Identification of Low Molecular Weight Pyroglutamate A Oligomers in Alzheimer Disease A NOVEL TOOL FOR THERAPY AND DIAGNOSIS*□

Oliver Wirths, Christian Erck, Henrik Martens, Anja Harmeier, Constanze Geumann, Sadim Jawhar, Sathish Kumar , Gerd Multhaup, Jochen Walter , Martin Ingelsson**, Malin Degerman-Gunnarsson**, Hannu Kalimo, Inge Huitinga, Lars Lannfelt**, and Thomas A. Bayer From the Department of Molecular Psychiatry and Alzheimer, Graduate School, University Medicine Goettingen, 37075 Goettingen, Germany, Synaptic Systems, 37079 Göttingen, Germany, the Institute of Chemistry and Biochemistry, Free University of Berlin, 14195 Berlin, Germany, the Department of Neurology, Molecular Cell Biology, University of Bonn, 53127 Bonn, Germany, the **Department of Public Health and Caring Sciences Rudbeck Laboratory, Uppsala University, SE 75185 Uppsala, Sweden, the Department of Pathology, University of Helsinki, FI-00014 Helsingin yliopisto, Finland, and the Netherlands Brain Bank, Netherlands Institute for Neuroscience, 1105 BA Amsterdam, Netherlands

N-terminally truncated A␤ peptides starting with pyroglutamate (A␤pE3) represent a major fraction of all A␤ peptides in the brain of Alzheimer disease (AD) patients.A␤pE3 has a higher aggregation propensity and stability and shows increased toxicity compared with full-length A␤.In the present work, we generated a novel monoclonal antibody (9D5) that selectively recognizes oligomeric assemblies of A␤pE3 and studied the potential involvement of oligomeric A␤pE3 in vivo using transgenic mouse models as well as human brains from sporadic and familial AD cases.9D5 showed an unusual staining pattern with almost nondetectable plaques in sporadic AD patients and non-demented controls.Interestingly, in sporadic and familial AD cases prominent intraneuronal and blood vessel staining was observed.Using a novel sandwich ELISA significantly decreased levels of oligomers in plasma samples from patients with AD compared with healthy controls were identified.Moreover, passive immunization of 5XFAD mice with 9D5 significantly reduced overall A␤ plaque load and A␤pE3 levels, and normalized behavioral deficits.These data indicate that 9D5 is a therapeutically and diagnostically effective monoclonal antibody targeting low molecular weight A␤pE3 oligomers.
Alzheimer disease (AD) 3 represents the most frequent form of dementia and is characterized by the presence of extracellular amyloid plaques composed of amyloid-␤ (A␤) surrounded by dystrophic neurites and neurofibrillary tangles.The discovery that certain early-onset familial forms of AD may be caused by enhanced levels of A␤ peptides have led to the hypothesis that amyloidogenic A␤ is intimately involved in the AD pathogenic process (1).In the past extracellular A␤ has been regarded as the major culprit, whereas more recent evidence now points to toxic effects of A␤ in intracellular compartments (2)(3).In addition, other concepts propose that the soluble oligomers and the ␤-sheet containing amyloid fibrils are the toxic forms of A␤ (4 -6).Supporting this notion, it has been demonstrated that soluble oligomeric A␤42, but not plaque-associated A␤, correlates best with cognitive dysfunction in AD (7)(8).Oligomers are formed preferentially intracellulary within neuronal processes and synapses rather than extracellularly (9 -10).Besides full-length A␤ peptides starting with an aspartate at position 1, a variety of different N-truncated A␤ peptides have been identified in AD brains.Ragged peptides including phenylalanine at position 4 of A␤ have been reported as early as 1985 by Masters et al. (11).In contrast, no N-terminal sequence could be obtained from cores purified in a sodium dodecyl sulfate-containing buffer, which led to the assumption that the N terminus could be blocked (12)(13).The presence of A␤pE3 (N-terminally truncated A␤ starting with pyroglutamate) in AD brain was subsequently shown using mass spectrometry of purified A␤ peptides, explaining at least