Highly Specific and Sensitive Target Binding by the Humanized pS396-Tau Antibody hC10.2 Across a Wide Spectrum of Alzheimer’s Disease and Primary Tauopathy Postmortem Brains

Abstract

Background:

Deposits of hyperphosphorylated tau fibrils are hallmarks of a broad spectrum of tauopathies, including Alzheimer’s disease (AD).

Objective:

To investigate heterogeneity of tau pathology across brain extracts from a broad selection of different tauopathies and examine the binding properties of the humanized pS396-tau antibody hC10.2 and six other anti-tau antibodies.

Methods:

76 individual tauopathy tissue samples were analyzed in a battery of assays: immunohistochemistry, ELISA, tau aggregation assay, western blot, [³H]PI-2620 and [³H]MK-6240 tau tracer binding, and aggregated seeding activity in RD_P301S HEK293T Biosensor cells. The efficiency of seven anti-tau antibodies to engage with pathological tau species was directly compared.

Results:

Our data indicate that a strong correlation existed between the tau tracer binding, amount of tau aggregates, pS396-tau phosphorylation, and seeding activity. The hC10.2 antibody, which has entered clinical development, effectively engaged with its epitope across all individual cases of mid-stage and late AD, and primary tauopathies. hC10.2 was superior compared to other phospho- and total tau antibodies to prevent seeded tau aggregation in the biosensor cells. hC10.2 effectively depleted hyperphosphorylated and aggregated tau species across all tauopathy samples proportionally to the amount of tau aggregates. In AD samples, hC10.2 bound to ghost tangles which represent extracellular pathological tau species.

Conclusion:

S396 hyperphosphorylation is a feature of the formation of seeding-competent tau across different tauopathies and it is present both in intra- and extracellular pathological tau. hC10.2 represents an excellent candidate for a hyperphosphorylation-selective therapeutic tau antibody for the treatment of AD and primary tauopathies.

Keywords

Alzheimer’s disease human brain tissue humanized antibody PET tracer seeding and spreading tau aggregate tauopathy therapeutic monoclonal antibody

INTRODUCTION

Pathological species of the microtubule-associated protein tau (MAPT) form intracellular fibrillar deposits in neurons and glial cells. These deposits are hallmarks of a broad spectrum of age-related neurodegenerative diseases collectively referred to as tauopathies [1 –3]. The tau aggregates are detergent-insoluble and contain predominantly hyperphosphorylated tau [4]. Human tau, encoded by a single gene, has six isoforms generated by alternative splicing of exon 2, 3, and 10. These tau isoforms consist either of 3 repeats (3R) or 4 repeats (4R) of the microtubule binding (MTB) domains located in the C-terminal half of the tau protein, and zero to two inserts located in the N-terminal part [5]. The longest isoform of tau (2N4R) contains 85 putative phosphorylation sites (Ser, Thr, or Tyr) of which half have been confirmed experimentally [6]. Under normal conditions, cytoplasmic tau contains on average two to three phosphorylated sites. Pathological tau is hyperphosphorylated with at least seven sites phosphorylated [7, 8]. Alzheimer’s disease (AD), the most prevalent of the tauopathies, is categorized as a secondary tauopathy. The prevalent hypothesis [9] suggests that accumulation of amyloid-β peptides triggers a pathological cascade resulting in formation of pathological tau species. This is considered critical as the severity of cognitive decline correlates most strongly with the burden of tau deposits in form of neurofibrillary tangles [10]. The neuropathological diagnosis of AD is based on these intraneuronal tau lesions classified as Braak staging. In AD, accumulation of tau aggregates is detected predominantly in neurons and consist of mixed 3R/4R tau isoforms.

In primary tauopathies, a sub-population of frontotemporal lobar degenerative diseases (FTLD-tau), tau aggregation is the main pathological driver [11]. FTLD presents with a range of different clinical symptoms and can be biochemically defined by tau lesions predominantly composed of 3R (e.g., Pick’s disease, PiD) or 4R tau isoforms (e.g., corticobasal degeneration (CBD) and progressive supranuclear palsy (PSP)) (for review, see [12, 13]). Investigation of the role of tau pathology in the onset and progression of tauopathies is now possible due the recent development of tau-specific tracers for use with positron emission tomography (PET), including second-generation compounds such as [¹⁸F]MK-6240 and [¹⁸F]PI-2620 [14]. Both tau tracers demonstrated high affinity binding to neurofibrillary tangles in brain regions with plausible tau lesions in AD subjects [15, 16]

Diverse isoform compositions have an impact on the structure of the fibril cores (and vice versa), as demonstrated for the recently solved cryo-EM structures for tau fibrils derived from different human tauopathies [17 –23]. Difference in fibril core structures between tauopathies indicate that epitopes on tau fibrils can be differentially displayed, depending on the disease and post-translational modifications of the tau protein. The latter has been extensively studied in AD, but less in primary tauopathies. The complexity of post-translational modifications of pathological tau makes it difficult to discern functional implications of individual modifications in AD. This contributes to a heterogeneity in epitope display of individual tau species which in return can have implications for the ongoing efforts to develop tau immunotherapies. Based on the hypothesis of spatiotemporal progression of tau pathology through neuronal networks and along major afferents [24], seeding-competent pathological tau species may propagate pathology from one neuron to interconnected cells, and spreading could represent a viable therapeutic target. Several therapeutic tau antibodies, including the pS396 tau specific antibody hC10.2, are in clinical development for AD and PSP [25, 26]. These programs are based on experimental data showing that pathological tau can transmit from cell to cell, and recruit and convert normal tau to disease associated species by templating [27]. Interneuronal spreading of aggregated hyperphosphorylated tau in various animal models [24, 28], induction of seeded tau aggregation in tau-expressing cells [29], and small tau fibrils as the main seeding-competent tau species have been demonstrated [30]. Subsequently, phospho-dependent and -independent tau antibodies were reported to reduce seeded tau aggregation, and propagation of tau pathology in different models [25, 31].

This study investigated the heterogeneity of tau pathology across brain extracts from a broad selection of different tauopathies, and the binding properties of hC10.2 compared to six other anti-tau antibodies. A battery of diverse histological, biochemical, and functional assays was applied to directly compare tau pathology from more than 70 tauopathy cases and the efficiency of hC10.2 to engage with pathological tau species. S396-tau phosphorylation of intraneuronal lesions was prominent across different brain regions and Braak stages from early to late AD. Additionally, hC10.2 bound to extra-neuronal ghost tangles, suggesting that the anti-pS396 antibody has a therapeutic potential to prevent tau seeding and spreading in tauopathies

MATERIALS AND METHODS

Human tissue

Frozen and fixed samples of frontal cortex from 25 cases of Braak stage III/IV AD and 25 cases of Braak stage V/VI AD were obtained from the Arizona Study of Aging and Neurodegenerative Disease and Brain and Body Donation Program at Banner Sun Health Research Institute (Table 2). The donated tissues were collected by rapid autopsy and complemented with data from clinical assessment and histopathological characterization [32]. Braak staging was assessed according to topographical progression of Gallyas-stained neurofibrillary changes.

Table 1

Antibodies used

Antibody	Tau epitope/Ag	Isotype	Source
HJ8.5	25–30	chIgG1	Kfoury et al., 2012 [71]
MC1	5–15/313–322	hIgG4	Jicha et al., 1997 [72]
AT8	P199/p202/p205	mIgG1	Thermofisher (MN1020)
Tau5	210–241	mIgG1	Thermofisher (AHB0042)
7C30	240–261	hIgG1	Lundbeck
DC8E8	151–391	hIgG1	Kontsekova et al., 2014 [73]
hC10.2	pS396	hIgG1	Lundbeck, Rosenqvist et al., 2018 [25]
IgG (B12)	Control	hIgG1	Burton et al., 1994 [34]
DAKO	243–441	r-polyclonal	Agilent (A002401-2)

ch, chimeric; h, human; m, mouse; r, rabbit.

Table 2

Demographics of human AD cases: frozen samples

Case No.	Gender	Braak score	Age at death (y)	Disease duration (y)	PMI (h)	Case No.	Gender	Braak score	Age at death (y)	Disease duration (y)	PMI (h)
AD1	F	IV	82	9	4.5	AD26	M	V	83	7	3
AD2	M	IV	93	1	2.75	AD27	M	V	82	8	2.95
AD3	F	IV	77	3	2	AD28	F	V	86	4	2.52
AD4	F	IV	88	3	3.25	AD29	M	VI	63	10	5.17
AD5	M	IV	86	7	2.16	AD30	M	VI	74	4	2.58
AD6	F	IV	91	6	1.83	AD31	M	VI	85	14	3.58
AD7	F	III	96	11	2	AD32	F	V	76	7	3.08
AD8	F	IV	96	8	3	AD33	M	VI	80	20	3
AD9	M	IV	82	9	3	AD34	F	V	85	8	3.11
AD10	M	IV	78	7	3.33	AD35	F	VI	92	17	3.73
AD11	F	IV	103	10	2.6	AD36	F	VI	85	8	3.42
AD12	M	IV	96	5	4.5	AD37	M	V	75	7	3.83
AD13	M	IV	77	9	3.25	AD38	M	VI	77	12	3.62
AD14	F	IV	91	1	3	AD39	M	V	85	11	3.5
AD15	F	IV	93	3	2.15	AD40	M	V	80	12	4.23
AD16	M	IV	82	2	2.67	AD41	M	V	82	4	3.98
AD17	M	IV	97	2	2.98	AD42	M	VI	73	11	4.83
AD18	M	IV	91	7	2.67	AD43	F	VI	81	10	3.85
AD19	F	IV	97	1	5	AD44	M	V	83	8	4
AD20	F	IV	97	3	3.22	AD45	M	V	80	19	3.83
AD21	F	IV	101	1	2.78	AD46	F	V	83	10	3.62
AD22	M	IV	88	8	2.83	AD47	F	VI	84	1	3.33
AD23	M	IV	97	2	1.5	AD48	M	V	83	5	3.58
AD24	F	IV	85	5	2.25	AD49	M	VI	83	2	3.08
AD25	M	III	83	3	4.16	AD50	M	VI	91	14	2.97

In addition, fixed sections of four different brain regions from each of 10 Braak III/IV cases and 10 Braak V/VI cases of AD were received (Table 3). A cohort characterization is shown in Fig. 1B.

