Brain tissue samples and compliance with ethical standards

Anonymized post-mortem brain samples were obtained from the Netherlands Brain Bank (NBB), Netherlands Institute for Neuroscience, Amsterdam (open access: www.brainbank.nl), as well as from commercial sources (Additional file 1: Tables S2, S3, S4), including Tissue Solutions Ltd (Glasgow, UK), Analytical Biological Services Inc. (Delaware, USA), Banner sun health research institute (Arizona, USA), Discovery life sciences (Alabama, USA), Folio Biosciences (by now merged to discovery life sciences). All materials were collected from donors for whom a written informed consent for a brain autopsy and the use of the material and clinical information for research purposes had been obtained.

Sarkosyl-insoluble tau preparation

Tissue extraction was carried out on ice or at 4 °C, unless stated otherwise. Frozen human brain tissue samples were thawed on ice and homogenized in lysate buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 20 mM NaF, 1 mM Na3VO4, 0.5 mM MgSO4, cOmplete protease inhibitor with EDTA (Roche, #11,697,498,001) and PhosSTOP phosphatase inhibitor (Roche, #04906845001) at 5 μl/mg tissue using a cooled bead disruptor (FastPrep-24 5G System, MP Biomedicals). Resulting brain homogenates were frozen in liquid nitrogen and stored at − 80 °C. Sarkosyl-insoluble tau was extracted according to the protocol described in a previous publication [36]. In brief, 1.2 ml of brain homogenate was centrifuged at 27,000 × g for 20 min (min), after which the pellet was resuspended in 1.2 ml of 10 mM Tris/HCl pH 7.4, 1 mM EGTA, 0.8 M NaCl, 10% sucrose, cOmplete protease inhibitor and PhosSTOP phosphatase inhibitor using an Omni Bead Ruptor (Omni International, Inc) at room temperature (RT) and then spun at 27, 000 × g for 20 min at 4 °C. The supernatant was transferred to a new tube, sarkosyl added to a final concentration of 1%, and incubated for 1.5 h at RT with 750 rpm shaking. Subsequently, the sarkosyl solution was centrifuged at 150,000 × g for 45 min, the resulting pellet was washed once 200 μl TBS (25 mM Tris pH 7.4, 3 mM KCl, 140 mM NaCl) and resuspended in 120 μl TBS by vortexing and four sonication pulses (2 s each) with a probe sonicator (Sonoplus, Bandelin) at 35% amplitude. Finally, sarkosyl-insoluble tau extracts were flash-frozen in liquid nitrogen and stored at − 80 °C until further use.

Generation of tau pre-formed fibrils

2N4R tau with the P301L mutation in PBS (Gibco, #14,190) with 1 mM DTT (Serva, #20,710.3) was incubated for 15 min at 55 °C and subsequently cooled for 10 min on ice. Heparin (Sigma Aldrich, #H3393) was then added to a final concentration of 40 µM and fibrilization allowed for 7 days while shaking with 800 rpm at 37 °C. Tau fibrils where then collected by centrifugation with 20,817 × g for 2.5 h at 4 °C and resuspension of the pellet in 125 µl of PBS. Homogenization was achieved by sonification in a ChIP sonic digital sonifier (Branson, custom build, Germany) set to pulse on/off at 10 s / 10 s at an intensity of 70% for 30 min at 4 °C. Protein concentration was finally determined by bicinchoninic acid assay (BCA) (Pierce, #23,225) analysis. The tau fibrils were snap-frozen in liquid nitrogen and stored at − 80 °C until further use.

