Immunochemical characterization on pathological oligomers of mutant Cu/Zn-superoxide dismutase in amyotrophic lateral sclerosis
- 1.6k Downloads
Dominant mutations in Cu/Zn-superoxide dismutase (SOD1) gene cause a familial form of amyotrophic lateral sclerosis (SOD1-ALS) with accumulation of misfolded SOD1 proteins as intracellular inclusions in spinal motor neurons. Oligomerization of SOD1 via abnormal disulfide crosslinks has been proposed as one of the misfolding pathways occurring in mutant SOD1; however, the pathological relevance of such oligomerization in the SOD1-ALS cases still remains obscure.
We prepared antibodies exclusively recognizing the SOD1 oligomers cross-linked via disulfide bonds in vitro. By using those antibodies, immunohistochemical examination and ELISA were mainly performed on the tissue samples of transgenic mice expressing mutant SOD1 proteins and also of human SOD1-ALS cases.
We showed the recognition specificity of our antibodies exclusively toward the disulfide-crosslinked SOD1 oligomers by ELISA using various forms of purified SOD1 proteins in conformationally distinct states in vitro. Furthermore, the epitope of those antibodies was buried and inaccessible in the natively folded structure of SOD1. The antibodies were then found to specifically detect the pathological SOD1 species in the spinal motor neurons of the SOD1-ALS patients as well as the transgenic model mice.
Our findings here suggest that the SOD1 oligomerization through the disulfide-crosslinking associates with exposure of the SOD1 structural interior and is a pathological process occurring in the SOD1-ALS cases.
KeywordsAmyotrophic lateral sclerosis Cu/Zn-superoxide dismutase Protein misfolding Disulfide bond
Amyotrophic lateral sclerosis
- G1H mouse
A transgenic mouse carrying human SOD1 gene with G93A mutation
- loxG37R mouse
A transgenic mouse carrying SOD1 gene with G37R mutation
ALS with mutations in SOD1 gene
- WT mouse
A transgenic mouse carrying human wild-type SOD1 gene.
Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease, which associates with loss of motor neurons in the affected nervous tissues including motor cortex, brainstem, and spinal cords . After several years of disease onset, significant weakness of muscles is usually followed by death due to the failure in respiratory system. While most of the ALS cases are sporadic, dominant mutations in Cu/Zn-superoxide dismutase (SOD1) gene have been shown to cause familial forms of ALS (SOD1-ALS) . More than 150 types of pathogenic mutations in SOD1 gene have been identified , but importantly, no ALS-like phenotypes were confirmed in SOD1-knockout mice . SOD1 is hence considered to gain toxic properties by pathogenic mutations. A common pathological hallmark in SOD1-ALS cases is the abnormal accumulation of mutant SOD1 proteins in motor neurons of affected nervous tissues . Pathogenic mutations have hence been proposed to facilitate “misfolding” of SOD1 into abnormal conformation(s) and thereby exert toxicities causing the disease.
SOD1 is a homodimeric metalloprotein that binds copper and zinc ions and also forms an intramolecular disulfide bond . A folded conformation of enzymatically active SOD1 is significantly stabilized through the metal binding and the disulfide formation . Indeed, dissociation of metal ions and reduction of the disulfide bond are known to decrease the conformational stability of SOD1 and thereby facilitate its misfolding in vitro; for example, demetallated (apo) SOD1 forms cross-linked oligomers through the shuffling of the disulfide bond , and further reduction of the disulfide bond in apo-SOD1 leads to the formation of the amyloid-like fibrillar aggregates . Abnormal SOD1 trimers have also been recently shown to form in vitro at acidic pH and exhibit toxicities toward cultured cells . Moreover, the structural dynamics of immature SOD1 has been extensively characterized in the atomic level [11, 12, 13]. Increasing numbers of recent in vitro studies have revealed various misfolding pathways of SOD1 proteins; however, it still remains obscure how SOD1 changes its conformation under the pathological conditions in vivo.
Actually, quite limited information is available on the biochemical/structural properties of pathological SOD1 species in human SOD1-ALS cases, partly because most of the motor neurons, which are the most affected cell types in ALS, are usually lost at autopsies. Therefore, any changes of SOD1 occurring specifically in affected motor neurons could not become evident in the biochemical experiments using the homogenates of spinal cords. Nonetheless, in transgenic ALS-model mice overexpressing mutant SOD1, the enzymatic activation of SOD1 has been shown to be retarded in spinal cords but not in the control tissues such as kidney and liver . In other words, immature forms of SOD1 are expected to accumulate specifically in the spinal cord as misfolded proteins. Actually, the amyloid-like aggregates, which are composed of SOD1 lacking the disulfide bond, have been detected in the spinal motor neurons of transgenic ALS-model mice , while there has been no evidence to support the formation of amyloid-like aggregates in human SOD1-ALS cases . Alternatively, we have previously detected the disulfide-crosslinked SOD1 oligomers in the spinal cord but not in the liver of ALS-model mice . Pathological roles of the disulfide-crosslinked oligomers have been examined mainly in cultured cells [18, 19] but still remain less characterized in the transgenic mice and also in human SOD1-ALS cases.
