Discovery and Functional Characterization of hPT3,a Humanized Anti-Phospho Tau Selective Monoclonal Antibody

Abstract

Background:

As a consequence of the discovery of an extracellular component responsible for the progression of tau pathology, tau immunotherapy is being extensively explored in both preclinical and clinical studies as a disease modifying strategy for the treatment of Alzheimer’s disease.

Objective:

Describe the characteristics of the anti-phospho (T212/T217) tau selective antibody PT3 and its humanized variant hPT3.

Methods:

By performing different immunization campaigns, a large collection of antibodies has been generated and prioritized. In depth, in vitro characterization using surface plasmon resonance, phospho-epitope mapping, and X-ray crystallography experiments were performed. Further characterization involved immunohistochemical staining on mouse- and human postmortem tissue and neutralization of tau seeding by immunodepletion assays.

Results and Conclusion:

Various in vitro experiments demonstrated a high intrinsic affinity for PT3 and hPT3 for AD brain-derived paired helical filaments but also to non-aggregated phospho (T212/T217) tau. Further functional analyses in cellular and in vivo models of tau seeding demonstrated almost complete depletion of tau seeds in an AD brain homogenate. Ongoing trials will provide the clinical evaluation of the tau spreading hypothesis in Alzheimer’s disease.

Keywords

Alzheimer’s disease immunotherapy monoclonal antibodies tau protein

INTRODUCTION

Despite the long-established identification of amyloid-β plaques and tau neurofibrillary tangles as the main neuropathological hallmarks of Alzheimer’s disease (AD) [1], it remains unclear how these proteinaceous deposits exert their neurotoxic effect. As a result of genetic research, a linear amyloid/tau cascade has been the major working hypothesis [2]. Driven by more recent genome-wide association studies pointing to involvement of neuroinflammation [3, 4] and the pre-clinical demonstration of prion-like seeding by amyloidogenic proteins (including tau) with elevated β-sheet propensity [5 –8], this model evolved into a more complex mechanism where plaque deposition is not positioned as trigger but as an accelerator of the spreading of tau pathology [9 –11]. Since tau deposition is, to date, the best predictor of cognitive decline in AD [12 –14], therapeutic strategies to attenuate or even stop the spreading of tau aggregation are an important focus for the next decade [15].

Tau is an intracellular microtubule-associated protein encoded by the MAPT gene from which six major splice variants are described in the mature human central nervous system, mainly restricted to neurons [16]. In these cells, tau functions as a regulator of protein transport across axons, and its association/dissociation from the microtubules is regulated by phosphorylation and dephosphorylation cycles [17]. In addition to the detachment from microtubules, increased tau phosphorylation has been postulated to facilitate its misfolding [18] into species that could act as prion-like seeds. These recruit additional tau monomers to form toxic insoluble tau aggregate species [7 , 19–22]. Although these seeds act intracellularly, the continuous release of tau species as recently described [23] assumes that these tau seeds are also released in the extracellular environment contributing to the spatiotemporal progression of tau pathology as described by Braak [24, 25]. Many pre-clinical models confirm the seeding potential of tau aggregates, and one of the most advanced therapeutic approaches targets the extracellular tau seeds by active- or passive immunization, which has been shown to be efficacious in preclinical models and is currently under clinical investigation [15].

Here we describe the generation of a novel phospho-tau selective monoclonal antibody PT3 as a result of immunization using human AD brain-derived paired helical filaments (PHFs) as described previously [26]. This molecule (and its humanized variant hPT3) showed pM affinity to human AD PHFs but also to non-aggregated phospho tau in brain homogenates from different species with exception from Cynomolgus monkey. Functional assays demonstrated neutralization by PT3 and hPT3 on AD tau seeds in vitro and in vivo.

MATERIALS AND METHODS

All in vivo experiments were conducted in strict accordance with the guidelines of the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC), with the European Communities Council Directive of 24 November 1986 (86/609/EEC) and with protocols approved by the local Institutional Animal and Use Ethical Committee.

Antibodies

Except where indicated otherwise, antibodies used in these studies are obtained by recombinant expression as mouse IgG2a variants. Antibody N(a) has been derived based on variable sequences from heavy and light chain described in [27]. PT82 is an antibody which is generated internally [28].

Immunizations

Immunizations in BALB/c mice (PT3) and Tau KO mice (The Jackson Laboratory, Bar Harbor, ME, USA) (PT79) with PHFs were performed using standard hybridoma technology [29]. Briefly, mice were immunized with 50μg of human PHF tau for each immunization. Before intraperitoneal injection, antigen was mixed with incomplete Freund’s adjuvant (Millipore Sigma, St Louis, MO, USA). Mice were boosted every two weeks with aggregated tau preparation in incomplete Freund’s adjuvant. Four days before spleen extraction, mice were boosted with aggregated tau preparation in saline.

Following this immunization regimen, spleen cells were harvested and fused with myeloma cells to generate hybridomas. Hybridomas were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum (Hyclone, Europe), Hybridoma Fusion Cloning Supplement (2%) (Roche Diagnostics, Mannheim, Germany), 2% HT (Sigma), 1 mM sodium pyruvate, 2 mM L-glutamine and penicillin (100 U/ml), and streptomycin (100μg/ml). Antibody variable regions were cloned from hybridoma cells, selected by a direct PHF ELISA screen, sequenced using standard methods, and subcloned into expression vectors for mAb and Fab. Mab was produced on a mouse IgG2a/κ background and expressed and purified by affinity chromatography (protein A). Fab was produced as chimeric versions with the mouse variable domains fused to human IgG1/κ constant domains and a His tag at the C-terminus of the heavy chain. Fab was transiently expressed in HEK293F cells and purified by affinity chromatography (HisTrap).

Humanization of PT3

The anti-tau mouse antibody PT3 was humanized using the Human Framework Adaptation (HFA) method [30]. To find the best combination of humanized heavy chain (HC) and light chain (LC), several HC and LC V-region sequences were tested as acceptor frameworks for complementarity-determining region (CDR) grafting. Selection of human germlines was based solely on the overall sequence similarity to the mouse antibody in the framework region (FR). Neither the CDR sequences, nor their length or canonical structures, were considered in this selection. The CDR definition used in HFA is described in [30]. The CDRs, as defined by the Chothia numbering scheme [31], are: HCDR1, 26–35; HCDR2, 50–58; HCDR3, 95–102; LCDR1, 24–34; LCDR2, 50–56; LCDR3, 89–97. After CDR grafting onto the acceptor human frameworks, all heavy and light chain pairs were generated as mAbs and tested for binding to ePHF-Tau and the best binding pair was chosen as the humanized lead molecule, hPT3.

SPR analysis

The interaction of anti-tau mAbs with PHF-tau was analyzed by ProteOn (BioRad, Hercules, CA, USA) as described previously [32]. The interaction of anti-tau Fabs or mAbs with recombinantly expressed control tau (human tau isoform 2N4R) was studied with a Biacore T200 (GE Healthcare, Marlborough, MA, USA). A biosensor surface was prepared by coupling an anti-mouse IgG Fc- or Fab-domain specific antibody to the surface of a CM5 sensor chip using the manufacturer’s instructions for amine-coupling chemistry (∼6500 RU). The coupling buffer was 10 mM sodium acetate, pH 4.5. The anti-tau Fabs or mAbs were diluted in the running buffer and injected to obtain a capture of at least 5 RU. Capture of anti-tau mAbs or Fabs was followed by injection of recombinantly expressed control tau in solution (0.12 to 75 nM in 5-fold dilutions). The association was monitored for 3 min (150μL injected at 50μL/min). The dissociation was monitored until the signal decreased by at least 5% for reasonable off-rate determination. Regeneration of the sensor surface was obtained with 0.85% phosphoric acid followed by 50 mM NaOH. The data for both mAbs and Fabs were fit using a 1:1 Langmuir binding model if binding was observed.

Epitope mapping

To reconstruct epitopes of the target molecule, a library of peptides (20-mers with an overlap of 18 amino acids) covering the Tau 441 sequence was performed using Pepscan’s proprietary Chemically Linked Peptides on Scaffolds (CLIPS) technology (Pepscan Presto B.V., Lelystad, the Netherlands). The binding of antibodies (recombinantly expressed as IgG2a) to each of the synthesized peptides was tested in a Pepscan-based ELISA. The peptide arrays were incubated with primary antibody solution (overnight at 4°C). After washing, the peptide arrays were incubated with a 1/1000 dilution of a peroxidase conjugated anti mouse antibody for 1 h at 25°C. After washing, the peroxidase substrate 2,2’-azino-di-3-ethylbenzthiazoline sulfonate (ABTS) and 20μl/mL of 3% H₂O₂ were added. After 1 h, the color development was quantified with a charge coupled device (CCD)-camera and an image processing system.

Detailed phosphopeptide mapping

Peptides for epitope mapping were synthesized by New England Peptide and contained biotin-PEG4 moiety at the N-terminus. The panel of peptides covered tau residues 202–225 and included combinations of phosphorylation at the following sites: T205, S208, S210, T212, S214, T217, T220.