Tau seed amplification assay (SAA)

Full-length 0N3R tau (Fig. 2a) cysteine free (C322S) (SeNostic Health GmbH, Germany) in water, stored at a concentration of 1.0 mg/ml at − 80 °C, was thawed at RT and centrifuged with 21,000 × g at 4 °C for 30 min to remove potential preformed aggregates. The supernatant was adjusted to 0.1 mg/ml 0N3R tau in 40 mM HEPES pH 7.4, 400 mM NaF, 40 μM heparin (Sigma Aldrich, #H3393) and 10 μM Thioflavin T (ThT) (Sigma #T3516-25G), as reported by Metrick et al. [32]. After gentle mixing with the described SAA reaction buffer, 47 μl of the mixture was added to each well of a 384-well plate with an optically clear bottom (Greiner Bio-one, #781,906). Crude brain homogenates were thawed from − 80 °C storage at 16.7% (w/v) in PBS pH 7.4 supplemented with 1 × phosphatase inhibitors (Roche, #04906837001) and 1× protease inhibitors (Roche, # 04693124001) and serially diluted tenfold using a buffer containing 1 × N-2 Supplement (Gibco, #17,502–048), 10 mM HEPES and 0.53% crude brain homogenate from 8-month-old tau knockout mice (KO; B6.Cg-Mapt < tm1(EGFP) Klt > Tg(MAPT)8cPdav/J WT, Jackson Laboratories). The diluted mixture was then homogenized in TBS pH 7.4 with 1 × phosphatase inhibitors (Roche, #04906837001) and 1 × protease inhibitors (Roche, # 04693124001). Three µl of human brain homogenate dilutions were seeded into triplicates or quadruplicates in a reaction volume of 50 μl per well. We analyzed the seeding capacity in crude brain homogenates to preserve the integrity and activity of tau seeds and to avoid potential biochemical alterations, e.g. by extraction with detergents. Plates were sealed with fluorescence-permeable, clear adhesive sealing tape (Applied Biosystems, #4,311,971) and incubated in an Omega FLUOStar plate reader (bmg) set to 42 °C and with rounds of 1-min double orbital shaking at 500 rpm and 1 min resting, while fluorescence reads (450 excitation, 480 emission) were taken every 15 min. Please note that we did not observe evaporation due to leakage of the sealing tape during our assay runtime. Seed amplification kinetics were quantified using the time-to-threshold value (TTT) i.e., the time when the aggregation curve reached a ThT fluorescence threshold of 1500 relative fluorescence units (rfu). This threshold was empirically chosen aiming at sensitive detection of tau seeding activity while differentiating AD from non-AD cases. Lower TTT values indicate faster seeding kinetics due to higher levels of tau seeding activity in the biospecimen. To reflect the inverse relationship of TTT and seeding kinetics, we plotted the TTT values on an inverted axis throughout this study.

Cellular tau seeding assay

HEK-293 T cells that stably expressed the tau microtubule-binding repeat domains (aa 244–342, numbering referred to 2N4R tau (Fig. 2a)) with the p301S mutation fused to yellow fluorescence protein (YFP) at their c-terminus were used as cellular tau seeding assay [2]. Stable cell lines, cultured in D-MEM (Gibco #31,966–047), 10% FBS and 100 µg/ml gentamicin at 8% CO2, were plated at a density of 2,000 cells/well in a black, poly-D-lysine coated 384 well plate (Corning # 354,663). At ~ 60% confluency, 24 h after plating, cells were transduced in quadruplicates with fibrillar tau seeds or human brain tissue that had been diluted in PBS with 0.1% Pluronic F68. These transduction complexes were made by combining Lipofectamine 2000 (Invitrogen #11,668–019) in Opti-MEM medium (Gibco #51,985–026) with the respective analyte dilution in a ratio of 1:1 for 10 min at RT. Subsequently, 20 µl liposome preparations were added directly to 20 µl cell culture medium per well and incubation allowed for 48 h. Next, cells were fixed and the nucleus stained by addition of 4% PFA (VWR #20,909.290), 4% sucrose (Sigma #S9378) and 2.5 µg/ml Hoechst-33342 (Thermo, #H3570) solved in DPBS in a final volume of 80 µl per well for 30 min at RT. Finally, fixed cells were washed three times with DPBS and stored at 4 °C until analysis in an array scan (Illumina) with an 20 × objective. Analysis was set up for 25 fields per well, exposure of 0.08 s and 25% target for Hoechst-33342 and 0.2 s with 40% target for the YFP measurement. Finally, tau-YFP spots were counted and the % of cells with spots determined by comparison to the number of Hoechst-33342-stained nuclei.