In this study, we prepared and characterized the antibodies recognizing the disulfide-crosslinked SOD1 oligomers in vitro to test their pathological relevance in SOD1-ALS. Compared to the previous studies, we have more extensively examined the reactivity of our antibodies against purified SOD1 proteins in various metallation/disulfide states and ensured the recognition specificities of our antibodies to the disulfide-crosslinked SOD1 oligomers. In the ALS-model mice, the immunoreactivities with our antibodies were evident from their pre-symptomatic stage and also specifically in their spinal cords. More importantly, the immunoreactivities with our antibodies were detected in spinal motor neurons of the human SOD1-ALS cases. We thus propose that the disulfide-crosslinked SOD1 oligomers possess an immunological epitope selective for the SOD1 species with abnormal conformations occurring in the pathological conditions.
Protein preparation and purification
Introduction of mutations was performed by an Inverse PCR method using a KOD-FX-neo DNA polymerase (TOYOBO) and confirmed by DNA sequencing. Escherichia coli SHuffle™ (NEB) was transformed with a pET-15b plasmid (Novagen) containing cDNA of human SOD1, and the protein expression was induced in the shaking culture with 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) at 20 °C for 20 h. Cells were lysed with ultrasonication in PBS containing 2% Triton X-100, DNase I, and MgSO4, and the supernatant after centrifugation at 20,000 x g for 15 min. was loaded on a HisTrap HP column (1 mL, GE Healthcare). SOD1 proteins were eluted with a buffer containing 50 mM sodium phosphate (Na-Pi), 100 mM NaCl, and 250 mM imidazole at pH 7.0. Metal ions bound to SOD1 proteins were removed by two-step dialysis first against a buffer containing 50 mM sodium acetate, 100 mM NaCl, and 10 mM EDTA at pH 4.0 at 4 °C for 16 h and then against a buffer containing 100 mM Na-Pi, 100 mM NaCl, and 5 mM EDTA at pH 7.4 (called NNE buffer). The proteins were treated with thrombin (GE Healthcare) to remove an N-terminal His-tag and further purified by size-exclusion chromatography using a Cosmosil 5Diol-300-II column (nacalai tesque).
For the epitope mapping of antibodies, we have prepared the following eight peptides as a fusion protein with glutathione-S-transferase, which was further N-terminally tagged with a 6 x His tag: Ala 1 – Lys 23 (Pepexon1), Glu 24 – Ala 55 (Pepexon2), Gly 56 – Arg 79 (Pepexon3), His 80 – Val 118 (Pepexon4), Val 119 – Gln 153 (Pepexon5), Glu 24 – His 43 (Pep1), Ser 34 – Asn 53 (Pep2), and Gly 44 – His 63 (Pep3). All of those fusion proteins were overexpressed in E. coli BL21(DE3) with shaking at 20 °C for 20 h in the presence of 0.1 mM IPTG. As described above, the cells were lysed, and the fusion proteins were purified from the soluble supernatant with a HisTrap HP column.
Preparation and purification of anti-SOD1olig antibody
Demetallated SOD1 with A4V mutation as purified above (5 mg/mL) was incubated in the NNE buffer at 37 °C for five days, by which soluble SOD1(A4V) oligomers were prepared. Those SOD1(A4V) oligomers were emulsified with either complete Freund's adjuvant (DIFCO) in the initial injection or incomplete Freund's adjuvant and injected subcutaneously into a female New Zealand White rabbit at intervals of 2 – 4 weeks. Antisera were sampled at two weeks after the fifth or sixth injection, and immunoglobulins specific to antigen were affinity-purified using CNBr-activated Sepharose 4B (GE Healthcare) conjugated with the SOD1(A4V) oligomers.
To isolate the antibodies recognizing SOD1 oligomers but not the folded proteins, the affinity-purified immunoglobulins were washed with Ni2+-affinity resins that bind His-tagged wild-type SOD1S-S proteins (SOD1-resins). SOD1-resins were prepared by adding 100 μL of His-SELECT nickel affinity gel (Sigma) to 500 μL of 200 μM His-tagged wild-type SOD1S-S in a buffer containing 50 mM Tris and 100 mM NaCl at pH 7.4 and incubated at 4 °C for an hour. The resins were washed with PBS and then incubated with the affinity-purified immunoglobulins in PBS with rotation at 4 °C for an hour. The resins were spun down, and again, the freshly prepared SOD1-resins were added to the supernatant and rotated at 4 °C for an hour. After repeating this absorption procedure four times, Ni2+-affinity resins were added to the supernatant in order to remove His-tagged SOD1 proteins detached from the SOD1-resins. Concentrations of purified antibodies were then determined by Micro BCA Protein Assay kit (Thermo).
Preparation and purification of anti-SOD1int antibody
Production of a polyclonal antibody to a peptide of SOD1 (Gly 44 – Asn 53) was performed by Eurofins Genomics. Briefly, the peptide, H2N-CG44FHVHEFGDN53-COOH, was conjugated through its N-terminal Cys with keyhole limpet hemocyanin, with which a rabbit was immunized in the 42-day protocol. The sera were then purified using a Sulfo-Link Coupling Resin (Thermo) with the peptide, Gly 44 – Asn 53, Gly 44 – Glu 49, His 46 – Gly 51, or His 48 – Asn 53, by which anti-SOD144–53, anti-SOD144–49, anti-SOD146–51, or anti-SOD148–53 antibody was purified, respectively. All of the peptides have an additional Cys residue at the N-terminus for its conjugation with the resin. For preparation of anti-SOD1int antibody, anti-SOD144–53 antibody was first loaded on a Sulfo-Link Coupling Resin (Thermo) cross-linked with purified apo-SOD1S-S proteins, and the flow-through fraction was collected, concentrated, and then purified using a Sulfo-Link resin conjugated with a His 48 – Asn 53 peptide. Concentrations of purified antibodies were determined by Micro BCA Protein Assay kit (Thermo).