Table 3

Demographics of human AD cases: fixed samples

Case No.	Gender	Braak score	Age at death (y)	Disease duration (y)	PMI (h)
AD1	F	IV	82	9	4.5
AD4	F	IV	88	3	3.25
AD5	M	IV	86	7	2.16
AD6	F	IV	91	6	1.83
AD7	F	III	96	11	2
AD9	M	IV	82	9	3
AD11	F	IV	103	10	2.6
AD51	F	IV	91	10	2.5
AD52	M	III	90	2	3.75
AD53	M	II	86	1	2.75
AD54	M	V	81	13	3
AD55	F	V	96	2	2.75
AD56	M	V	86	21	3
AD57	F	V	87	6	3
AD58	F	V	89	9	2.33
AD59	M	V	70	11	2.33
AD60	F	V	86	9	2.5
AD61	M	VI	77	7	3.5
AD62	F	V	86	5	2.5
AD63	M	V	89	11	2.2

Fig. 1

A) Schematic diagram illustrating binding sites of the tau-specific antibodies. B) Clinical and histopathological characteristics of postmortem frozen brain samples, n = 25 stage III/IV, n = 25 stage V/VI. Braak staging was assessed according to topographical progression of Gallyas-stained tangles [32]. Tangle and plaque scores represent density scores accumulated from 5 brain regions, with a maximum score of 3/region. Mann-Whitney U-test: ¤¤¤¤p < 0.0001. Unpaired t-test: ^**p < 0.01, ^***p < 0.001, ^****p < 0.0001. C) Immunohistochemistry of prefrontal cortex sections. Representative images from end stage case AD29 are shown. A positive signal was obtained with all anti-tau antibodies in AD cases with detectable levels of tau pathology. The IgG control did not stain tau. Scale bar = 50μm.

For tauopathy studies, frozen and fixed samples of frontal cortex from five cases each of PSP, CBD, and PiD, and three age-matched controls were donated from Banner Sun Health Research Institute. Frozen and fixed frontal cortex samples from seven cases of FTD, six cases of CBD, and two cases of primary progressive aphasia (PPA) were generously donated from London Neurodegenerative Diseases Brain Bank (Table 4). To maintain patient anonymity, all samples included in this study were blinded and designated with a code. Approval for this study was obtained from the MRC London Neurodegenerative brain bank and from Banner Sun Health Research Institute.

Table 4

Demographics of human non-AD tauopathy cases: frozen and fixed samples

Banner Sun Health Research Inst.				MRC London Neurodegenerative Diseases Brain Bank
Case No.	Gender	Age at death (y)	PMI (h)	Case No.	Gender	Age at death (y)	PMI (h)	Pathological diagnosis	Mutation
PiD1	F	74	3.75	CBD3	M	73	n.k.
PiD2	F	61	4.25	CBD4	F	73	54
PiD3	M	70	3.85	CBD5	F	84	50
PiD4	F	69	3.43	CBD6	M	73	46
PiD5	M	75	2.88	CBD7	F	63	36
PSP1	F	73	2.71	FTDL1	M	83	37	FTD	His362Tyr
PSP2	F	64	2.63	FTDL2	M	71	5	FTD	Tau 10 + 16
PSP3	M	90	3.5	FTDL3	M	48	65	FTD	Tau 10 + 16
PSP4	M	77	3.22	FTDL4	M	67	35	FTD	Tau 10 + 16
PSP5	M	86	2.55	FTDL5	F	63	22	FTD	Tau 10 + 16
CBD1	M	73	2.75	FTDL6	F	62	30.5	FTDP-17	MAPT
CBD2	M	84	4.3	FTDL7	M	62	n.k.	FTDL^*
HC1	F	82	2.07	PPA1	M	72	5.6
HC2	M	79	4.3	PPA2	F	77	34
HC3	M	84	4.3

n.k., not known. ^*FTDL favoring CBD. CBD, corticobasal degeneration; FTD, frontotemporal dementia; FTDL, frontotemporal lobar dementia; FTDP-17, frontotemporal dementia with parkinsonism-17; HC, healthy control; PiD, Pick’s disease; PPA, primary progressive aphasia; PSP, progressive supranuclear palsy.

Recombinant antibody production

Non-commercial antibodies were generated based on sequences derived from patents (WO2016079597A1, WO2014008404A1, WO2016137811A1). Synthetic genes for the heavy and light chains were cloned into the pTT5 vector (NRC-CNRC) and used for transient expression in HEK293 6E cells. Antibodies were purified from clarified culture media by protein-G Sepharose (GE healthcare) affinity chromatography followed by Q-sepharose chromatography in flow through mode (buffers selected according to pI of antibody). Purified antibodies were analyzed for purity by SDS-PAGE by both NuPage gels (Thermofisher) and the P230 chip on Bioanalyzer 2000 for quantification of purity; Aggregation by UPLC-SEC; concentration by NanoDrop (Thermofisher) using calculated extinction coefficient [33] and A280; and endotoxin by Endosafe PTS (Charles River). Confirmation of the antibody identity was done by analyzing the intact mass of the heavy chain and light chain by RP-uplc (BioResolve RP mAb Polyphenyl in 0.1% FA; ACN gradient) followed by ESI–MS on a VION QTOF instrument (Waters) and comparing the expected mass with the observed mass. Multicharged spectra were processed using the MaxEnt 1 deconvolution algorithm.

hC10.2 is the humanized version of the mouse monoclonal antibody specific for pS396-tau as described in [25]. The mouse monoclonal 7C30 antibody was generated by five monthly immunizations of Tg4510 mice with P30-pS262-tau peptide. Hybridoma clones were selected based on binding activity in plates coated with phosphorylated and non-phosphorylated S262-tau (240–270) peptide, and AD-P3 brain material (previously described in [25]). 7C30 proved to be a phospho-independent antibody and the current study was conducted with a mouse/human chimeric version of 7C30. An IgG raised against HIV-1 served as negative control antibody throughout [34].

Immunohistochemistry

Formalin-fixed paraffin-embedded sections from AD cases were deparaffinized, rehydrated and subjected to antigen retrieval by boiling in 10 mM Citrate buffer, pH 6, and pretreated with 99% formic acid for 7 min. The sections were incubated o/n at 4°C with the antibodies listed in Table 1. The concentration of the individual antibodies was titrated to yield the strongest immunoreactive signal (HJ8.5, 7C30, hC10.2 at 1μg/ml; MC1, DC8E8 at 2μg/ml; AT8 at 1:1000; Tau5 at 1:2000). Following thorough rinsing, sections were incubated with biotinylated anti-human (Sigma #B1140, 1:200) or anti-mouse (Dako #E0354, 1:500) antibodies. Immunoreactivity was visualized using the Avidin-Biotin-Complex (Vector) and 0.05% 3,3’-diaminobenzidine. Paraffin sections from tauopathy tissue were processed the same way and reacted with hC10.2 at 1μg/ml.

Immunohistochemical co-localization

Tissue sections were processed and analyzed by Gubra (Hørsholm, Denmark). Paraffin sections of four brain regions; locus coeruleus (LC), entorhinal cortex (EC), hippocampus CA1 region, and prefrontal cortex (mediofrontal gyrus; PFC) from each of 10 Braak III/IV and 10 Braak V/VI cases were pretreated for antigen retrieval as described above. One section per brain region was analyzed. Double immunohistochemical staining was performed by simultaneous incubation with primary antibodies hC10.2 (5μg/ml) and AT8 (1:1000).

Secondary Cy3-conjugated anti-human (Jackson #709-165-149, 1:200) and AF488-conjugated anti-mouse (Life Technology #A21202, 1:400) antibodies were applied. The immunohistochemical procedures were performed using the Dako autostainer. The sections were counterstained with Dapi, and autofluorescence was quenched by 10 min incubation in 0.3% Sudan Black B in 70% ethanol. The slides were digitalized using the Olympus VS120 fluorescence scanner. A semi-quantitative assessment was performed by delineating regions of interest using the newCAST software (Visiopharm, Denmark) and counting the number of positive immunostained neurons in a fixed number of field of views for each region. Double positivity was not caused by spectral bleed-through. All layers were analyzed in EC and PFC. In hippocampal sections, analysis was performed in the CA1 region. LC neurons were identified by their neuromelanin content. The data are presented as relative densities and ratios of single-stained and double-stained neurons. Number of neurons counted per section: LC 10–14; EC 335–750; CA1 160–420; PFC 84–640.

Prefrontal cortex homogenization and depletion

Crude homogenates were prepared from grey matter samples of AD tissues and from grey and white matter samples of tauopathy tissues as described by Clavaguera et al. [24]. Homogenates were immunodepleted as previously described [25]. In brief, 18μg of the respective antibodies was immobilized to 100μl of magnetic Dynabead suspension (Immunoprecipitation Kit Dynabeads Protein G Novex). After washing, the beads were mixed with 100μl crude tissue homogenate (3–5 mg/ml total protein) and incubated at room temperature for 10 min. The homogenates were separated from the magnetic beads and stored at –80°C until use.

