Introduction

Tau is a phosphoprotein expressed in the central nervous system (CNS) and found to modulate the dynamics of microtubule polymerization1,2,3,4,5,6,7. Through alternative splicing of exons 2, 3 and 10, six major tau isoforms are produced in the adult human CNS which result in tau proteins with 0,1, or 2 N-terminal inserts and 3 or 4 repeats in the microtubule binding domain8,9,10. Under physiological conditions, tau is highly soluble with minimal secondary structure observed throughout the molecule3,11,12,13. Under pathological conditions still not entirely understood, tau misfolds and polymerizes into insoluble filaments composed of hyperphosphorylated tau in a group of neurodegenerative diseases known as tauopathies14.

Different tauopathies are hallmarked by the aberrant accumulation of aggregated tau protein which affects stereotypical brain regions thought to contribute to the specific clinical symptomology of disorders such as Alzheimer’s disease (AD), corticobasal degeneration (CBD), progressive supranuclear palsy (PSP), and Pick’s disease (PiD)14,15,16,17,18,19,20. Neuropathological features of these diseases can differ significantly based on the cell types vulnerable to tau accumulation, morphology, tau isoform selectivity, and neuroanatomical regions affected20,21. For example, in AD, neurons are prone to the accumulation of tau filaments composed of all 6 isoforms in the form of neurofibrillary tangles, dystrophic neurites, and neuropil threads in limbic and neocortical regions9,22,23,24,25,26,27. The neuropathology of AD is profoundly different from that of both PSP and CBD. Glial cells, such as astrocytes and oligodendrocytes, are typically spared in AD, while in PSP and CBD, 4-repeat tau isoforms preferentially accumulate in cortical and subcortical regions in the form of tufted astrocytes and astrocytic plaques, respectively20,28,29,30,31,32. In PiD, neurons are heavily affected by the aggregation of 3-repeat tau isoforms in circumscribed inclusions known as Pick bodies in the cortex and hippocampus19,20.

Tau can be altered by several covalent post-translational modifications (PTMs) such as acetylation, ubiquitination, and phosphorylation33. Phosphorylation of tau occurs at physiological levels, but during disease, tau is hyperphosphorylated which is observed biochemically by reduced electrophoretic mobility and immunochemical labelling with phosphorylation dependent tau antibodies which readily detect neuropathological lesions unique to each tauopathy23,24,34,35,36,37. Additionally, phosphorylation has also been demonstrated to negatively affect the ability of tau to bind and promote the polymerization of microtubules38,39,40,41,42. Together with the neuropathological and functional significance of this PTM, specific phosphorylation sites such as pThr181 and more recently pThr205, pThr217, and pThr231 are also found to be elevated during the progression of AD43,44,45,46,47,48,49,50. Phosphorylation at these sites also display strong correlations with other fluid and imaging biomarkers reflecting authentic AD pathology, further indicating significant clinical value47,51,52,53. It has also been demonstrated that pThr217 performs on par with or better than other gold standard fluid and imaging biomarkers in terms of distinguishing AD from other neurodegenerative diseases, correlation with abnormal PET status, or post-mortem confirmed neuropathology54,55,56,57,58,59. Indeed, phosphorylation of tau at specific epitopes, especially pThr217, may be potent instigators of tau aggregation and a viable therapeutic target. Potentially targeting pThr217 immunologically may lower the underlying pathological burden and improve or prevent the onset of clinical symptoms as seen in patients with AD. While the presence and utility of pThr217 in biological fluids from clinical and post-mortem confirmed cohorts has been established, the neuropathological profile of this site is yet to be fully determined.

In the study described here, we generated and characterized a novel mouse monoclonal antibody immunoreactive for tau pThr217. We report the immunolabeling of tau neuropathology in a mouse model of tauopathy and in post-mortem human brain tissue from individuals diagnosed with AD, CBD, PSP, and PiD. Our data indicates that this novel antibody will be valuable in deciphering the role pThr217 has in mediating tau pathogenicity in models of tau aggregation, with potential applications as a tool in biomarker and immunotherapeutic research.