Enzyme-linked immunosorbent assay (ELISA)
To prevent adventitious binding of contaminant metal ions to SOD1 proteins in ELISA, we used Tris-buffered saline (TBS) that was treated with Chelex® 100 Resin (Bio-Rad). For the assay of E,E-SOD1 proteins, a strong chelator for divalent metal ions, EDTA (5 mM), was further included in TBS, by which an artificial supply of divalent metal ions (zinc ions, in particular) from buffers could be prevented. SOD1 variants with distinct metallation and thiol-disulfide status (5 μg/well) were coated on 96-well plates (Nunc-Immuno™ Plate CII, Thermo) overnight at 4 °C. After three washes with TBS containing 0.05% (v/v) Tween 20 (TBS-T), the plates were blocked with TBS containing 0.5% (w/v) BSA for an hour at room temperature. After six washes with TBS-T, either antibody purified in this study, polyclonal anti-human SOD1 (FL-154, Santa Cruz Biotechnology), USOD (#SPC-205, StressMarq Bioscience), or SEDI (#SPC-206, StressMarq Bioscience) antibody was added as a primary antibody (0.2 μg/mL) and incubated for an hour at room temperature, which was then followed by secondary antibody with horseradish peroxidase (goat anti-rabbit IgG, 1:1,000; Thermo Scientific) for an hour at room temperature. As the substrate solution, O-phenylenediamine and 0.012% H2O2 in a buffer containing 100 mM sodium citrate at pH 5.0 were used. The absorbance was read at 490 nm using a plate reader (Epoch, BioTek).
For sandwich ELISA, a plate (Nunc-Immuno™ Plate CII, Thermo) was coated with the capture antibodies (0.2 μg/mL anti-SOD1int or 0.02 μg/mL anti-SOD1 (FL-154, Santa Cruz Biotechnology) antibodies) overnight at 4 °C and blocked with 1% BSA for an hour. Soluble extracts of the tissue samples containing 10 μg of total proteins were then applied and incubated at room temperature for an hour. The captured SOD1 proteins were detected by sheep anti-SOD1 (1:2,000, Calbiochem) and HRP-conjugated rabbit anti-sheep (1:1,000, Bio-Rad) antibodies as the detection and secondary antibodies, respectively. As the substrate solution, O-phenylenediamine and 0.012% H2O2 in a buffer containing 100 mM sodium citrate at pH 5.0 were used. The absorbance was read at 490 nm using a plate reader (Epoch, BioTek).
Transgenic mice carrying human SOD1 gene with G93A mutation (B6.Cg-Tg(SOD1*G93A)1Gur/J in a C57BL/6 background) and human wild-type SOD1 gene (B6.Cg-Tg(SOD1)2Gur) were purchased from Jackson Laboratory (Bar Harbor, ME) and maintained heterozygous with a C57BL/6 background. Mice expressing human SOD1 with G37R were described previously . Mice were genotyped for human SOD1 using tail DNA as described previously . All experiments were reviewed and approved by the Animal Use and Care Committees of Keio University and Nagoya University, and care was taken to minimize suffering and limit the number of animals used.
Mice were deeply anesthetized with sodium pentobarbital and then perfused via the aortic cone with PBS, followed by 4% paraformaldehyde in a buffer containing 0.1 M Na-Pi at pH 7.4. The lumbar region of each spinal cord (ca. 2 cm) was removed and post-fixed in the same fixative overnight at 4 °C, after which it was immersed in 20% sucrose in 0.1 M Na-Pi, pH 7.4, overnight at 4 °C. The tissue was then frozen in OCT compound (Sakura Finetek) and sectioned at 40 μm on a cryostat. Rabbit anti-SOD1olig (0.02 μg/mL) and mouse monoclonal anti-human SOD1 (0.02 μg/mL, clone 1G2, MBL) antibodies were used for immunohistochemistry as a primary antibody, and biotinylated anti-mouse IgG (H + L) (1:200 dilution, Vector Laboratories, Inc.) was used as a secondary antibody. The immunoreaction was amplified using the VECTASTAIN ABC HRP Kit (Vector Laboratories, Inc.) according to the manufacturer’s direction. The free-floating sections were processed using diaminobenzidine (DAB) as the chromogen followed by counter-staining with hematoxylin . Stained sections were then examined using a microscope (BX51, Olympus).
The human cases examined in this study included three SOD1-ALS cases with C111Y mutation, four sporadic ALS cases with TDP-43-positive inclusions (negative for SOD1 mutations), and three non-ALS controls. All tissues from ALS patients and non-ALS controls were obtained by autopsy with informed consent at Matsumoto Medical Center in Japan, and information on the cases was summarized in Additional file 1: Table S1. The collection of tissues and their use in this study were approved by the institutional review board for research ethics of Matsumoto Medical Center and Keio University, Japan.
