly inhibits VDAC1 conductancein a mouse model of inherited ALS. Neuron. 2010;67(4):575–87.21. Dal Canto MC, Gurney ME. Development of central nervous systempathology in a murine transgenic model of human amyotrophic lateralsclerosis. Am J Pathol. 1994;145(6):1271–9.22. Higgins CM, Jung C, Xu Z. ALS-associated mutant SOD1G93A causesmitochondrial vacuolation by expansion of the intermembrane space andPickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 13 of 14by involvement of SOD1 aggregation and peroxisomes. BMC Neurosci.2003;4:16.23. Kong J, Xu Z. Massive mitochondrial degeneration in motor neuronstriggers the onset of amyotrophic lateral sclerosis in mice expressing amutant SOD1. J Neurosci. 1998;18(9):3241–50.24. Mattiazzi M, et al. Mutated human SOD1 causes dysfunction of oxidativephosphorylation in mitochondria of transgenic mice. J Biol Chem.2002;277(33):29626–33.25. Tradewell ML, et al. Calcium dysregulation, mitochondrial pathology andprotein aggregation in a culture model of amyotrophic lateral sclerosis:mechanistic relationship and differential sensitivity to intervention.Neurobiol Dis. 2011;42(3):265–75.26. Magrane J, et al. Abnormal mitochondrial transport and morphology arecommon pathological denominators in SOD1 and TDP43 ALS mousemodels. Hum Mol Genet. 2014;23(6):1413–24.27. Li Q, et al. ALS-linked mutant superoxide dismutase 1 (SOD1) altersmitochondrial protein composition and decreases protein import. Proc NatlAcad Sci U S A. 2010;107(49):21146–51.28. Israelson A, et al. Macrophage migration inhibitory factor as a chaperoneinhibiting accumulation of misfolded SOD1. Neuron. 2015.29. Ayers JI, et al. Experimental transmissibility of mutant SOD1 motor neurondisease. Acta Neuropathol. 2014;128(6):791–803.30. Bergh J, et al. Structural and kinetic analysis of protein-aggregate strains invivo using binary epitope mapping. Proc Natl Acad Sci U S A. 2015;112(14):4489–94.31. Prudencio M, Borchelt DR. Superoxide dismutase 1 encoding mutationslinked to ALS adopts a spectrum of misfolded states. Mol Neurodegener.2011;6:77.32. Howland DS, et al. Focal loss of the glutamate transporter EAAT2 in atransgenic rat model of SOD1 mutant-mediated amyotrophic lateralsclerosis (ALS). Proc Natl Acad Sci U S A. 2002;99(3):1604–9.33. Chan PH, et al. Overexpression of SOD1 in transgenic rats protectsvulnerable neurons against ischemic damage after global cerebral ischemiaand reperfusion. J Neurosci. 1998;18(20):8292–9.34. Pickles S, Arbour N, Vande Velde C. Immunodetection of outer membraneproteins by flow cytometry of isolated mitochondria. J Vis Exp. 2014;91:51887.35. Vassall KA, et al. Equilibrium thermodynamic analysis of amyotrophic lateralsclerosis-associated mutant apo Cu, Zn superoxide dismutases.Biochemistry. 2006;45(23):7366–79.36. Vassall KA, et al. Decreased stability and increased formation of solubleaggregates by immature superoxide dismutase do not account for diseaseseverity in ALS. Proc Natl Acad Sci U S A. 2011;108(6):2210–5.37. Hwang YM, et al. Nonamyloid aggregates arising from mature copper/zincsuperoxide dismutases resemble those observed in amyotrophic lateralsclerosis. J Biol Chem. 2010;285(53):41701–11.38. Hayward LJ, et al. Decreased metallation and activity in subsets of mutantsuperoxide dismutases associated with familial amyotrophic lateral sclerosis.J Biol Chem. 2002;277(18):15923–31.39. Mullaney PF, Dean PN. Cell sizing: a small-angle light-scattering method forsizing particles of low relative refractive index. Appl Opt. 1969;8(11):2361–2.40. Chen Y, Dorn 2nd GW. PINK1-phosphorylated mitofusin 2 is a Parkinreceptor for culling damaged mitochondria. Science. 2013;340(6131):471–5.41. Metivier D, et al. Cytofluorometric detection of mitochondrial alterations inearly CD95/Fas/APO-1-triggered apoptosis of Jurkat T lymphoma cells.Comparison of seven mitochondrion-specific fluorochromes. Immunol Lett.1998;61(2–3):157–63.42. Xu X, Arriaga EA. Qualitative determination of superoxide release at bothsides of the mitochondrial inner membrane by capillary electrophoreticanalysis of the oxidation products of triphenylphosphonium hydroethidine.Free Radic Biol Med. 2009;46(7):905–13.43. Mukhopadhyay P, et al. Simple quantitative detection of mitochondrialsuperoxide production in live cells. Biochem Biophys Res Commun. 