MSD ELISA

Synthetic peptides were dissolved in carbonate/bicarbonate buffer, pH 9.4, to generate a stock solution of 1 mg/mL. Stock solutions were diluted to 5μg/mL for each peptide. Thirty microliters of each peptide were incubated with Streptavidin Gold Plates (MSD Cat# L15SA-1, Gaithersburg, MD) for 1 h at room temperature. One-hundred fifty microliters of 5% MSD Blocker A buffer was added to each well and incubated for 1 h at room temperature. Plates were washed three times with 0.1 M HEPES buffer, pH 7.4, followed by the addition of 40 nM to 320 pM Ruthenium (Ru)-labeled PT3 mAb (Lot # 121128-CP05879m) in a 25μL volume per well. Plates were washed three times with HEPES buffer, pH 7.4 followed by the addition of 150μL per well of diluted MSD Read buffer T and analyzed using a SECTOR imager.

SPR ProteOn phosphopeptide binding analysis

The kinetic rate constants and affinity of PT3 mAb and Fab were determined for 14 peptides using a Bio-Rad ProteOn XPR36. All peptides were from the panel listed in Table 1 and contained a biotin PEG4 moiety at the N-terminus. Biotinylated peptide was captured on a neutravidin-coated NLC biosensor chip and PT3 mAb or Fab was flowed over the surface to measure kinetic parameters. All experiments were performed at 25°C using PBS-T as both running buffer and sample dilution buffer. Prior to running samples, the NLC chip was conditioned by running PBST buffer over the chip surface for 1 h. Approximately 5–10 RU of peptide was captured on the chip surface by diluting peptide to 10 ng/mL in PBST and injecting over the flow channels at 30μL/min for 100 s. Serial dilutions of PT3 Fab (1.1 nM to 90 nM) and PT3 mAb (0.62 nM to 50 nM) were prepared, including a buffer only control for a total of 6 concentrations.

Table 1

Epitope mapping of PT3. From the 202-205 sequence in the 441 amino acid tau isoform, different variants were used for epitope mapping of PT3. Phosphorylated threonines are highlighted in yellow and phosphorylated serines are highlighted in red. Binding of PT3 to the indicated phospho-peptides was determined by MSD ELISA (mAb signals) and surface plasmon resonance (SPR) (Fab and mAb affinity)

After capture of biotinylated peptides, the antibody titration was injected at 60μL/min for 3 min (association phase), followed by 300 s of buffer only (dissociation phase). The chip surface was regenerated using a single injection of 0.85% phosphoric acid at 30μL/min for 100 s contact time, followed by 4 PBST injections before the next antibody titration injection.

The data processing and kinetic analysis were done using the instrument software. Measurements were performed in duplicate except for peptide NPT-2. The data were double referenced by subtraction of the interspot response and the curves generated by the buffer only injection. The data were analyzed using a simple Langmuir 1:1 binding model.

Western blotting

Brain tissue from WT (C57Bl/6J) and Tau KO mice, rat, rabbit, beagle dog, mini pig, marmoset, and cynomolgus monkey were obtained according to procedures approved by the local ethical committee and national institutions adhering to AAALAC guidelines. Human brain tissue was obtained from the IBB Biobank. For total homogenates, tissue was weighed and homogenized in 10 volumes of buffer H (10 mM Tris, 800 mM NaCl, 1 mM EGTA, and 10% sucrose/ pH 7.4). The homogenate was centrifuged at 27000×g for 20 min, and the supernatant was aliquoted and frozen at –80°C. To prepare heat-stable extract (HSE), homogenate was boiled for 5 min (100°C) and cooled on ice for 10 min. After ultracentrifugation (150000×g; 4°C; 2 h) supernatants containing HSE were aliquoted and frozen at –80°C. Samples were diluted in sample buffer and loaded on SDS (4–12%) or native (3–12%) PAGE (Life Technologies, Thermo Scientific) according to the manufacturer’s instructions. After the separation, the gel was blotted on a nitrocellulose membrane which was blocked with TBS-T containing 5% Non-fat dry milk. Blots were incubated with non-labelled primary antibody solutions (overnight at 4°C) and detected by HRP-labelled anti mouse antibodies. Incubation with HRP-labelled tau antibodies was performed during 2 h at RT. In both cases, detection was done with West Dura (Thermo Scientific).

Fab crystallization

PT3 Fab was expressed and purified at Sino Biologicals in a final buffer of 20 mM Tris, 50 mM NaCl, pH 7.4. hPT3 Fab was produced in-house in a final buffer of 20 mM MES, 0.2 M NaCl, pH 6.0. PT3 epitope peptide for crystallization with sequence Ac-SRpTPSLPpTPPTRE-OH (residues 210–222 of tau-441, pT212/T217) was synthesized by New England Peptides. Lyophilized peptide was dissolved in 100 mM Tris pH 8.5 to ∼55 mg/mL. Fabs were mixed with peptide at a 9.3 or 10-fold molar excess of peptide and crystallization trials were performed with custom screens and PEGs (Qiagen) using the Mosquito crystallization robot (TTP Labtech). A crystal of PT3 Fab/peptide was obtained from 20% PEG 3350, 0.2 M NH₄H₂PO₄), which was immersed in reservoir solution supplemented with 20% glycerol and flash-cooled in liquid nitrogen. A crystal of hPT3 Fab/peptide was obtained from 0.1 M Sodium Acetate pH 5.5, 37% PEG200 and flash-cooled in liquid nitrogen. Frozen crystals were used for X-ray diffraction collection as described.

X-ray crystallography

X-ray data for PT3 Fab/peptide were collected at the Advanced Photon Source (Argonne, IL) IMCA-CAT beamline 17-ID-B at 100 K. Diffraction intensities were collected on a Pilatus 6M detector over a 180° rotation with an exposure time of 0.5 s per 0.5° image. The data were processed with the reduction package XDS [33] to the maximum resolution of 2.0 Å. The structure was determined by molecular replacement with the program Phaser using human germline antibody IGHV3–23/IGKV1–39 (PDB ID:5I19) [34] as search model. Structure refinement was performed with PHENIX applying non-crystallography symmetry (NCS) restraints [35] (Supplementary Table 1). Model adjustments were carried out using the program Coot [36]. X-ray data for hPT3 Fab/peptide were collected and processed similarly to a maximum resolution of 2.6 Å.

Immunohistochemistry

Brains of sacrificed mice were removed from the skull and fixed overnight in a formalin-based fixative, followed by paraffin embedding and sectioning of 5μm thick sagittal sections. After deparaffinization, rehydration, quenching of endogenous peroxidase and heat induced epitope retrieval in citrate buffer (pH 6), primary antibody (1μg/mL) was applied to the sections for 1 h. Extensive rinsing was followed by 30 min incubation with anti-mouse secondary antibody (Envision, DAKO, Glostrup, Denmark) and chromogenic labelling using 3,3-diaminobenzidine (DAKO). Following hematoxylin counterstaining, sections were dehydrated and mounted. Sections were imaged with the Hamamatsu NanoZoomer slidescanner (Hamamatsu Photonics, Shizuoka, Japan). For colocalization studies, sections from WT mice were incubated with primary antibodies PT3 (2μg/mL) expressed as mouse IgG2a (produced at Janssen) and anti-MAP2 (20μg/ml) (Thermo Fisher, # PA5-86569). Anti-mouse IgG2a Alexa 488 (Thermo Fisher, A21131) and anti-rabbit Alexa 647 (Thermo Fisher, A21244 or 21246) were used to visualize PT3 and anti-MAP2 signals respectively. Confocal images were acquired using a Zeiss Airyscan 2 LSM 900, featuring a Plan Apochromat 40×/1.3 Oil lens and using a Scan zoom factor of 1.3. Laser used were 405, 488, and 640 for respectively DAPI, Alexa 488, and Alexa 647.

Cryopreserved human brain tissue (Newcastle brain tissue resource) was sliced with a cryostat (20μm thickness) and stored at –80°C before use. Sections were dried, followed by formalin fixation, blocking of endogenous peroxidase with 3% hydrogen peroxide (DAKO, Glostrup, Denmark, S2023) and permeabilization in PBS1x + 0.3% Triton X-100 during 1 h. Primary antibodies (0.4μg/ml) were diluted in antibody diluent with background reducing components (DAKO, S3022) and applied to the sections for 1 h. After extensive washing, slides were incubated with HRP-conjugated anti-mouse secondary antibody (Envision, DAKO, K4000), followed by chromogenic DAB labelling (DAKO, K4368). Slides were counterstained with hematoxylin, dehydrated and mounted with organic mounting medium (Vectamount, Vector labs, Burlingame, CA, USA, H-5000). Imaging was performed with a Hamamatsu NanoZoomer 2.0 RS (Hamamatsu Photonics, Shizuoka, Japan).

Cellular assay

In the immunodepletion assays, tau seeds were incubated with test antibody and removed from the solution with protein G-coupled magnetic beads (Life Technologies, Thermo Scientific). The depleted supernatant was tested for residual seeding capacity in a monoclonal K18 (P301S) FRET biosensor HEK-293 T cell line and analyzed by FACS as previously described [22 , 38]. Homogenates containing tau seeds for immunodepletion were generated from spinal cords of 22- to 23-week-old P301S transgenic mice [39] or from cryopreserved human AD brain tissue obtained from the Newcastle brain tissue resource. After depletion, the human AD supernatant was tested in the presence of the transfection reagent Lipofectamine2000 (Life Technologies, Thermo Scientific). Immunodepleted fractions from P301S spinal cord extracts were added to the cells without transfection.

