Address correspondence and reprint requests to Martin Ingelsson, Department of Public Health and Caring Sciences, Rudbeck Laboratory, 751 85 Uppsala, Sweden. E-mail: email@example.com
Inclusions of intraneuronal alpha-synuclein (α-synuclein) can be detected in brains of patients with Parkinson's disease and dementia with Lewy bodies. The aggregation of α-synuclein is a central feature of the disease pathogenesis. Among the different α-synuclein species, large oligomers/protofibrils have particular neurotoxic properties and should therefore be suitable as both therapeutic and diagnostic targets. Two monoclonal antibodies, mAb38F and mAb38E2, with high affinity and strong selectivity for large α-synuclein oligomers were generated. These antibodies, which do not bind amyloid-beta or tau, recognize Lewy body pathology in brains from patients with Parkinson's disease and dementia with Lewy bodies and detect pathology earlier in α-synuclein transgenic mice than linear epitope antibodies. An oligomer-selective sandwich ELISA, based on mAb38F, was set up to analyze brain extracts of the transgenic mice. The overall levels of α-synuclein oligomers/protofibrils were found to increase with age in these mice, although the levels displayed a large interindividual variation. Upon subcellular fractionation, higher levels of α-synuclein oligomers/protofibrils could be detected in the endoplasmic reticulum around the age when behavioral disturbances develop. In summary, our novel oligomer-selective α-synuclein antibodies recognize relevant pathology and should be important tools to further explore the pathogenic mechanisms in Lewy body disorders. Moreover, they could be potential candidates both for immunotherapy and as reagents in an assay to assess a potential disease biomarker.
One of the main neuropathological features in Parkinson's disease (PD) and dementia with Lewy bodies (DLB) is the intraneuronal presence of Lewy bodies in certain brain areas. Alpha-synuclein (α-synuclein) is the main constituent of the Lewy body, which has been described to also contain a large number of other molecules (Leverenz et al. 2007).
Alpha-synuclein is mainly expressed in the central nervous system (CNS), although the protein can be detected also in the peripheral nervous system and other tissues outside the CNS (reviewed in (Wakabayashi et al. 2010)). It is normally localized to pre-synaptic terminals but, in the diseased brain, α-synuclein starts to accumulate in cell bodies, processes, and synapses (Roy 2009). In the cell, α-synuclein seems to predominantly occur as a disordered monomer (Fauvet et al. 2012), although one study reported that it adopts a folded tetrameric structure in human erythrocytes (Bartels et al. 2011). Its normal physiological function is unclear, but it has been proposed to promote the formation of soluble N-ethylmaleimide-sensitive factor attachment protein receptors complexes (Burre et al. 2010) and over-expression of α-synuclein was shown to reduce dopamine release (Nemani et al. 2010).
The process by which α-synuclein forms fibrils has been well described. A conformational shift of the monomeric structure results in partially folded and aggregation-competent species that can start to form oligomers and protofibrils of increasing sizes. Such soluble pre-aggregates eventually become insoluble and accumulate as fibrils into Lewy bodies (Uversky et al. 2001).
The term oligomer refers to smaller aggregates that can be either on their way of forming fibrils, i.e., on-pathway oligomers, or are locked in a conformation that does not permit fibrils to be formed, i.e., off-pathway oligomers. Protofibrils, however, are larger multimeric species that are always on the pathway of forming fibrils (Volles et al. 2001; Cole et al. 2005). Both oligomers and protofibrils are probably present in the diseased human brain as well as in transgenic mouse brain.
Although there is an increased amount of α-synuclein aggregates in DLB as compared with non-neurological control brains (Campbell et al. 2000; Paleologou et al. 2009), the number of Lewy bodies does not correlate well with either duration or severity of the disease (Harding and Halliday 2001). Biochemically, only detergent-insoluble α-synuclein was found to be elevated in DLB as compared with non-diseased brains, whereas the total amount of α-synuclein did not differ between the two groups (Kahle et al. 2001; Klucken et al. 2006).
Genetic evidence suggests that the α-synuclein A30P mutant, causing dominantly inherited PD, may be pathogenic by accelerating oligomerization of α-synuclein (Conway et al. 2000). Recent research has also indicated that large α-synuclein oligomers exhibit neurotoxic effects via disruption of cellular membranes (Danzer et al. 2007; Winner et al. 2011) or by neuroinflammatory mechanisms (Wilms et al. 2009). In addition, such oligomers may cause dopaminergic cell loss, as indicated in a rat model expressing various oligomerization–prone α-synuclein mutants (Winner et al. 2011).
Alpha-synuclein oligomers can be generated in vitro from recombinant protein with the help of various substances, such as epigallocatechin gallate (Ehrnhoefer et al. 2008) and ethanol (Danzer et al. 2007). Many different types of oligomers have been described. Certain globular or annular species cause an increase in intracellular calcium, whereas others seem to directly enter the cell and influence α-synuclein aggregation (Danzer et al. 2007). Yet other in vitro-generated oligomers are seemingly off-pathway species that can neither aggregate into fibrils nor seed the fibril formation (Souza et al. 2000; Zhu et al. 2004; Cappai et al. 2005; Konno et al. 2005; Ehrnhoefer et al. 2008).
Large, stable α-synuclein oligomers induced by reactive aldehydes cause mitochondrial toxicity in vitro (Nasstrom et al. 2009, 2011) as well as impairment of long-term potentiation (Diogenes et al. 2012; Martin et al. 2012). Such oligomeric species, produced in the presence of either 4-oxo-2-nonenal (ONE) or 4-hydroxy-2-nonenal (HNE), have been described as amorphous or curved-like structures with lengths varying between 100 and 200 nm and a molecular weight of approximately 2000 kDa (Nasstrom et al. 2009, 2011). Although these large oligomeric species do not form fibrils, they bear morphological resemblance to protofibrils.