2007;358(1):203–8.44. Robinson KM, et al. Selective fluorescent imaging of superoxide in vivo usingethidium-based probes. Proc Natl Acad Sci U S A. 2006;103(41):15038–43.45. Wang J, Xu G, Borchelt DR. High molecular weight complexes of mutantsuperoxide dismutase 1: age-dependent and tissue-specific accumulation.Neurobiol Dis. 2002;9(2):139–48.46. Brotherton TE, et al. Localization of a toxic form of superoxide dismutase 1protein to pathologically affected tissues in familial ALS. Proc Natl Acad SciU S A. 2012;109(14):5505–10.47. Westermark P. Aspects on human amyloid forms and their fibrilpolypeptides. FEBS J. 2005;272(23):5942–9.48. Zerovnik E, et al. Mechanisms of amyloid fibril formation–focus on domain-swapping. FEBS J. 2011;278(13):2263–82.49. Kato S, et al. New consensus research on neuropathological aspects offamilial amyotrophic lateral sclerosis with superoxide dismutase 1 (SOD1)gene mutations: inclusions containing SOD1 in neurons and astrocytes.Amyotroph Lateral Scler Other Motor Neuron Disord. 2000;1(3):163–84.50. Wang J, et al. Fibrillar inclusions and motor neuron degeneration intransgenic mice expressing superoxide dismutase 1 with a disruptedcopper-binding site. Neurobiol Dis. 2002;10(2):128–38.51. Furukawa Y, et al. Complete loss of post-translational modifications triggersfibrillar aggregation of SOD1 in the familial form of amyotrophic lateralsclerosis. J Biol Chem. 2008;283(35):24167–76.52. Rotunno MS, et al. Identification of a misfolded region in superoxidedismutase 1 that is exposed in amyotrophic lateral sclerosis. J Biol Chem.2014;289(41):28527–38.53. Shoshan-Barmatz V, Golan M. Mitochondrial VDAC1: function in cell life anddeath and a target for cancer therapy. Curr Med Chem. 2012;19(5):714–35.54. Tan W, et al. Small peptides against the mutant SOD1/Bcl-2 toxicmitochondrial complex restore mitochondrial function and cell viability inmutant SOD1-mediated ALS. J Neurosci. 2013;33(28):11588–98.55. Rambaran RN, Serpell LC. Amyloid fibrils: abnormal protein assembly. Prion.2008;2(3):112–7.56. Brotherton TE, Li Y, Glass JD. Cellular toxicity of mutant SOD1 protein islinked to an easily soluble, non-aggregated form in vitro. Neurobiol Dis.2013;49:49–56.57. Redler RL, et al. Non-native soluble oligomers of Cu/Zn superoxidedismutase (SOD1) contain a conformational epitope linked to cytotoxicity inamyotrophic lateral sclerosis (ALS). Biochemistry. 2014;53(14):2423–32.58. Matsumoto G, et al. Structural properties and neuronal toxicity ofamyotrophic lateral sclerosis-associated Cu/Zn superoxide dismutase 1aggregates. J Cell Biol. 2005;171(1):75–85.59. Martinelli P, Rugarli EI. Emerging roles of mitochondrial proteases inneurodegeneration. Biochim Biophys Acta. 2010;1797(1):1–10.60. Hashimoto M, et al. Role of protein aggregation in mitochondrialdysfunction and neurodegeneration in Alzheimer’s and Parkinson’s diseases.Neuromolecular Med. 2003;4(1–2):21–36.61. Baker MJ, Tatsuta T, Langer T. Quality control of mitochondrial proteostasis.Cold Spring Harb Perspect Biol. 2011;3(7).62. Vijayvergiya C, et al. Mutant superoxide dismutase 1 forms aggregates inthe brain mitochondrial matrix of amyotrophic lateral sclerosis mice. JNeurosci. 2005;25(10):2463–70.63. Kawamata H, et al. Lysyl-tRNA synthetase is a target for mutant SOD1toxicity in mitochondria. J Biol Chem. 2008;283(42):28321–8.64. Oladzad Abbasabadi A, et al. Disruption of mitochondrial membraneintegrity induced by amyloid aggregates arising from variants of SOD1. Int JBiol Macromol. 2013;61:212–7.65. Salehi M, et al. Mitochondrial membrane disruption by aggregationproducts of ALS-causing superoxide dismutase-1 mutants. Int J BiolMacromol. 2015;75c:290–7.66. Weisiger RA, Fridovich I. Mitochondrial superoxide simutase. Site of synthesisand intramitochondrial localization. J Biol Chem. 1973;248(13):4793–6.67. Field LS, et al. Factors controlling the uptake of yeast copper/zincsuperoxide dismutase into mitochondria. J Biol Chem. 2003;278(30):28052–9.68. Kawamata H, Manfredi G. Different regulation of wild-type and mutant Cu,Zn superoxide dismutase localization in mammalian mitochondria. HumMol Genet. 2008;17(21):3303–17.69. Pickles S, Vande Velde C. Misfolded SOD1 and ALS: zeroing in onmitochondria. Amyotroph Lateral Scler. 2012.Pickles et al. Acta Neuropathologica Communications  (2016) 4:43 Page 14 of 14


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