In vivo co-injection

Frozen parietal/frontal cortex tissue from 5 histopathologically confirmed cases of AD (Braak V-VI) (tissue provided by University of Pennsylvania Perelman School of Medicine, PA, USA) was used to prepare one pool of PHF preps as described in [26]. For injection studies, transgenic Tau-P301L mice expressing the longest human tau isoform with the P301L mutation (tau-4R/2N-P301L) [40, 41] were used for surgery at the age of 3 months. All experiments were performed in compliance with protocols approved by the local ethical committee and national institutions adhering to AAALAC guidelines. For stereotactic surgery, the mice received a unilateral (right hemisphere) injection of 2 pmoles of AD-derived PHFs in the hippocampus. For co-injection studies, equimolar concentrations of antibodies and PHFs were used. Two months after injection, tissue from the contralateral hemisphere was analyzed with light sheet microscopy (LSM) staining using AT8 (Fig. 6A) while the ipsilateral hemisphere was processed to extract total and sarcosyl insoluble fractions as described below.

Sarcosyl extraction from mouse brain

Tissue was weighed and homogenized in 600μL of buffer H per 100 mg tissue (10 mM Tris, 800 mM NaCl, 1 mM EGTA, and 10% sucrose/pH 7.4). The homogenate was centrifuged at 27 000×g for 20 min and 1% N-lauroylsarcosine was added to the supernatant. After 90 min, the solutions were again centrifuged at 184 000×g for 1 h. The supernatants were kept as sarcosyl-soluble fraction, whereas the pellet containing the sarcosyl-insoluble material was resuspended in homogenization buffer.

Biochemical analysis MesoScale Discovery (MSD)

Coating antibodies (AT8, PT3, PT51) were diluted in PBS (1μg/mL) and aliquoted into MSD plates (30μL per well) (L15XA, MSD, Rockville, MD, USA), which were incubated overnight at 4°C. After washing with 5×200μL of PBS/0.5% Tween-20, the plates were blocked with 0.1% casein in PBS and washed again with 5×200μl of PBS/0.5% Tween-20. After adding samples and standards (both diluted in 0.1% casein in PBS), the plates were incubated overnight at 4°C. Subsequently, plates were washed with 5×200μL of PBS/0.5% Tween-20, and SULFO-TAG™ conjugated detection antibodies (AT8, PT82, PT51) in 0.1% casein in PBS were added and incubated for 2 h at room temperature while shaking at 600 rpm. After a final wash (5×200μL of PBS/0.5% Tween-20), 150μL of 2×buffer T (MSD) was added, and plates were read with an MSD imager. Raw signals were normalized against a standard curve consisting of 16 dilutions of a sarcosyl-insoluble prep from postmortem AD brain (ePHF) and were expressed as arbitrary units (AU) ePHF. Statistical analysis (ANOVA with Bonferroni post test) was performed with the GraphPad prism software.

Light sheet microscopy

After dissection, contralateral mouse brain hemispheres were incubated overnight in a formalin-based fixative and further washed in PBS (3×15 min) before storage in PBS-Azide 0.1%. Tau labelling and brain-clearing were based on the iDisco + protocol [42] and as previously described [43]. Tau aggregation was evaluated by staining with AT8, labelled in-house with a near-infrared label following manufacturer’s recommendations (VivoTag 680XL, PerkinElmer), denoted AT8-680. Brains were incubated with AT8-680 (11.5μg/mL) for 2 weeks at 37°C prior to brain clearing. Cleared hemispheres were scanned with an Ultramicroscope II (Lavision Biotec GmbH) using 10μm z-steps and a 1.6 total magnification factor. The analysis of the resulting AT8 signal was performed as previously described [43]. The computed LabelRatio (AT8-positive voxels per brain region volume) were plotted for each mouse.

RESULTS

PT3 binds to phosphorylated tau in AD PHF and in non-AD soluble tau.

From different AD PHF immunization campaigns, as described earlier [26], >100 monoclonal antibodies have been generated via hybridoma technology. In an initial western blot experiment, anti tau antibodies were divided in groups based on their binding preference toward recombinant 2N4R and pathological PHF tau. Phospho tau selective antibodies AT8 and AT100 [32 , 45], the non-phospho tau selective antibody BT2 [46] and the total tau antibody HT7 are used for comparison. From the blots shown in Fig. 1A, it was observed that PT3 displayed the same specificity compared to AT8 and AT100, suggesting that this antibody is selective for PHF and does not bind recombinant tau.

Fig. 1

A) Western blot analysis of recombinant 2N4R tau and AD brain-derived PHFs with the indicated antibodies demonstrate PHF selectivity for AT8, AT100, and PT3 while BT2 shows preference for 2N4R tau and HT7 detects both 2N4R tau and PHF. B) Further western blot profiling of PT3 and other indicated anti tau mAbs was performed by applying brain homogenates from WT mouse (1), tau KO mouse (2), dog (3), Cynomolgus (4), and a non-AD human brain heat-stable extract (5) and a human AD brain-derived PHF sample (6). C) Direct comparison of AT8 and PT3 mAbs and Fab binding* to PHF. * Note: Representative sensorgrams of at least two independent experiments of a selection of tau mAbs to PHF is shown. Different colors reflect the concentration of injected antibody (pink: 75 nM; cyan: 15 nM; dark blue: 3 nM; green: 0.6 nM and magenta: 0.12 nM). D) Compilation of Pepscan epitope mapping and SPR experiments for PT3, AT8 and AT120 (pS/pT: phosphorylated Serine/Threonine).

In a subsequent western blot experiment, it was demonstrated that PT3 not only binds to hyperphosphorylated PHF tau but also to tau in brain extracts from non-diseased species. The blots in Fig. 1B show PT3 binding to tau in a total brain extract from WT mouse and dog and in a heat-stable extract from non-AD human brain but not from Cynomolgus monkey. For the phospho tau-selective antibody AT8, the selectivity for PHF was more apparent in the western blot (Fig. 1B), but the lack of signal in brain homogenates from non-AD brain could be a consequence of its lower (nM range) intrinsic affinity for its phospho-peptide ₁₉₇YSpSPGpSPGpTP₂₀₆ [32]. This was in accordance with the observed lower affinity for PHF (Fig. 1C). As expected, the total tau antibody AT120 [47] reacted with tau from all different species but with a weaker signal for tau from Cynomolgus brain while PT79 [26] showed selectivity for primate and human brain (Fig. 1A).

From these western blot experiments, it was concluded that PT3 binds to hyperphosphorylated PHF tau and not to recombinant non-phosphorylated tau; yet, the binding to tau in brain extracts from non-AD brain (mouse, dog, human) suggests that in addition to its strong binding to PHF tau, PT3 also binds to non-aggregated phosphorylated tau. This was confirmed by a linear peptide mapping screen (Pepscan) where PT3 was shown to bind ₂₁₁RpTPpSLPpTPPTR₂₂₁ but not to the corresponding non-phosphorylated peptide ₂₁₁RTPSLPTPPTR₂₂₁. The same screen was used to perform a first mapping of all internal anti tau monoclonal antibodies and some reference antibodies including AT8 and AT120 (Fig. 1D). Although the phosphorylation-dependency is assessed in this type of screen, the data provide incomplete information on the optimal combination of different phosphorylated S/T residues as has been determined previously for AT8 [32].

Therefore, to determine the PT3 epitope, a more extensive phospho tau peptide mapping ELISA and SPR screen was performed with peptides outlined in Table 1. In these experiments, PT3 showed strongest binding to the tau peptide phosphorylated at T212 and T217 but other combinations did show binding as well. Clear reductions in binding were observed to T212 and T217 monophosphorylated peptides while the effect of phosphorylation at S210 and S214 seemed limited.

The structure of the PT3 Fab in complex with the pT212/pT217 tau phosphopeptide was determined by X-ray crystallography at 2.0 Å resolution (Fig. 2A). The Fab heavy and light chains form a shallow binding pocket, into which the peptide fits (Fig. 2B). The tau phosphopeptide is in an extended conformation with polyproline II-helix character and contains intrapeptide hydrophobic sidechain contacts (pT212 methyl –P213, pT212 methyl –L215 sidechain, pT217 methyl –P218). Residues of the PT3 Fab that are involved in the interaction with the peptide are shown schematically in Fig. 2C. Hydrophobic and electrostatic interactions make up the interface between PT3 and its epitope peptide, and the interactions extend from residues 211–221. The structure shows that the epitope includes the phosphates of pT212 and pT217. The heavy chain Y32 hydroxyl group forms an important hydrogen bond with a phosphate oxygen of pT212 (Supplementary Figure 2). Heavy chain residues W99 and Y32 form and hydrophobic interactions with the side chain of L215 (Supplementary Figure 2). Heavy chain K53 forms a key salt bridge interaction with pT217 (Supplementary Figure 2). Heavy chain residue W104 has extensive interactions with the peptide and also forms part of the VH/VL interface (Supplementary Figure 2). The W104 side chain stacks onto P218 for a hydrophobic interaction and simultaneously contacts the methyl group of pT217. In addition, a hydrogen bond is formed between the W104 indole amide and the sidechain hydroxyl group of T221. W104 sidechain also forms part of the VH/VL interface by packing against light chain residues F94 and L96. While the electrostatic interactions with the phosphates of pT212 and pT217 are critical for the selectivity of PT3 for phospho-tau, the numerous hydrophobic interactions can partly explain its weak reactivity for unphosphorylated peptide (Table 1, NPT-C).