Several mouse models over-express α-synuclein (reviewed in (Beal 2010)). In homozygous (Thy-1)-h [A30P] α-synuclein transgenic mice (Kahle et al. 2001; Neumann et al. 2002; Freichel et al. 2007; Schell et al. 2009), the transgene is expressed throughout the CNS at a roughly twofold higher level compared with the endogenous expression (Kahle et al. 2000). In these mice, proteinase K-resistant α-synuclein pathology appears in the brainstem, midbrain, and spinal cord from 12 months of age (Neumann et al. 2002), whereas hyperphosphorylated α-synuclein (Schell et al. 2009) occurs at a later stage. Behaviorally, the mice suffer from cognitive impairment at 12 months of age, whereas motor deficits are usually seen from 17 months of age.
The presence of α-synuclein oligomers in transgenic mouse brain has recently been investigated. In transgenic mice expressing human A53T-mutated α-synuclein, the levels of α-synuclein oligomers started to increase in the endoplasmic reticulum (ER) at the onset of symptoms in these mice (Colla et al. 2012a,b). Interestingly, other subcellular compartments did not display any alterations in the levels of oligomers, indicating that the accumulation of α-synuclein oligomers in the ER may be a central pathogenic event, at least in this mouse model (Colla et al. 2012a,b).
As α-synuclein oligomers/protofibrils seem to play a central pathophysiological role in Lewy body disorders, it is of great importance to develop reliable methods for the detection of such species. Thus, in this study, monoclonal antibodies selective for α-synuclein oligomers were generated followed by a demonstration of their utility for immunohistochemistry of α-synuclein brain pathology on human and transgenic mouse brain as well as quantitative analysis, by ELISA, of biological samples from transgenic mouse brain.
Materials and methods
HNE, 10 mg/mL, was supplied in 99% ethanol (Cayman Chemicals, Ann Arbor, MI, USA).
The experiments involving animals were performed in compliance with the local animal ethics committee (Stockholm, Sweden, N 417/08). All efforts were made to minimize pain and discomfort. The ARRIVE guidelines have been followed. The mice were housed in groups of two to four animals per cage in a temperature- and humidity-controlled room and were given access to food and water ad libitum.
Generation of recombinant α-synuclein
Human α-synuclein cDNA was cloned into a pET-17b vector (Novagen, Madison, WI, USA) and transformed into BL21 (DE3) competent cells (Invitrogen, Carlsbad, CA, USA). After induction, the protein was extracted and purified on a Q sepharose fast flow anion exchange column (GE Healthcare, Uppsala, Sweden), followed by desalting and equilibration with 50 mM disodium hydrogen phosphate, pH 8.5. The purity of the α-synuclein preparation was determined to be > 95%, as assessed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and size-exclusion–high-performance liquid chromatography (SEC-HPLC). Finally, the protein concentration was measured using a bicinchoninic acid protein assay (Pierce Biotechnology, Rockford, IL, USA) and aliquots (140 μM) were stored at −20°C prior to use.
Generation of α-synuclein oligomers
The HNE-induced α-synuclein oligomers were generated as previously described (Nasstrom et al. 2009).
Immunization with HNE-induced α-synuclein oligomers to generate monoclonal antibodies
Balb/C mice (Jackson laboratories, Bar Harbor, ME, USA) were repeatedly immunized subcutaneously with HNE-induced α-synuclein oligomers. Mice with high plasma titers were sacrificed after which the spleen was removed and fused with Sp2/0 myeloma cells to generate antibody-producing hybridomas according to standard techniques (de StGroth and Scheidegger 1980). The antibodies produced by the hybridomas were screened for α-synuclein reactivity by ELISA and thereafter subcloned to produce monoclonal cultures. For isotype determination, an IsoStrip kit (Roche Diagnostics, Basel, Switzerland) was used. Three different antibodies, named mAb38F, mAb38E2 and mAb15, were purified from the conditioned media with affinity chromatography using Protein G-Sepharose (GE Healthcare).
Characterization of antibody affinity and selectivity
Inhibition ELISA: Ninety-six-well high-binding EIA/RIA plates (Corning Inc, Corning, NY, USA) were coated with α-synuclein monomers and blocked with 1% bovine serum albumin (BSA) for 1 h. The three generated antibodies (mAb38F, mAb38E2, mAb15) and Syn-1 were incubated with serially diluted α-synuclein monomers or HNE-induced oligomers in low-binding ELISA plates for 1 h at 20°C. Next, the pre-incubated antibodies were added to the monomer-coated plate for 10 min at 20°C. The alkaline phosphatase–coupled anti-mouse-IgG (MabTech, Nacka, Sweden) was used as secondary antibody and incubated for 1 h at 20°C before addition of the substrate pNPP (Sigma-Aldrich, St. Louis, MO, USA) dissolved in diethanol buffer. The OD was read at 405 nm.
To investigate if mAb38F or mAb38E2 recognizes a general oligomeric/protofibrillar epitope, the binding of the antibodies to α-synuclein oligomers and amyloid-beta (Aβ) protofibrils, (Englund et al. 2007) involved in Alzheimer's disease, was compared using the inhibition ELISA described above.
As the transgenic mouse model used in this study expresses A30P α-synuclein, we also wanted to evaluate whether mAb38F and mAb38E2 bind wild-type α-synuclein and A30P α-synuclein oligomers equally well. Large α-synuclein oligomers were generated by incubating HNE with recombinant A30P α-synuclein, followed by analysis with the inhibition ELISA as described above.
For all ELISA measurements, the concentration of oligomers was calculated from the starting concentration of the α-synuclein monomers.
