Background
Alzheimer’s disease (AD) is the most prevalent form of dementia worldwide. Compared to other deadly diseases, Alzheimer’s is the only disease that cannot yet be prevented, cured or slowed down. While death rates for other major diseases, such as heart diseases, cancer, AIDS etc., have declined, death rates from Alzheimer’s disease have risen by 71% since 2000 (
www.alz.org). Since the “amyloid cascade hypothesis” was proposed, a primary focus of therapeutic approaches for AD have involved reducing amyloid-β (Aβ) levels in the brain [
1,
2]. Anti-Aβ immunotherapy is considered as one of the most promising AD treatments under investigation, as it is currently being tested in many clinical trials [
3‐
7]. Unfortunately, none of the attempts reported to date have shown positive clinical outcome in pivotal phase 3 trials. However, lessons learned from these trials indicate that, primarily, to be effective, anti-Aβ immunotherapy should be initiated before cognitive decline and severe pathological changes have occurred. Secondly, clearing Aβ at the late stages may be insufficient to halt the progression of AD. According to multiple reports, tau pathology correlates much better with the degree of dementia than Aβ plaque burden [
8‐
14]. Targeting tau is now considered a promising approach for treatment of advanced AD stages. Additionally, tau is a common pathological marker for several neurodegenerative disorders collectively referred to as tauopathies, e.g. frontotemporal dementia with parkinsonism linked to chromosome 17 (FTDP-17) and Pick’s disease, corticobasal degeneration, progressive supranuclear palsy (PSP), amyotrophic lateral sclerosis (ALS), guam parkinsonism dementia complex, and dementia pugilistica. Therefore, development of safe and effective immunotherapy targeting pathological tau may become universal for the treatment of many diseases.
In this study, we generated and characterized monoclonal humanized version of a select mouse antibodies targeting a non-phospho-epitope located in N-terminal region of tau spanning aa 2–18 (tau
2-18, known as PAD, phosphatase activation domain). This domain is normally hidden in a paperclip-like conformation of native protein, but becomes exposed in aggregated pathological tau [
15]. It has been shown that exposed PAD plays an important role in the inhibition of anterograde fast axonal transport (FAT) as well as in polymerization of tau [
16‐
21]. Immunohistochemical studies of human postmortem tissues and immunoreactivity with AD brain extracts with PAD specific antibodies (TNT-1) demonstrated that exposure of N-terminal region of Tau (TNT-1 immunoreactivity) is an early event in AD that is becoming more and more revealed in the severe stages of AD [
16,
20‐
22]. Importantly, it was shown that appoptosin-mediated caspase-3 activation observed in several chronic neurodegenerative diseases including PSP, AD, FTD-T,HD, and PD, leads to tau cleavage at D421 [
23‐
25]. Cleaved c-Tau is more prone to form aggregates/fibrils [
25‐
29], associates with both early and late markers of NFTs in AD and is correlated with cognitive decline [
29], strongly supporting the strategy aimed to target N-terminus of Tau exposed in caspase-cleaved c-Tau. Recently Bright et al. demonstrated the presence of extracellular N-terminal tau fragments secreted by iPSC cortical neurons from AD patients. Based on
in vitro data they suggested that secreted extracellular Tau negatively impacts the neurons by inducing their hyperactivity and may elevate Aβ production in AD brain [
30]. It is assumed that neutralization of these species can potentially slow the clinical progression of dementia.
Based on above mentioned data we hypothesized that antibodies generated against tau2–18 epitope may recognize pathological, but not normal tau at the early stages of tauopathy, and prevent/decrease the polymerization of this molecule.
We report the discovery of mouse mAbs targeting N-terminus of Tau leading to generation and characterization of a humanized anti-tau antibody, and culminating with development of a CHO cell line producing >1.5 g/l of this therapeutically potent Mab. Humanized antibody was termed Armanezumab according to “Programme on International Nonproprietary Names (INN), Quality Assurance and Safety: Medicines, Essential Medicines and Pharmaceutical Policies (EMP)” of WHO. Pending safety and therapeutic efficacy assessment, Armanezumab could be a candidate for clinical trials in mild to moderate AD patients.