Western blot analysis

Input and immunodepleted AD homogenates were analyzed by western blot as previously described [25]. A phospho-independent total tau polyclonal rabbit antibody (Dako, A002401, 1:1000) was applied to detect total tau; and the pS396 antibody hC10.2 (1μg/ml) was used for detection of hyperphosphorylated tau species and detected using the goat anti-rabbit: IRDye 800RD (cat no. 926-32211) and goat anti-mouse: IRDye 680RD (cat no. 926-68070) for detection. For each case, the input and immunodepleted material were included on the same blot. The six major tau bands were quantified for signals from both detection antibodies using Azur instrument and the Image studio software (LI-COR biosciences). After scanning the membranes, the blots were re-probed using a GAPDH antibody (Abcam ab9484) and the signal was detected using the IRDye 680RD. The tau signal was normalized to the GAPDH signal. The effect of tau depletion is presented in relation to the input material, which was set to 100%.

MSD measurement to determine pS396 tau in AD brain homogenates

High sensitivity sandwich immunoassay on the MSD® S-PLEX assay platform using biotionylated hC10.2 as capture antibody bound to streptavidine-coated plates was established and carried out by MSD®. Captured tau antigens from standard and brain homogenate samples were detected with sulfo-tagged antibodies specific for N-terminal tau (HT7 - Thermo Scientific MN1000 and Tau 12 - Biolegend 806502) and mid-domain tau (7C30). An MSD-calibrator (E. coli-derived total-tau calibrator) was used as standard curve to calibrate tau concentrations in homogenate samples from AD cases.

Measurements of tau aggregates in brain homogenates via TR-FRET

All human brain homogenates were assessed in the commercial HTRF Cisbio tau aggregation immunoassay. This assay is based on the principle of using a mixture of the same tau antibody labelled with 1) a FRET donor or 2) a FRET acceptor. A FRET signal is only obtained when the epitope is in a multimer form, thus quantifying aggregates selectively. Samples were measured at 6-, 18-, and 54-fold dilutions in Cisbio buffer D to ensure that aggregates were measured on the correct side of the ‘hook’ according to the manufacturer’s protocol. In brief, 9μl sample was mixed with 4.5μl Tb and 4.5μl D2 labelled antibody. After 20 h incubation, homogenous TR-FRET signal was measured using a Pherastar FSX multiplate reader equipped with appropriate filters. TR-FRET data were plotted for the 54-fold dilution calculated according to manufacturer’s instruction, normalized to BCA and plotted as Relative FRET. For the AD sample measurements, a signal was detected in all samples. Thus, a cut-off of 3 times the standard deviation of background was applied to define the lower limit of quantitation (LLOQ) of background. All samples were immunodepleted. However, only samples above LLOQ were plotted as it was questionable if samples below the LLOQ had any aggregates. This was reflected in the depletion results which showed an extreme degree of variability for samples below LLOQ. Based on this learning, an arbitrary cut-off of a relative tau aggregation level of 200 was applied to all tauopathy samples to select samples with enough tau aggregates to allow for analysis, and with a higher probability of having seeding activity.

Seeding assay

Seeding studies were performed with human brain homogenates from all 50 AD brains and 26 primary tauopathy brains in HEK293T Biosensor cells expressing Tau RD_P301S (ATCC CRL3275) according to slightly modified previously published protocols [29]. In brief, 35,000 Biosensor cells were plated in 130μl DMEM +10% FCS on collagen-coated black clear-bottom Cell-carrier plates (PerkinElmer) and incubated o/n in a cell incubator (37°C, 5% CO₂). The next day, 8μl crude brain homogenate was added to 72μl OptiMEM and mixed with 80μl lipofectamine Mix (10μl lipofectamine 2000 and 70μl OptiMEM). Twenty μl homogenate/transfection mix per well was gently added to the cells after a 20 min preincubation period and the cells were placed in the cell incubator for 48 h. Cells were fixed for 10 min at room temperature with 4% PFA/4% Sucrose in PBS, washed twice in PBS. FRET was measured by a plate reader (Pherastar FSX) using excitation/emission filters (FRETraw: 430 nm/530 nm, CFP: 430 nm/470 nm and YFP: 485 nm/520 nm) and reading the plate from the bottom. Background values (signal from cells with no fluorophore) were subtracted from all values and FRET signal was corrected for CFP and YFP channel crosstalk. FRET was calculated from formula: Net FRET = FRETraw – (a^*YFP signal) – (b^*CFP signal), where a is a constant set on YFP-only cells using the same instrument as a ratio of CFP(Ex)YFP(Em)/YFP(Ex)YFP(Em); and b is a constant measured on CFP-only cells using a ratio of CFP(Ex)YFP(Em)/CFP(Ex)CFP(Em) [35]. The Net FRET signal was normalized to CFP to account for any potential well-to-well variability in cell density and plotted as relative FRET = Net FRET/CFP(Ex)CFP(Em). For comparison of immunodepleted samples, all depleted samples were normalized to IgG controls. Samples with low seeding activity showed highly variable antibody effects. An arbitrary cut-off was applied to eliminate these samples from the analysis, i.e., data was plotted for samples showing robust seeding. Subsequently, 14 immunodepleted AD homogenates (AD3,5,11,20,26,27,28,29,31,32,34,37,38,46) and 17 primary tauopathy homogenates were included in the seeding assay.

Tau PET tracer binding assay

PI-2620 and MK-6240 were synthesized as described in the literature [15, 16]. [³H]PI-2620 and [³H]MK-6240 were prepared by Pharmaron (Cardiff, UK) by treating PI-2620 and MK-6240 with tritium gas followed by purification using HPLC. [³H]PI-2620 was obtained with a radiochemical purity of > 99.9% and a specific activity of 32 Ci/mmol, whereas [³H]MK-6240 was obtained with a radiochemical purity of 99.3% and a specific activity of 20 Ci/mmol.

Brain homogenates of frontal cortex grey matter were prepared according to [16].∥In brief, 80 mg wet weight tissue was homogenized in 1 ml ice cold PBS containing protease and phosphatase inhibitor cocktails (Roche, cat no. 11836170001, 04906845001) using an Ultra-Turrax^® homogenizer at 24.000 rpm. The homogenate was further diluted with ice cold PBS to 30 mg wet weight tissue per 1 ml, homogenized for an additional minute and stored at –80°C until use. Non-specific binding was determined by cross-blocking the binding with 1μM unlabeled ligand. Before the binding experiment, the tissues were quickly thawed and homogenized using an Ultra-Turrax^® and were further diluted to 1 mg/ml in binding buffer (PBS, pH 7.4, containing 0.1% BSA and protease and phosphatase inhibitor cocktails).∥Aliquots consisting of 200μl tissue, 25μl radioligand (1 nM final) and either 25μl buffer (total binding) or 25μl compound (non-specific binding, 1μM final) were incubated for 1 h at room temperature.∥Bound ligand was separated from free ligand by filtration using a FilterMate harvester (PerkinElmer) onto GF/C uni-filters pretreated with 0.1 % PEI for 1 h. Filters were subsequently washed twice with ice-cold buffer (50 mM TRIS, pH 7.4, 120 mM NaCl, 5 mM KCl, 4 mM MgCl₂-6H₂O and 1 mM EDTA). The filters were dried 20 min (37°C) before addition of OptiPhase SuperMix (Perkin Elmer) and counted in a MicroBeta^® TriLux (TriLux 1450, Perkin Elmer Wallac) counter for 5 min per well. Specific binding is presented as CPM and calculated by subtracting non-specific binding from total binding. In addition, the specific binding of either ligand was determined in a pooled AD homogenate (AD cases 26, 36, 39, 40) immunodepleted with hC10.2 or IgG control antibody. As control, we demonstrated that either tracer could block binding of the other tracer to the homogenate in a cross-blocking study (Fig. 2D).

Fig. 2

A, B) Tau PET tracer binding to AD cortex homogenates. Bars represent mean counts per minute±SD from 200μg total protein homogenate and 1 nM radioligand. [³H]MK-6240 and [³H]PI-2620 showed similar binding patterns and revealed a heterogeneous binding capacity across AD cases. C) The p396-tau content in the cortex homogenates measured by MSD show a linear correlation with tau PET tracer binding. Tracer binding and MSD were performed on two different homogenate preparations. D) Tracer binding of [³H]MK-6240 and [³H]PI-2620 was performed as in A, but in competition with cold PI-2620 or MK-6240, respectively. E) A pool of four Braak V/VI AD homogenates was immuno-depleted with control and hC10.2 antibodies and assayed for tau tracer binding. Bars represent means±SD. F) Tracer binding of [³H]MK-6240 and [³H]PI-2620 was performed as in A, but homogenates were preincubated with up to 0.5μM hC10.2 or control IgG for 1 h prior to tracer binding.

RESULTS

The pS396-tau hC10.2 antibody detects tau pathology in a large cohort of AD brains

We generated a highly specific and selective anti-pS396-tau antibody, C10.2, targeting seeding-competent tau [25]. C10.2 was humanized and hC10.2 has entered clinical development for the potential treatment of AD. In this study, the effective target binding of hC10.2 was evaluated in a large cohort of AD postmortem brains, representative of tau pathology at different progression stages. Binding of the hC10.2 antibody was compared with a panel of phospho and total tau (non-phosphorylated) antibodies targeting N- and C-terminals, as well as the mid-domain of tau (Fig. 1A, Table 1). This enabled us to compare the display of epitopes in a larger cohort of mid-stage Braak III/IV (n = 25) and late-stage Braak V/VI (n = 25) AD cases (Fig. 1B, Table 2) by applying an array of different assays to the same tissue samples. Immunohistochemistry revealed that all tau antibodies recognized the hallmarks of tau pathology from AD postmortem brains (Fig. 1C). At the concentrations and conditions applied, a strong immunoreactive signal was observed for the C-terminal hC10.2 and tau antibodies targeting mid-domain (AT8, Tau5, 7C30), while HJ8.5 and MC1 targeting the N-terminal displayed lower signals. The MTBR-specific antibody DC8E8 only bound to pathological tau after pretreatment of tissue sections with formic acid, indicating that DC8E8 recognized a cryptic epitope hidden within the neurofibrillary tangles.