For immunohistochemical examination, the spinal cord was fixed in 10% buffered formalin, and multiple tissue blocks were embedded in paraffin. Deparaffinized 4-μm-thick sections were immunostained by the streptavidin-biotin method using rabbit anti-SOD1olig (0.02 μg/mL), rabbit anti-SOD1int (0.3 μg/mL), mouse monoclonal anti-human SOD1 (0.5 μg/mL, clone 1G2, MBL) antibodies, and the corresponding biotin-conjugated secondary antibodies. The sections were processed with HRP-conjugated streptavidin and DAB as the chromogen and further stained for nuclei with hematoxylin. For double immunofluorescence, deparaffinized sections were first incubated with Sudan Black B to suppress auto-fluorescence and then stained with the primary antibodies followed by the corresponding FITC- or Cy3-labeled secondary antibodies (Jackson Labs, Pittsburgh, PA).
Sample preparations for biochemical analysis on human and mouse tissues
For human cases, the ventral and dorsal horns were separately excised from the frozen thoractic spinal cord samples. Frozen mice tissues (lumbar spinal cord, cervical spinal cord, cerebellum, and brainstem) were also separately prepared. The tissues were then homogenized and ultrasonicated in PBS containing 1% NP-40, 100 mM iodoacetamide, 5 mM EDTA, and EDTA-free Complete Protease inhibitor cocktail (Roche). The homogenates were centrifuged (20,000 x g, 30 min. 4 °C) to prepare the soluble supernatants and then examined for their total protein concentrations by using Micro BCA Assay Kit (Thermo Scientific).
Western blotting analysis
Soluble proteins in the tissue extracts (15 μg/lane) were separated using 12.5% polyacrylamide gels and blotted onto a PVDF membrane (0.2 μm, Wako). The membrane was treated with a blocking solution containing 5% (w/v) dried milk and 0.01% (v/v) Tween 20 in PBS at pH 7.4. The blots were probed with rabbit anti-SOD1 antibody (1:10,000; FL-154, Santa Cruz Biotechnology) and HRP-conjugated goat anti-rabbit IgG antibody (1:10,000; Thermo Scientific), visualized using ImmunoStar LD (Wako), and then observed in the LumiCube (Liponics). To validate an equal loading of tissue extracts, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an internal marker. The membranes were treated with the WB Stripping Solution (nacalai tesque) for 1 h at 37 °C and reprobed with rabbit anti-GAPDH antibody (1:5,000; FL-335, Santa Cruz Biotechnology).
All statistical tests were performed using Statcel 3 software (OMS Publishing Inc.). After the determination of normality, multiple group comparisons were performed using a one-way ANOVA followed by the Tukey–Kramer post-hoc test.
Preparation of soluble and disulfide-crosslinked oligomers of SOD1 in vitro
Purification of an antibody recognizing the disulfide-crosslinked SOD1 oligomers
A rabbit was first immunized with the soluble and disulfide-crosslinked oligomers of A4V-mutant SOD1 prepared in vitro, and then the polyclonal antibodies were affinity-purified using those oligomers. Nonetheless, the purified antibody was not selective to the oligomers; a natively folded SOD1 protein (Cu,Zn-SOD1S-S) significantly reacted with the antibody (data not shown). To increase the specificity of antibodies to the oligomers, the purified antibodies were washed with Ni2+-affinity resins on which wild-type SOD1S-S was immobilized through its N-terminal His tag (see Methods). As shown in Fig. 1b and c, the finally purified antibody (called anti-SOD1olig antibody) exhibited significantly higher ELISA signals to the soluble disulfide-crosslinked SOD1(A4V) oligomer than those to wild-type (WT) and A4V-mutant E,E-SOD1S-S proteins (P < 0.01). It is also important to note that anti-SOD1olig antibody can recognize the soluble and disulfide-crosslinked oligomers of G37R-mutant SOD1 but not E,E-SOD1(G37R)S-S (P < 0.01: Fig. 1b and c). Our anti-SOD1olig antibody was hence found to exclusively recognize the soluble and disulfide-crosslinked SOD1 oligomers.
Immunohistochemical detection of pathological SOD1 species by anti-SOD1olig antibody
At the end stage of the disease (140 and 160 days of age), however, immunostaining with anti-SOD1olig antibody was significantly reduced in the lumbar spinal cord of G1H mice (Fig. 2c and d),. This was not described by the changes in total amounts of soluble SOD1 in the lumbar spinal cord, which was actually increased during aging (; also see below). We previously showed that the number of the ChAT-positive motor neurons at 140 days of age was reduced down to one third of those at 60 days of age in G1H mice . The reduced immunostaining with anti-SOD1olig antibody might thus indicate loss of motor neurons at the disease end-stage. Nonetheless, mutant SOD1 has been also known to accumulate as inclusions in the surviving motor neurons of diseased G1H mice . Actually, the diffuse staining of SOD1 was observed in the lumbar spinal cord of pre-symptomatic G1H mice (60 and 100 days of age), and the SOD1-positive inclusions became evident in the disease end-stage (140 and 160 days of age) (Additional file 4: Figure S3). Our anti-SOD1olig antibody is hence expected to have little immunoreactivity toward the SOD1-positive inclusions formed in terminally ill G1H mice. Based upon these results, we suggest that anti-SOD1olig antibody specifically detects the pathological SOD1 species occurring in the lumbar spinal cords of the pre-symptomatic G1H mice.