Results

Characterization of novel phosphorylation-dependent monoclonal Tau antibody 1F10

We generated a novel mouse monoclonal antibody using a synthetic tau peptide consisting of residues 208–225 with phosphorylation sites at Thr212, Ser214, and Thr217 hereafter referred to as clone 1F10. We used recombinant tau protein, encoding the longest human brain tau isoform (2N/4R) and synthetic tau peptides with single, double, and no phosphorylation sites to map the epitope by enzyme-linked immunosorbent assay (ELISA)(Table 1). 1F10 reacted with the immunization peptide, 3P, but was not reactive for the non-phosphorylated peptide (noP) or 2N/4R tau indicating that 1F10 is phosphorylation-dependent (Fig. 1a). Of the tau peptides with single phosphorylation sites (pThr212, pSer214, or pThr217), 1F10 showed the strongest immunoreactivity for pThr217 (Fig. 1a). 1F10 also reacted with the dual phosphorylated peptides including pThr212, pSer214 dual phosphorylated peptide despite 1F10’s minimal immunoreactivity for the single pThr212 and pSer214 peptides (Fig. 1a). This data demonstrated that 1F10 is selective for phosphorylation at Thr217, but it also displayed reactivity if dual phosphorylation sites are proximal. Additional pThr217 commercial antibodies were investigated in this study. Rabbit monoclonal antibody E9Y4S did not react with recombinant 2N/4R tau but did react with all phosphorylated peptides as well as the noP peptide, which indicated that this antibody is not phosphorylation-dependent nor specific for pThr217 (Fig. 1b). Rabbit polyclonal antibody 44–744 did not display immunoreactivity for recombinant 2N/4R tau and modest binding was observed for the noP and the phosphorylated peptides that did not have phosphorylation at Thr217. The highest immunoreactivity for 44–744 was observed for all the peptides containing pThr217 (Fig. 1c).

Table 1 Summary of peptides used in this study.
Full size table
Fig. 1
figure 1

ELISA characterization of novel monoclonal antibody 1F10. ELISA was performed as described in “Materials and methods” using peptides listed in Table 1 or uncoated (Blank) and incubation with (a) 1F10, (b) pThr217 clone E9Y4S, or (c) pThr217 polyclonal antibody 44–744. Data are shown as mean +/- SEM.

Full size image

Glycogen synthase kinase-3β promotes phosphorylated Tau and the 1F10 epitope

Previous reports have shown that tau is a substrate of glycogen synthase kinase-3β (GSK3β), and that several amino acid residues, including Thr217, are phosphorylated by GSK3β60,61,62. Therefore, we expressed 0N/4R human tau in HEK293T cells with or without the overexpression of GSK3β and assessed the antibody immunoreactivities by immunoblotting. In lysates from cells that expressed human tau, the levels detected by total tau antibody 3026 were similar in the absence and presence of GSK3β (Fig. 2a). Notably, a reduction in electrophoretic mobility of tau is observed in the presence of GSK3β, indicating that tau is in a hyperphosphorylated state (Fig. 2a). 7F2 was used as a positive control for detection of phosphorylation at Thr205 of tau promoted by GSK3β since it had previously been shown 7F2 immunoreactivity of tau is enhanced in the presence GSK3β63. Immunoreactivity was observed when 0N/4R tau was transfected in the absence and presence of GSK3β, but signal was greatly increased in the presence of GSK3β (Fig. 2b). Tau immunoreactivity detected by 1F10 was observed only when GSK3β was co-expressed with tau, indicating that endogenous phosphorylation of tau at the 1F10 epitope is low (Fig. 2c). Non-specific bands both of higher and lower molecular weight than tau were observed, indicating 1F10 is not completely specific to tau in this model (Fig. 2c). Rabbit monoclonal antibody E9Y4S, which we found was not specific for pThr217 by ELISA (Fig. 1b), also bound to tau in the absence and presence of GSK3β, but signal was greater in the presence of GSK3β (Fig. 2d). Both 5E2 and 2F12 are conformation-dependent tau antibodies, which require residues Pro218-Glu222 in human tau, but were previously shown to have enhanced immunoreactivity in the presence of authentic and mimicking phosphorylation sites upstream and adjacent to their minimal epitope64. Both 5E2 and 2F12 detected tau only in the presence of GSK3β, indicating that GSK3β affects the phosphorylation state of sites adjacent to the 5E2 and 2F12 epitope (Fig. 2e, f). Immunoblots were also probed with anti-HA and anti-actin antibodies as controls (Fig. 2g, h).