Tau aggregation correlates well with pS396-tau concentration in AD brains

The diversity of tau pathology between individual donors was evaluated across Braak stages in crude homogenates of frontal cortex, a region where pathological tau accumulates at an advanced disease stage. Neurofibrillary tangle load measured by Braak staging correlated well with in vitro binding of tau tracers [³H]MK-6240 and [³H]PI-2620 in the 50 AD brain homogenates (Fig. 2A, B). These PET tracers bind primarily to aggregated tau [15, 16]. Additionally, tau tracer binding correlated with the concentration of pS396-tau in the homogenates (Fig. 2C), indicating a link between tau aggregation and phosphorylation at the S396 epitope in AD. In a cross-blocking study, it was demonstrated that either tracer could completely block the capacity of the other tracer to bind the target (Fig. 2D).

Tau aggregates were quantified using the commercial Cisbio HTRF tau aggregation immunoassay in individual homogenates at the same dilution (1:54) and normalized to total protein. The amount of tau aggregates varied between samples, but generally correlated with Braak staging (Fig. 3A, input). A strong linear correlation was also observed with the concentration of pS396-tau (Fig. 3C). Thus, a link was established between neurofibrillary tangle load determined by Braak staging and tau tracer binding, tau aggregation, and pS396-tau concentration in the corresponding AD brain homogenates.

Fig. 3

A) Heterogenous aggregation level across AD cohorts. Immuno-depleted homogenates from 50 AD cortex samples were analyzed in the Cisbio tau aggregation assay and plotted as the mean±SEM. Input material represents the crude homogenate. The AD cases were organized in the plot according to aggregation levels. LLOD defines lower level of detection; LLOQ indicates lower level of quantification calculated as LLOD+3XSD. B) Tau aggregation in immuno-depleted homogenates shown as percentage of control IgG. Only cases with a positive signal in non-depleted homogenates were included. Data are the mean±SEM of n = 11 and n = 23 of Braak III/IV and V/VI. One-way ANOVA with Dunnett’s multiple comparisons test versus IgG ^****p < 0.0001, ^***p < 0.001, ^**p < 0.01. C) The p396-tau content in the cortex homogenates measured by MSD was plotted against the aggregation level of the input material measured in (A) and data was fitted by a linear regression. Both assays were performed on the same homogenate preparation.

Efficient depletion of hyperphosphorylated and aggregated tau species from AD brains by hC10.2

Pooled AD homogenates were immunodepleted with hC10.2 which resulted in strongly decreased [³H]MK-6240 and [³H]PI-2620 binding signals (Fig. 2E). The reduced tracer signal was not caused by interference of hC10.2 with the PET ligands since concurrent incubation or preincubation of either tracer with up to 1μM hC10.2 did not hinder tracer binding to tau aggregates (Fig. 2F).

The efficacy of hC10.2 to bind to tau aggregates was assessed with the tau aggregation assay. The 50 crude AD homogenates were immunodepleted with hC10.2 and in parallel with the different other tau-specific antibodies and compared to IgG-depleted samples and input material. A lower limit of quantitation (LLOQ) of the average Cisbio signal+3xSD (99.7% confidence interval) from three control brain homogenates was applied to identify which AD samples contained measurable tau aggregates (Fig. 3A). Of the 25 + 25 AD samples, 11 Braak III/IV samples and 23 Braak V/VI samples generated positive signals above the set LLOQ. Tau antibody-depleted samples were normalized to their respective IgG controls and plotted as the average of all positive samples (Fig. 3B). hC10.2 efficiently depleted tau aggregates from all cases by approximately 90%, both for Braak III/IV and V/VI. This suggests that S396 phosphorylation is a general and prominent characteristic of pathological tau in mid- and end-stage AD (Fig. 3B). Similarly, HJ8.5 (N-terminal), AT8 (mid-domain p-tau), and MC1 (conformational) efficiently depleted tau aggregates. 7C30, which flanks the MTBR, only partly depleted tau aggregates. DC8E8, targeting the MTBR, did not deplete aggregates. Tau5 (targeting non-phosphorylated mid-domain) also failed to efficiently deplete tau aggregates (Fig. 3B). All immunodepleted brain samples were analyzed by western blot (Fig. 4A), demonstrating total and pS396 hyperphosphorylated tau species, the latter displayed as the AD typical three tau bands being mobility shifted on SDS page [1, 4]. The residual tau signals after immunodepletion of the 11 tau aggregate positive Braak III/VI and 23 Braak V/VI samples were plotted for both total tau and pS396 tau (Fig. 4B, C). The phospho-specific antibodies hC10.2 and AT8 efficiently depleted pS396-positive hyperphosphorylated tau, whereas they did not deplete total tau above the control IgG level. Phospho-independent tau antibodies depleted most of the total tau signal but varied in the efficiency to deplete hyperphosphorylated pS396-tau. HJ8.5 efficiently depleted hyperphosphorylated pS396-tau, whereas Tau5 and DC8E8 were less efficacious (Fig. 4B, C). MC1 depleted total tau efficiently but showed little discrimination between total and hyperphosphorylated pS396 + tau.

Fig. 4

A) AD homogenates were immuno-depleted as indicated and the amount of residual tau was analyzed by western blot using the total tau Dako (upper panel) or the pS396-specific hC10.2 antibody (lower panel). A representative blot from a single case (AD46) is shown. X, a non-disclosed antibody that is not included in the present study. B, C) The major tau bands were quantified as indicated by squares in A and normalized to GAPDH. Cases with a positive pS396 signal in the input material are included in the graphs. Data are the mean±SEM of n = 34. Phospho-specific antibodies (AT8 and hC10.2) did not deplete the normal tau bands, as demonstrated in the Dako western blot. One-way ANOVA with Tukey’s multiple comparisons test versus IgG ^****p < 0.0001, ^***p < 0.001.

hC10.2 efficiently depletes seeding-competent tau species from AD brains

The seeding potential of crude brain homogenates from the 50 AD samples were measured in the cellular seeding assay in HEK293T biosensor cells expressing tau RD P301S* using BioFRET as a readout for seeded tau aggregation. This offered a functional assay readout that was independent of detection antibodies. A considerable variation of induction of seeded tau aggregation was observed between individual cases, but with an overall good correlation to the Braak stages (Fig. 5A). Also, there was a linear correlation between seeding activity and tau aggregates measured by the tau aggregation assay (Fig. 5C). We have previously demonstrated that hyperphosphorylated forms represent seeding-competent tau species [25]. Fifteen individual sets of immunodepleted brain homogenates selected to have high and medium aggregate loads were tested for seeding activity. All antibodies, apart from DC8E8, showed some capacity to remove seeding-competent tau species. The numerically largest depletion of seeding activity was achieved with hC10.2 (Fig. 5B), indicating superiority of hC10.2 over the other anti-tau antibodies to engage with seeding-competent hyperphosphorylated tau species. A negative correlation between seeding activity of the brain homogenates and both age of death and disease onset of the AD patients was observed (Fig. 5D, E).

Fig. 5

Heterogeneous tau seeding potential. A) Cortical homogenates from 50 AD samples were tested for tau seeding activity in tau RD P301S-expressing HEK293 cells. Tau seeding was analyzed by lipofectamine-transfection of homogenates into Biosensor cells. B) Immunodepletion of homogenates prior to lipofectamine-transfection. Seeding activities are presented as aggregation normalized to total protein, and in relation to the respective IgG-depleted samples. Bars represent the mean of 14 AD cases±SEM. Input represents the non-depleted crude homogenate. One-way ANOVA with Dunnett’s multiple comparisons test versus IgG ^****p < 0.0001. C) The seeding potential correlated positively with tau aggregation measured in the Cisbio assay, and (D) negatively with the age of death and (E) age at disease onset. The age of onset was calculated as age of death - disease duration in years.

Prominent pS396-tau pathology in AD progression and detection of ghost tangles by hC10.2

Seeding-competent hyperphosphorylated tau species are hypothesized to transmit between neurons and template normal tau to pathological forms in AD. We studied the progression of pS396-tau pathology load along the hypothesized tau seeding and spreading pathway in AD [36]. The density of hC10.2-positive neurons in four brain regions from each of 10 mid- and 10 late-stage AD cases were analyzed and compared to AT8-immunoreactivity levels. Accumulation of AT8 antibody staining of specific brain regions forms the basis of Braak staging [37]. Relative densities of immunoreactive neurons were analyzed by AT8 and hC10.2 double immunofluorescence staining (Fig. 6A). Regions early in the seeding and spreading pathway (LC, EC) harbored comparable pS396-tau levels in mid and late stage brains. In the hippocampus CA1 subregion and PFC, the density of hC10.2-positive neurons increased in late-stage brains compared to mid-stage (Fig. 6A). Throughout the brain regions, hC10.2 IHC reflected the tau pathology levels detected by AT8 IHC (Fig. 6A). These IHC findings confirmed that phosphorylation of S396-tau was an early and prominent characteristic of pathological tau during AD progression in the affected brain regions.

Fig. 6

Progression of tau pathology along the hypothesized spreading pathway. A) The densities of hC10.2- and AT8-positive neurons were analyzed in four brain regions per AD case. Densities are relative and comparable between mid and end Braak stages; mean±SEM of n = 10 cases per region of each Braak stage. B) Double immunohistochemical staining of four brain regions per case with AT8 and hC10.2 antibodies. Representative images of end stage cases are shown. Neurons stained by the hC10.2 alone are predicted to represent ghost tangles (arrows). Scale bar = 50μm. C) The relative densities of immunoreactive neurons were semi-quantified for each region from mid and end stage AD (n = 10 cases per Braak stage, 1 section per case).