partially initial difficulties in sequencing A␤ peptides purified from human brain tissue (14).The authors reported that only 10 -15% of the total A␤ isolated by this method begins at position 3 with A␤pE3.Saido et al. (15) subsequently showed that A␤pE3 represents a dominant fraction of A␤ peptides in senile plaques of AD brains.Recently, we generated a new mouse model selectively expressing A␤pE3-42 in neurons, and demonstrated for the first time that this peptide is neurotoxic in vivo leading to neuron loss and an associated neurological phenotype (16).Recently, it has been demonstrated that the N-terminal pEformation can be catalyzed by glutaminyl cyclase (QC), which can be pharmacologically inhibited by QC inhibitors, both in vitro (17) and in vivo (18).QC expression was found up-regulated in the cortex of patients with AD and correlated with the appearance of pE-modified A␤.Oral application of a QC inhibitor resulted in reduced A␤pE3-42 burden in two different transgenic mouse models of AD as well as in a transgenic Drosophila model.Interestingly, treatment of these mice was accompanied by reductions in A␤x-40/42, diminished plaque formation and gliosis, as well as improved performance in context memory and spatial learning tests (18).Thus, A␤pE3-42 variants are promising targets in both therapeutic and diagnostic strategies of AD.
Western Blotting of Synthetic Peptides-For Western blot analysis, 1.5 g of peptides were loaded on 4 -12% vario gels (Anamed), transferred to 0.45 m nitrocellulose membranes and detected using the primary antibodies W0 -2 (1 g/ml) and 9D5 (10 g/ml) in blocking buffer.The blots were developed using enhanced chemiluminescence.
Thioflavin T Aggregation Assay-Peptides were solubilized in 10 mM NaOH at a concentration of 1 mg/ml, sonicated for 5 min, frozen in liquid nitrogen, and stored at Ϫ80 °C until use.Aggregation of A␤ peptides was investigated online using ThT aggregation assay (Varian fluorescence spectrophotometer) using an excitation wavelength of 446 nm and emission wavelength of 482 nm.Samples contained 55 M A␤, 50 mM sodium phosphate buffer (pH 7.4), 50 mM NaCl, 20 M ThT, and 0.01% sodium azide.The samples were incubated at 37 °C in a peltier adapter with stirring.Data points were recorded every 10 min during the assay.
Toxicity of Peptides on Neuroblastoma Cells-Toxicity was verified as previously published (6).Briefly, SH-SY5Y neuroblastoma cells were routinely cultured.After 48 h, medium was replaced by medium containing freshly dissolved peptides, each at 2 M concentration in the presence or absence of 1 ng/l 9D5 antibody and incubated for 12 h.Cell viability was determined using MTS assay (Promega), according to the manufacturer's instructions compared with vehicle-treated control cells.
ELISA of A␤pE Oligomers in Plasma-Plasma samples (stored at Ϫ70 °C) from patients with AD and healthy controls (HC) were analyzed.The patients were recruited at the Memory Clinic at the Department of Geriatrics, Uppsala University Hospital.All AD patients were diagnosed according to DSM IV and NINCDS-ADRDA.A␤pE oligomer levels in human plasma samples were measured by ELISA according to standard methods.Briefly 9D5 antibody was coated as capture antibody and blocked with 5% skimmed milk, 0.05% Tween in PBS.Biotinylated antibody 2-48 against pE-A␤ in combination with streptavidin-HRP and the chromogen TMB (Pierce) was used for detection.
ELISA of A␤pE Levels in Brain-Frozen brains were homogenized in a Dounce homogenizer in TBS (120 mM NaCl, 50 mM Tris pH 8.0 containing complete protease inhibitor (Roche)), and subsequently centrifuged at 27.000 g for 20 min at 4 °C.The resulting pellets were resuspended in 2% SDS, sonified and centrifuged for 15 min.A␤pE levels were measured by ELISA according to standard methods.NT78 (against A␤1-16) antibody was coated as capture antibody.Biotinylated antibody 2-48 (against N-terminal A␤pE3) was used as detection antibody in combination with streptavidin-HRP and the chromogen TMB (Pierce).