Discussion
Although Aβ may be the primary trigger in AD pathogenesis, it is clear that pathological tau also plays an important role in AD [
38]. By the time the clinical signs of AD appear first, there is already substantial tau pathology in the brain [
39,
40], which may become self-propagating [
41‐
44]. Observations showed that tau oligomers are directly toxic to neurons, and tau pathology better correlates with clinical cognitive decline in AD [
14,
45,
46]. Therefore, several groups have proposed anti-tau immunotherapy (passive and active vaccinations) as an effective therapeutic approach [
4,
47‐
49].
Importantly, therapeutics aimed at eliminating pathological tau may also be beneficial for treatment of a group of neurodegenerative disorders other than AD, categorized as tauopathies. These diseases include ALS, FTD with parkinsonism linked to chromosome 17, Pick’s Disease, PSP, Creutzfeldt-Jakob Disease, Dementia Pugilistica, Down’s Syndrome and others.
Tau-targeting immunotherapy was questioned initially due to the intracellular nature of tau protein. However, data demonstrating that during disease progression in humans tau-related pathology spreads from the affected areas of the brain to the healthy regions [
50], and the discovery of trans-cellular propagation of tau aggregates [
51‐
55] suggested a possible mechanism of action for antibody-mediated reduction of tau pathology observed in mouse models, and supported the feasibility of using anti-tau immunotherapy in a clinical setting. Although the exact mechanism of antibody-mediated reduction of pathological tau is currently unknown, several studies suggest different mechanisms of action, possibly depending on the particular antibody, as well as type and size of tau aggregates [
56‐
61]. Recently it was shown that tau aggregates-antibody complexes taken up by cultured neuronal cells bind to cytosolic protein TRIM21 that triggered clearance of tau particles [
58]. A mechanism of action where neuronal FcγRs mediate antibody uptake by neurons, followed by engagement of intracellular tau tangles by the internalized antibody, has been proposed [
56,
57]. Another suggested mechanism is that antibodies bind extracellular Tau and prevent its uptake by adjacent or post-synaptic neurons [
59,
60]. It is also shown that anti-tau antibodies can promote microglial uptake and clearance of tau [
59,
61], although, recently another group reported that effector function is not required for antibody efficacy [
60].
An early immunotherapeutic study showed that immunization of wild type mice with recombinant full-length tau is not entirely safe, and tau-specific polyclonal antibodies induced NFT-like pathology, glial activation, axonal damage, and even infiltration of mononuclear cells into the brains [
62]. However, several subsequent studies utilizing both passive vaccinations [
43,
57,
63‐
72] and short peptide-based active vaccinations [
73‐
79] targeting various phosphorylated, non-modified linear or conformational epitopes of tau showed that these antibodies are safe and efficacious for reducing tau pathology and improving tau-associated functional deficits in tau/Tg mouse models. Currently, two active vaccines (AADvac1 [
80,
81] and ACI-35 [
79]) are being tested in clinical trials with AD patients, and two humanized versions of anti-tau antibodies have been reported to have advanced to clinical testing in subjects with progressive supranuclear palsy (NCT02494024, NCT02460094, NCT02658916).
In this study, we report a discovery and characterization of a humanized antibody targeting N-terminal region of tau, which, after successful preclinical safety studies will be ready to advance into clinical testing. As our starting point, a mouse mAb was selected based on the ability of binding to recombinant tau with high affinity, as well as binding to oligomeric/fibrillar tau in brain homogenates and NT/NFT in the brain tissue from AD cases. Importantly, our data reveals this mAb’s high specificity to pathological tau species characterized by an accessible N-terminal region, and it’s inability to effectively bind normally folded tau in non-diseased brain. Translation of monoclonal antibodies to a clinical setting requires humanization to make them non-immunogenic or at least less immunogenic in humans. For humanization of mouse mAb 1C9, we adopted an approach involving CDR-grafting into selected human frameworks from human Immunoglobulin Germline V gene database based on similarity to the parent mouse mAb, combined with phage display to identify the optimal humanized variant. Constructs encoding humanized VH and VL were fused with constant regions from IgG1 human immunoglobulin to generate full human antibody. The optimal humanized antibody (Armanezumab) was characterized in different in vitro/ex vivo and in vivo assays to test for activity relative to the parent mouse mAb. We demonstrated that Armanezumab (i) possesses a tau-binding affinity, which is only slightly lower than parental mouse mAb 1C9; (ii) recognizes the same tau4–8 epitope as 1C9; (iii) efficiently binds to pathological tau in brain sections from AD, FTD, Pick’s Disease cases, without binding to brain sections from control non-demented subjects; (iv) significantly inhibits the seeding ability of misfolded tau in brain extracts from P301S Tg as evidenced by decreasing integrated FRET density in a cellular assay of tau aggregation; (v) Inhibits neurotoxic effects of aggregated tau; vi) acutely reduces tau pathology following intracranial administration in Tau Tg mice.