These results show that our anti-SOD1olig antibody for the disulfide-crosslinked oligomers was able to detect pathological SOD1 species, but a major concern in the preparation of this antibody is a quite low yield (7.5 mL of 0.2 μg/mL antibody from one immunized rabbit) after several purification procedures. Actually, we have little amounts of anti-SOD1olig antibody left, and constant reproduction of anti-SOD1olig antibody would also be quite difficult due to its polyclonal nature. To deal with those troubles on anti-SOD1olig antibody, we attempted to first determine the epitope of anti-SOD1olig antibody and then produce another oligomer-specific antibody by immunizing rabbits with the peptide covering that epitope.
Anti-SOD1olig antibody recognizes interior of the SOD1 folded structure
Antibody recognizing the structural interior of SOD1 exhibits the specificity to the disulfide-crosslinked oligomers
To further check the specificity of our anti-SOD1int antibody in vitro, we prepared various forms of WT and ALS-mutant (A4V, G37R, and G85R) SOD1 proteins including E,E-SOD1SH, E,E-SOD1S-S, E,Zn-SOD1SH, E,Zn-SOD1S-S, Cu,E-SOD1S-S, Cu,Zn-SOD1S-S, and soluble oligomers with disulfide crosslinks (see Fig. 1a). Insoluble and amyloid-like SOD1 aggregates were also prepared by shaking E,E-SOD1SH with 1,200 rpm at 37 °C . Among those, only the soluble and disulfide-crosslinked oligomers but none of the others were recognized by anti-SOD1int antibody (Fig. 5b and Additional file 6: Figure S5B). Based upon these results in vitro, therefore, anti-SOD1int antibody can detect the disulfide-crosslinked SOD1 oligomers, in which the protein interior is significantly exposed to the solvent.
Actually, there are several precedents of the antibodies recognizing the protein interior of SOD1, which include USOD and SEDI polyclonal antibodies raised against the peptides, GG-L42HGFHVH48-GG and GG-R143LACGVIGI151-GG, respectively (also see Discussion) [16, 28]. Unfortunately, canonical USOD and SEDI antibodies reported by Chakrabartty and co-workers were not available, but the polyclonal antibodies raised against the same peptides as above were commercially available. We thus characterized those commercially available “USOD-like” and “SEDI-like” antibodies; indeed, USOD-like and SEDI-like antibodies were confirmed to specifically recognize Pepexon2 and Pepexon5, respectively (Additional file 7: Figure S6). As shown in Fig. 5c and d, USOD-like and SEDI-like antibodies exhibited reactivities toward almost all states examined except wild-type SOD1S-S that is fully or partially metallated and would thus be selective to mutant SOD1 proteins. Nonetheless, the recognition specificity of our anti-SOD1int antibody toward the disulfide-crosslinked oligomers was significantly higher than those of USOD/SEDI-like antibodies (Fig. 5b, c and d). These data thus emphasize the unprecedented recognition specificity of our anti-SOD1int antibody toward the disulfide-crosslinked SOD1 oligomers.
Disulfide-crosslinked SOD1 oligomers as an early pathological species in spinal cords of ALS-model mice
To check the availability/specificity of anti-SOD1int antibody for the detection of pathological SOD1 in vivo, we first examined the immunohistochemical analysis of G1H mice as well as non-transgenic mice. Unfortunately, however, the lumbar spinal cords of non-transgenic mice were immunostained with anti-SOD1int antibody (data not shown). This is in contrast to anti-SOD1olig antibody showing no immunostaining in non-transgenic mice (Additional file 3: Figure S2A and B). Those two antibodies hence appear to have distinct specificities in the immunohistochemical examination on mouse tissues. Nonetheless, anti-SOD1int as well as anti-SOD1olig antibody can exclusively recognize disulfide-crosslinked SOD1 oligomers in vitro in ELISA (Fig. 1b and 5b). We thus further examined homogenates from model mice for disulfide-crosslinked SOD1 oligomers by sandwich ELISA with anti-SOD1int antibody.
It should also be noted in Fig. 6a (red circles) that the ELISA signals from the lumbar spinal cords increase from 30 to 140 days of age but significantly drop from 140 to 160 days of age (P < 0.01). The reduction in the ELISA signals at 160 days of age was not described by the changes in total amounts of soluble SOD1 in the lumbar spinal cord (Fig. 6b, red circles). This is consistent with the decreased anti-SOD1olig immunostaining of the lumbar spinal cords of G1H mice at 140 and 160 days of age (Fig. 2c and d, Additional file 4: Figure S3C and D), while slightly different age-dependency in immunochemical response between anti-SOD1int and anti-SOD1olig antibodies would reflect their distinct immunological properties. As described later, however, the ELISA signals with anti-SOD1int antibody did not decrease in the other mouse model. The signal reduction in the disease end-stage may thus be a phenomenon specific to the lumbar spinal cord of G1H mice, but an exact reason for this remains obscure. Instead, we would like to emphasize that anti-SOD1int antibody can detect conformationally abnormal SOD1 species in model mice with sandwich ELISA.