Fig. 2

Overall structure of PT3 Fab + pT212/pT217-tau peptide. Peptide is pink, colored by atom type, heavy chain is in green, light chain is in magenta. Fab is shown as a ribbon structure (A) and space-filling model (B). The tau phosphopeptide is in an extended conformation with characteristics consistent with polyproline-II helix secondary structure. C) Interaction diagram of PT3 Fab with its target peptide based on crystallography data. Phospho-tau residues are in pink boxes, PT3 residues are in green (VH) and magenta (VL) boxes. Dashed lines represent hydrogen bonds, single lines represent hydrophobic contacts.

PT3 binds to phosphorylated tau in rodent and human brain sections.

Based on western blotting and epitope mapping, it was demonstrated that PT3 binds tau within the proline-rich domain (PRD) to a stretch encompassing phosphorylated T212/T217 residues, which seems to be present in both hyperphosphorylated/aggregated tau, present in PHFs, and also in non-aggregated tau in non-diseased brain (e.g., WT mouse brain homogenate). In a series of immunohistochemical staining experiments, it was first verified whether PT3 binds to aggregated and non-aggregated tau in human brain cryosections. From the images shown (Fig. 3A) it can be concluded that PT3 binds not only to aggregated tau in AD brain but also to non-aggregated tau from non-AD and AD brain. However, compared to staining with AT120 on similar sections, the PT3 signals on non-aggregated tau in non-AD brain appeared to be weaker, which is in line with the western blotting results (Fig. 1A).

Fig. 3

A) IHC profiling on cryosections from human non-AD and AD sections showing a homogeneous staining for AT120 reacting with soluble and aggregated tau (indicated by yellow arrows) and for the phosphorylation selective PT3 antibody showing preference for aggregated tau but weak detection of non-aggregated tau (in non-AD and AD) brain was observed as well. B) IHC profiling of AT8 and PT3 on paraffin sections (sagittal) from WT and P301S tau Tg mice. The phospho-selective antibody PT3 detects aggregated tau and mislocalized soluble tau in cortex from WT/P301S mouse. AT8 appeared to be more selective for aggregated tau which could be driven by differences in affinity. C) IHC profiling on paraffin sections (sagittal) from WT mouse brain in hippocampus and cortex. PT3 detects soluble tau localized in soma and dendrites. The non-phospho tau selective antibody Tau 1 seems specific for axonal tau and AT120 detects axonal and somatodendritic tau. D) IHC profiling on paraffin sections (sagittal) from WT and tau KO mouse brain in cortex. Phospho-selective antibody PT3 detects soluble tau localized in soma and dendrites. The non-phospho tau selective antibody Tau 1 seems specific for axonal tau and AT120 detects axonal and somatodendritic tau. None of these antibodies bound to sagittal sections from tau –/–mice. E) Confirmation of the colocalization of PT3 (green) and MAP2 (red) signals in cortical and hippocampal sections from WT mice.

In formalin-fixed brain sections, PT3 detects non-aggregated tau in WT mice and both non-aggregated and aggregated tau in aged P301S tau Tg mice. AT8 showed a similar binding pattern in P301S tau Tg mouse brain sections but no signal in sections from WT mouse (Fig. 3B). From these data, one could speculate that AT8 binds more specifically to phosphorylated tau compared to PT3. However, based on epitope mapping experiments, PT3 only showed very weak binding to non-phosphorylated tau suggesting that the PT3 signal in these sections is phospho-tau dependent. Since phosphorylation has been shown to promote dissociation of tau from the microtubules, a difference in subcellular localization could be expected for phosphorylated (PT3) and non-phosphorylated tau (Tau1). Indeed, non-phosphorylated tau is mainly present in axons and virtually absent from neuronal somata and dendrites, as shown in Fig. 3 (C, D) which is consistent with previous work [16].

Staining of the same WT mouse brain sections with PT3 resulted in an almost opposite localization compared to Tau1. The high affinity of PT3 for phosphorylated tau enables the detection of non-aggregated phosphorylated tau in the wild type mouse brain, which was mainly observed in the somatodendritic compartment without axonal labelling (Fig. 3C, D). This was confirmed by the observed co-localization with MAP2 (Fig. 3E). Consistently, AT120, an antibody that recognizes both non-phosphorylated and phosphorylated tau, displayed tau labelling in both axons and the somatodendritic compartment (Fig. 3C, D). For all antibodies, selectivity for tau was confirmed by the lack of signals on sections from tau KO mouse brain (Fig. 3D).

To summarize, PT3 is an anti-phospho-tau selective monoclonal antibody showing picomolar affinity to PHF and pT212/pT217 phosphorylated monomer tau. Besides its strong binding to pathological tau species, the high affinity enables binding to phosphorylated tau species in WT mice. These properties were used to select this antibody for V-region cloning (Supplementary Figure 1) and humanization.

hPT3 is a humanized version of PT3 with highly similar in vitro binding properties

Using the HFA procedure described in [30], four heavy chain human germlines (IGHV3-23*01, IGHV3-33*01, IGHV3-11*01, and IGHV1-3*01) and three light chain human germlines (IGKV1-16*01, IGKV1-39*01, and IGKV2-24*01) as acceptors for CDR grafting were identified. In addition, we introduced a D56S mutation in CDRL2 in order to mitigate an isomerization motif DG at the junction of CDRL2 and FR3 (framework 3). A matrix of 12 mAbs was produced and based on affinity to PHF, the VH/VL pair of IGHV3-33*01 and IGKV1-16*01 germlines was chosen as the final lead molecule and subsequently named hPT3. The affinity of hPT3 Fab for PHF-tau was greater than 200 pM and only approximately 2.5-fold less than the affinity of the mouse parent PT3 Fab. hPT3 Fab bound tightly to the pT212/pT217 epitope peptide (∼300 pM), and the co-structure with peptide confirms that the epitope is the same as the mouse parent (Supplementary Figure 3).

Further in vitro profiling using an MSD ELISA showed similar binding profiles between PT3 and hPT3. In addition to the species investigated in Fig. 1, reactivity to tau in brain extracts from different non-human species were included. Binding curves in Fig. 4 displayed strong binding to tau in brain extracts from mouse, rat, dog, minipig, marmoset, and a heat-stable extract from non-AD human brain. Interestingly, similar to the western blotting experiment in Fig. 1, no reactivity was observed with tau derived from Cynomolgus brain extract. Comparing the sequence homology of the PT3/hPT3 epitope within the different species revealed identical sequences with the exception of a single amino acid difference (T220 is mutated to A220) in Cynomolgus tau (Table 3). Since crystallography data (Fig. 2) demonstrated that T220 interacts with 5 amino acid residues from the light chain (Y91, D92, E93, F94, L96) and 1 residue from the heavy chain (W104), the T220A mutation explains the lack of PT3 cross reactivity towards tau in Cynomolgus brain extracts. Identical species cross reactivity was observed for hPT3 confirming that, despite the minor loss in affinity (Table 2), the humanization did not impact epitope specificity of the antigen-binding site.

Fig. 4

Estimation of reactivity of PT3 and hPT3 to tau from different species. Different dilutions of brain homogenates from indicated sources are analyzed with a sandwich MSD assay where PT3 of hPT3 are used as coating antibody and sulfo-tagged PT82 is used as detection antibody. Data show signal/background (S/N) values in function of the sample dilution. Dotted line indicates the limit of detection (S/N = 5).

Table 2

Comparison PT3 and hPT3. Compilation SPR binding data showing affinities of PT3 and hPT3 mAbs and Fabs for PHF

Antibody	Kd (PHF) mAb pM	Kd (PHF) Fab pM	Kd (NPT6) mAb pM	Kd (NPT6) Fab pM
PT3	<16	68	29	87
hPT3	<30	167	190	305

Table 3

Summary of species cross-reactivity assessment for PT3 and hPT3 based on MSD (Fig. 4A): Lack of reactivity to tau from cynomolgus monkey brain, is likely a result of a point-mutation at T220 (highlighted in yellow)

Brain sample		Sequence homology
Mouse WT	+	₂₁₁RTPSLPTPPTR₂₂₁
Rat	+	₂₁₁RTPSLPTPPTR₂₂₁
Rabbit	+	₂₁₁RTPSLPTPPTR₂₂₁
Dog (beagle)	+	₂₁₁RTPSLPTPPTR₂₂₁
minipig	+	₂₁₁RTPSLPTPPTR₂₂₁
marmoset	+	₂₁₁RTPSLPTPPTR₂₂₁
cyno	–	₂₁₁RTPSLPTPPAR₂₂₁
Human (non-AD)	+/–	₂₁₁RTPSLPTPPTR₂₂₁
Human ePHF (AD)	++	₂₁₁RTPSLPTPPTR₂₂₁

Comparison of N-terminal anti tau mAbs, PT3 and hPT3, in depletion of AD tau seeds in vitro and in vivo

Functional comparison between PT3 and hPT3 was performed in cellular and in vivo seeding assays described in [26]. Consistent with that work, all antibodies showed nearly complete depletion of tau seeds derived from P301S tau Tg mouse spinal cord. On the other hand, N-terminal antibodies N(a) (26±9% remaining seeding activity normalized to non-depleted AD brain homogenate) and N(b) (24±8% remaining seeding activity) only partially depleted tau seeds derived from a human AD brain homogenate while PT3 and hPT3 depleted almost all seeding activity from these homogenates with <10% of the remaining seeding activity after immunodepletion. Both antibodies also performed better than AT8, which displayed a partial maximal depletion (17±3%) suggesting that the ₁₉₇YSpSPGpSPGpTP₂₀₆ epitope is lower in abundancy compared to the epitope of PT3 and hPT3 (Fig. 5A, B).