In addition, pharmacokinetics analysis of Armanezumab revealed no differences in clearance rates from the blood circulation of wildtype and tau/Tg PS19 mice, with about 9 days of half-life in both strains. This similarity in pharmacokinetics of mouse and humanized Mab could be explained by negligible concentrations of tau protein in blood of tau transgenic mice that express human tau selectively in neurons, leading to the same non-target-mediated profile of elimination seen in wild-type mice. Despite in general higher binding affinity of human Fc-fragment to mouse neonatal FcRn [
82], half-life of Armanezumab is close to theoretical mouse IgG half-life. In fact, it was shown that higher affinity for FcRn does not correlate with extended
in vivo half-lives of antibodies [
83].
Recently, various groups developed antibodies specific to different epitopes located at N-terminal region of Tau. It was demonstrated that both intracerebroventricular infusion and peripheral administration of antibodies specific to Tau
25–30 are therapeutic in P301S mouse model of tauopathy [
43,
72], although very high concentration (50 mg/kg) of systemically administered (i.p.) antibodies are needed to achieve significant decrease in insoluble tau, prevent brain atrophy, and see improvement of motor/sensorimotor function [
72]. On the other hand, intravenous injection of only 15 μg (~750 μg/kg) of antibody specific to Tau
6–18 markedly reduced tau pathology and rescued cognitive impairment of vaccinated 3xTg-AD mice [
69,
84]. Such large differences in effective dose of above mentioned antibodies could be attributable to differences in administration route, targeted epitope or the aggressiveness of mouse model of tauopathy. This assumption is supported by another study showing that single i.v. administration of 60 μg antibodies specific to oligomeric tau for 60 d was sufficient to reverse both locomotor and memory deficits in a not very aggressive Htau mouse model of tauopathy [
85].
Overall, these data suggest that ideally various routes of delivery, variable doses and different mouse models should be tested to properly assess the antibody’s ability to reduce the tau pathology. However, even after such careful testing it is difficult to predict the efficacy of antibodies in humans. The majority of mouse models of tauopathies constitutively overexpress mutated human tau and rapidly develop severe pathology, although not completely resembling human disease. We believe that the modest efficacy of many anti-tau antibodies currently tested in preclinical studies might be associated with very low amount of antibody reaching CNS for clearing such severe pathology. Therefore, various groups suggest testing the efficacy of anti-tau biologics and drugs in
in vitro or
ex vivo experimental systems such as (i) reduction of aggregation of tau; (ii) inhibition of toxicity of oligomeric tau and inhibition of propagation of tau [
86]. As mentioned above, the efficacy of Armanezumab was shown in all these assays, therefore, to save time and reduce the cost of non-clinical studies we decided to evaluate the ability of Armanezumab to reduce pathological tau after intracranial injection of mice. These
in vivo data demonstrated that 2 μg/mouse of Armanezumab in 5 days could reduce not only total tau, but also various pathological tau molecules phosphorylated at positions S212, S214, Ser396/404 (PHF1), Thr212/Ser214 (AT100), Thr231 (AT180), Ser202/Thr205 (AT8). Fine epitope mapping revealed the epitope PRQEF at positions tau 4–8 with the most essential amino acids at positions 6&7, indicating that Armanezumab is different from antibodies specific to N-terminal tau region reported previously by other groups [
43,
69]. Based on the fact that only a tiny amount of circulating antibody passes the BBB and enters the brain [
87‐
89] we think that more feasible route of administration for antibodies might be intranasal route as a non-invasive and efficient approach of delivering therapeutics to the brain that circumvents systemic extraction/alteration [
90,
91].