To exclude the possibility that the antibodies bind to HNE-modified epitopes, 1 μg/ml HNE coupled to BSA was coated in a 96-well high-binding EIA/RIA plate (Corning Inc,). Next, mAb38F, mAb38E2 or mAb15 was added to the HNE–BSA-coated plate followed by the use of an horseradish peroxidase (HRP)-coupled anti-mouse IgG secondary antibody (Southern Biotechnology, Birmingham, AL, USA). An antibody directed against HNE was used as a positive control (Abcam, Cambridge, UK).
HNE-induced α-synuclein oligomers were generated as described above. Tau and Aβ aggregates were induced by preparing either recombinant Tau-441 2N4R (R-peptide, Bogart, GA, USA) at a concentration of 21 μM in 50 mM MES, pH 6.8, 100 mM NaCl, and 0.5 mM EGTA, or Aβ1-42 (American Peptide, Sunnyvale, CA, USA) dissolved at 496 μM in 10 mM NaOH and diluted with Tris-buffered saline (TBS) to a final concentration of 89 μM. The samples were incubated in a non-binding polystyrene 96-well plate (Greiner Bio ONE, Frickenhausen, Germany) and agitated at 900 rpm on a Titramax 101 (Heidolph Instruments, Schwabach, Germany) for 4 days at 37°C. Islet amyloid polypeptide (IAPP, Bachem, Bubendorf, Switzerland) was dissolved in sodium phosphate, pH 7.4 at a concentration of 25 μM and quiescently incubated for 4 h at 37°C. Lyophilized β- and γ-synuclein (R-peptide) was dissolved at 70 μM in 20 mM Tris/HCl, pH 7.4.
Fifty nanograms of protein was applied onto a nitrocellulose membrane (Bio-Rad, Hercules, CA, USA) and the membrane was blocked with Odyssey blocking buffer (Li-Cor, Lincoln, NE, USA) overnight at 4°C. The blots were probed with the following primary antibodies for 2 h at 20°C: α-synuclein (mAb38F, 0.5 μg/mL, mAb38E2 0.5 μg/mL, mAb15 0.5 μg/mL), β- and γ-synuclein (1977-1, 1/5000, and 1978-1, 1/2000, rabbit monoclonals, Epitomics, Burlingame, CA, USA), Tau (Tau-1, 0.2 μg/mL, mouse monoclonal, Millipore, Billerica, MA, USA), Aβ (6E10, 1/2000, mouse monoclonal, Covance, Princeton, NJ, USA), and IAPP (IAPP 25-37, T-4157, 1/5000, rabbit polyclonal, Bachem). After extensive washing with 50 mM Tris/HCl pH 7.4, 150 nM NaCl, 0.05% Tween 20, the blots were incubated with a goat-anti-rabbit or donkey-anti-mouse Dylight 800 antibody (0.1 μg/mL, Pierce, Thermo Scientific) for 1 h at room temperature. Finally, the blots were washed with 50 mM Tris/HCl pH 7.4, 150 nM NaCl, 0.05% Tween 20 and visualized using an Odyssey imaging system (Li-Cor).
Immunohistochemistry of human brain tissue
For immunostaining, brains from two PD patients, two DLB patients and three non-neurological control subjects were included. For the PD brains, sections from the mesencephalon/substantia nigra were stained, whereas both mesencephalon and neocortex were analyzed for the DLB and control brains. All brains were from the University Hospitals in Helsinki or Turku, and the research project had been approved by the regional ethical committees.
Paraffin-embedded human brain tissues were sectioned (5 μm), deparaffinized in xylene, and hydrated in a decreasing ethanol series (99%, 95%, 70%), followed by subsequent rinsing in water and phosphate-buffered saline (PBS). For antigen retrieval, sections were brought to boil in 25 mM sodium citrate buffer (pH 7.3), directly followed by 20 min cooling. After a quick rinse with PBS, sections were treated with 70% formic acid for 5 min and washed under running tap water for 10 min. Endogenous peroxidases were blocked by treatment with 0.8% H2O2 (in methanol) for 30 min. Next, sections were incubated with Background sniper (Biocare Medical, Concord, CA, USA) for 1 h at 20°C to avoid non-specific binding. After permeabilization of the tissues in PBS 0.4% Triton X-100 for 5 min, sections were incubated overnight with the primary antibody in a humidified chamber at 4°C. The primary α-synuclein antibodies (mAb38F, mAb38E2 or mAb15) were diluted in PBS 0.1% Tween 20 to a final concentration of 0.4 μg/ml. The α-synuclein antibodies Syn-1 (BD Biosciences, Franklin lake, NJ, USA) and 4B12 (Covance) were diluted in PBS 0.1% Tween 20 to a final concentration of 0.4 μg/mL and 10 μg/mL, respectively.
The following morning, sections were rinsed carefully with PBS to remove unbound primary antibody before incubating with the biotinylated secondary antibody anti-mouse PK-2200 (1 : 250, M.O.M. Immunodetection Kit, Vector Laboratories, Burlingame, CA, USA) or anti-rabbit BA-1000 (1 : 100, Vector Laboratories) for 30 min. Next, the sections were thoroughly washed with PBS and incubated with Streptavidin-HRP (1 : 30, 3310-9, Mabtech) for 30 min. Thereafter, the bound HRP was visualized using NovaRed substrate kit (SK-4800, Vector Laboratories). The sections were counter-stained with hematoxylin, washed in dH2O, dehydrated and mounted with DPX (VWR, Stockholm, Sweden). As negative control experiments, sections were incubated with primary or secondary antibodies alone or with isotype-matched control antibodies.
Immunohistochemistry of transgenic mouse brain tissue
(Thy-1)-h [A30P] α-synuclein transgenic mice were anesthetized at 4 (n = 3), 12 (n = 3), and 18 months (n = 4) of age with 5% Avertin. Mice were perfused with 0.9% NaCl, after which the brains were dissected and fixed in 4% phosphate-buffered formaldehyde. One hemisphere was embedded in paraffin and sagitally sectioned (6 μm). Sections were stained with the generated α-synuclein antibodies mAb38E2, mAb38F and mAb15 (0.4 μg/mL), as well as with the phospho-α-synuclein antibody PS129 (1 : 100, ab59264, Abcam) and the non- selective α-synuclein antibody Syn-1 (0.4 μg/mL, BD Biosciences).