Ghost tangles are the remnants of dead neurons with neurofibrillary tangle pathology exposed to the extracellular environment. They represent a small, but potentially therapeutically important pool of extracellular tau species that are difficult to isolate in pure form with current fractionation methods. However, tau tangles lacking AT8-immunoreactivity have been reported to constitute a pool of extracellular and truncated tau [38] that can be detected by C-terminal tau antibodies. To obtain a measure of extracellular tau species in form of ghost tangles, double IHC combining hC10.2 and AT8 was carried out on 10 mid and 10 late-stage AD cases (Fig. 6B). Densities of neurons and neuronal structures positive for both antibodies or either antibody alone were estimated in a semi-quantitatively manner (Fig. 6C). Whereas a minor fraction (0–2%) of neurons was stained only by AT8, 30–60% of neuronal structures were stained by hC10.2 alone. The proportion of these hC10.2-positive neuronal structures persisted across brain regions and disease stages (Fig. 6C). This observation of hC10.2 binding to ghost tangles suggests that this antibody could effectively engage with extracellular pathological tau species.

Efficient depletion of seeding-competent hyperphosphorylated tau species from diverse primary tauopathy brains with hC10.2

Next, a collection of non-AD cases representing a heterogenous pool of 3R, 4R and 3R/4R tauopathies was studied (Table 3). Frontal cortex samples from each of five cases of PiD and PSP, and each of seven cases of CBD and FTLD were investigated. Two cases of PPA were included: one exhibiting AD-like tau inclusions (PPA1) and the other PSP-like pathology (PPA2). As negative and positive controls, tissues from three HC cases and a pooled AD homogenate prepared from five late-stage AD cases were used. hC10.2 recognized pathological tau across all tauopathies and stained glial as well as neuronal deposits (Fig. 7A). [³H]MK-6240 and [³H]PI-2620 tracer binding to homogenates from all tauopathy cases was also tested. Little or no binding with either tracer could be detected despite aggregation levels comparable to that of control AD material (Fig. 7B, C). Only the PPA case with AD pathology was able to elicit a clear signal. Thus, at the conditions applied, the two tau PET tracers were only engaging with mixed 3R/4R tau aggregates from AD and not with pure 3R or 4R tau aggregates from primary tauopathies.

Fig. 7

The hC10.2 antibody binds tau aggregates in non-AD tauopathies. A) Representative images of brain sections stained with hC10.2 showing neuronal and glial pathological hallmarks of the respective non-AD tauopathies. The PSP image represents substantia nigra, the rest are cortical regions. Scale bar = 50μm. B, C) Tau PET tracer binding displayed positive signals in homogenates from brain material exhibiting AD pathology (PPA1 and AD samples). Data are presented as means±SD n = 3. D) Tauopathy brain homogenates were analyzed for aggregation levels in the Cisbio tau aggregation assay. E) Homogenates from cases with relative tau aggregation signal above 200 units were immuno-depleted with hC10.2 and compared to IgG-depleted samples in the Cisbio assay. Data are means±SEM of n = 7 FTD, n = 6 CBD, n = 3 PiD, n = 2 PPA, n = 3 PSP, n = 1 AD pooled. F) hC10.2-depleted homogenates were analyzed by western blot and detected by DAKO total tau or hC10.2 antibodies and normalized to GAPDH. Homogenates with a positive hC10.2 signal in the input material were included. Data are means±SEM of n = 5 FTD, n = 4 CBD, n = 3 PiD, n = 2 PPA, n = 3 PSP, n = 1 AD pooled, n = 1 HC. One-way ANOVA with Dunnett’s multiple comparisons test versus IgG ^****p < 0.0001. G) Crude homogenates were lipofectamine-transfected into HEK293 cells expressing tau RD P301S* and analyzed for induction of tau seeding by CFP/YFP FRET and plotted as mean±SD of n = 6. An arbitrary limit of seeding activity above which samples were included in immunodepletion analysis is indicated. H) hC10.2-depleted and HJ8.5-depleted homogenates were analyzed for seeding potential and compared to respective IgG-depleted samples. Data was pooled and plotted as mean±SD of PSP n = 3, CBD n = 4 and AD n = 1. One-way ANOVA with Tukey’s multiple comparisons test versus IgG ^****p < 0.0001, ^**p < 0.01, ^*p < 0.05, ^§§§§ < 0,0001. FTD, frontotemporal dementia; CBD, corticobasal degeneration; PPA, primary progressive aphasia; PiD, Pick’s disease; HC, healthy control; AD, Alzheimer’s disease.

Despite variation, it was possible to identify individual cases within each tauopathy that contained measurable levels of aggregated tau using the tau aggregate HTRF assay (Fig. 7D). The considerable individual variation detected between cases was reflected by the tau pathology levels observed in IHC. Homogenates exhibiting robust signals above 200 relative tau aggregate counts were immuno-depleted with hC10.2 and compared to control IgG-depleted homogenates. hC10.2 removed tau aggregates efficiently from all tauopathy subtypes (Fig. 7E). This was confirmed by western blot analysis showing that hC10.2 removed both hyperphosphorylated 3R or 4R tau species, but only partly reduced the signal of total tau (all tau species). This suggests that hC10.2 contributed only minimal binding to normal tau (Fig. 7F). Homogenates prepared from 4R tauopathies (CBD, PSP, PPA, and FTLD) induced seeded tau aggregation in HEK293T biosensor cells. In contrast, 3R-tau homogenates isolated from PiD cases were not able to induce seeded tau aggregation in the 4R-tau expressing cells to any significant degree (Fig. 7G). The 10 homogenates exhibiting the strongest seeding activity were immuno-depleted with hC10.2 and in parallel with HJ8.5. In all cases, this resulted in prevention of seeded tau aggregation in the biosensor cells (Fig. 7H). hC10.2 exhibited the strongest effect, indicating superior engagement of seeding-competent hyperphosphorylated tau species from different primary tauopathies.

DISCUSSION

We have generated a comprehensive data set demonstrating high sensitivity and specificity of hC10.2 engaging with pathological tau species by applying a battery of assays to the same samples of brain tissue from a large number of AD and primary tauopathy cases. This enabled us to establish a correlation of tau PET tracer binding (in AD cases), tau aggregation and S396-tau hyperphosphorylation levels with the in vitro seeding potential within individual cases. In summary, hC10.2 efficiently depleted hyperphosphorylated, aggregated, and seeding-competent tau species from tauopathy samples and was superior in the prevention of seeded tau aggregation in human tau expressing biosensor cells. S396-tau hyperphosphorylation was prominent in the spatial-temporal progression both in mid-stage and late AD, and hC10.2 bound to ghost tangles representing extracellular pathological tau.

Consistent S396-tau hyperphosphorylation in a heterogenous AD population

The present findings underline the diversity of tau pathology load that can be observed across a population of AD patients, even in cases having reached the final stages of the disease. By comparing prefrontal cortex homogenates from 25 Braak stage III/IV and 25 Braak stage V/VI AD brains [39], we detected a high inter-case variability of tau aggregation and hyperphosphorylation that consistently translated to PET tracer binding efficiency and seeding competence. This suggests that tau fibrils from individual AD cases are coherent in tracer sensitivity and transmission of pathology regardless of the level of tau pathology and progression state. We found that the concentration of pS396-tau strongly correlated with tau aggregation and hyperphosphorylation levels, and PET tracer binding. This observation strengthens the position of pS396-tau as a viable immunotherapeutic target for treatment at all stages of AD and encourages the promotion of therapeutic antibodies directed toward this epitope. Seeding activity across the individual AD samples was strongly correlated to tau aggregation and hyperphosphorylation levels. Interestingly, we found a negative correlation of tau aggregation and seeding activity with age of disease onset and age of death in the individual donors. This finding is consistent with recent reports [40, 41] where it was suggested that low abundance of seeding-competent tau species in long-lived AD patients may result from a slow replication of tau conformers, or an increased clearance efficiency. Along this line, a recent study showed that later age of onset was associated with lower tau-PET burden and atrophy rates [42].

[¹⁸F]MK-6240 and [¹⁸F]PI-2620 are known as second generation tau PET tracers, showing high affinity binding to tau tangles and less propensity for off-target binding (including MAO-A, MAO-B, and basal ganglia) compared to first generation tracers (for a review, see [43]). Initial clinical data on both tracers demonstrate high binding to brain regions with plausible tau deposition in AD subjects [44, 45]. In AD homogenates, we observed identical binding patterns of the two PET tracers when represented by their tritiated analogues. MK-6240 could displace [³H]PI-2620 and vice versa, suggesting similar binding sites. Several binding sites for tau tracers on tau tangles from AD material have been hypothesized from computational studies but whether MK-6240 and PI-2620 bind to similar sites has not specifically been discussed [46, 47]. Further, we demonstrated a strong correlation of tracer binding with pS396-tau levels and a lack of binding to hC10.2-depleted AD homogenates (IgG depleted 10% of signal). This suggests a relationship between PET tracer binding and pS396-tau pathology, positioning both tracers as valuable markers for therapeutic intervention studies targeting pS396-tau. Importantly, we observed no indication that the binding of hC10.2 to tau aggregates interfered with subsequent tracer binding in vitro, suggesting both [¹⁸F]MK-6240 and [¹⁸F]PI-2620 as useful PET tracers for AD intervention studies with hC10.2.