Methods
Mice, peptide epitope vaccine, and immunizations
Female, 6–8 weeks old B6SJL (H-2
bxs haplotype) and PS19 Tau/Tg mice were obtained from the Jackson Laboratory. THY-Tau22 Tau/Tg mice (generous gift of Dr. Luc Buee, Inserm) were bred at UCI animal facility. All animals were housed in a temperature and light-cycle controlled facility, and their care was under the guidelines of the National Institutes of Health and an approved IACUC protocol at University of California, Irvine. B6SJL mice were immunized with MultiTEP platform-based vaccine, AV-1980R, as previously described [
92]. Briefly, mice were injected subcutaneously (SC) with 50 μg/mouse dose of AV-1980R formulated in Quil-A adjuvant (Sigma-Aldrich, MA). All mice were boosted at 2-week intervals. Sera were collected on day 12 after each immunization. Mice were terminated 4 days after the 4
th immunization and splenocytes were used for fusion and selection of hybridomas producing anti-Tau antibodies.
Peptides were synthesized by standard chemical peptide synthesis and purity was analyzed by HPLC (GenScript, NJ).
Generation of hybridoma
Mouse spleen was aseptically removed, cells were washed out from the spleen by perfusion with 10 ml of sterile serum-free DMEM/F12. Splenocyte suspension was passed through sterile Falcon 0.70 μm cell filter followed by two wash steps. Splenocytes were fused with SP2/0 myeloma (Sp2/0-Ag14 ATCC® CRL-1581) cells per standard protocols and selected in HAT media. Hybridoma supernatants were screened by ELISA for reactivity with tau2–18 peptide. Positive hybridomas were subcloned by limiting dilution. Fusion procedure and subcloning have been performed in Recombinant Protein Production Core (rPPC), CLP at Northwestern University. Twenty-eight clones were selected and screened for binding to brain sections from the AD case. Clone 1C9 has been selected and used for further characterization and humanization.
Monoclonal antibody sequencing
Total RNA was isolated from hybridoma cells following the technical manual of TRIzol® Plus RNA Purification System and analyzed by agarose gel electrophoresis. RNA was reverse transcribed into cDNA using isotype-specific anti-sense primers or universal primers following the technical manual of SuperScriptTM III First-Strand Synthesis System. The antibody fragments of VH, VL, CH and CL were amplified according to the standard operating procedure for RACE at GenScript (Genscript, Piscataway, NJ). Amplified antibody fragments were separately cloned into a standard cloning vector using standard molecular cloning procedures. Clones with inserts of correct sizes were identified by PCR screening. No less than five single colonies with inserts of correct sizes were sequenced for each antibody fragment.
Humanization of mouse antibody
1C9 antibody was humanized by Genscript using proprietary technology. Variable domain sequences were blasted against human germline and several FR1, FR2, and FR3 were selected independently from human FRs, which share the highest identity with the mouse antibody. Selected FRs were assembled with 1C9 CDRs using overlapping PCR and phage display library was constructed for expression of Fab fragments. High tau protein-binding phages were selected after three rounds of panning using Genscript’s proprietary FASEBA screening methodology. Selected Fab genes were amplified from phage DNA. Genes encoding Fab were fused with genes encoding the appropriate constant regions of human IgG1 in order to generate whole IgG. Resulting light chain and heavy chain constructs were cloned into the mammalian expression vectors pcDNA3 and pGN-M, respectively.
Purification of Armanezumab
Stable DG44 cell line expressing Armanezumab mAb was generated by clone selection with increased concentrations of methotrexate (MTX). DG44 cells of selected clone were re-suspended at a density of 5x105 cells/ml and 600 ml of cell suspension was seeded into 1 L shaker flasks. 10% (v/v) Feed B solution was added to the culture on day 2, day 4, day 6, day 8. Glutamine was added to maintain the concentration range from 1 mM to 4 mM, and glucose was added to the culture to maintain the concentration range from 4 g/L to 10 g/L. Cell counting was conducted daily to determine the cell density and viability. The collected cell culture supernatants were used for purification when the cell viability was below 50%. Cell culture broths were centrifuged and followed by filtration. Filtered supernatant was loaded onto Protein A CIP column 65 ml (GenScript, Cat.No.L00433) at 8.0 ml/min. After washing with PBS, elution with 50 mM Citric acid, pH 3.0 and neutralization with 1 M Tris–HCl, pH 9.0, the eluted fractions were desalted by HiPrep 26/10 Desalting column with PBS. The purified antibody was analyzed by SDS-PAGE, Western-blot and HPLC by using standard protocols for molecular weight, yield and purity measurements.