Fig. 5

Concentration-dependent efficacy of PT3 and hPT3 in an immunodepletion assay on human AD brain and P301S spinal cord extracts (A). Depleted fractions from P301S spinal cord extracts are directly added to the medium and human AD seeds are added by transfection. In both cases, functional analysis was done in a cellular FRET biosensor cell model and are averages ± SD of % remaining seeding signal normalized to the condition without antibody which was taken as 100%. Graph in panel B shows efficacy of the indicated antibodies at 300 nM (maximum depletion efficacy).

Anti-tau antibodies also have been shown to neutralize AD tau seeds in vivo but, in accordance with cellular models of tau seeding, the degree of efficacy is strongly related to the epitope [48, 49]. In a first in vivo seeding study, PHF seeding in P301L tau Tg mice [40] was performed in the presence/or absence of PT3 (Fig. 6A) which was co-administered with PHF. Two months after seeding, right (injected) hemispheres were used for biochemical analysis while the left (non-injected) hemisphere was fixed in DMFA for LSM analysis using detection with fluorescently labelled AT8 as described in materials and methods. Co-injected PT3 significantly reduced PHF-induced tau aggregation in the injected hemisphere (Fig. 6B).

Fig. 6

Efficacy of PT3 in an in vivo co-injection model. A) Human AD brain–derived PHFs are injected in the right hippocampus of 3-month-old P301L mice in the presence of PT3 (N = 19) or an IgG2a isotype control antibody (N = 13). Two months after injection (at an age of 5 months), mice are sacrificed, and injected hemisphere is used for biochemical quantification of tau aggregation load expressed as arbitrary units (AU) of PHF (A). Brain homogenates (Th) are analyzed with AT8/AT8 MSD assay (AT8/AT8 Th) while sarcosyl insoluble fractions are analyzed with a PT51/PT51 selective MSD assay (PT51/PT51 INS) and show a strong reduction in the injected hemisphere of mice co-injected with PT3. C) For a subgroup of mice (10 per group), the effect of PT3 has been confirmed in the non-injected hemisphere by the use of light sheet microscopy (LSM) on cleared brain stained with AT8. (***p < 0.001; ****p < 0.0001, student T-test). From a few IgG2a-treated mice, both non-injected and injected hemispheres were analyzed with LSM. A representative section from each hemisphere is shown.

Analysis of total homogenates with the AT8/AT8 aggregate selective MSD assay revealed 81 (±7)% reduction by PT3 compared to IgG2a co-injected mice (p < 0.001), while analysis of insoluble fractions using PT51/PT51 MSD showed 74 (±12)% reduction (p < 0.001). In addition to the analysis of the injected hemispheres with biochemical assays, tau aggregation was also determined in the non-injected hemisphere by using LSM as described previously [43]. This technique was applied to a subgroup of the study and revealed AT8⁺ pathology in different regions within the non-injected hemisphere. Figure 6C shows that in these regions (CA1, hippocampal region, isocortex) but also in the overall hemisphere, AT8 signals were lower in the contra-lateral hemisphere of PT3 co-injected mice. These data confirm that PHF seeding results in tau aggregation in both ipsi-and contra-lateral hemispheres and that co-injection with PT3 can dampen this signal in both hemispheres.

A similar study was performed to compare a chimeric mIgG2a form of hPT3 with the efficacy of its parent molecule PT3 but also with the mouse IgG2a variant of mAb N(a) as outlined in Fig. 7A. Consistent with immunodepletion data in Fig. 5, all tau antibody treated groups showed a significant reduction compared to the IgG2a co-injected group (p < 0.0001), and no difference in efficacy was observed between PT3 (79 (±13)%) and hPT3 (81 (±12)%) (p > 0.99). In addition, the difference between N-terminal mab N(a) (48 (±24)% inhibition) and PT3 or hPT3 is recapitulated in the PHF seeding model (Fig. 7A). As hPT3 showed dose-dependent activity after peripheral dosing (Supplementary Figure 4), the effect by intraperitoneally dosed mouse IgG2a chimeric variants of hPT3 and mAb N(a) were compared in this model. In accordance with the co-injection study (Fig. 7A), hPT3 (63 (±5)% inhibition) and mAb N(a) (49 (±13)% inhibition) showed differences in efficacy after intraperitoneal dosing (Fig. 7B). These data demonstrated that antibodies present in plasma can attenuate intracranial PHF seeding and that differences in efficacy observed in cellular immunodepletion assays and in vivo PHF co-injection studies can predict the efficacy of an antibody after peripheral administration in a seeding model using AD brain-derived PHFs.

Fig. 7

Efficacy of PT3, hPT3 and mAb N(a) in an in vivo co-injection model. A) Human AD brain-derived PHFs are injected in the right hippocampus of 3-month-old P301L mice in the presence of PT3 (N = 14), hPT3 (N = 16), mAb N (N = 16), or an IgG2a isotype control antibody (N = 14). B) IgG2a, hPT3, or mAb N(a) were dosed twice a week (20 mg/kg) with two injections before seeding with AD PHF, continuing after seeding. Two months after injection (at an age of 5 months), mice are sacrificed, and injected hemisphere is used for biochemical quantification of tau aggregation load expressed as arbitrary units (AU) of PHF. Sarcosyl insoluble fractions (INS) are analyzed with a PT51/PT51 selective MSD assay. Compared to mAb N(a), PT3 and hPT3 show improved inhibition of AD PHF-induced seeding. (**p < 0.01; ****p < 0.0001, Log transformed data were used for One Way ANOVA with Bonferroni correction for multiple testing).

DISCUSSION

As there is a substantial correlation between tau aggregation and the clinical manifestation of cognitive decline [13], therapeutic strategies to attenuate tau aggregation and spreading for the treatment of neurodegenerative diseases, are intensively explored [50, 51]. Since all known biological functions of tau occur inside the neuron [52], neutralization of extracellular tau seeds by passive or active immunization, could be expected to be a safe strategy with little incidence of adverse events in the clinic [53 –55]. Nevertheless, as some evidence of intraneuronal uptake of tau antibodies has been reported [56, 57], selectivity for more pathological forms of tau could be a way to avoid potential undesirable effects associated with such an antibody uptake. In this respect, hPT3, the humanized variant of PT3, was selected as a clinical candidate based on 1) its high affinity for AD tau PHF, 2) its selectivity for phosphorylated tau, and 3) its strong efficacy in pre-clinical cellular and in vivo models of AD tau seeding.

Although phosphorylation of tau has been proposed to be a marker for pathological tau [58], there is clear evidence that tau phosphorylation also takes place during normal physiological conditions and drives the association/dissociation kinetics of microtubule binding, which is the main biological function of tau [17, 59]. Accordingly, PT3 and hPT3 displayed binding to phosphorylated tau monomers in homogenates from non-diseased brain from most of the investigated species. The subcellular distribution of PT3-reactive tau revealed the intriguing observation that the epitope was restricted to cell bodies and dendrites without notable detection of axonal tau. An opposite distribution was seen when the same sections were stained with the non-phospho tau selective antibody Tau1 [16, 60], which is in line with the different subcellular localization of phosphorylated and de-phosphorylated tau.

In P301S mouse brain, the PT3⁺ tau pool has similar localization but is already increased at an age of 2 months, a timepoint at which limited or no aggregation is observed [39]. Also, under these conditions, the AT8 epitope seems detectable on non-aggregated tau. The PT3⁺ tau pool could be representative for the amount of substrate that can be incorporated into an aggregate during initial tau seeding events [20] suggesting that changes in the phospho-tau/non-phospho-tau balance affects the magnitude of induced tau aggregation induction. Together with the suggested increase in aggregation propensity [18], the degree of somatodendritic/phosphorylated tau seems to be indicative for the magnitude of tau pathology induction. The PT3 antibody seems able to detect the physiological T212/T217 tau phosphorylation events at early stages which is likely driven by its high affinity for this phosphorylated epitope.