Fig. 2
figure 2

GSK3β promotes immunoreactivity at 1F10, 5E2 and 2F12 epitopes. HEK293T cells were transfected to express GSK3β-HA, wild type 0N/4R human tau, both co-transfected, or empty vector (pcDNA3.1) as (-) controls (Ctl) (N = 3). Cell lysates were probed with total tau antibody (a) 3026, (b) 7F2, (c) 1F10, (d) pThr217 clone E9Y4S, (e) 5E2, (f) 2F12, as well as (g) anti-HA, and (h) anti-actin antibodies. The relative mobilities of the molecular weight marker are indicated on the left side of the immunoblots. Asterisk indicates non-specific protein band detected with anti-HA antibody. 

Full size image

1F10 labels neurofibrillary tangle-like inclusions in the SPAM mouse model of tauopathy

The SPAM (S320F P301S aggregating mutations) mouse model of tauopathy develops neurofibrillary tangle (NFT) like-inclusions throughout the brain which temporally increase from 2 to 6 months of age in brain regions affected during the neuropathological progression of AD65. 1F10 labelling of NFT-like inclusions was observed at 2 and 6 months of age in regions such as the hippocampus, cortex, and amygdala, similar to our positive control 7F2. As with 7F2, the burden of inclusions labelled by 1F10 increased with age. Minimal 1F10 immunoreactivity was found in both non-transgenic (nTg) and Tau knockout (KO) mice, but some nuclear labelling was observed (Fig. 3).

Fig. 3
figure 3

Antibody 1F10 labels NFT-like inclusions in a mouse model of tauopathy. Immunohistochemical staining using monoclonal antibodies 1F10 and 7F2 of brain tissue sections from 2-month old SPAM mice and 6-month old SPAM Tg, nTg, and Tau KO mice as described in “Materials and methods”. Hpc = hippocampus, Ctx = cortex, Amy = amygdala. Scale bar of low magnification images represents 300 μm and 50 μm for high magnification images.

Full size image

Hallmark Alzheimer’s disease pathology is detected by 1F10

Tau immunoreactive neurofibrillary tangles, neuropil threads and dystrophic neurites surrounding amyloid-β (Aβ) plaques, termed neuritic plaques, are defining neuropathological features of AD21. 7F2 and AT180 are specific for pThr20563 and pThr23166, respectively, which are both phosphorylation sites elevated in AD cerebrospinal fluid (CSF) and blood and have been shown to correlate with other established AD biomarkers50,51,52,67,68,69,70. The AT100 antibody is specific for tau phosphorylated at both Thr212 and Ser21471,72,73, which is close to the epitope for 1F10. We performed immunohistochemical labelling of AD and control frontal cortex tissue (Supplementary Tables S1, S2) using established tau antibodies 7F2, AT180 and AT100 as well as 1F10. In AD frontal cortex, 7F2, AT180, and AT100 labelled NFTs, neuropil threads, and neuritic plaques (Fig. 4a). Detection of NFTs, neuropil threads, and neuritic plaques was also observed with the novel antibody 1F10 indicating the epitope for 1F10 is conserved in post-mortem confirmed AD brain tissue (Fig. 4a). Semi-quantification of AD frontal cortex tissue with all four antibodies revealed similar levels of antibody sensitivity for detection of AD neuropathology (Fig. 4b).

Fig. 4
figure 4

Immunohistochemical labelling of Alzheimer’s disease neuropathology. Immunohistochemical staining was performed as described in the “Materials and methods”. (a) Staining of AD and control frontal cortex tissue using monoclonal antibodies 7F2, AT180, 1F10, and AT100 is shown. Thin arrows indicate NFTs, arrow heads demonstrate neuritic plaques, thick arrows highlight neuropil threads. Scale bar of low magnification images is 100 μm, inset is 15 μm. (b) Semi-quantitative analysis of AD tau pathology in the frontal cortex is described in the “Materials and methods”. Data is presented as mean +/- SEM.