Epitope mapping of tau-specific antibodies
Epitope mapping of anti-tau antibodies was performed by “alanine scanning” using competitive ELISA. Briefly, 17 peptides spanning tau2–18 sequence, but possessing one alanine substitution in each position were synthesized. 96-well plates (Immulux HB; Dynex Technologies, Inc., VA) were coated with 1 μg/well (in 100 μl; Carbonate-Bicarbonate buffer, pH 9.6, o/n at 4 °C) tau2–18 peptide (GenScript, NJ). Next day coated plates were blocked with blocking buffer (3% dry, non-fat milk in TBST, 300 μl/well). Serial dilutions of reference wild type (tau2–18), or mutated test peptides (corresponding to 0.02 μM, 0.1 μM, 0.5 μM, 2.5 μM and 12.5 μM final concentrations) were incubated with anti-tau antibodies (0.04 mg/ml final concentration) for 1.5 h at 37 °C. After incubation 100 μl of antibody/peptide mixture were added into the wells. HRP-conjugated goat anti-mouse IgG (1:2500; Jackson ImmunoResearch Laboratories, PA) in case of 1C9 and HRP-conjugated goat anti-human IgG (1:2500; Jackson ImmunoResearch Laboratories, PA) in case of Armanezumab were used as secondary antibodies. The reaction was developed by adding 3,3’,5,5’tetramethylbenzidine (TMB) (Pierce, IL) substrate solution and stopped with 2 M H2SO4. The optical density (OD) was read at 450 nm (Biotek, Synergy HT, VT). The percent of binding of 1C9 and Armanezumab antibodies blocked with wild type or mutated peptides to tau2–18 was calculated relative to the binding of 1C9 and Armanezumab antibodies without competing peptides to tau2–18 as 100%. The half maximal inhibitory concentration (IC50) for each peptide was calculated by Excel.
Recombinant tau protein preparation
Gene encoding 0N4R tau protein was amplified from human whole brain Marathon®-Ready cDNA library (Clontech) using primers 5’-
catatggctgagccccgccaggagttcgaagtgatg (forward) and 5’-
ctcgagtcacaaaccctgcttggccagggaggcagac (reverse) and cloned into the pET24a + E.coli expression vector in frame with 6xHis-tag at the C-terminus using restriction sites NdeI and XhoI. The sequence of the cloned tau protein was verified by DNA sequence analysis. Gene encoding tau, containing deletion of 2–18 region was amplified from plasmid carrying full-length tau gene and cloned into the pET24a + vector. Both recombinant proteins were purified from
E. coli BL21 (DE3) cells transformed with pET24a+/tau or pET24a+/tauΔ2–18 plasmids. Cells were grown up to the optical density (OD) 0.7–0.8 at 600 nm in the Luria Bertani medium containing 100 μg/ml of kanamycin. Gene expression was induced by adding isopropyl-β-D-1-thiogalactopyranoside (IPTG) at a final concentration of 1 mM and incubating for 4 h at 28 °C. Cells harvested by centrifugation were re-suspended in the B-PER reagent (Pierce) disrupting the integrity of cells. NaCl was added to a final concentration of 500 mM to the lysate and solution was boiled for 20 min. Most of the proteins precipitated following heat denaturation, while tau remained in the solution. Denatured proteins were removed by centrifugation and 6xHis-tagged tau was purified from the final supernatant using Ni-NTA columns (Qiagen, CA). Positive fractions were combined and concentrated by centricon filters (Millipore, MA). Concentrated protein fractions were analyzed by 10% Bis-Tris gel electrophoresis (NuPAGE Novex Gel, Invitrogen, CA) (Fig.
1).
Preparation of oligomeric and fibrillar tau
Fibrillar and oligomeric forms of tau (cross-linked and non-cross-linked) have been kindly provided by Dr. Kayed (UTMB Neurology). Oligomeric Tau used in neurotoxicity assay were prepared according to protocol described in [
93], using seed oligomers obtained from Dr. Kayed. Briefly, 7 μl of tau oligomers 0.6 mg/ml were added to 1 ml of recombinant tau protein solution of the same concentration and incubated for 1, 2, 4.5 and 24 h on an orbital shaker.