The same protocol as for human tissue sections were used for analysis of the mice brain sections, with the following exceptions: blocking of non-specific binding to the sections were performed with Mouse IgG blocking reagent (M.O.M. Immunodetection Kit, PK-2200, Vector Laboratories) and the primary α-synuclein antibodies were diluted in M.O.M. diluent (Vector Laboratories). The mouse sections were not counter-stained with hematoxylin. In addition, sections from wild-type mice (C57/Bl6) were used as control tissues and treated as described above. Sections were also incubated with primary or secondary antibodies alone or with irrelevant isotype-matched antibodies.
Homogenization of transgenic mouse brain
Samples of one cerebral hemisphere (olfactory bulb and cerebellum excluded) and the upper part of the spinal cord (10 mm) of (Thy-1)-h [A30P] α-synuclein transgenic mice at 3 (n = 6) and 15 (n = 6) months of age were homogenized in 1 : 10 (tissue weight:extraction volume ratio) in TBS (20 mmol/l Tris and 137 mmol/l NaCl, pH 7.6) with complete protease inhibitor cocktail (Roche) using a tissue grinder with teflon pestle (2 × 10 strokes on ice). The slurry was centrifuged at 16 000 g for 1 h at 4°C. The supernatant (TBS-soluble α-synuclein) was saved and the resultant pellets were homogenized as described above, but with 10 strokes on ice in TBS 0.5% Triton X-100 (Sigma-Aldrich). The Triton X-100 slurry was centrifuged at 16 000 g for 1 h at 4°C. The supernatant (membrane bound α-synuclein) was saved and the pellet was resuspended and homogenized as described above, but in TBS 1% sodium dodecyl sulfate (SDS) (Sigma-Aldrich). The resulting pellet was resuspended in 70% formic acid and centrifuged as described above before the samples were aliquoted and stored at −70°C. One non-transgenic mouse (16 months old) brain hemisphere was used as control and prepared as described above. All samples were analyzed with ELISA to assess levels of α-synuclein oligomers/protofibrils.
In addition, one of the hemispheres from the mice used for subcellular fractionation (see below) was homogenized for analysis of total α-synuclein.
Brain tissue from the one hemisphere (olfactory bulb and cerebellum excluded) was collected from a separate set of (Thy-1)-h [A30P] α-synuclein transgenic mice of the following ages: 6, 12, and 18 months (n = 5–6/age group). Frozen brain tissue was homogenized with Lysis Buffer (Sucrose 250 mM, HEPES 20 mM, KCl 10 mM, MgCl2 1,5 mM, EDTA 2 mM) pH 7.5 with complete protease inhibitor (Roche) and fractionated by centrifugation in accor-dance with a previously published protocol (Colla et al. 2012a,b). Samples containing nuclei, mitochondriae and ER were collected and stored in −20°C until they were analyzed with the oligomer/protofibril-selective ELISA.
The other hemisphere from each mouse was prepared as described under Homogenization of transgenic mouse brain and analyzed with ELISA to yield an estimate of total α-synuclein levels in these brains (see below). To correlate the oligomer/protofibril levels from the various subcellular fractions with total α-synuclein, we used the generated TBS fraction. It has previously been reported (Cox and Emili 2006; Colla et al. 2012a,b) that a centrifugation step at 100 000 g is needed to pellet the actual organelles. Thus, as the samples used for measuring total α-synuclein were centrifuged at 16 000 g it is likely that both the nuclei, mitochondriae and ER remain in the supernatant, i.e., in the TBS fraction.
High-binding 96-well EIA/RIA plates (Corning Inc) were coated with 1 μg/mL of mAb38F in PBS overnight at 4°C. Each plate was blocked with 1% BSA for 1 h after which the samples were added and incubated for 2 h at 20°C with shaking (900 rpm). The mice brain tissue samples were diluted 1 : 4 for TBS and Triton X-100 fractions and 1 : 50 for the SDS fraction in incubation buffer (PBS with 0.1% BSA, 0.05% Tween, and 0.15% Kathon). For detection, biotinylated mAb38F (1 μg/mL) was added for 1 h at 20°C with shaking, followed by incubation for 45 min with streptavidin-HRP (diluted 1 : 5000, Mabtech). Wells were washed three times in ELISA-washing buffer (phosphate-buffered NaCl with 0.1% Tween 20 and 0.15% Kathon) between each step. All dilutions were performed in incubation buffer. K-blue Aqueous TMB substrate (Neogen Corporation, Lansing, MI, USA) was added and the reaction was stopped after 30 min with 2 M H2SO4.
Total α-synuclein ELISA
Total levels of α-synuclein were measured in the formic acid fraction (reflecting the amount of fibrillar α-synuclein) as well as in the TBS fraction from mice used in the subcellular fractionation. High-binding 96-well EIA/RIA plates (Corning Inc) were coated with 0.5 μg/mL Syn-1 (BD) in PBS over night at 4°C. Unspecific binding was blocked with 1% BSA for 1 h, after which the homogenized samples were added and incubated for 2 h at 20°C while shaking. As detection antibody, a rabbit IgG anti-α-synuclein polyclonal antibody, FL140 (1 : 1000, Santa Cruz Biotechnology, Santa Cruz, CA, USA), was used for 1 h at 20°C with agitation. Thereafter, an HRP-coupled goat-anti-rabbit IgG antibody (1 : 10000, Thermo Fisher Scientific Inc. Rockford, IL, USA) was added for 1 h at 20°C while shaking. The aqueous TMB substrate (Neogen Corporation) was added and the reaction was stopped after 15 min with 2 M H2SO4. Between each step, the wells were washed three times with ELISA washing buffer. Samples and antibodies were diluted in ELISA incubation buffer.