Human Brain Samples-Human brain samples were obtained from the Netherlands Brain Bank (NBB), Hopital del la Salpetrie `re (a generous gift of Prof. Dr. Charles Duyckaerts and Dr. Veronique Sazdovitch), University Hospital Helsinki and from Uppsala University.Definite diagnosis was based on established criteria and informed consent was obtained from all subjects.
Transgenic Mice-APP/PS1KI (23) and 5XFAD (24) female bigenic mice have been described previously.All mice were backcrossed for more than 10 generations on a C57BL/6J genetic background and housed at a 12-h day/12-h night cycle with free access to food and water.For passive immunization 4.5-month-old 5XFAD mice were weekly injected with 250 g of 9D5 or PBS intraperitoneally for 6 weeks.9D5 antibody was purified by protein G-agarose.250-ml culture supernatant was applied to the column and allowed to drain through.The column was washed with 200 ml of PBS, and the antibodies were eluted with 0.1 M glycine (pH 2.5), neutralized with 100 l of 1.5 M Tris/HCL buffer (pH 8.8).Samples were measured at 280 nm.Eluted IgG fractions containing the highest absorptions were pooled and dialyzed with PBS.PBS injection has been used previously as a control for treatment effects of passive immunization of different AD mouse models with a variety of A␤ antibodies (25)(26)(27)(28)(29).All animals were handled according to German guidelines for animal care and studies were approved by the local legal authorities (LAVES).
Immunohistochemistry-Human and mouse tissue were processed as described previously (19).In brief, 4-m paraffin sections were pretreated with 0.3% H 2 O 2 in PBS to block endogenous peroxidases, and antigen retrieval was achieved by boiling sections in 0.01 M citrate buffer pH 6.0, followed by 3-min incubation in 88% formic acid.Primary antibodies were incubated overnight, followed by incubation with biotinylated secondary antibodies (DAKO) before staining was visualized using the ABC method with the Vectastain kit (Vector Laboratories) and diaminobenzidine (DAB) as chromogen.Alternatively, fluorochromated secondary antibodies (anti-mouse AlexaFluor594 and anti-rabbit AlexaFluor488, Invitrogen) were used for immunofluorescence detection.
Quantification of Plaque Load-Extracellular A␤ load (4G8, G2-10, G2-11, 2-48) was evaluated in cortex and hippocampus using an Olympus BX-51 microscope equipped with an Olympus DP-50 camera and the ImageJ software (V1.41,NIH).Serial images of 40ϫ magnification (hippocampus) and 100ϫ (cortex) were captured on six sections per animal, which were 30 m afar from each other.Using ImageJ the pictures were binarized to 16-bit black and white images and a fixed intensity threshold was applied defining the DAB staining.
Behavioral Testing-Anxiety levels were assessed using an elevated plus maze as described previously (30).The percentage of the time spent in the open arms to the overall time and the ratio of the open arms to the total arms entries were measured using an automatic video tracking system (Video-Mot2, TSE-Systems).
Statistical Analysis-Statistical differences were evaluated using one-way ANOVA followed by Bonferroni post-hoc test or unpaired t test as indicated.All data are given as means Ϯ S.E. of the mean (S.E.).All statistics were calculated using GraphPad Prism V5.00 software.