Preparation of human and mouse brain homogenates
Frozen blocks of post mortem human brains from AD (Braak stage VI) and control (Braak stage 0) cases were received from the Brain Bank and Tissue Repository, MIND, University of California, Irvine. 0.2 g of each brain tissue were homogenized in 0.4 ml TBS buffer with Halt™ Protease and Phosphatase Inhibitor Cocktail (100X, Thermo Scientific, CA), then centrifuged at 6400xg for 15 min at +4 °C. Supernatants were collected and stored at−80 °C for further analysis as soluble fractions.
Mouse brain homogenates were prepared exactly as described in [
37], More specifically, brain tissue was suspended in 10% (wt/vol) ice-cold TBS containing cOmplete protease inhibitors (Roche). Tissue was homogenized at 4 °C using a probe sonicator (Omni Sonic Ruptor 250) at 30% power, receiving 25 pulses. Lysates were centrifuged at 21,000 × g for 15 min to remove cellular debris and large, insoluble material. Supernatants were aliquoted and stored at−80 °C until further use.
Western blot
Western Blot analyses of recombinant full-length 4R/0 N tau, tauΔ2–18 and soluble fractions from brain homogenates were performed to confirm the specificity and binding ability of 1C9 and Armanezumab to pathological human tau. Commercial HT7 and TNT-1 antibodies have been used as positive controls. Briefly, brain homogenates containing equal amounts of total protein in SDS sample buffer (non-reducing conditions without heat incubation) were subjected to electrophoresis in 10% Bis-Tris polyacrylamide gel in MES buffer (Invitrogen, CA), then electro-transferred onto nitrocellulose membrane (GE Healthcare, NJ). The membranes were blocked overnight with 5% fat-free dry milk in TBS with 0.05% Tween following by detection of tau using 1C9, Armanezumab, HT7 (Life Technology, CA) or TNT-1 (Millipore, MA) monoclonal antibody and appropriate HRP-conjugated secondary antibody. All primary antibodies were used at concentration 1 μg/ml. Proteins were visualized with enhanced chemiluminescence detection using Luminol reagent (Santa Cruz Biotechnology, CA).
Dot blot
Binding to recombinant tau
1 μl of each form of tau protein (monomeric, oligomeric and fibrillar) was applied to nitrocellulose membrane (GE Healthcare, NJ). The membrane was air-dried, blocked overnight with 5% fat-free dry milk in TBS with 0.05% Tween following by detection of tau using 1C9 monoclonal antibody (1 μg/ml) and HRP-conjugated anti-mouse secondary antibody (0.2 μg/ml).
Binding to tau in human brain tissues
Soluble fractions of brain extracts from AD cases and controls were applied to membrane (2 μg of total protein in 1 μl volume) and proteins were detected using mouse 1C9, TNT-1 (Millipore, MA), HT7 (Life Technology, CA) monoclonal and rabbit polyclonal anti-Tau antibodies (BioLegend, CA). All primary antibodies were used at concentration of 1 μg/ml. Bovine anti-mouse and mouse anti-rabbit HRP-conjugated secondary antibodies were used at concentration of 0.2 μg/ml (Santa Cruz Biotechnology, CA). Proteins were visualized with enhanced chemiluminescence detection using Luminol reagent (Santa Cruz Biotechnology, CA).
Binding of Armanezumab to pathological tau in human brain tissues
1C9 or Armanezumab (1.0 μg/ml) were tested for the ability to bind to tau tangles in the human brain using 50 μm brain sections of formalin-fixed cortical tissues from severe AD (inferior parietal gyrus,
n = 6), FTD (midfrontal cortex,
n = 3) and Pick’s Disease (midfrontal cortex,
n = 3) cases and normal control brains (inferior parietal gyrus,
n = 2), all received from Brain Bank and Tissue Repository, MIND, UC Irvine, using immunohistochemistry as described previously [
92]. A digital camera (Olympus) was used to capture the representative images at 40x (for IC9) and 60x (for Armanezumab, PHF1, AT8, AT100, TNT-1, and HT7) original magnifications.
Surface plasmon resonance (SPR) analyses
SPR binding studies were performed on the BIAcore T200. Antibodies were immobilized on the surface of the biosensor chip CM5 series S (GE Healthcare) through Fc capture. Serial dilutions of tau recombinant protein in the running buffer containing 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, 0.005% Tween 20, pH 7.4, were injected at 5 mL/min over immobilized antibody, and the kinetics of binding/dissociation were measured as a change in the SPR signal (in resonance units). Each injection was followed by a regeneration step consisting of a 12-s pulse of 50 mM HCl. Fitting of experimental data was done with Biacore T200 evaluation software, version 1.0 using a 1:1 interaction model to determine apparent binding constants.