To test for statistically significant differences, groups were analyzed with unpaired Student′s t-test. Probability values < 0.05 were considered significant using a two-tailed confidence interval.
Generation of HNE-induced α-synuclein oligomers
Large α-synuclein oligomers were generated by adding HNE to α-synuclein monomers followed by incubation overnight. SEC-HPLC analysis showed that HNE-induced α-synuclein oligomers elute with a volume corresponding to a molecular weight of ~ 2000 kDa, whereas native α-synuclein elute as a monomeric peak corresponding to a molecular weight of ~ 47 kDa (Nasstrom et al. 2011). No monomeric protein could be detected in the HNE-induced α-synuclein oligomer samples (data not shown). The concentration of the generated HNE-induced oligomers was calculated as the starting concentration of the monomer. The resulting oligomers were used both for immunization and as standard in the oligomer selective ELISA.
Affinity and selectivity of the generated antibodies
Several antibody-producing hybridomas were generated. An inhibition ELISA was set up to characterize their selectivity and affinity for HNE-induced α-synuclein oligomers vs. α-synuclein monomers. Two of the antibodies generated, mAb38F and mAb38E2, were found to be strongly selective for the HNE-induced α-synuclein oligomers as compared with monomers (Fig. 1a, b). Another antibody, mAb15, was found to bind HNE-induced α-synuclein oligomers and monomers equally well (Fig. 1c). The commercially available antibody Syn-1 was used as a reference and displayed a stronger preference for the monomeric form of α-synuclein (Fig. 1d).
IC50 values were calculated as the concentrations of either α-synuclein monomers or HNE-induced oligomers needed to quench half of the ELISA signal and was used as an estimate of the antibody's affinity for the investigated antigen (Neri et al. 1996). For the oligomer-selective antibody mAb38F, the IC50 value for oligomers was 3 nM and for monomers 1 μM. For the other oligomer-selective antibody, mAb38E2, the IC50 value for oligomers was 10 nM and for monomers 1.5 μM. For the non-selective antibody mAb15, the IC50 value for oligomers was 8 nM and for monomers 5 nM. Taken together, mAb38F and mAb38E2 showed very strong affinity for HNE-induced α-synuclein oligomers as compared with monomers.
In this system, we also evaluated the binding of mAb38F and mAb38E2 to HNE-induced α-synuclein oligomers with the A30P mutation. We found that the antibodies bind almost equally well to oligomers of mutated α-synuclein and wild-type α-synuclein (Fig. 2a, b). In a separate inhibition ELISA, mAb38F and mAb38E2 were found to not recognize Aβ protofibrils (Fig. 3a, b).
To further evaluate the selectivity of the antibodies and to ensure that they do not merely recognize HNE-modified proteins, the reactivity of the antibodies against HNE coupled to BSA was evaluated. However, none of the antibodies was found to bind HNE-modified BSA (data not shown).
Antibody selectivity for α-synuclein versus other related proteins
Aggregated species of Aβ, tau, IAPP, as well as native monomers of β-synuclein and γ-synuclein, were blotted onto a nitrocellulose membrane to evaluate if the antibodies also bind other amyloidogenic or homologous proteins. The oligomer-selective antibodies (mAb38F and mAb38E2) were found to only bind α-synuclein (Fig. 4), whereas the non-selective antibody mAb15 was found to bind α-synuclein and to some extent also β-synuclein, but not any of the other proteins (Fig. 4).
Immunostaining of human brain with α-synuclein oligomer/protofibril-selective antibodies
Brain sections from PD (n = 2) and DLB (n = 2) patients were examined and compared with control subjects (n = 3) with immunohistochemical staining, using the two oligomer-selective antibodies (mAb38E2 and mAb38F) as well as two non-oligomer-selective α-synuclein antibodies (mAb15 and Syn-1). In substantia nigra of all PD and DLB cases, the number of pigmented cells was reduced as compared with control cases. Moreover, probing with mAb38E2 and mAb38F resulted in distinct staining of Lewy bodies and neurite-like structures in both substantia nigra and cortex of the PD and DLB cases (Fig. 5a–c and e–g). In the same cases, also Syn-1 resulted in stained Lewy bodies and neurite-like structures, albeit with a more pronounced background staining (Fig. 5m–o). The antibody mAb15 gave rise to a similar staining pattern as mAb38E2 (Fig. 5i–k). In the control brains, neuromelanin could be easily observed and none of the antibodies showed any α-synuclein immunoreactivity (Fig. 5d, h, l, p). The sections incubated with primary or secondary antibodies alone, or with irrelevant isotype-matched antibodies, did not show any immunoreactivity (data not shown).
In addition, we also stained consecutive sections with mAb38F and with 4B12, a C-terminal-reactive α-synuclein antibody. By doing so, we were able to stain the same Lewy bodies with both antibodies, clearly showing that they recognize the same α-synuclein deposits in substantia nigra of this PD brain (Fig. 6).
Immunostaining of transgenic mouse brain with α-synuclein oligomer/protofibril-selective antibodies
Brain sections from (Thy-1)-h [A30P] α-synuclein mice at 4, 12, and 18 months of age were examined with immunohistochemistry using five different α-synuclein antibodies. Apart from the oligomer/protofibril-selective antibodies mAb38E2 and mAb38F, we included the non-oligomer-selective antibodies mAb15 and Syn-1 as well as PS129, targeting phosphorylated α-synuclein. Remarkably, mAb38E2 and mAb38F recognized structures in the upper part of the brainstem already at 4 months of age (Fig. 7a, d). In mid-aged mice, at 12 months of age, the staining was stronger and included the lower parts of the brainstem, the mid-brain (Fig. 7b, e), and the white matter of the cerebellum (data not shown). In 18-month-old mice, the staining with mAb38E2 and mAb38F was even more pronounced in all areas and was also present throughout the brainstem (Fig. 7c, f).