The phosphopeptide binding data and the structure of PT3 Fab in complex with a tau phosphopeptide shows that the epitope includes the phosphate of pT212 and pT217. Differences in mAb affinity toward PHF and pT212/pT217 monomeric phosphopeptide could be explained by avidity, which would underscore the high PT3 epitope density in PHFs. The tau phosphopeptide epitope is in an extended conformation with polyproline II-helix character. Because the affinities of the Fab for PHF and for pT212/pT217 peptide are similar (within 2-fold), the conformation that the peptide adopts in the crystal structure is likely representative of the physiological structure in PHF. Therefore, a conformational component in the interaction between PT3 and PHF (the antigen used for immunization) could be proposed. X-ray crystallography data indeed revealed a polyproline-II helical structure [61] which is also observed in hyperphosphorylated tau [62] suggesting that the choice of PHF as immunization antigen resulted in an increased intrinsic affinity for this pathological form of tau.

As outlined above, passive immunization likely affects the extracellular tau pool, so little impact on the intracellular biological function(s) of this microtubule-associated protein is expected. Selecting a phospho-tau epitope would de-risk this further as this approach would target only the dissociated/mislocalized tau pool. More important to evaluate is the epitope density/availability on the tau species responsible for seeding and further propagation. Although the biochemical and biophysical properties of these prion-like seeds are still elusive, small misfolded monomers, oligomers, and larger fibrils have been described to have seeding potential [20 , 63–67].

Given the reactivity of PT3 toward phosphorylated monomers and larger filaments, one could expect binding to both oligomeric and fibrillary tau seeds. Indeed, in our cellular model where tau seeds are depleted from total brain homogenates with tau antibodies, PT3 and hPT3 almost completely depleted seeding activity with little difference in efficacy toward tau seeds from P301S tau Tg mice and from human AD brain. Similar results were seen for other phosphorylation selective antibodies AT180 and PHF1, while for the pS409 Tau selective antibody PT84, the neutralization of AD tau seeds was less efficient suggesting that while phosphorylated epitopes are in enriched in AD tau seeds, relative abundancy of each of these phospho-sites may differ. A reduced epitope density was also observed for antibodies targeting the N-terminal epitope which are typically showing incomplete or limited efficacy toward human AD tau seeds [26, 49].

Indeed, comparison of PT3 and hPT3 with mAb N(a) which targets the N-terminal region in tau [68] in the cellular depletion model demonstrated the incomplete depletion of AD tau seeds by N(a). This difference was also observed after comparison of the same antibodies in the PHF seeding model. In this model, a sarcosyl insoluble prep from AD brain [26, 44] is injected in the CA1 hippocampal region of 3-month-old P301L tau Tg mice. Co-incubation of PHFs with PT3 substantially reduced the seeding potential which was confirmed by biochemical analysis in homogenates from the injected hemisphere. Use of LSM [43] allowed the analysis of tau seeding events in the non-injected hemisphere where the PT3 effect also was apparent. Whether the signals observed in these regions are a result of a secondary tau seeding event or propagation of the initial inoculated fibrils remains speculative.

Although a PHF preparation does not comprise smaller oligomeric tau seeds, it still represents the main seeding competent fraction from a total brain homogenate [66]. Also the AD tau preparation causing tau seeding in WT mouse brain [67] is derived from a detergent insoluble fraction. Furthermore, alignment of immunodepletion data with antibody efficacy in the PHF seeding model confirms the ranking of antibody performance in the prevention of seeding. However, the mechanism underlying incomplete reduction in seeding by N-terminal antibodies needs careful investigation. One could explain these findings as a result of extensive processing of PHF tau seeds as proposed earlier [26, 70]. In aging AD brain, ghost tangles are truncated at both N- and C-termini [71], and in particular the N-terminus is available in the PHF structure [72]. Since tau fibrils present in brain extracts from tau Tg mice have a much shorter in vivo existence, truncation could be absent explaining the almost complete depletion of murine brain-derived tau seeds with N-terminal anti tau mAbs. On the other hand, N-terminal truncation of tau has been described independently from incorporation into filamentous structures [73]. Therefore, one could also speculate that truncation is important for the aggregation of WT tau [74] while fibrillization of mutated Tg Tau is driven more by the mutation of P301L and does not require truncation. Ultrasensitive techniques to detect and characterize AD tau seeds in human CSF/ISF will be crucial to determine which tau epitopes are optimal target for tau immunotherapy.

Conclusion

The current work describes the pre-clinical development and functional characterization of the T212/T217 phosphorylation-selective anti tau mAb hPT3. Based on its strong affinity for PHF and efficacy in cellular and in vivo seeding models, the molecule was selected for further development. Different ongoing clinical trials with anti-tau antibodies targeting various epitopes will contribute to the evaluation of clinical proof of concept for tau immunotherapy.

Footnotes

ACKNOWLEDGMENTS

We thank Werner Schillebeeckx, Willy Janssens, Annemie Wellens, Bruno Vasconcelos, Jan Verrelst, and Michel Mahieu for their excellent assistance during the in vivo experiments and Javier Franco (Open Analytics, Antwerp Belgium) for statistical support. Tissue to generate PHFs for immunizations was kindly provided by Dr. Ralph Nixon, Nathan Kline Institute (NY, USA) and human tissue for cellular experiments in this communication was provided by the Newcastle Brain Tissue Resource which is funded in part by a grant from the UK Medical Research Council (G0400074), by NIHR Newcastle Biomedical Research Centre and Unit awarded to the Newcastle upon Tyne NHS Foundation Trust and Newcastle University, and by a grant from the Alzheimer’s Society and Alzheimer’s Research Trust as part of the Brains for Dementia Research Project. Funding was provided by VLAIO (AD immunotherapy project 150882).

The study was funded by Janssen Research & Development, LLC, and a number of authors were employees of Janssen Research & Development, LLC at the time the study was conducted and own stock/stock options in Johnson & Johnson.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

DeTure

, Dickson

(2019) The neuropathological diagnosis of Alzheimer’s disease. Mol Neurodegener 14, 32.

Karran

, Mercken

, De Strooper

(2011) The amyloid cascade hypothesis for Alzheimer’s disease: An appraisal for the development of therapeutics. Nat Rev Drug Discov 10, 698–712.

Villegas-Llerena

, Phillips

, Garcia-Reitboeck

, Hardy

, Pocock

(2016) Microglial genes regulating neuroinflammation in the progression of Alzheimer’s disease. Curr Opin Neurobiol 36, 74–81.

Zhou

, Chen

, Mok

, Zhao

, Chen

, Hardy

, Li

, Fu

AKY

, Guo

, Ip

; Alzheimer’s Disease Neuroimaging Initiative (2018) Identification of genetic risk factors in the Chinese population implicates a role of immune system in Alzheimer’s disease pathogenesis. Proc Natl Acad Sci U S A 115, 1697–1706.

Goedert

, Falcon

, Clavaguera

, Tolnay

(2014) Prion-like mechanisms in the pathogenesis of tauopathies and synucleinopathies. Curr Neurol Neurosci Rep 14, 495.

Clavaguera

, Bolmont

, Crowther

, Abramowski

, Frank

, Probst

, Fraser

, Stalder

, Beibel

, Staufenbiel

, Jucker

, Goedert

, Tolnay

(2009) Transmission and spreading of tauopathy in transgenic mouse brain. Nat Cell Biol 11, 909–913.

Clavaguera

, Akatsu

, Fraser

, Crowther

, Frank

, Hench

, Probst

, Winkler

, Reichwald

, Staufenbiel

, Ghetti

, Goedert

, Tolnay

(2013) Brain homogenates from human tauopathies induce tau inclusions in mouse brain. Proc Natl Acad Sci U S A 110, 9535–9540.

Lewis

, Dickson

(2016) Propagation of tau pathology: Hypotheses, discoveries, and yet unresolved questions from experimental and human brain studies. Acta Neuropathol 131, 27–48.

Gomes

, Hipp

, Rijal Upadhaya

, Balakrishnan

, Ospitalieri

, Koper

, Largo-Barrientos

, Uytterhoeven

, Reichwald

, Rabe

, Vandenberghe

, von Arnim

CAF

, Tousseyn

, Feederle

, Giudici

, Willem

, Staufenbiel

, Thal

(2019) Aβ-induced acceleration of Alzheimer-related τ-pathology spreading and its association with prion protein. Acta Neuropathol 138, 913–941.

10.

Vasconcelos

, Stancu

, Buist

, Bird

, Wang

, Vanoosthuyse

, Van Kolen

, Verheyen

, Kienlen-Campard

, Octave

, Baatsen

, Moechars

, Dewachter

(2016) Heterotypic seeding of tau fibrillization by pre-aggregated Abeta provides potent seeds for prion-like seeding and propagation of tau-pathology in vivo . Acta Neuropathol 131, 549–569.

11.

, Guo

, McBride

, Narasimhan

, Kim

, Changolkar

, Zhang

, Gathagan

, Yue

, Dengler

, Stieber

, Nitla

, Coulter

, Abel

, Brunden

, Trojanowski

, Lee

VMY

(2018) Amyloid-β plaques enhance Alzheimer’s brain tau-seeded pathologies by facilitating neuritic plaque tau aggregation. Nat Med 24, 29–38.

12.

Cho

, Choi

, Hwang

, Lee

, Kim

, Lee

, Lyoo

, Ryu

, Lee

(2016) Tau PET in Alzheimer disease and mild cognitive impairment. Neurology 87, 375–383.