Blocking of brain lysate seeding activity
Measuring of RD-CFP/YFP co-aggregation by FRET was described in detail previously [
36,
37]. Briefly, HEK293 cells were plated at 250,000 cells/well in a 12-well plate and co-transfected with plasmids encoding P301S tau repeat domain (RD) containing ΔK280 mutation and fused to cyan (RD(ΔK280)-CFP) or yellow (RD(ΔK280)-YFP) fluorescent protein at proportions 1:3 as described earlier. 15 h later, cells were harvested with 0.05% trypsin for 3 min at 37 °C, and then re-plated in a 96-well plate in quadruplicate. After 15 h brain lysates of P301S Tg mice untreated (prepared in 1X TBS with protease, phosphatase inhibitors) or pre-incubated for 16 h at 4 °C under constant rotation with Armanezumab (10 μg/ml) were added. Cells were then cultured for additional 24 h before FRET analysis by flow cytometry. For determination of baseline level of RD(ΔK280)-CFP/YFP endogenous aggregation, cells were cultured without addition of brain lysates or antibody. FRET flow cytometry was performed using MACSQuant VYB (Miltenyi). Integrated FRET density was calculated as percent of FRET-positive cells x median fluorescence intensity (MFI).
Brain lysate seeding activity after immunodepletion
Armanezumab and control antibodies (50 μg) were added to agarose beads with protein A and G, then covalently coupled using Crosslink Immunoprecipitation Kit (Pierce) according to the provided protocol. Then brain lysates were immunodepleted four times with column elution between each immunodepletion. The final immunodepleted lysate was normalized to 0.97 μg/ml protein (measured by BCA), co-incubated with lipofectamine® reagent and added to HEK293T cell line (kindly provided by Dr. Diamond) constitutively expressing P301S tau repeat domain (RD) fused with either cyan protein (CFP) or yellow protein (YPF). Overall 10.3 μl of lysate, 1.25 μl of lipofectamine and 13.45 μl of optimem were used for each well. The incubations of lipofectamine were done according to the recommended protocol for DNA transfection. After 24 h incubation cells were harvested by trypsinization and tested for FRET by flow cytometry. Integrated FRET density was calculated.
Neurotoxicity assay
Neurotoxicity of oligomeric tau was assessed in SHSY-5Y cell culture and in primary neurons using MTT (3-[4,5-dimethylthiazol-2-yl]-2.5-diphenyltetrazolium bromide) assay performed as described previously [
31,
32] in the presence of tau oligomers plus or minus test antibody (Armanezumab). Incubation time of cells with tau oligomers and tau oligomers/antibodies was 16 h, and the final concentrations of protein and antibodies were 0.5 μM and 1 μM, respectively.
Intracranial injection of antibodies and quantitative image analysis
Armanezumab and control IgG were delivered to the hippocampus and cortex using a stereotaxic apparatus. Briefly, tau/Tg mice were anesthetized with isofluorane, placed in the stereotaxic frame and injected with Armanezumab (2 μg/μl) into the hippocampus (1 μl/injection) and into the cortex (1 μl/injection) by a 5 μl Hamilton microsyringe (30-gauge) at an injection rate of 0.5 μl/min, using the following coordinates relative to Bregma: AP: −2.06, ML: ±1.75; DV: −1.95 pocket 1 μl, and −1.00 1 μl. Control IgG was injected into the contralateral hemisphere using the same technique. The right side of each brain received Armanezumab while the left side was injected with the control IgG. Mice were sacrificed five days after the injection via Nembutal overdose and transcardially perfused with ice-cold 0.01 M phosphate-buffered saline (PBS). Brains were rapidly removed and fixed in 4% paraformaldehyde in PBS (pH 7.4) for 24 h at 4 °C and then sunk in 20% sucrose (in PBS). Prior to sectioning, a notch was made in the bottom left cortex so that left and right sides could be readily distinguished after sectioning. Brains were then cut in 40 μm thick coronal sections on a Microtome slicing system (Leica) and stored in PBS with 0.05% NaN3 at 4 °C until used. Sections were mounted on the slides and dried. Slices were pretreated with sodium citrate buffer, pH 6.0 at 90 °C for 30 min. All sections were hydrogen peroxide quenched, blocked and incubated with primary antibody overnight at 4 °C. Total