In contrast, a diffuse staining pattern was observed at all ages when using Syn-1. A distinct staining, resembling that seen with the oligomer-selective antibody, was not visible at 4 months of age with Syn-1 (Fig. 7j), but had started to appear at 12 months of age (Fig. 7k), although not as prominently as with the oligomer-selective antibody. When staining with the non-selective α-synuclein antibody mAb15 (Fig. 7g–i), a similar pattern was observed as for Syn-1. The diffuse staining pattern seen with Syn-1 and mAb15 at 4 months is interpreted as specific staining of monomeric α-synuclein rather than unspecific background staining because incubation with the same antibodies on brain sections from wild-type mice did not generate any signals.
Apart from age-related alterations, there was a marked difference between the antibodies in terms of intracellular staining. The oligomer/protofibril-selective antibodies mAb38E2 and mAb38F gave rise to a rounded cytosolic staining, typically localized in the vicinity of the cell nuclei or in a neuritic staining pattern (Fig. 7b, c and e, f). When using either of the two non-oligomer-selective α-synuclein antibodies mAb15 and Syn-1, the staining pattern was generally more diffuse (Fig. 7h, i and k, l). With the well-characterized polyclonal antibody PS129 against phosphorylated α-synuclein, a faint dot-like immunostaining was seen throughout cortex, in hippocampus as well as in the olfactory bulb at 4 months of age, with a more pronounced staining at 12 and 18 months (data not shown). In addition, in 18-month-old mice an intense, neuritic-like staining could be observed throughout the brainstem and midbrain (Fig. 7o). The sections incubated with primary or secondary antibodies alone or with irrelevant isotype-matched antibodies did not show any immunoreactivity. A wild-type mouse (C57/Bl 6) was incubated with all of the above antibodies and did not show any staining (data not shown) (Fig. 7).
Measurement of α-synuclein oligomers/protofibrils in transgenic mouse brain
The oligomer-selective ELISA based on mAb38F, both as capturing and detecting antibody, was used to measure levels of α-synuclein oligomers/protofibrils in both brain and spinal cord homogenate from (Thy-1)-h [A30P] α-synuclein transgenic mice of different ages.
In 3-month-old (Thy-1)-h [A30P] mice, the mean concentration of α-synuclein oligomers/protofibrils in the brain fraction extracted with TBS was 118 ± 67 pM, with Triton X-100 194 ± 60 pM, and with SDS 1 ± 3 pM (Fig. 8a). At 15 months of age, oligomeric/protofibrillar levels in both the TBS and Triton X-100 fractions were found to be elevated (1021 ± 1240 pM and 460 ± 502 pM) as compared with levels in 3-month-old mice. In the SDS fraction from the 15-month-old mice, the levels of α-synuclein oligomers/protofibrils (1341 ± 2046 pM) showed large variability and two mice had elevated levels of SDS-resistant α-synuclein as compared with 3-month-old mice (Fig. 8b).
In spinal cord, the levels in young mice were higher in the TBS and Triton-X 100 fraction (321 ± 426 pM and 563 ± 99 pM) compared with the brain. No oligomers/protofibrils could be detected in the SDS fraction (Fig. 8c). The levels of oligomers/protofibrils in 15-month-old mice were 820 ± 915 pM in the TBS fraction, 1265 ± 874 pM in the Triton-X 100 fraction, and 2732 ± 5261 pM in the SDS fraction (Fig. 8d). In the SDS fraction from 15-month-old mice, the same two animals showed high signals both in brain and spinal cord. Thus, there was a marked difference in the brain and spinal cord levels of α-synuclein oligomers/protofibrils between different fractions and ages, with a trend toward higher levels of α-synuclein oligomers/protofibrils in older mice. No oligomers/protofibrils or only small amounts of such species (0 pM, 86 pM, and 0 pM in TBS, Triton and SDS fraction, respectively) could be detected in a non-transgenic mouse (data not shown).
Measurements of α-synuclein in the formic acid fraction
To further characterize the distribution of α-synuclein at different ages of the (Thy-1)-h [A30P] α-synuclein mice, the total amount of α-synuclein in the formic acid fraction, representing fibrillar protein, of 3- and 15-month-old mice was measured by total ELISA. In the brain-derived samples, the levels were significantly elevated in 15-month-old mice (15 ± 10 nM) as compared with 3-month-old mice (1 ± 0.6 nM), p = 0.0094 (Fig. 9a). A similar significant elevation was observed in samples from the spinal cord, where 15-month-old mice had elevated levels (24 ± 15 nM) compared with 3-month-old mice (1 ± 0.2 nM), p = 0.0036 (Fig. 9b).
Alpha-synuclein oligomers/protofibrils in subcellular compartments of transgenic mouse brain
The levels of α-synuclein oligomers/protofibrils in the ER, mitochondriae and nuclei from brain samples of (Thy-1)-h [A30P] α-synuclein mice were measured using the oligomer-selective ELISA. In the ER fraction, there was a significant elevation of oligomers in 12-month-old mice (13 ± 5 pM) compared with 18-month-old mice (7 ± 0.7 pM), p =0.0332, but not compared with 6-month-old mice (9 ± 2 pM)) (Fig. 10a). In the mitochondrial fraction, the levels were similar between the ages (78 ± 20 pM in 6-month-old mice and 80 ± 38 pM and 86 ± 37 in 12- and 18-month-old mice, respectively) (Fig. 10b). Also in the nucleus fraction, the levels of α-synuclein oligomers/protofibrils were similar between the three different age groups (43 ± 12 pM, 43 ± 17 pM and 34 ± 10 pM for 6-, 12-, and 18-month-old mice, respectively) (Fig. 10c).