To test if SOD1s were oligomerized with disulfide bonds in lumbar spinal cords of ALS-model mice, soluble fractions of the lumbar spinal cord homogenates of G1H mice were separated by non-reducing SDS-PAGE and analyzed by Western blotting. Soluble and disulfide-crosslinked SOD1 oligomers in vitro can be characterized by the reductant-sensitive smears in the high molecular weight region in SDS-PAGE gels (Additional file 2: Figure S1). As shown in Fig. 6c (left panel), smears in the high molecular weight region (>50 kDa) were evident, albeit weak intensity, as early as 60 days of age in G1H mice but not in WT mice (150 and 360 days) and non-transgenic mice (100 days; data not shown). Also importantly, those smears in G1H mice at 60 and 100 days of age disappeared when the soluble fractions were treated with β-mercaptoethanol (β-ME) prior to their loading on an SDS-PAGE gel (Fig. 6c, right panel), supporting the formation of disulfide-crosslinked SOD1 oligomers in the ALS-model mice even before the disease onset.
After the disease onset (at 140 and 160 days of age), in contrast, the reductant-sensitive SOD1 species in lumbar spinal cords of G1H mice were observed as more distinct bands in the high molecular weight region (>50 kDa, Fig. 6c). Also, even in the presence of β-ME, some SOD1-positive species were stuck on top of the separating gel. Therefore, we suppose different molecular properties of SOD1 oligomers between pre- and post-symptomatic stages of G1H mice, which might describe significant reduction of the SOD1 species immunoreactive to anti-SOD1int antibody at 160 days of age (Fig. 6a, red circles). Taken together, we speculate that anti-SOD1int antibody specifically detects the disulfide-crosslinked oligomers formed in the lumbar spinal cords of G1H mice from their pre-symptomatic stages.
We also tested the tissue-specificity in the formation of the disulfide-crosslinked oligomers; soluble supernatants from the homogenates of cervical spinal cord, brainstem and cerebellum of G1H mice were examined by sandwich ELISA using anti-SOD1int antibody. In ALS cases, the lumbar spinal cord is mainly affected, but the other regions of brains and spinal cords have also been shown to be involved in the pathology . In G1H mice, the lumbar spinal cord is the most severely damaged, and the changes occur later in the cervical spinal cords . Some pathological changes are reported in the brainstem , but the cerebellum is relatively spared . As shown in Fig. 6a, the ELISA signal intensities were significantly weaker in cervical spinal cord, brainstem, and cerebellum than those of lumbar spinal cord (P < 0.01 within the same age group, except at 160 days of age); in particular, almost no ELISA signals were observed in cerebellum. We confirmed similar levels of total SOD1 proteins among all of those tissues (Fig. 6b). Also, no obvious smears in the high molecular weight region were observed in the Western blots of soluble fractions of the cerebellum, while the brainstem lysates of G1H mice exhibited reductant-sensitive smears at 160 days of age, albeit with weak intensities (Additional file 9: Figure S8A and B). While toxic SOD1 species might appear everywhere but only afflict the spinal cord due to its vulnerability, amounts of SOD1 species probed with anti-SOD1int antibody were well correlated with the intensity of high-molecular-weight smears in the Western blots and also the severity of the damages in tissues (lumbar spinal cord > cervical spinal cord > brainstem > cerebellum) of G1H mice.
We have also examined the ALS model mice expressing human SOD1 with another mutation, G37R (loxG37R mice) . Compared to G1H mice, the expression level of mutant SOD1 is lower, and the disease progression is slower in loxG37R mice (the disease onset: ~350 days of age). The sandwich ELISA showed the age-dependent increase of the anti-SOD1int-positive SOD1 species in lumbar and cervical spinal cords but not in cerebellum of loxG37R mice (Additional file 10: Figure S9A), while amounts of the total soluble SOD1 remained almost constant during aging in loxG37R mice (Additional file 10: Figure S9B). When the spinal cord samples from loxG37R mice were pre-treated with DTT, the ELISA signals with anti-SOD1int antibody disappeared (data not shown). Furthermore, Western blotting analysis on loxG37R mice revealed the reductant-sensitive smears in lumbar spinal cord (Additional file 10: Figure S9C), albeit with significantly weaker intensities compared to those of G1H mice, but not in cerebellum (Additional file 9: Figure S8C). Taken together, these results suggest the formation of the soluble disulfide-crosslinked SOD1 oligomers as pathological changes also in loxG37R mice.
Anti-SOD1int antibody detects pathological SOD1 in SOD1-ALS cases
To reduce the chance of potential conformational changes of SOD1 during the preparation of sections and their immunostaining procedures, we prepared soluble fractions by centrifugation of the homogenates from either ventral or dorsal horn in the thoractic spinal cords and then examined those by sandwich ELISA using anti-SOD1int antibody. ALS associates with degeneration of motor neurons in the ventral horn with less involvement of sensory neurons in the dorsal horn of the spinal cord . As shown in Fig. 7d, the ventral horn of the SOD1-ALS patients with C111Y mutation (cases III-4 and IV-6; Additional file 1: Table S1) showed higher signal intensities of anti-SOD1int antibody than those of non-ALS controls (three cases). In contrast, weak ELISA signals of anti-SOD1int antibody were detected in the dorsal horn with almost no difference between the controls and the ALS patients (Fig. 7d). Also importantly, the ELISA signals obtained by using anti-SOD1int antibody disappeared upon pre-treatment of the samples with a reductant, DTT, suggesting the involvement of disulfide-crosslinks in the pathological SOD1 species (Fig. 7d). The amounts of total soluble SOD1 proteins in the ventral and dorsal horns were not different between the controls and the ALS patients (Fig. 7e). Collectively, our anti-SOD1int antibody is considered to detect the disulfide-crosslinked SOD1 oligomers as pathological species in SOD1-ALS patients.