Protein purification of hyperphosphorylated tau in Sf9 insect cells

Hyperphosphorytlated tau was generated and purified as described previously [37]. In brief, Sf9 cells were cultured in Sf900™ III SFM medium (Invitrogen, #10,658-027) with PenStrep (Invitrogen, #15,140-122) at 27.5 °C with a medium change every 4–5 days. Cells were diluted at 99% cell viability to 3–5 × 105 cells/ml one day before transfection. On the next day, 1 µg bacmid DNA encoding for human 2N3R tau (pFastBac-huTAU (2N3R)) in 100 µl of Graces Insect Medium (Thermo, #11,595–030) were mixed with 2 µl Cellfectin II Reagent (Thermo, #10,362-010) in 100 µl Graces Insect Medium and incubated at RT for 30 min. An additional 800 µl of insect medium were added to the bacmid-Cellfectin mixture before mixing it with the Sf9 cells that were harvested via aspiration after a wash step with Graces Insect Medium. Transfection was allowed for 5 h at 27.5 °C. Cells were then incubated for 5 days in a 6-well plate with 2 ml of Sf900™ III SFM medium (Thermo, #10,658-027), PenStrep (Thermo, #15,140-122) and 10% of FCS at 27.5 °C. Successful expression of tau in the transfected Sf9 cells was verified by Western blotting, using tau5 (Thermo, #AHB0042) as primary antibody, anti-mouse-alkaline phosphatase conjugate (Promega, #S3728) as secondary antibody and Western Blue (Promega, #S3841) as alkaline phosphatase substrate. To purify the phosphorylated 2N3R tau, cell pellets were resuspended in lysis buffer (50 mM Tris–HCl [pH7.4], 500 mM NaCl, 10% glycerol, 0.1% Nonidet-P40, 10 mM EGTA, 20 mM NaF, 5 mM dithiothreitol, 1 mM orthovanadate) and lysed by sonication. The lysates were then centrifuged to eliminate cell debris, resuspended in lysis buffer and boiled for 10 min. This treatment ensures that nearly all proteins are denatured and precipitated, except for tau, which stays soluble. Precipitates were removed by centrifugation, the resulting supernatant was concentrated, loaded onto a Superdex 200 (GE Healthcare, #17-1071-01) and tau proteins were eluted in PBS, pH7.4, 1 mM DTT. Tau containing fractions were determined by SDS-PAGE and combined for dialysis against 100 mM MES pH6.8, 2 mM DTT, 1 mM NaEGTA, 1 mM MgSO4. Dialyzed samples were then subjected to ion exchange chromatography using a Q Sepharose column (GE Healthcare, #17-0510-01), where tau protein eluted with 0.2 M NaCl To facilitate buffer exchange into final formulation of PBS pH 7.4, 1 mM DTT, and to further purify the protein, an additional Superdex 200 chromatography step was performed. Finally, successful purification of ptau was ensured by i) immunblotting against total tau utilizing the tau5 antibody (Thermo Fisher, #AHB0042), as well as ptau utilizing the anti-ptau Thr-231 antibody (Thermo, # MN1040); ii) analytical size exclusion chromatography with a Superdex 200 and iii) mass spectrometry. More than 56% of purified tau displayed six to 11 phosphate groups while 44% displayed one to five phosphate groups, indicating a successful generation of hyperphosphorylated tau.