The TBS fraction from the homogenates of the other hemisphere from the same mice was used to determine total α-synuclein levels. There were no statistical differences in total α-synuclein levels between the age groups (11 ± 2 μM, 13 ± 2 μM, and 11 ± 2 μM for 6-, 12-, and 18-month-old mice, respectively) (Fig. 10d).
Previous studies aimed at detecting α-synuclein oligomers/protofibrils have relied on either western blotting (Sharon et al. 2003) or ELISAs designed to measure a multitude of different α-synuclein species, ranging from dimers to multimeric aggregates (Paleologou et al. 2009; Tokuda et al. 2010). In this study, two novel monoclonal antibodies highly selective for large α-synuclein oligomers are described. Importantly, these antibodies are not cross-reacting with aggregated forms of other amyloidogenic proteins, such as Aβ, tau and IAPP, and do not recognize the other synuclein proteins, β-synuclein and γ-synuclein. Hence, they do not recognize a general amyloid epitope and can therefore be viewed as α-synuclein specific. Thus, these antibodies are considered to be useful to accurately assess the presence and/or quantity of α-synuclein oligomers/protofibrils in biological samples. A previously reported polyclonal antibody, A11, was also described to recognize α-synuclein oligomers, but was not able to discriminate between such species and oligomers of other proteins known to form deposits in various disorders (Kayed et al. 2010).
Another polyclonal antibody, recognizing both oligomers and ready-formed fibrils of α-synuclein (Fila-1), has been used to assess α-synuclein oligomers in post-mortem DLB brain with an indirect ELISA, in which the proteins were fixed to a surface at analysis (Paleologou et al. 2009). Moreover, a human single-chain antibody fragment, specifically binding oligomeric α-synuclein, has also been developed and was reported to inhibit aggregation and prevent oligomer-induced toxicity (Emadi et al. 2007).
To accurately evaluate antibody affinity for certain conformational or aggregational states, it is essential to adopt methods where physiologically relevant epitopes will be accessible for the antibodies. Applying inhibition ELISA (Englund et al. 2007), which detects oligomers and monomers in solution, both the mAb38F and mAb38E2 antibodies were shown to be truly selective with much higher affinity for α-synuclein oligomers than for monomers.
As it has been proven difficult to generate pure and stable oligomeric preparations of α-synuclein, recombinant α-synuclein was coincubated with an oligomer-promoting agent. Among the different molecules shown to induce α-synuclein oligomerization, HNE was selected. This reactive aldehyde can be formed from lipid peroxidation as a result of oxidative stress, a process thought to be relevant in the pathophysiology of Lewy body disorders. Increased presence of HNE has indeed been demonstrated in the PD brain (Yoritaka et al. 1996), indicating that the use of this agent may mimic a naturally occurring process by which oligomers are formed in the diseased brain. To further ensure that our antibodies did not have any affinity to HNE modifications per se, HNE-modified BSA was coincubated with the antibodies and no binding was detected.
In this study, the aldehyde-induced aggregates used as antigen are not on the pathway of forming fibrils and are hence described as large oligomers (Nasstrom et al. 2009, 2011). However, although the in vitro characterization of the antibodies were performed only on oligomers, the generated antibodies are likely to also recognize large on-pathway oligomers, i.e., protofibrils.
Applied for immunohistochemistry, the generated antibodies were shown to clearly recognize Lewy bodies and neurite-like structures in brains from patients with Lewy body disorders, thus indicating that they recognize physiologically relevant aggregates. The oligomer-selective antibodies mAb38E2 and mAb38F were able to distinguish intracellular, presumably pathological, accumulation of α-synuclein already at 4 months of age in (Thy-1)-h [A30P] α-synuclein transgenic mice. Interestingly, neither the Syn-1 antibody nor other previously tested α-synuclein antibodies (Freichel et al. 2007; Schell et al. 2009) could detect any pathology at this age in these mice. It can therefore be speculated that the oligomer/protofibril-selective antibodies enable detection of early pathology that cannot be identified by α-synuclein antibodies raised against linear epitopes. With increasing age, immunostaining with the oligomer-selective antibodies resulted in a more robust staining with the appearance of both punctate and neuritic structures in the mouse brain tissues. Interestingly, when comparing the staining pattern between mAb38E2/mAb38F and phospho-129 antibodies in 18-month-old mice, it appeared as if the respective antibodies did not stain exactly the same structures, which could indicate that most of the oligomeric α-synuclein is not phosphorylated.
On human brain tissue, the new antibodies were shown to recognize both Lewy bodies and Lewy neurites. There were no obvious differences in staining with the oligomer-selective mAb38E2 and mAb38F as compared with the non-selective mAb15. This might be explained by the fact that human tissues represent the end stage of the disease. Possibly, the oligomer-selective antibodies might be able to recognize earlier types of pathology than traditional antibodies on tissues from cases at less advanced disease stages.
The antibody mAb38F was also used in a sandwich ELISA to measure oligomeric/protofibrillar α-synuclein in (Thy-1)-h [A30P] α-synuclein transgenic mice. In addition to the brain, spinal cord samples were analyzed as there are studies reporting the presence of α-synuclein pathology in human spinal cord (Oinas et al. 2010; Del Tredici and Braak 2012). Both in brain and spinal cord, oligomers/protofibrils were detected at a high picomolar range already at 3 months of age. The levels of oligomers/protofibrils in the TBS and Triton-X fractions were increased in older mice and some mice had started to accumulate SDS-resistant oligomers/protofibrils at 15 months of age. Therefore, the levels of oligomers/protofibrils increase with the disease progression and soluble oligomeric/protofibrillar species seem to evolve into a more insoluble state as these mice age. To further verify that α-synuclein oligomers/protofibrils and aggregates increase with age, the total levels of α-synuclein in the formic acid fraction (representing aggregated forms of α-synuclein) were measured. As compared with 3-month-old mice, a significant elevation was seen in 15-month-old mice, which strengthens the observed increase seen in the oligomer/protofibril measurements.