SOD1-ALS cases are characterized mainly by abnormal accumulation of mutant SOD1 proteins in motor neurons of the affected spinal cords , while the pathological involvement of the other types of neurons and glia cells has been reported [20, 29, 34, 35]. The conformational stability of SOD1 is significantly compromised by most of the mutations , which triggers the formation of soluble oligomers and insoluble aggregates of SOD1 in vitro. In this study, we successfully prepared anti-SOD1olig/int antibodies exclusively recognizing the disulfide-crosslinked SOD1 oligomers in vitro and then found that those antibodies detect the pathological SOD1 species in spinal cords of the SOD1-ALS patients (C111Y) as well as transgenic model mice (G1H and loxG37R mice).
Depending upon experimental conditions in vitro and in vivo, nonetheless, SOD1 is known to misfold in distinct pathways [8, 9, 10, 51, 52, 53, 54]. Therefore, we could not exclude the possibility that our antibodies detect certain misfolded species other than the disulfide-crosslinked oligomer. It has been known that several misfolded conformations of SOD1 can be reproduced in vitro simply by metal dissociation and/or disulfide reduction. Actually, we have shown that reduction of the disulfide bond drastically increases fluctuation of the loops IV and VII (Fig. 8b), temporarily “peels” those loops off from the β-barrel scaffold, and thus potentially exposes the epitope of anti-SOD1int antibody . As shown in Fig. 5b, however, disulfide-reduced and/or demetallated forms of SOD1 were not recognized by our anti-SOD1int antibody. Therefore, the reversible conformational fluctuation increased by disulfide reduction/demetallation is probably not sufficient to allow the antibody to access the epitope buried inside the protein. The disulfide-reduced and demetallated SOD1 has been shown to irreversibly form the amyloid-like aggregates , which were again not recognized by anti-SOD1int antibody (Fig. 5b). We further examined ELISA to test the reactivities of anti-SOD1int antibody toward E,E-SOD1S-S and E,E-SOD1SH that were misfolded/unfolded with either guanidine hydrochloride (6 M) or acidic buffer (pH 3.0) in vitro, but no signals were observed (Additional file 11: Figure S10A). In sharp contrast, USOD-like and SEDI-like antibodies were found to react with various non-native forms of SOD1 proteins (Fig. 5c and d, Additional file 11: Figure S10B, C, and D). Taken together, those extensive tests reveal quite high recognition specificity of our anti-SOD1int antibody toward the disulfide-crosslinked SOD1 oligomer.
Anti-SOD1int antibody detected pathological SOD1 species in vivo, and also, the reductant-sensitive smears of SOD1 were observed in the Western blots of the affected tissues of model mice (Fig. 6, Additional file 10: Figure S9). We hence speculate that the disulfide-crosslinked SOD1 oligomer is involved in the pathology of SOD1-ALS. Actually, disulfide-crosslinked oligomers of SOD1 have been reproduced in cultured cells [18, 57]. While the reducing environment of the cytoplasm might be unfavorable for crosslinking proteins via disulfide bonds, our preliminary in vitro experiments confirmed the formation of disulfide-crosslinked SOD1 oligomers in the presence of 5 mM reduced glutathione with 0.5 mM oxidized glutathione (data not shown), which is a feasible redox condition of the cytoplasm . In our proposed mechanism for the formation of SOD1 oligomers , the disulfide bond is not newly introduced but rather shuffled among the Cys residues in SOD1 (Fig. 8c). More specifically, the disulfide shuffling will break the canonical Cys 57 - Cys 146 disulfide bond and detain SOD1 in the misfolded conformation where the loops IV and VII are peeled off from the β-barrel scaffold with the persistent exposure of the epitope region (Fig. 8c). Such a disulfide shuffling may be more robust against reducing environment than de novo formation of disulfide bonds.
We have also shown here that immunoreactivities of our anti-SOD1olig/int antibodies are quite exclusive to the human SOD1-ALS cases as well as the transgenic mice (G1H and loxG37R mice) but not to the controls without SOD1 mutations. Their immunoreactivities were, furthermore, well correlated with the severity of degeneration in the model mice (lumbar spinal cord > cervical spinal cord > brainstem > cerebellum; Fig. 6a, Additional file 10: Figure S9A) and in the SOD1-ALS cases (ventral horn > dorsal horn; Fig. 7d). As described in the Results section, nonetheless, we should note significant reduction in the immunoreactivities of anti-SOD1olig/int antibodies toward the G1H mice in the disease end stage (Figs. 2 and 6a). While such decline will be partly because of the concomitant loss of motor neurons , mutant SOD1 is also known to accumulate as amyloid-like aggregates in the spinal cord of the model mice after the appearance of motor symptoms [9, 15]. Given that our anti-SOD1int antibody was not able to react with the amyloid-like SOD1 aggregates in vitro (Fig. 5b), the declined immunoreactivity of anti-SOD1olig/int antibodies at the end-stage might indicate the formation of amyloid-like SOD1 aggregates due to quite high expression of G93A SOD1 proteins in G1H mice.