Antibody 9D5 Inhibited Aggregation and Toxicity in Vitro-The aggregation of monomeric A␤1-42 and A␤pE3-42 peptides (55 M) was investigated using a ThT fluorescence assay.While A␤1-42 showed the expected aggregation profile with a pronounced lag phase before fibril growth, A␤pE3-42 showed very rapid formation of intermediate oligomeric assemblies.Interestingly, elongation rates of A␤pE3-42 were much slower as that of A␤1-42 (Fig. 1D).These data indicate that A␤pE3-42 rapidly formed intermediate oligomeric assemblies, but has decreased propensity to form larger fibrils, a behavior that clearly differs from that of that of A␤1-42.Notably, the presence of antibody 9D5 efficiently decreased the formation of higher aggregates of the A␤pE3-42 peptide at a 1:76 (9D5:A␤) ratio, but not the rapid formation of lower oligomers, further demonstrating the specificity of this antibody for lower oligomeric species of A␤pE3 and its efficiency in the inhibition of further peptide aggregation (Fig. 1E).This observation suggests that 9D5 inhibits the formation of higher A␤ aggregates by binding to LMW oligomers as indicated in SEC and Western blot experiments.We next studied the toxicity of A␤1-42 and A␤pE3-42 peptides in SH-SY5Y neuroblastoma cells.To determine whether the toxic effect of A␤pE3-42 can be influenced by 9D5 antibody, we incubated neuroblastoma cells either with A␤1-42, A␤pE3-42 or with peptides and 9D5.Application of both, A␤pE3-42 and 9D5 completely abolished the toxic effect of A␤pE3-42.This effect is highly specific, as application of A␤1-42 together with 9D5 caused the same effect on cell viability as A␤1-42 incubation alone (Fig. 1F).
Antibody 9D5 Shows a Specific Staining Profile in Alzheimer Brain-9D5 was used to characterize the distribution of oligomeric A␤pE in human post-mortem brain tissue (frontal cortex and hippocampus from sporadic AD, familial AD (FAD) and non-demented individuals) (Fig. 2, supplemental Table S1 and Fig. S7).While none of the non-demented controls showed plaque staining with 9D5, some specimen showed plaque staining using 4G8 (against A␤17-24).Occasionally weak 9D5 blood vessel immunostaining was observed.This observation demonstrated that plaques in healthy controls do not harbor the 9D5 epitope and indicates that plaques in nondemented controls do not contain oligomeric A␤pE3.In contrast, most of the sporadic AD and FAD cases demonstrated high abundance of intraneuronal and cerebral amyloid angiopathy (CAA) staining with 9D5 (Fig. 2, A and B), clearly differing from the 4G8 pattern (e.g.sporadic case 1 and 2, Fig. 2A).FAD cases having mutations in the APP gene (Swedish or arctic mutation) revealed abundant 9D5 immunoreactivity.Of interest, all analyzed FAD cases harboring mutations in the presenilin-1 gene (P264L, L418F, PS1⌬exon9) showed prominent intraneuronal 9D5 immunoreactivity (Fig. 2B, supplemental Table S1).Oligomeric A␤pE3-42 antibody 9D5 showed a specific staining pattern different from the staining pattern with an antibody against the N terminus of A␤pE3-42 (supplemental Fig. S3).

Lower Levels of Oligomeric A␤pE3 in Plasma of Alzheimer
Patients-To assess the diagnostic potential of oligomeric A␤pE3 variants and antibody 9D5, we established a novel ELISA and tested plasma of AD patients and healthy controls (HC).Interestingly, levels of 〈␤p⌭3 oligomers were significantly reduced in AD patients by 46% as compared with healthy controls (p Ͻ 0.05) (Fig. 2C).We have previously published that the level of IgM autoantibodies in plasma directed against A␤pE3 was significantly decreased in AD patients as compared with healthy controls.In good agreement with these observations, the signal of A␤pE3 oligomers detected by 9D5 was significantly lower in plasma of AD patients again pointing out that 9D5 can be used as a biomarker tool for AD diagnosis (36).We hypothesize that lower levels of A␤pE3 oligomers in plasma are due to development of cerebral amyloid angiopathy and increased accumulation within neurons (supplemental Fig. S6).Aggregates were very rapidly generated from A␤pE3-42, indicating an instant seeding of the aggregation process.A␤1-42 showed a typical lag phase, i.e. the phase in which oligomers and protofibrils are slowly formed, whereas A␤pE3-42 rapidly formed intermediate oligomeric assemblies, but has decreased propensity to form larger fibrils, a behavior that clearly differs from that of that of A␤1-42.E, accelerated increase after the inflection point at 200 min was efficiently blocked by addition of 9D5 together with A␤pE3-42.F, toxicity measurements of SH-SY5Y neuroblastoma cells incubated with A␤1-42 and A␤pE3-42 in addition with 9D5 antibody compared with vehicle control.Whereas A␤1-42 (with and without 9D5) and A␤pE3-42 displayed high toxic effects, A␤pE3-42 in the presence of 9D5 is not toxic (n ϭ 3-5; ANOVA, p ϭ 0.0001; followed by t test, p ϭ 0.0048).