To further decrease the risk of targeting monomeric protein, the sandwich ELISA employs mAb38F both as the capture and the detection antibody. As ELISA standard, HNE-induced α-synuclein oligomers were used, which have been shown to be stable in vitro (Nasstrom et al. 2011) and are therefore unlikely to dissociate during analysis. Consequently, the OD signal from the samples can be related to an accurate concentration of α-synuclein oligomers/protofibrils. As the (Thy-1)-h [A30P] α-synuclein transgenic mice model used in this study is based on the over-expression of A30P α-synuclein, the binding of mAb38F to HNE-induced A30P α-synuclein oligomers was also evaluated using an inhibition ELISA. The antibody was then found to bind almost equally well to oligomers, regardless of whether these consist of wild-type α-synuclein or the α-synuclein A30P mutant.
To explore the intracellular distribution of α-synuclein oligomers/protofibrils, a subcellular fractionation protocol was used. Interestingly, α-synuclein oligomers/protofibrils were increased in the ER fraction in 12-month-old mice compared with 18-month-old mice, although the difference compared with 6-month-old mice did not quite reach statistical significance. As we could not detect any differences in total α-synuclein between the age groups in these mice, we believe that the observed increase of oligomers/protofibrils in 12-month-old mice can not be explained by a general increase in total α-synuclein at this age. For the other subcellular samples, i.e., the mitochondrial and nuclear fractions, the levels of oligomers/protofibrils did not change with age. The observed increase of α-synuclein oligomers/protofibrils in the ER in 12-month-old mice is intriguing, as it coincides with the time point when these mice develop behavioral disturbances (Schell et al. 2009). Similar to our results, a previous study also found an increase in α-synuclein oligomers in the ER at the onset of symptoms in that mouse model (Colla et al. 2012a,b).
There is a need for novel therapeutics against Lewy body disorders, as existing therapies only replace neurotransmitter deficiency to alleviate the symptoms. Given its central pathogenic role, α-synuclein should be an attractive target and a recent study found that intraperitoneally administered α-synuclein antibodies decreased the levels of calpain-cleaved α-synuclein in the brain of transgenic mice (Masliah et al. 2011). Yet another recently published study describes the involvement of microglia in the antibody-assisted clearance of α-synuclein (Bae et al. 2012). However, as the etiological process in Lewy body disorders is likely to start years before the age at onset of symptoms, a successful immunotherapy needs to be directed against early pathogenic α-synuclein species. With more advanced protein pathology, i.e., when aggregated α-synuclein species have begun to deposit, extensive neuronal loss may already have occurred. As their presence seems to precede the formation of fibrillar species, oligomers/protofibrils of α-synuclein should be a suitable target and antibodies directed against such protein forms could be useful for immunotherapy.
In addition to their potential use for therapy, oligomer-selective antibodies could also serve to detect a relevant biomarker for the actual disorders. A previous study found elevated levels of α-synuclein oligomers in CSF from PD patients (Tokuda et al. 2010), although the ELISA method was based on non-oligomer/protofibril-selective antibodies. In addition, several studies (Tokuda et al. 2006; Mollenhauer et al. 2008; Parnetti et al. 2011) have shown decreased levels of total α-synuclein in patients with synucleinopathies, as compared with non-neurological control subjects. If it could be established that the levels of presumably early pathogenic protein forms are elevated in patients, immunoassays selectively measuring such species could serve as tools for early diagnosis and to guide the decision on when preventive therapy should be initiated.
To conclude, this study demonstrates the generation of monoclonal antibodies highly selective against oligomeric/protofibrillar forms of α-synuclein present in the brains of patients with Lewy body disorders, as well as in transgenic α-synuclein mice. Antibodies against such protein forms, believed to be central in the pathogenesis, could prove to be useful both for immunotherapy and in assays to assess a novel disease biomarker.
We thank Valentina Screpanti Sundquist, Monica Ekberg, and Sofie Ingvast for valuable help with the study. The work was supported financially by grants from Swedish Research Council (2006-2822(LL); 2006-6326 and 2006-3464(MI)), Uppsala Berzelii Technology Center for Neurodiagnostics, Swedish Brain Foundation, Swedish Alzheimer Foundation, Swedish Parkinson Foundation, Swedish Society of Medicine, Hans-Gabriel and Alice Trolle Wachtmeister's Foundation for Medical Research, Lundbeck Foundation, private donations from Lennart and Christina Kalén, Hans and Helen Danielsson, Stohne's Foundation, Söderström-Königska Foundation, Swedish Dementia Foundation, Björklund's Foundation for ALS research, Magn Bergwall Foundation, Thore Nilsson Foundation, Old Servants' Foundation, Åhlén Foundation, Loo and Hans Osterman's Foundation, Jeansson's Foundation, Larsson-Röst's Foundation, Golje's Foundation, and Göransson Sandvikens Foundation.
We have read the journal's policy and have the following disclosures: EN, AL, SMET, CS, AK, JA, MH, CM, and PG are employed by BioArctic Neuroscience AB. LL is Chairman of the Board of BioArctic Neuroscience AB.
MI, JB, LL, PG, CM, MH, EN, JB, and CS conceived and designed the experiments; TF, VL, EN, AL, SMET, XS, AK, JA, HK, HW, TN, JB, and CS performed the experiments; TF, VL, EN, AL, SMET, XS, JA, HK, JB, and MI analyzed and interpreted the data; TF, MI, VL, and EN wrote the paper; PJK, HS, and HK critically revised the manuscript and gave conceptual advice; final approval of the version to be published was obtained from all authors.