In contrast, the spinal motor neurons of the SOD1-ALS patients in their disease end-stage were immunostained with anti-SOD1olig/int antibodies (Figs. 3 and 7), suggesting distinct properties of pathological SOD1 species between G1H mice and the patients. Because the inclusions in SOD1-ALS patients exhibited no reactivity to an amyloid-diagnostic dye, Thioflavin-S , the amyloid-like SOD1 aggregates would form only in the end-stage G1H mice but not in human SOD1-ALS cases. While it needs to be tested whether the SOD1 species detected by our antibodies was an on-pathway intermediate for the formation of amyloid-like SOD1 aggregates, the pathologies in the autopsied human cases might not proceed into the terminal stage as in the G1H mice at 160 days of age. Compared to the amyloid-like SOD1 aggregates in the end stage of G1H mice, we speculate that SOD1 species detected by anti-SOD1olig/int antibodies in the pre-symptomatic G1H mice have more significance in the pathogenicity of SOD1-ALS.
It is also important to note moderate immunoreactivities of anti-SOD1int antibody in the spinal cords of WT mice at 360 days but not at 150 days of age (Additional file 8: Figure S7A). WT mice do not develop severe motor phenotypes but show several neuropathological and symptomatic changes in their advanced age including mitochondrial vacuolization in spinal cords (>210 days), impaired motor performance (>410 days), and motor neuron death (~2 years) . No immunoreactivities of C4F6 and SEDI antibodies were reported in the spinal cords of WT mice at 215 and 100 days of age, respectively [28, 43]. Instead, the immunoreactivity of anti-SOD1int antibody appears to match with the neuropathological changes in WT mice and would hence probe the misfolded SOD1 proteins with toxicity toward motor neurons. Involvement of wild-type SOD1 proteins in the ALS pathogenesis still remains controversial , and no immunostaining by anti-SOD1int antibody was confirmed in our sporadic ALS cases without SOD1 mutations (Additional file 5: Figure S4C and D). Nonetheless, we speculate that SOD1 could become toxic to motor neurons by assuming the anti-SOD1olig/int antibody-positive conformation such as the disulfide-crosslinked oligomers even in the absence of any pathogenic mutations.
In summary, we successfully prepared the antibodies exclusively recognizing the disulfide-crosslinked SOD1 oligomers in vitro. Pathological SOD1 species in the affected tissues of SOD1-ALS patients as well as transgenic mice provided the immunological epitope to those antibodies. While it is possible that the epitope of our antibodies becomes available upon misfolding of SOD1 through some other mechanisms, we propose that the oligomerization via shuffling of the disulfide bond has pathological significance in SOD1-ALS.
This work was supported by Grants-in-Aid 16H04768 for Scientific Research (B) (to YF), 15H01566 for Scientific Research on Innovative Areas (to YF), 15 K14480 for Challenging Exploratory Research (to YF), 15H06588 for Young Scientists (Start-up) (to ET) from the Ministry of Education, Culture, Sports, Science and Technology of Japan. Preparation of anti-SOD1olig antibody was supported by Comprehensive Brain Science Network (CBSN), Japan.
Availability of data and materials
The datasets during and/or analysed during the current study available from the corresponding author on reasonable request.
YF directed the project, analyzed the data and wrote the manuscript. ET, IA, TN, and KT prepared the purified protein samples and purified the antibodies. ET and TN performed Western blotting experiments and ELISA. MW prepared the anti-SOD1olig antibody, and immunohistochemical examination on mice was performed by SW, KY, YM and HM. Clinical data collection as well as immunohistochemical examination on human cases was performed by SO. All authors read and approved the manuscript.
The authors declare that they have no competing interests.
Consent for publication
Ethics approval and consent to participate
All procedures performed in studies involving human participants were in accordance with the ethical standards of the Matsumoto Medical Center and Keio University research committees and also with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. All procedures performed in studies involving animals were in accordance with the ethical standards of Keio University and Nagoya University where the studies were conducted.
- 22.Ichikawa T, Ajiki K, Matsuura J, Misawa H. Localization of two cholinergic markers, choline acetyltransferase and vesicular acetylcholine transporter in the central nervous system of the rat: in situ hybridization histochemistry and immunohistochemistry. J Chem Neuroanat. 1997;13:23–39.CrossRefPubMedGoogle Scholar
- 24.Morisaki Y, Niikura M, Watanabe M, Onishi K, Tanabe S, Moriwaki Y, et al. Selective Expression of Osteopontin in ALS-resistant Motor Neurons is a Critical Determinant of Late Phase Neurodegeneration Mediated by Matrix Metalloproteinase-9. Sci Rep. 2016;6:27354.CrossRefPubMedPubMedCentralGoogle Scholar
- 39.Sabado J, Casanovas A, Hernandez S, Piedrafita L, Hereu M, Esquerda JE. Immunodetection of disease-associated conformers of mutant cu/zn superoxide dismutase 1 selectively expressed in degenerating neurons in amyotrophic lateral sclerosis. J Neuropathol Exp Neurol. 2013;72:646–61.CrossRefPubMedGoogle Scholar
- 59.Jaarsma D, Haasdijk ED, Grashorn JA, Hawkins R, van Duijn W, Verspaget HW, et al. Human Cu/Zn superoxide dismutase (SOD1) overexpression in mice causes mitochondrial vacuolization, axonal degeneration, and premature motoneuron death and accelerates motoneuron disease in mice expressing a familial amyotrophic lateral sclerosis mutant SOD1. Neurobiol Dis. 2000;7:623–43.CrossRefPubMedGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the 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.