13.

Nelson

, Alafuzoff

, Bigio

, Bouras

, Braak

, Cairns

, Castellani

, Crain

, Davies

, Del Tredici

, Duyckaerts

, Frosch

, Haroutunian

, Hof

, Hulette

, Hyman

, Iwatsubo

, Jellinger

, Jicha

, Kovari

, Kukull

, Leverenz

, Love

, Mackenzie

, Mann

, Masliah

, McKee

, Montine

, Morris

, Schneider

, Sonnen

, Thal

, Trojanowski

, Troncoso

, Wisniewski

, Woltjer

, Beach

(2013) Correlation of Alzheimer disease neuropathologic changes with cognitive status: A review of the literature. J Neuropathol Exp Neurol 71, 362–381.

14.

Schöll

, Lockhart

, Schonhaut

, O’Neil

, Janabi

, Ossenkoppele

, Baker

, Vogel

, Faria

, Schwimmer

, Rabinovici

, Jagust

(2016) PET imaging of tau deposition in the aging human brain. Neuron 89, 971–982.

15.

Sigurdsson

(2009) Tau-focused immunotherapy for Alzheimer’s disease and related tauopathies. Curr Alzheimer Res 6, 446–450.

16.

Binder

, Frankfurter

, Rebhun

(1985) The distribution of tau in the mammalian central nervous system. J Cell Biol 101, 1371–1378.

17.

Cho

, Johnson

GVW

(2003) Glycogen synthase kinase 3β phosphorylates tau at both primed and unprimed sites: Differential impact on microtubule binding. J Biol Chem 278, 187–193.

18.

Despres

, Byrne

, Qi

, Cantrelle

, Huvent

, Chambraud

, Baulieu

, Jacquot

, Landrieu

, Lippens

, Smet-Nocca

(2017) Identification of the tau phosphorylation pattern that drives its aggregation. Proc Natl Acad Sci U S A 114, 9080–9085.

19.

Iba

, Guo

, McBride

, Zhang

, Trojanowski

, Lee

VMY

(2013) Synthetic tau fibrils mediate transmission of neurofibrillary tangles in a transgenic mouse model of Alzheimer’s-like tauopathy. J Neurosci 33, 1024–1037.

20.

Mirbaha

, Chen

, Morazova

, Ruff

, Sharma

, Liu

, Goodarzi

, Pappu

R V

, Colby

, Mirzaei

, Joachimiak

, Diamond

(2018) Inert and seed-competent tau monomers suggest structural origins of aggregation. Elife 7, e36584.

21.

Mirbaha

, Holmes

, Sanders

, Bieschke

, Diamond

(2015) Tau trimers are the minimal propagation unit spontaneously internalized to seed intracellular aggregation. J Biol Chem 290, 14893–14903.

22.

Holmes

, Furman

, Mahan

, Yamasaki

, Mirbaha

, Eades

, Belaygorod

, Cairns

, Holtzman

, Diamond

(2014) Proteopathic tau seeding predicts tauopathy in vivo . Proc Natl Acad Sci U S A 111, E4376–E4385.

23.

Sato

, Barthélemy

, Mawuenyega

, Patterson

, Gordon

, Jockel-Balsarotti

, Sullivan

, Crisp

, Kasten

, Kirmess

, Kanaan

, Yarasheski

, Baker-Nigh

, Benzinger

TLS

, Miller

, Karch

, Bateman

(2018) Tau kinetics in neurons and the human central nervous system. Neuron 97, 1284–1298.e7.

24.

Braak

, Braak

(1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol 82, 239–259.

25.

Braak

, Thal

, Ghebremedhin

, Del Tredici

(2011) Stages of the pathologic process in Alzheimer disease: Age categories from 1 to 100 years. J Neuropathol Exp Neurol 70, 960–969.

26.

Vandermeeren

, Borgers

, Van Kolen

, Theunis

, Vasconcelos

, Bottelbergs

, Wintmolders

, Daneels

, Willems

, Dockx

, Delbroek

, Marreiro

, Ver Donck

, Sousa

, Nanjunda

, Lacy

, Van De Casteele

, Van Dam

, De Deyn

, Kemp

, Malia

, Mercken

(2018) Anti-tau monoclonal antibodies derived from soluble and filamentous tau show diverse functional properties in vitro and in vivo . J Alzheimers Dis 65, 265–281.

27.

Griswold-Prenner

, Stagliano

, Dang

(2017) Methods of treating a tauopathy.World Intellectual Property Organization, WO/2014/028777

28.

Van Kolen

, Mercken

, Barone

, Lacy

, Nanjunda

, Wheeler

, Luo

(2019) Anti-PHF-Tau Antibodies and Uses Thereof. United States Patent Application, Pub. No. US 2019/0270793 A1.

29.

Köhler

, Milstein

(2005) Pillars article: Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 1975, 256 (5517): 495-497. J Immunol 174, 2453–2455.

30.

Fransson

, Teplyakov

, Raghunathan

, Chi

, Cordier

, Dinh

, Feng

, Giles-Komar

, Gilliland

, Lollo

, Malia

, Nishioka

, Obmolova

, Zhao

, Swanson

R V

, Almagro

(2010) Human framework adaptation of a mouse anti-human IL-13 antibody. J Mol Biol 398, 214–231.

31.

Chothia

, Lesk

(1987) Canonical structures for the hypervariable regions of immunoglobulins. J Mol Biol 196, 901–917.

32.

Malia

, Teplyakov

, Ernst

, Wu

, Lacy

, Liu

, Vandermeeren

, Mercken

, Luo

, Sweet

, Gilliland

(2016) Epitope mapping and structural basis for the recognition of phosphorylated tau by the anti-tau antibody AT8. Proteins 84, 427–434.

33.

Kabsch

(2010) Integration, scaling, space-group assignment and post-refinement. Acta Crystallogr D Biol Crystallogr 66, 133–144.

34.

Teplyakov

, Obmolova

, Malia

, Luo

, Muzammil

, Sweet

, Almagro

, Gilliland

(2016) Structural diversity in a human antibody germline library. MAbs 8, 1045–1063.

35.

Terwilliger

, Grosse-Kunstleve

, Afonine

, Moriarty

, Zwart

, Hung

, Read

, Adams

(2008) Iterative model building, structure refinement and density modification with the PHENIX AutoBuild wizard. Acta Crystallogr D Biol Crystallogr 64, 61–69.

36.

Emsley

, Lohkamp

, Scott

, Cowtan

(2010) Features and development of Coot. Acta Crystallogr D Biol Crystallogr 66, 486–501.

37.

Furman

, Holmes

, Diamond

(2015) Sensitive detection of proteopathic seeding activity with FRET flow cytometry. J Vis Exp 106, e53205.

38.

Holmes

, DeVos

, Kfoury

, Li

, Jacks

, Yanamandra

, Ouidja

, Brodsky

, Marasa

, Bagchi

, Kotzbauer

, Miller

, Papy-Garcia

, Diamond

(2013) Heparan sulfate proteoglycans mediate internalization and propagation of specific proteopathic seeds. Proc Natl Acad Sci U S A 110, E3138–E3147.

39.

Allen

, Ingram

, Takao

, Smith

, Jakes

, Virdee

, Yoshida

, Holzer

, Craxton

, Emson

, Atzori

, Migheli

, Crowther

, Ghetti

, Spillantini

, Goedert

(2002) Abundant tau filaments and nonapoptotic neurodegeneration in transgenic mice expressing human P301S tau protein. J Neurosci 22, 9340–9351.

40.

Terwel

, Lasrado

, Snauwaert

, Vandeweert

, Van Haesendonck

, Borghgraef

, Van Leuven

(2005) Changed conformation of mutant tau-P301L underlies the moribund tauopathy, absent in progressive, nonlethal axonopathy of tau-4R/2N transgenic mice. J Biol Chem 280, 3963–3973.

41.

Peeraer

, Bottelbergs

, Van Kolen

, Stancu

ICC

, Vasconcelos

, Mahieu

, Duytschaever

, Ver Donck

, Torremans

, Sluydts

, Van Acker

, Kemp

, Mercken

, Brunden

, Trojanowski

, Dewachter

, Lee

VMY

, Moechars

(2015) Intracerebral injection of preformed synthetic tau fibrils initiates widespread tauopathy and neuronal loss in the brains of tau transgenic mice. Neurobiol Dis 73, 83–95.

42.

Renier

, Adams

, Kirst

, Wu

, Azevedo

, Kohl

, Autry

, Kadiri

, Umadevi Venkataraju

, Zhou

, Osten

, Tessier-Lavigne

(2016) Mapping of brain activity by automated volume analysis of immediate early genes. Cell 165, 1789–1802.

43.

Detrez

, Maurin

, Van Kolen

, Willems

, Colombelli

, Lechat

, Roucourt

, Van Leuven

, Baatout

, Larsen

, Nuydens

, Timmermans

, De Vos

(2019) Regional vulnerability and spreading of hyperphosphorylated tau in seeded mouse brain. Neurobiol Dis 127, 398–409.

44.

Mercken

, Vandermeeren

, Lübke

, Six

, Boons

, Van de Voorde

, Martin

, Gheuens

(1992) Monoclonal antibodies with selective specificity for Alzheimer tau are directed against phosphatase-sensitive epitopes. Acta Neuropathol 84, 265–272.