ELISA

Total tau, aggregated tau, and ptau212/214 in human brain homogenates were quantified by ELISA, using tau-12 (Biolegend, #806,502), HT7 (Thermo, # MN1000), and AT100 (Thermo, #MN1060) as capture antibodies, respectively. Afterwards, biotin-HT7 (Thermo, # MN1000B) was utilized as detection antibody for each ELISA experiment. Tau 2N4R (Sigma T0576) was used as calibrator for the total tau ELISA. Pre-formed fibrils of 2N4R tau (P301L) were used as calibrator for the aggregated tau ELISA and hyperphosphorylated 2N3R tau for the ptau212/214 ELISA, each as described above. For plate coating with the respective capture antibody, 100 μl/well of 2 μg/ml antibody in PBS with 20% glycerol was adsorbed to Maxisorp NUNC-Immuno Plates (Thermo, #442,404) at 4 °C overnight. All following steps were performed at RT. The wells were washed three times with 250 µl PBST and blocked with 2% BSA (Serva #11,926) in PBST and 20% glycerol for 1.5 h. As a next step, all wells were washed three times with 250 µl PBST. Subsequently, calibrators (Total tau: 15.6–1000 pg/ml; aggregated tau: 78.13–5000 pg/ml; ptau212/214: 1562–100,000 pg/ml) and samples were diluted in PBST with 0.1% BSA and applied in duplicate. Plates were incubated for 2 h. After three additional washes of each well with 250 µl PBST, 100 µl of detection antibody biotin-HT7 at a concentration of 0.2 μg/ml was added and incubated for 1 h and subsequently washed again three times with 250 µl PBST. Next, 0.05 μg/ml Streptavidin Poly-HRP (Thermo, #21,140) in 0.1% BSA in PBST was added for 1 h, followed by another three washes with 250 µl PBST. Finally, 100 μl/well of TMB substrate (KEM-EN-TEC Diagnostics, #4800A) was added and the reaction stopped after 10 min by adding 100 μl 0.18 M sulphuric acid. The absorbance at 450 nm was then immediately measured in a Ledetect 96 plate reader (Anthos, Mikrosysteme GmbH). The lower limit of detection (LOD) was determined by the mean blank value plus two-fold standard deviation. A factor to be considered was that the phosphorylated tau (ptau) calibrator was not directly targeted to be phosphorylated at the AT100 sides, amino acids 212 and 214 (see above). Hence, only a portion of the total amount of ptau calibrator used to generate the standard curve was phosphorylated at the ptau 212/214 site. Consequently, ptau concentrations in the brain analytes were overestimated, when compared to e.g. total tau or aggregated tau ELISA concentrations.

Singulex Erenna

Ptau181 capture mAb (Abcam, #ab236458) were conjugated to streptavidin Dynabeads following the Erenna Capture Kit manufacturer’s instructions (catalog #03–0077-02) to a final concentration of 0.25 mg/ml in bead buffer, provided in the kit. ptau standard (see section ELISA) (0.8–1000 pg/ml), blank and Brain samples diluted 1:25,000 or 1:50,000 in assay buffer (Merck #02–0474-00) were added to a 96-well plate (Axygen, #P-96-450V-C) in a final volume of 100 µl per well. Each analyte was then mixed with 100 µl of 0.05 mg/ml conjugated bead suspension in assay buffer (Merck #02–0474-00) and incubated for 2 h at RT on a shaking device (Boekel). After magnetic isolation, beads were washed with 1 × Erenna wash buffer (Merck, #02–0111-03) with a HydroFlex plate washer (Tecan Group AG). Subsequently, 20 µl of Alexa-647 fluorescently labeled tau12 detection antibody (500 ng/ml) (Biolegend, #806,502) were added to each well and the plate then placed on a Jitterbug shaker at RT for 1 h. After extensive washing, beads were eluted by incubation in 9.5 µl elution buffer (Merck 02–0297-00) for 20 min at RT on a Jitterbug shaker, transferred to a transparent 384-well plate (Thermo Fisher, #264,573) and the solution neutralized by addition of 10 µl/well neutralization buffer D (Merck #02-0368-00). Finally, the 384-well plate was heat-sealed with a variable temperature sealer (Thermo Fisher) and analyzed with the Singulex Erenna immunoassay system (Millipore). The LOD was defined as the lowest back interpolated standard and a coefficient of variance (CV) 20%.

Statistical analysis

GraphPad Prism was used to perform statistical analyses. Differences among two groups were analyzed by Mann–Whitney test, while multiple comparison were performed by non-parametric one-way ANOVA Kruskal–Wallis test. A p-value of < 0.05 was considered statistically significant.