Full size image

Differential detection of PSP neuropathology by antibody 1F10

Primary tauopathies such as PSP and CBD are characterized by the accumulation of 4-repeat tau that is found aggregated in glial cells20,28. Astrocytic pathology is differentiated in these diseases by morphology; in PSP tau immunoreactive tufted astrocytes are detected while in CBD, tau immunoreactive astrocytic plaques are observed20,28. Striatum tissue of control, PSP, and CBD cases (Supplementary Tables S2–S4) were immunohistochemically stained using 7F2, AT180, as well as 1F10. Neuropathological labelling of PSP pathology using 7F2 and AT180 was observed in caudate and putamen regions of the striatum, while with 1F10, we observed that signal was remarkably lower compared to both 7F2 and AT180 (Fig. 5a). This observation is reflected by semi-quantitative analysis of caudate and putamen regions immunohistochemically stained with 7F2, AT180, and 1F10 (Fig. 5b). Detection of PSP pathology by AT180 was also modestly lower compared to 7F2 (Fig. 5b). CBD pathology was detected by 7F2, AT180, and 1F10 in both the caudate and putamen (Fig. 5a). 7F2 and AT180 had similar sensitivity for detection of CBD pathology, while 1F10 immunoreactivity was slightly lower depending on the case (Fig. 5b). We also performed immunohistochemical labelling of PiD cases (Supplementary Table S5), a predominantly 3-repeat tauopathy, where the neuropathological hallmarks include aggregated tau inclusions termed Pick bodies74,75,76,77,78. We neuropathologically labeled the hippocampus, a region found to accumulate Pick bodies in the granule cells of the dentate gyrus, with 7F2, AT180 and 1F10. Pick bodies were positively immunostained with 7F2, AT180, and 1F10 tau antibodies (Fig. 6). Due to the low number of cases (N = 2), the quantification of these lesions was not performed, but the sensitivity for Pick pathology appeared to be similar for all three antibodies.

Fig. 5
figure 5

Immunohistochemical labelling of PSP and CBD neuropathology. Immunohistochemical staining was performed as described in the “Materials and methods”. (a) Staining of PSP and CBD striatal tissue using monoclonal antibodies 7F2, AT180, and 1F10 is shown. Scale bar is 80 μm. (b) Semi-quantitative analysis of PSP and CBD tau pathology in the caudate and putamen is described in the “Materials and methods”. Data is presented as mean +/- SEM.

Full size image
Fig. 6
figure 6

Immunohistochemical labelling of PiD neuropathology. Immunohistochemical staining was performed as described in the “Materials and methods”. Staining of PiD hippocampal tissue using antibodies 7F2, AT180, and 1F10 is shown. Scale bar of low magnification images is 100 μm, inset is 15 μm.

Full size image

Discussion

Tauopathies are a group of progressive neurological disorders characterized by diverse clinical and pathological presentations for which disease-modifying therapies are currently inadequate14. Several PTMs have established disease relevance; phosphorylation in particular, has been demonstrated to be a major modification conserved across multiple tauopathies including AD, PSP, CBD, and PiD24,35,79,80. Phosphorylation as a core biochemical feature of aggregated tau protein in AD and other tauopathies is readily detected in hallmark lesions. Using immunological and mass spectrometry methods, phosphorylation sites that characterize certain pathological conditions are found to cluster in the proline-rich region and C-terminus of the tau molecule81. The generation of monoclonal antibodies which target specific phosphorylation sites have provided further information on the relevance of phosphorylation site proximity and the nature of phosphorylation sites potentially arising in tandem with one another (e.g. AT8 (pSer202/pThr205) and AT100 (pThr212/pSer214)71,72,73,82. Recent mass spectrometry studies have further highlighted that double and triple phosphorylated peptides are highly represented in AD brain tissue compared to controls and other primary tauopathies83suggesting that utilizing an antibody capable of binding to multiple phospho-tau species may prove beneficial in monitoring or targeting aggregated tau along the AD continuum. However, it is still unknown the extent to which hyperphosphorylation is a key driver of pathological tau accumulation in distinct topographical brain regions and cell types which characterize these neurodegenerative diseases. Nevertheless, the identification of sensitive and specific biomarker sites elevated in biological fluids has further enabled investigations of the AD pathophysiological process in vivo. For example, pThr217 is one of many sites originally found in brain tissue33,84,85,86,87 but has now been uniquely established as a high-performing biomarker in both CSF and blood46,48,88. Increased concentrations of CSF and blood pThr217 are found to reflect neuropathological burden due to high correlations with Aβ plaque and tau tangle pathology determined by PET imaging and by post-mortem neuropathological assessment51,54,57,70,89. pThr217 has also been widely demonstrated as a critical biomarker site across several assay types (antibody-based and mass spectrometry) and shown to accurately differentiate AD from other neurodegenerative disorders53,54,56,58,70,89,90,91,92. Furthermore, both CSF and plasma levels of pThr217 are highly correlated88,93 indicating that authentic changes in tau pathobiology detected in the periphery are likely stemming from the CNS. Even so, the presence of tau pThr217 in neuropathological lesions in tauopathies has not been as thoroughly established, therefore we generated and characterized a novel mouse monoclonal antibody with strong immunoreactivity for pThr217.