45.

Zheng-Fischhöfer

, Biernat

, Mandelkow

, Illenberger

, Godemann

, Mandelkow

(1998) Sequential phosphorylation of Tau by glycogen synthase kinase-3β and protein kinase A at Thr212 and Ser214 generates the Alzheimer-specific epitope of antibody AT100 and requires a paired-helical-filament-like conformation. Eur J Biochem 252, 542–552.

46.

Rösner

, Rebhan

, Vacun

, Vanmechelen

(1994) Expression of a paired helical filament tau epitope in embryonic chicken central nervous system. Neuroreport 5, 1164–1166.

47.

Vandermeeren

, Mercken

, Vanmechelen

, Six

, Van de Voorde

, Martin

, Cras

(1993) Detection of proteins in normal and Alzheimer’s disease cerebrospinal fluid with a sensitive sandwich enzyme-linked immunosorbent assay. J Neurochem 61, 1828–1834.

48.

Albert

, Mairet-Coello

, Danis

, Lieger

, Caillierez

, Carrier

, Skrobala

, Landrieu

, Michel

, Schmitt

, Citron

, Downey

, Courade

, Buée

, Colin

(2019) Prevention of tau seeding and propagation by immunotherapy with a central tau epitope antibody. Brain 142, 1736–1750.

49.

Courade

, Angers

, Mairet-Coello

, Pacico

, Tyson

, Lightwood

, Munro

, McMillan

, Griffin

, Baker

, Starkie

, Nan

, Westwood

, Mushikiwabo

, Jung

, Odede

, Sweeney

, Popplewell

, Burgess

, Downey

, Citron

(2018) Epitope determines efficacy of therapeutic anti-Tau antibodies in a functional assay with human Alzheimer Tau. Acta Neuropathol 136, 729–745.

50.

Jadhav

, Avila

, Schöll

, Kovacs

, Kövari

, Skrabana

, Evans

, Kontsekova

, Malawska

, de Silva

, Buee

, Zilka

(2019) A walk through tau therapeutic strategies. Acta Neuropathol Commun 7, 22.

51.

Khanna

, Kovalevich

, Lee

VMY

, Trojanowski

, Brunden

(2016) Therapeutic strategies for the treatment of tauopathies: Hopes and challenges. Alzheimers Dement 12, 1051–1065.

52.

Sotiropoulos

, Galas

, Silva

, Skoulakis

, Wegmann

, Maina

, Blum

, Sayas

, Mandelkow

, Spillantini

, Sousa

, Avila

, Medina

, Mudher

, Buee

(2017) Atypical, non-standard functions of the microtubule associated Tau protein. Acta Neuropathol Commun 5, 91.

53.

Qureshi

, Tirucherai

, Ahlijanian

, Kolaitis

, Bechtold

, Grundman

(2018) A randomized, single ascending dose study of intravenous BIIB092 in healthy participants. Alzheimers Dement (N Y) 4, 746–755.

54.

Boxer

, Qureshi

, Ahlijanian

, Grundman

, Golbe

, Litvan

, Honig

, Tuite

, McFarland

, O’Suilleabhain

, Xie

, Tirucherai

, Bechtold

, Bordelon

, Geldmacher

, Grossman

, Isaacson

, Zesiewicz

, Olsson

, Muralidharan

, Graham

, O’Gorman

, Haeberlein

, Dam

(2019) Safety of the tau-directed monoclonal antibody BIIB092 in progressive supranuclear palsy: A randomised, placebo-controlled, multiple ascending dose phase 1b trial. Lancet Neurol 18, 549–558.

55.

Medina

(2018) An overview on the clinical development of tau-based therapeutics. Int J Mol Sci 19, 1160.

56.

, Lin

, Gu

, Sigurdsson

(2018) Dynamic assessment of tau immunotherapies in the brains of live animals by two-photon imaging. EBioMedicine 35, 270–278.

57.

Collin

, Bohrmann

, Göpfert

, Oroszlan-Szovik

, Ozmen

, Grüninger

(2014) Neuronal uptake of tau/pS422 antibody and reduced progression of tau pathology in a mouse model of Alzheimer’s disease. Brain 137, 2834–2846.

58.

Grundke-Iqbal

, Iqbal

, Tung

, Quinlan

, Wisniewski

, Binder

(1986) Abnormal phosphorylation of the microtubule-associated protein tau (tau) in Alzheimer cytoskeletal pathology. Proc Natl Acad Sci U S A 83, 4913–4917.

59.

Hernández

, García-García

, Avila

(2013) Microtubule depolymerization and tau phosphorylation. J Alzheimers Dis 37, 507–513.

60.

Biernat

, Mandelkow

, Schröter

, Lichtenberg-Kraag

, Steiner

, Berling

, Meyer

, Mercken

, Vandermeeren

, Goedert

(1992) The switch of tau protein to an Alzheimer-like state includes the phosphorylation of two serine-proline motifs upstream of the microtubule binding region. EMBO J 11, 1593–1597.

61.

Adzhubei

, Sternberg

MJE

, Makarov

(2013) Polyproline-II helix in proteins: Structure and function. J Mol Biol 425, 2100–2132.

62.

Bielska

, Zondlo

(2006) Hyperphosphorylation of tau induces local polyproline II helix. Biochemistry 45, 5527–5537.

63.

Audouard

, Houben

, Masaracchia

, Yilmaz

, Suain

, Authelet

, De Decker

, Buée

, Boom

, Leroy

, Ando

, Brion

(2016) High-molecular-weight paired helical filaments from alzheimer brain induces seeding of wild-type mouse tau into an argyrophilic 4R tau pathology in vivo . Am J Pathol 186, 2709–2722.

64.

Sanders

, Kaufman

, DeVos

, Sharma

, Mirbaha

, Li

, Barker

, Foley

, Thorpe

, Serpell

, Miller

, Grinberg

, Seeley

, Diamond

(2014) Distinct tau prion strains propagate in cells and mice and define different tauopathies. Neuron 82, 1271–1288.

65.

Takeda

, Commins

, DeVos

, Nobuhara

, Wegmann

, Roe

, Costantino

, Fan

, Nicholls

, Sherman

, Trisini Lipsanopoulos

, Scherzer

, Carlson

, Pitstick

, Peskind

, Raskind

, Li

, Montine

, Frosch

, Hyman

(2016) Seed-competent high-molecular-weight tau species accumulates in the cerebrospinal fluid of Alzheimer’s disease mouse model and human patients. Ann Neurol 80, 355–367.

66.

Jackson

, Kerridge

, Cooper

, Cavallini

, Falcon

, Cella

C V

, Landi

, Szekeres

, Murray

, Ahmed

, Goedert

, Hutton

, O’Neill

, Bose

(2016) Short fibrils constitute the major species of seed-competent tau in the brains of mice transgenic for human P301S tau. J Neurosci 36, 762–772.

67.

Guo

, Narasimhan

, Changolkar

, He

, Stieber

, Zhang

, Gathagan

, Iba

, McBride

, Trojanowski

, Lee

VMY

(2016) Unique pathological tau conformers from Alzheimer’s brains transmit tau pathology in nontransgenic mice. J Exp Med 213, 2635–2654.

68.

Bright

, Hussain

, Dang

, Wright

, Cooper

, Byun

, Ramos

, Singh

, Parry

, Stagliano

, Griswold-Prenner

(2015) Human secreted tau increases amyloid-beta production. Neurobiol Aging 36, 693–709.

69.

Greenberg

, Davies

(1990) A preparation of Alzheimer paired helical filaments that displays distinct tau proteins by polyacrylamide gel electrophoresis. Proc Natl Acad Sci U S A 87, 5827–5831.

70.

Zhou

, Shi

, Chu

, Hu

, Guan

, Gong

, Iqbal

, Liu

(2018) Relevance of phosphorylation and truncation of tau to the etiopathogenesis of Alzheimer’s disease. Front Aging Neurosci 10, 27.

71.

Endoh

, Ogawara

, Iwatsubo

, Nakano

, Mori

(1993) Lack of the carboxyl terminal sequence of tau in ghost tangles of Alzheimer’s disease. Brain Res 601, 164–172.

72.

Fitzpatrick

AWP

, Falcon

, He

, Murzin

, Murshudov

, Garringer

, Crowther

, Ghetti

, Goedert

, Scheres

SHW

(2017) Cryo-EM structures of tau filaments from Alzheimer’s disease. Nature 547, 185–190.

73.

Derisbourg

, Leghay

, Chiappetta

, Fernandez-Gomez

, Laurent

, Demeyer

, Carrier

, Buée-Scherrer

, Blum

, Vinh

, Sergeant

, Verdier

, Buée

, Hamdane

(2015) Role of the Tau N-terminal region in microtubule stabilization revealed by new endogenous truncated forms. Sci Rep 5, 9659.

74.

Wang

, Biernat

, Pickhardt

, Mandelkow

(2007) Stepwise proteolysis liberates tau fragments that nucleate the Alzheimer-like aggregation of full-length tau in a neuronal cell model. Proc Natl Acad Sci U S A 104, 10252–10257.