Antibody 9D5 Shows a Specific Staining Profile in Transgenic Alzheimer Mouse Models-We next asked the question whether oligomeric A␤pE peptides could also be detected in transgenic mouse models for AD.Staining of 3-month-old 5XFAD mice did not show any immunoreactivity (Fig. 3A), whereas considerable staining was detected in the subiculum of 6-month-old 5XFAD mice (Fig. 3B), showing a dramatic increase at the age of 12 months (Fig. 3C).9D5 detected only intracellular and no plaque-associated staining corroborating the staining pattern in AD patients.In addition, other brain areas like cortex, pons or brainstem nuclei stained strongly positive at that time point (not shown).A very similar age-dependent accumulation of A␤pE3 was also observed in APP/ PS1KI mice, another model with robust neuron loss and associated behavioral deficits (23, 37) (supplemental Fig. S4).Double-staining using 9D5 (red) and the astrocytic marker GFAP (green) in the subiculum of 12-month-old 5XFAD mice revealed almost no co-localization in astrocytes (Fig. 3D).On the other hand, double-staining using 9D5 (red) and the microglia/macrophage marker Iba-1 (green) showed a strong co-localization, suggesting oligomeric A␤pE3 variants are internalized by microglia (arrowheads) (Fig. 3E).In addition, strong intraneuronal 9D5-immunoreactivity could be demonstrated (Fig. 3F, arrows).The finding of intraneuronal 9D5 staining is corroborated by strong 9D5-immunoreactivity in spinal cord motor neurons of aged APP/PS1KI mice (supplemental Fig. S4).
Therapeutic Effect of Passive Immunization with 9D5 in 5XFAD Mice-Next we studied a potential therapeutic effect of 9D5 using a passive immunization approach.4.5month-old 5XFAD (female) mice were weekly injected with 250 g of 9D5 intraperitoneally for 6 weeks.9D5 treatment significantly reduced generic A␤, A␤42, A␤40, and A␤pE3 plaque load in hippocampus (HC) and cortex (Ctx) (Fig. 4,  A and B).In good agreement, 9D5 treatment significantly stabilized the performance in the elevated plus maze (Fig. 4C).Confirming the plaque load data, a significant reduction of A␤pE3 levels was observed in both the TBS and SDS fraction of brain lysates after 9D5 immunization of 5XFAD mice (Fig. 4D).Passive immunization of 5XFAD mice with 9D5 showed reduction of the intracellular A␤pE3-42 oligomers (supplemental Fig. S5).

DISCUSSION
Soluble oligomers (also described as ADDLs and/or protofibrils) of A␤ have been discussed to be causally involved in synaptic and cognitive dysfunction in the early stages of AD (38 -39).However, there is no consensus on which aggregation state exerts the highest toxicity in AD.Nanomolar concentrations of small diffusible A␤ oligomers (17-27 kDa) cause neuronal death in hippocampal slice cultures (40) and A␤ dimers that were either cell-derived or extracted from AD brains impair synaptic plasticity (41).On the other hand, dodecameric A␤56* oligomers extracted from the brain of APP transgenic mice interfere with learning and memory performance in rat (42).Analysis of neurotoxicity of oligomers ranging from monomers to tetramers of synthetic A␤ peptides demonstrated that tetramers have the strongest effect (43).The conclusion that oligomers are more potent candidates as pathogens is based primarily on experimental evidence demonstrating that natural and synthetic A␤ oligomers impair synaptic plasticity (40 -41, 44), memory (33,42,44) and inducing loss of synapses (34,45) when applied exogenously into rat cerebral ventricle, cultured brain slices, or dissociated neurons.It has been shown that soluble oligomeric A␤42 and not plaque-associated A␤ correlate best with cognitive dysfunction (7)(8).A␤ oligomers are formed preferentially intracellularly within neuronal processes and synapses rather than within the extracellular space (9 -10).Tomiyama et al. generated APP transgenic mice expressing the E693⌬ mutation, which causes neuronal cell death and cognitive impairment by enhanced A␤ oligomerization without fibrillization.The mice displayed age-dependent accumulation of intraneuronal A␤ oligomers from 8 months but no extracellular amyloid deposits even at 24 months.Hippocampal synaptic plasticity and memory were impaired at 8 months of age (46).A␤ protofibril levels correlate with spatial learning in AD transgenic mice expressing human APP with the arctic mutation (47) facilitating early intraneuronal A␤ aggregation (48).Despite the difficulty to compare the different studies on oligomeric A␤ species there seems to be converging evidence that they: 1) are primarily formed within neurons, 2) oligomeric A␤ species are more neurotoxic than monomeric or fibrillar A␤ in vitro, and 3) oligomeric A␤ species decrease synaptic activity.