The considerable variation in tau pathology levels may reflect heterogeneity between individual cases, accentuated by prefrontal cortex being a tissue where tau pathology accumulates late in the progression of AD. Another contributing factor to variation could be inhomogeneous distribution of tau deposits across a tissue block. For preparation of the homogenates, smaller samples were dissected from a 1g tissue block. Whereas tau aggregation, tau hyperphosphorylation, pS396-tau and cellular seeding assays were performed from the same homogenates, tau tracer binding was performed on separate homogenates prepared from the same tissue block. All readouts showed a strong linear correlation to each other. Also, pathology loads detected by IHC (sections from the opposite hemisphere) correlated well with the measured aggregation and hyperphosphorylation levels. This collectively suggests that the intraregional variation may play a minor role in interpretation of the data.

Detection of extracellular tau species as ghost tangles with the pS396-tau hC10.2 antibody

The concept of spatiotemporal progression of tau pathology through neuronal networks or along major afferents [36, 39] has been supported by the demonstration of interneuronal spread of tau fibrils in animal models [24 , 48–51]). We analyzed pS396-tau in anatomically connected regions resembling early (LC) to late (PFC) events in AD tau pathology and established that immunostaining with the hC10.2 antibody correlated to that of AT8 (targeting mid-domain phospho-tau). This finding indicates that these two phospho-tau antibodies were binding equally well to tau pathology without displaying any spatio-temporal difference. This supports S396-phosphorylation as a general feature of tau fibrils, and a therapeutic potential of hC10.2 in AD.

Extracellular tau species may represent a therapeutically important pool of tau for immunotherapy. It is not possible to separate extracellular from intracellular tau species in brain homogenates. However, one way to display extracellular tau deposits in human brain is through the detection of extra-neuronal ghost tangles, remnants of degenerating or dead neurons [38]. Ghost tangles are considered a source of tau seeds in human brain [52]. To our knowledge, ghost tangles represent the main measurable in vivo extracellular pool of tau with an intact MTBR and are potentially seeding-competent if solubilized (although that is yet to be proven). It has been repeatedly demonstrated that ghost tangles are not stained by N-terminal and mid-domain tau antibodies. On the other hand, some C-terminal epitopes are displayed on ghost tangles and recognized, e.g., by pS396-tau antibodies [53 –56]. This can also be observed with phospho-independent C-terminal antibodies suggesting that ghost tangles are N-terminally truncated [31 , 58]. We employed double immunofluorescence of AT8 and hC10.2 and semi-quantitatively determined the fraction of tau deposits exclusively positive for hC10.2 from those stained by both hC10.2 and AT8. The solely hC10.2-positive constituted 30–60% of the immuno-stained neuronal structures analyzed across the four regions from each of 10 mid and 10 end stage AD cases. Similar results have been reported previously when analyzing AT100 or AT8 and pS396 tau immunostaining in AD hippocampus [55, 56]. Although the results were hampered by the limited number of sections quantified, the detection of extracellular tau in the form of ghost tangles in all four regions was indicative of an ongoing neurodegeneration across AD stages. The transition of tau tangles from the intra- to the extracellular compartment during neuronal death makes pathological tau species available for immunotherapeutic intervention and confirms pS396-tau as a highly relevant target.

Efficiency of target binding by hC10.2

The present work has demonstrated that the pS396 epitope is accessible in aggregated, hyperphosphorylated, and seeding-competent tau species. Although C-terminal truncation can occur [59 –62], the proximity of the S396-epitope to MTB domains was advantageous for inhibition of seeded tau aggregation. Furthermore, hyperphosphorylation of S396 is considered an early event in the AD etiology since the pS396 epitope was present in regions affected at the earliest stages of disease as demonstrated in this study. When analyzing tau aggregates after immunodepletion with different anti-tau antibodies, the N-terminal-dependent antibodies HJ8.5 and MC1 were on par with the C-terminal hC10.2 and the mid-domain AT8 antibody in depleting tau aggregates. To appropriately interpret our results, it should be highlighted that the Cisbio tau aggregation assay is based on detection of an N-terminal tau epitope; and only full-length tau species were analyzed in western blot. Since N-terminally truncated tau species would not be detected, the efficiency of hC10.2 to deplete this category of tau aggregates in comparison to the other antibodies was not addressed. However, using an unbiased approach by assessing cellular seeding activity of AD homogenates in an assay independent of detection antibodies, we observed a differentiation in the efficacy of the antibodies. Although the difference was small, depletion with hC10.2 (80% reduction) was significantly more efficacious on tau seeding activity compared to the N-terminal antibodies HJ8.5 and MC1 (60% reduction) and compared to the mid domain phospho-specific antibody AT8 (70% reduction). By a different approach, Courade et al. [31] demonstrated neutralization of seeding activity when combining antibodies and tau seeds in solution and applying the Cisbio tau aggregation assay for detection of tau aggregates. The authors reported that at the highest concentration tested, antibodies targeting epitopes equivalent to AT8, MC1, and HJ8.5 blocked seeded tau aggregation to ranges similar to the present study (30–40%). Despite detecting pathological tau in IHC, the mid-domain total tau antibody Tau5 performed poorly in depleting tau aggregates for reasons yet unclear. Antibodies directed to the MTB core (DC8E8 and to some extent 7C30) repeatedly showed reduced efficiency to deplete tau aggregates. This indicates that core epitopes may be embedded within the fibril structure of larger aggregates and are less accessible for antibody binding, similar to what has been observed by others [63]. In support of this, the DC8E8 antibody only bound to pathological tau in IHC when the sections were pretreated with formic acid as previously reported for other MTB-targeting antibodies [64].

Considering the limited uptake of antibodies into the brain parenchyma, binding competition of anti-tau antibodies to non-pathological tau species would reduce the availability of the antibody for therapeutic intervention. The findings in the present study that phospho-dependent antibodies (hC10.2 and AT8) did not bind soluble normal tau species may be advantageous. However, it must be considered that the brain homogenates used represent intra- as well as extracellular tau species, likely with the largest contribution from intracellular tau. The extent of monomeric tau outside cells and the extent of post-translational modification of extracellular tau is an area of investigation [65]. Both the phospho-specific antibodies showed little variation in their depletion efficiency across all samples with hyperphosphorylated and aggregated tau. This indicates, that although the amount of tau aggregates varied across patient samples there was little variation between the samples regarding the post-translational modifications investigated in this study.

hC10.2 efficiently engages with 3R and 4R pathological tau

Primary tauopathies represent a group of diseases with a broad spectrum of clinical phenotypes and with a diverse pattern of disease progression (see [13] for review). The mechanism of pathology spreading in primary tauopathies remains to be elucidated. Nonetheless, in vivo seeding in mouse brains has been demonstrated with 4R-tauopathy brain extracts with a recapitulation of the hallmark neuronal and glial lesions [24, 66] supporting the possibility of a spatiotemporal disease propagation.

By in vitro seeding in biosensor cells, we showed that 4R and 3R/4R tau aggregates were able to induce seeded tau aggregation in HEK293T cells expressing tau RD 301S*. In contrast, 3R aggregates from PiD homogenates did not readily template the 4R tau expressed in the biosensor cells, suggesting a conformational barrier and confirming previous reports by others [31 , 68].

We were unable to demonstrate clear binding of [³H]MK-6240 and [³H]PI-2620 to non-AD tauopathy extracts. Only a single case of PPA that presented with AD pathology generated a positive signal underlining the apparent preference for 3R/4R tau aggregates of both tracers. By autoradiography, [¹⁸F]MK-6240 has previously been reported to bind AD tau deposits, but not 3R or 4R tau [69], whereas [¹⁸F]PI-2620 was shown to bind PSP and PiD brain homogenates and PSP sections [15]. Preliminary clinical PET imaging studies with [¹⁸F]PI-2620 show retention in PSP and CBD cases in basal ganglia. In CBD, cortical binding was reported, however, with a lower binding than in AD [13]. In the previous in vitro studies, tau tracer binding was mainly detected in subcortical structures whereas our binding studies were conducted on homogenates prepared from frontal cortex tissue samples. Although the presence of tau aggregates at levels comparable to AD was confirmed in the present study, a diversity of tau fibrils may explain the underlying differences. This is substantiated by cryo-EM studies having shown that tau assemblies can adopt different filament conformations that are distinct between individual 3R/4R, 3R and 4R tauopathies [17 , 22]. This may create different binding pockets and explain the lack of high affinity binding of certain tau tracers to 3R and 4R tau reported here. In further support, a difference in argyrophilic grain preference towards 3R, 4R and 3R/4R has also been demonstrated [70].

In summary, we analyzed hC10.2 binding to pathological tau species in prefrontal cortex samples from a selection of 3R/4R- (AD), 3R- (PiD), and 4R-tauopathies (CBD, PSP, FTLD, and PPA). The demonstration that hC10.2 engaged efficiently with 3R and 4R tau isoforms and removed aggregated, hyperphosphorylated and seeding-competent tau species efficiently positions this antibody as a therapeutic candidate for AD and primary tauopathies.

Footnotes

ACKNOWLEDGMENTS

The authors would like to acknowledge the excellent technical support from Lene Nielsen, Pia Møller Carstensen, Anette Bredal Christensen, Karina Lynge Laursen, Bo Albrechtslund, Anette Bjørn, Kirsten Jørgensen, and Kirsten Assing. The authors would also like to acknowledge the excellent work of the London neurodegenerative disease brain bank and Claire Troakes for providing brain tissue. The London Neurodegenerative Diseases Brain Bank receives funding from the UK Medical Research Council (MR/L016397/1) and as part of the Brains for Dementia Research programme, jointly funded by Alzheimer’s Research UK and the Alzheimer’s Society.