tau was detected with HT7 recognizing epitopes 159–163, (0.4 μg/ml), phosphorylated tau was detected with T212 (pT212, 1 μg/ml), S214 (pS214, 0.4 μg/ml), AT100 (pT212/pS214, 0.4 μg/ml), PHF1 1 μg/ml (pS396/pS404, 2 μg/ml), AT8 (pS202/T205, 0.4 μg/ml) (all from ThermoFisher Scientific, Waltham, MA except for PHF1 from Peter Davis), and AT180 (pT231, 0.4 μg/ml (Abcam, Cambridge, United Kingdom). Anti-human IgG antibody was used to determine the antibody diffusion area in the brain parenchyma and anti-NeuN antibody was used to define the neuron density in injected sites. Incubations with appropriate biotinylated secondary antibody and ABC for 1 h were performed followed by color development using DAB (3,3’-diaminobenzidine) substrate kit (both from Vector Labs, CA). Images were captured by an Olympus microscope. Immunostaining profiles were observed by the means of a Sony high resolution CCD video camera (XC-77) and NIH Image version 1.59b5 software. For every animal, the images of the CA1 areas of three sections per each antibody were captured with the 10X objective. For the quantitative analysis, the images were cropped to the size of 1000 × 800 pixels at 600 dpi resolution, imported using ImageJ software (Scion) and converted to black and white using automatic binarization script. Thresholds were calculated for each picture separately using the same algorithm, and the number of pixels expressing staining density was determined in CA1 (injection) area for both ipsilateral and contralateral hemispheres. Density is expressed in the percentage units calculated using formula I/(I + C)x100 and C/(I + C)x100 where I-ipsilateral, C-contralateral. Average and standard deviation for each hemisphere between all mice were calculated and compared.
Pharmacokinetic data analysis
C57BL6 and PS19 tau/Tg mice received a single 600 μg/mouse intravenous (IV) dose of Armanezumab. Blood samples were collected by retro-orbital bleeding for PK at the following time points: 1, 4, 7, 11, 14, 18, 21, 28 and 54 days. Three mice per group underwent bleeding for each time point so that blood was drawn from an individual mouse only for every fourth time point. Concentrations of Armanezumab in the sera were determined by ELISA as described in [
94] except that plates were coated with 2.5 μM recombinant tau protein and purified 1C9 mAb and Armanezumab were used for calibration curve. PK parameters were calculated by non-compartmental analysis of the mean concentration values for each mouse using WinNonlin, version 5.2 (Pharsight). The following standard pharmacokinetic parameters were determined: (i) the maximum concentration (C
max); (ii) the time of C
max occurrence (t
max); (iii) area under the sera concentration-time curve from time zero to the time of the last measured serum level (AUC); (iv) t
1/2, calculated by linear regression of the logarithm of plasma concentration-time curve; (v) elimination constant kel = 0.693/t
1/2; (vi) apparent volume of distribution (V
d) obtained by calculating V
d = dose/AUC x k
el.
Statistical analysis
All statistical parameters [mean, standard deviation (SD), significant difference, etc.] used in experiments were calculated using Prism 6 software (GraphPad Software, Inc.). Statistically significant differences were examined using unpaired t-test or Ordinary one-way ANOVA Tukey’s multiple comparisons test (P value less than 0.05 was considered as statistically different).
Acknowledgements
We thank Dr. Izolda Popova (Recombinant Protein Production Core (rPPC), CLP Northwestern University) for generation of hybridomas; Dr. Anatoly Kiyatkin (Yale University, School of Medicine) for help in interpretation of SPR data; Dr. Grigor Mamikonyan (Clinartis LLC) for his assistance in calculation of pK parameters; Dr. Rakez Kayed and Urmi Sengupta (UTMB, Department of Neurology, Mitchell Center for Neurodegenerative Disorders) for providing oligomeric and fibrillar forms of Tau and technical assistance in conducting neurotoxicity assay with primary neurons; Dr. Marc Diamond (UT Southwestern, Center for Alzheimer’s and Neurodegenerative Diseases) for help in performing seeding activity inhibition assay and Dr. Guriqbal Basi for editing the text and valuable comments.