In the present study, we have identified LMW A␤pE3 oligomers, which can be detected by 9D5, a novel mouse monoclonal antibody.9D5 did not cross react with any A␤1-42 species indicating that these oligomers present a unique and novel epitope.The therapeutic potential of 9D5 was demonstrated in passively immunized 5XFAD mice as plaque load and A␤ levels were reduced and behavioral deficits were normalized.In an ELISA using 9D5 as capture antibody, we could show that the signal was significantly lower in plasma of AD patients as compared with non-demented controls.We believe that our observation represents a novel therapeutic mechanism rescuing AD pathology and related behavioral deficits.Several studies demonstrated that N-terminal specific A␤ antibodies showed significant beneficial effect in AD mouse models.Bard et al. (26) and Buttini et al. (49) studied the optimal antibody response for reducing neuropathology in PDAPP transgenic mice.Immune sera with reactivity against different A␤ epitopes and monoclonal antibodies with different isotypes were examined for efficacy and showed that antibodies against the N-terminal regions of A␤ were able to invoke beneficial effects.Saido et al. (15) suggested that A␤pE3-42 is generated step-by-step from its precursor A␤1-42 by N-truncation and glutamate to pyroglutamate formation.We therefore assume that reducing A␤1-42 by passive immunization (reviewed in (50)) will also reduce A␤pE3-42 levels and the resulting oligomeric forms.We think that A␤pE3 oligomers represent an important pathological step appearing at a time point when behavioral deficits occur.Interrupting this toxic pathway by specifically reducing these oligomers also has an impact on other A␤ peptides as shown for example in reducing general plaque load.In conclusion, we have therefore demonstrated for the first time that oligomeric A␤pE peptides represent a novel A␤ entity, which mice showing that the signal starts to appear at 6 months.D, double staining using 9D5 (red) and the astrocytic marker GFAP (green) in the subiculum of a 12-month-old 5XFAD mouse revealed no colocalization in astrocytes.E, in contrast, double-staining using 9D5 (red) and the microglia/macrophage marker Iba-1 (green) showed a strong colocalization in the subiculum of a 12-month-old 5XFAD mouse (arrowheads).F, strong intraneuronal 9D5-immunoreactivity could be demonstrated in the pons of a 12-month-old 5XFAD mouse.
can be detected by specific antibodies serving as promising tools for diagnosis and therapeutic intervention of AD.

FIGURE 3 .
FIGURE 3. Intracellular age-dependent staining of 〈␤p⌭3 oligomers.Staining with 9D5 in the subiculum of (A) 3-, (B) 6-, and (C) 12-month-old 5XFAD mice showing that the signal starts to appear at 6 months.D, double staining using 9D5 (red) and the astrocytic marker GFAP (green) in the subiculum of a 12-month-old 5XFAD mouse revealed no colocalization in astrocytes.E, in contrast, double-staining using 9D5 (red) and the microglia/macrophage marker Iba-1 (green) showed a strong colocalization in the subiculum of a 12-month-old 5XFAD mouse (arrowheads).F, strong intraneuronal 9D5-immunoreactivity could be demonstrated in the pons of a 12-month-old 5XFAD mouse.