The Arizona Study of Aging and Brain and Body Donation Program has been supported by the National Institute of Neurological Disorders and Stroke (U24 NS072026 National Brain and Tissue Resource for Parkinson’s Disease and Related Disorders), the National Institute on Aging (P30 AG19610 Arizona Alzheimer’s Disease Core Center), the Arizona Department of Health Services (contract 211002, Arizona Alzheimer’s Research Center), the Arizona Biomedical Research Commission (contracts 4001, 0011, 05-901 and 1001 to the Arizona Parkinson’s Disease Consortium) and the Michael J. Fox Foundation for Parkinson’s Research.

Authors’ disclosures available online ().

References

Lee

, Goedert

, Trojanowski

(2001) Neurodegenerative tauopathies. Annu Rev Neurosci 24, 1121–1159.

Leyns

CEG

, Holtzman

(2017) Glial contributions to neurodegeneration in tauopathies. Mol Neurodegener 12, 50.

Spillantini

, Goedert

(2013) Tau pathology and neurodegeneration. Lancet Neurol 12, 609–622.

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.

Goedert

, Spillantini

, Jakes

, Rutherford

, Crowther

(1989) Multiple isoforms of human microtubule-associated protein tau: Sequences and localization in neurofibrillary tangles of Alzheimer’s disease. Neuron 3, 519–526.

Hanger

, Seereeram

, Noble

(2009) Mediators of tau phosphorylation in the pathogenesis of Alzheimer’s disease. Expert Rev Neurother 9, 1647–1666.

Kanemaru

, Takio

, Miura

, Titani

, Ihara

(1992) Fetal-type phosphorylation of the tau in paired helical filaments. J Neurochem 58, 1667–1675.

Kopke

, Tung

, Shaikh

, Alonso

, Iqbal

, Grundke-Iqbal

(1993) Microtubule-associated protein tau. Abnormal phosphorylation of a non-paired helical filament pool in Alzheimer disease. J Biol Chem 268, 24374–24384.

Selkoe

, Hardy

(2016) The amyloid hypothesis of Alzheimer’s disease at 25 years. EMBO Mol Med 8, 595–608.

10.

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

(2012) Correlation of Alzheimer disease neuropathologic changes with cognitive status: A review of the literature. J Neuropathol Exp Neurol 71, 362–381.

11.

Arendt

, Stieler

, Holzer

(2016) Tau and tauopathies. Brain Res Bull 126, 238–292.

12.

Dickson

(2001) Neuropathology of Pick’s disease. Neurology 56, S16–20.

13.

Rosler

, Tayaranian Marvian

, Brendel

, Nykanen

, Hollerhage

, Schwarz

, Hopfner

, Koeglsperger

, Respondek

, Schweyer

, Levin

, Villemagne

, Barthel

, Sabri

, Muller

, Meissner

, Kovacs

, Hoglinger

(2019) Four-repeat tauopathies. Prog Neurobiol 180, 101644.

14.

Leuzy

, Chiotis

, Lemoine

, Gillberg

, Almkvist

, Rodriguez-Vieitez

, Nordberg

(2019) Tau PET imaging in neurodegenerative tauopathies-still a challenge. Mol Psychiatry 24, 1112–1134.

15.

Kroth

, Oden

, Molette

, Schieferstein

, Capotosti

, Mueller

, Berndt

, Schmitt-Willich

, Darmency

, Gabellieri

, Boudou

, Juergens

, Varisco

, Vokali

, Hickman

, Tamagnan

, Pfeifer

, Dinkelborg

, Muhs

, Stephens

(2019) Discovery and preclinical characterization of [(18)F]PI-2620, a next-generation tau PET tracer for the assessment of tau pathology in Alzheimer’s disease and other tauopathies. Eur J Nucl Med Mol Imaging 46, 2178–2189.

16.

Walji

, Hostetler

, Selnick

, Zeng

, Miller

, Bennacef

, Salinas

, Connolly

, Gantert

, Holahan

, O’Malley

, Purcell

, Riffel

, Li

, Balsells

, JA

, Melquist

, Soriano

, Zhang

, Ogawa

, Xu

, Joshi

, Della Rocca

, Hess

, Schachter

, Hesk

, Schenk

, Struyk

, Babaoglu

, Lohith

, Wang

, Yang

, Fu

, Evelhoch

, Coleman

(2016) Discovery of 6-(Fluoro-(18)F)-3-(1H-pyrrolo[2,3-c]pyridin-1-yl)isoquinolin-5-amine ([(18)F]-MK-6240): A positron emission tomography (PET) imaging agent for quantification of neurofibrillary tangles (NFTs). J Med Chem 59, 4778–4789.

17.

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.

18.

Falcon

, Zhang

, Murzin

, Murshudov

, Garringer

, Vidal

, Crowther

, Ghetti

, Scheres

SHW

, Goedert

(2018) Structures of filaments from Pick’s disease reveal a novel tau protein fold. Nature 561, 137–140.

19.

Falcon

, Zhang

, Schweighauser

, Murzin

, Vidal

, Garringer

, Ghetti

, Scheres

SHW

, Goedert

(2018) Tau filaments from multiple cases of sporadic and inherited Alzheimer’s disease adopt a common fold. Acta Neuropathol 136, 699–708.

20.

Zhang

, Falcon

, Murzin

, Fan

, Crowther

, Goedert

, Scheres

(2019) Heparin-induced tau filaments are polymorphic and differ from those in Alzheimer’s and Pick’s diseases. Elife 8, e43584.

21.

Zhang

, Tarutani

, Newell

, Murzin

, Matsubara

, Falcon

, Vidal

, Garringer

, Shi

, Ikeuchi

, Murayama

, Ghetti

, Hasegawa

, Goedert

, Scheres

SHW

(2020) Novel tau filament fold in corticobasal degeneration. Nature 580, 283–287.

22.

Arakhamia

, Lee

, Carlomagno

, Duong

, Kundinger

, Wang

, Williams

, DeTure

, Dickson

, Cook

, Seyfried

, Petrucelli

, Fitzpatrick

AWP

(2020) Posttranslational modifications mediate the structural diversity of tauopathy strains. Cell 180, 633–644e612.

23.

Scheres

, Zhang

, Falcon

, Goedert

(2020) Cryo-EM structures of tau filaments. Curr Opin Struct Biol 64, 17–25.

24.

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.

25.

Rosenqvist

, Asuni

, Andersson

, Christensen

, Daechsel

, Egebjerg

, Falsig

, Helboe

, Jul

, Kartberg

, Pedersen

, Sigurdsson

, Sotty

, Skjodt

, Stavenhagen

, Volbracht

, Pedersen

(2018) Highly specific and selective anti-pS396-tau antibody C10.2 targets seeding-competent tau. Alzheimers Dement (N Y) 4, 521–534.

26.

Sandusky-Beltran

, Sigurdsson

(2020) Tau immunotherapies: Lessons learned, current status and future considerations. Neuropharmacology 175, 108104.

27.

Jucker

, Walker

(2013) Self-propagation of pathogenic protein aggregates in neurodegenerative diseases. Nature 501, 45–51.

28.

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.

29.

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, E4376–4385.

30.

Jackson

, Kerridge

, Cooper

, Cavallini

, Falcon

, Cella

, 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.

31.

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.

32.

Beach

, Adler

, Sue

, Serrano

, Shill

, Walker

, Lue

, Roher

, Dugger

, Maarouf

, Birdsill

, Intorcia

, Saxon-Labelle

, Pullen

, Scroggins

, Filon

, Scott

, Hoffman

, Garcia

, Caviness

, Hentz

, Driver-Dunckley

, Jacobson

, Davis

, Belden

, Long

, Malek-Ahmadi

, Powell

, Gale

, Nicholson

, Caselli

, Woodruff

, Rapscak

, Ahern

, Shi

, Burke

, Reiman

, Sabbagh

(2015) Arizona Study of Aging and Neurodegenerative Disorders and Brain and Body Donation Program. Neuropathology 35, 354–389.

33.

Pace

, Vajdos

, Fee

, Grimsley

, Gray

(1995) How to measure and predict the molar absorption coefficient of a protein. Protein Sci 4, 2411–2423.

34.

Burton

, Pyati

, Koduri

, Sharp

, Thornton

, Parren

, Sawyer

, Hendry

, Dunlop

, Nara

, et al. (1994) Efficient neutralization of primary isolates of HIV-1 by a recombinant human monoclonal antibody. Science 266, 1024–1027.

35.

Trinkle-Mulcahy

, Sleeman

, Lamond

(2001) Dynamic targeting of protein phosphatase 1 within the nuclei of living mammalian cells. J Cell Sci 114, 4219–4228.

36.

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.

37.

Braak

, Alafuzoff

, Arzberger

, Kretzschmar

, Del Tredici

(2006) Staging of Alzheimer disease-associated neurofibrillary pathology using paraffin sections and immunocytochemistry. Acta Neuropathol 112, 389–404.

38.

Braak

, Braak

, Mandelkow

(1994) A sequence of cytoskeleton changes related to the formation of neurofibrillary tangles and neuropil threads. Acta Neuropathol 87, 554–567.

39.

Braak

, Braak

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

40.

Aoyagi

, Condello

, Stohr

, Yue

, Rivera

, Lee

, Woerman

, Halliday

, van Duinen

, Ingelsson

, Lannfelt

, Graff

, Bird

, Keene

, Seeley

, DeGrado

, Prusiner

(2019) Abeta and tau prion-like activities decline with longevity in the Alzheimer’s disease human brain. Sci Transl Med 11, eaat8462.

41.

Dujardin

, Commins

, Lathuiliere

, Beerepoot

, Fernandes

, Kamath

, De Los Santos

, Klickstein

, Corjuc

, Dooley

, Viode

, Oakley

, Moore

, Mullin

, Jean-Gilles

, Clark

, Atchison

, Moore

, Chibnik

, Tanzi

, Frosch

, Serrano-Pozo

, Elwood

, Steen

, Kennedy

, Hyman

(2020) Tau molecular diversity contributes to clinical heterogeneity in Alzheimer’s disease. Nat Med 26, 1256–1263.

42.

La Joie

, Visani

, Baker

, Brown

, Bourakova

, Cha

, Chaudhary

, Edwards

, Iaccarino

, Janabi

, Lesman-Segev

, Miller

, Perry

, O’Neil

, Pham

, Rojas

, Rosen

, Seeley

, Tsai

, Miller

, Jagust

, Rabinovici

(2020) Prospective longitudinal atrophy in Alzheimer’s disease correlates with the intensity and topography of baseline tau-PET. Sci Transl Med 12, eaau5732.

43.

Okamura

, Harada

, Ishiki

, Kikuchi

, Nakamura

, Kudo

(2018) The development and validation of tau PET tracers: Current status and future directions. Clin Transl Imaging 6, 305–316.

44.

Lohith

, Bennacef

, Vandenberghe

, Vandenbulcke

, Salinas

, Declercq

, Reynders

, Telan-Choing

, Riffel

, Celen

, Serdons

, Bormans

, Tsai

, Walji

, Hostetler

, Evelhoch

, Van Laere

, Forman

, Stoch

, Sur

, Struyk

(2019) Brain imaging of Alzheimer dementia patients and elderly controls with (18)F-MK-6240, a PET tracer targeting neurofibrillary tangles. J Nucl Med 60, 107–114.

45.

Mueller

, Bullich

, Barret

, Madonia

, Berndt

, Papin

, Perrotin

, Koglin

, Kroth

, Pfeifer

, Tamagnan

, Seibyl

, Marek

, De Santi

, Dinkelborg

, Stephens

(2020) Tau PET imaging with (18)F-PI-2620 in patients with Alzheimer disease and healthy controls: A first-in-humans study. J Nucl Med 61, 911–919.

46.

Kuang

, Murugan

, Zhou

, Nordberg

, Agren

(2020) Computational insight into the binding profile of the second-generation PET tracer PI2620 with tau fibrils. ACS Chem Neurosci 11, 900–908.

47.

Murugan

, Chiotis

, Rodriguez-Vieitez

, Lemoine

, Agren

, Nordberg

(2019) Cross-interaction of tau PET tracers with monoamine oxidase B: Evidence from in silico modelling and in vivo imaging. Eur J Nucl Med Mol Imaging 46, 1369–1382.

48.

Dujardin

, Lecolle

, Caillierez

, Begard

, Zommer

, Lachaud

, Carrier

, Dufour

, Auregan

, Winderickx

, Hantraye

, Deglon

, Colin

, Buee

(2014) Neuron-to-neuron wild-type Tau protein transfer through a trans-synaptic mechanism: Relevance to sporadic tauopathies. Acta Neuropathol Commun 2, 14.

49.

de Calignon

, Polydoro

, Suarez-Calvet

, William

, Adamowicz

, Kopeikina

, Pitstick

, Sahara

, Ashe

, Carlson

, Spires-Jones

, Hyman

(2012) Propagation of tau pathology in a model of early Alzheimer’s disease. Neuron 73, 685–697.

50.

Liu

, Drouet

, Wu

, Witter

, Small

, Clelland

, Duff

(2012) Trans-synaptic spread of tau pathology in vivo. PLoS One 7, e31302.

51.

DeVos

, Corjuc

, Oakley

, Nobuhara

, Bannon

, Chase

, Commins

, Gonzalez

, Dooley

, Frosch

, Hyman

(2018) Synaptic tau seeding precedes tau pathology in human Alzheimer’s disease brain. Front Neurosci 12, 267.

52.

Lewis

, Dickson

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

53.

Garcia-Sierra

, Ghoshal

, Quinn

, Berry

, Binder

(2003) Conformational changes and truncation of tau protein during tangle evolution in Alzheimer’s disease. J Alzheimers Dis 5, 65–77.

54.

Luna-Munoz

, Peralta-Ramirez

, Chavez-Macias

, Harrington

, Wischik

, Mena

(2008) Thiazin red as a neuropathological tool for the rapid diagnosis of Alzheimer’s disease in tissue imprints. Acta Neuropathol 116, 507–515.

55.

Regalado-Reyes

, Furcila

, Hernandez

, Avila

, DeFelipe

, Leon-Espinosa

(2019) Phospho-tau changes in the human CA1 during Alzheimer’s disease progression. J Alzheimers Dis 69, 277–288.

56.

Furcila

, DeFelipe

, Alonso-Nanclares

(2018) A study of amyloid-beta and phosphotau in plaques and neurons in the hippocampus of Alzheimer’s disease patients. J Alzheimers Dis 64, 417–435.

57.

Guillozet-Bongaarts

, Garcia-Sierra

, Reynolds

, Horowitz

, Fu

, Wang

, Cahill

, Bigio

, Berry

, Binder

(2005) Tau truncation during neurofibrillary tangle evolution in Alzheimer’s disease. Neurobiol Aging 26, 1015–1022.

58.

Sato

, Barthelemy

, 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 98, 861–864.

59.

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.

60.

Garcia-Sierra

, Wischik

, Harrington

, Luna-Munoz

, Mena

(2001) Accumulation of C-terminally truncated tau protein associated with vulnerability of the perforant pathway in early stages of neurofibrillary pathology in Alzheimer’s disease. J Chem Neuroanat 22, 65–77.

61.

Flores-Rodriguez

, Ontiveros-Torres

, Cardenas-Aguayo

, Luna-Arias

, Meraz-Rios

, Viramontes-Pintos

, Harrington

, Wischik

, Mena

, Floran-Garduno

, Luna-Munoz

(2015) The relationship between truncation and phosphorylation at the C-terminus of tau protein in the paired helical filaments of Alzheimer’s disease. Front Neurosci 9, 33.

62.

Taniguchi-Watanabe

, Arai

, Kametani

, Nonaka

, Masuda-Suzukake

, Tarutani

, Murayama

, Saito

, Arima

, Yoshida

, Akiyama

, Robinson

, Mann

DMA

, Iwatsubo

, Hasegawa

(2016) Biochemical classification of tauopathies by immunoblot, protein sequence and mass spectrometric analyses of sarkosyl-insoluble and trypsin-resistant tau. Acta Neuropathol 131, 267–280.

63.

Roberts

, Sevastou

, Imaizumi

, Mistry

, Talma

, Dey

, Gartlon

, Ochiai

, Zhou

, Akasofu

, Tokuhara

, Ogo

, Aoyama

, Aoyagi

, Strand

, Sajedi

, Agarwala

, Spidel

, Albone

, Horie

, Staddon

, de Silva

(2020) Pre-clinical characterisation of E2814, a high-affinity antibody targeting the microtubule-binding repeat domain of tau for passive immunotherapy in Alzheimer’s disease. Acta Neuropathol Commun 8, 13.

64.

Dickson

, Ksiezak-Reding

, Liu

, Davies

, Crowe

, Yen

(1992) Immunocytochemistry of neurofibrillary tangles with antibodies to subregions of tau protein: Identification of hidden and cleaved tau eitopes and a new phosphorylation site. Acta Neuropathol 84, 596–605.

65.

Barthelemy

, Mallipeddi

, Moiseyev

, Sato

, Bateman

(2019) Tau phosphorylation rates measured by mass spectrometry differ in the intracellular brain vs. extracellular cerebrospinal fluid compartments and are differentially affected by Alzheimer’s disease. Front Aging Neurosci 11, 121.

66.

Narasimhan

, Guo

, Changolkar

, Stieber

, McBride

, Silva

, He

, Zhang

, Gathagan

, Trojanowski

, Lee

VMY

(2017) Pathological tau strains from human brains recapitulate the diversity of tauopathies in nontransgenic mouse brain. J Neurosci 37, 11406–11423.

67.

Woerman

, Aoyagi

, Patel

, Kazmi

, Lobach

, Grinberg

, McKee

, Seeley

, Olson

, Prusiner

(2016) Tau prions from Alzheimer’s disease and chronic traumatic encephalopathy patients propagate in cultured cells. Proc Natl Acad Sci U S A 113, E8187–E8196.

68.

Strang

, Croft

, Sorrentino

, Chakrabarty

, Golde

, Giasson

(2018) Distinct differences in prion-like seeding and aggregation between Tau protein variants provide mechanistic insights into tauopathies. J Biol Chem 293, 4579.

69.

Aguero

, Dhaynaut

, Normandin

, Amaral

, Guehl

, Neelamegam

, Marquie

, Johnson

, El Fakhri

, Frosch

, Gomez-Isla

(2019) Autoradiography validation of novel tau PET tracer [F-18]-MK-6240 on human postmortem brain tissue. Acta Neuropathol Commun 7, 37.

70.

Uchihara

, Tsuchiya

, Nakamura

, Akiyama

(2005) Argyrophilic grains are not always argyrophilic–distinction from neurofibrillary tangles of diffuse neurofibrillary tangles with calcification revealed by comparison between Gallyas and Campbell-Switzer methods. Acta Neuropathol 110, 158–164.

71.

Kfoury

, Holmes

, Jiang

, Holtzman

, Diamond

(2012) Trans-cellular propagation of Tau aggregation by fibrillar species. J Biol Chem 287, 19440–19451.

72.

Jicha

, Bowser

, Kazam

, Davies

(1997) Alz-50 and MC-1, a new monoclonal antibody raised to paired helical filaments, recognize conformational epitopes on recombinant tau. J Neurosci Res 48, 128–132.

73.

Kontsekova

, Zilka

, Kovacech

, Skrabana

, Novak

(2014) Identification of structural determinants on tau protein essential for its pathological function: Novel therapeutic target for tau immunotherapy in Alzheimer’s disease. Alzheimers Res Ther 6, 45.