Epitope mapping conducted by ELISA using phosphorylated and non-phosphorylated synthetic peptides and recombinant human 2N/4R tau protein revealed that 1F10 displayed stronger immunoreactivity for pThr217 compared to pThr212 or pSer214. Immunoreactivity for the non-phosphorylated peptide and recombinant human tau protein was not observed, indicating 1F10 is phosphorylation-dependent and pThr217 is an important component of the core epitope of 1F10. We also investigated the reactivity of two commercial antibodies marketed towards having specificity for pThr217. The rabbit monoclonal pThr217 E9Y4S was immunoreactive for all synthetic peptides independent of phosphorylation site. Unlike clone E9Y4S, pThr217 rabbit polyclonal antibody 44–744 was immunoreactive for synthetic peptides containing pThr217 at a much higher magnitude compared to noP, pThr212, pSer214, and dual phospho-peptide pThr212, pSer214. Both E9Y4S and 44–744 were immunoreactive for the non-phosphorylated peptide, but despite the non-phosphorylated peptide and 2N/4R tau sharing amino acid sequence identity, reactivity for recombinant 2N/4R tau was not observed.

Previous studies have found that GSK3β promotes phosphorylation of tau in vitro and in cultured cells, and SerPro/ThrPro motifs which are especially represent in the proline-rich region are subject to GSK3β enzymatic activity60,61,62,63,87,94,95,96. Thr217 is also an established site capable of being phosphorylated by GSK3β60,87,95 which prompted us to investigate the contributions of GSK3β towards forming the 1F10 epitope on tau in cultured cells. Compared to other antibodies investigated (7F2 and E9Y4S) which show some low levels of immunoreactivity in the absence of GSK3β, 1F10 immunoreactivity was observed only in the condition in which tau was co-transfected with GSK3β. In this mammalian cell culture model, we did find that 1F10 was not completely specific for tau and reacted with proteins of higher and lower molecular weights than that predicted for tau.

We previously reported on two conformation-dependent monoclonal antibodies 5E2 and 2F12 and found that their epitope comprised residues P218-E222 of tau. Both antibodies were immunoreactive for the non-phosphorylated tau peptide, but immunoreactivity for full-length recombinant tau was not observed. Interestingly, in the presence of adjacent phosphorylation-mimicking sites (S210E, T212E, S214E, and T217E), 5E2 and 2F12 signal was enhanced64. Furthermore, we show here that in HEK293T cells both antibodies 5E2 and 2F12 were exclusively immunoreactive for tau in the presence of GSK3β overexpression. Our data indicates that in the presence of GSK3β, phosphorylation at residues near P218-E222 is promoted considerably which yields the unmasking of the epitope, allowing it to be available for antibody binding.

Immunotherapeutic-mediated targeting of pThr217 has been investigated pre-clinically in mouse and cell models of tauopathy97,98. Treatment with monoclonal antibodies targeting pThr217 attenuated tau neuropathological burden in a transgenic mouse model and neutralized tau seeds in a model of prion-type induction of tau pathology, indicating targeting pThr217 can facilitate reductions in tau aggregation by multiple mechanisms of action97,98. We investigated tau pathology immunolabelling with 1F10 in our SPAM mouse model of tauopathy65. These mice develop tau neuropathology throughout the neuroaxis, resembling authentic neurofibrillary tangle pathology65. 1F10 immunolabelled NFT-like inclusions in several key brain regions at early and late time points of tau deposition which paralleled that of 7F2 immunoreactivity. Given the robust neuropathological burden throughout the neocortex and the association of pThr217 with early pathological tau deposition in our mouse model, appreciable reductions in pThr217 could potentially alleviate the overall pathological tau burden in both central and peripheral nervous systems of these mice.

Hallmark neuropathological lesions characteristic of AD such as neurofibrillary tangles, neuropil threads, and neuritic plaques were readily detected by antibody 1F10 with similar sensitivity to 7F2, AT180, and AT100. 7F2, AT180, and AT100 are mouse monoclonal antibodies with specificity for tau pThr205, pThr231, and pThr212/pSer214, respectively, and all are able to immunolabel AD neuropathology63,66,71,72,99,100. 7F2 was used as our standard antibody as we have previously demonstrated that 7F2 can also effectively detect tau astroglial pathology100,101. While the AT100 epitope is not identical to the 1F10 epitope, these antibodies are likely able to bind to similar tau species as both have the capacity to bind to the triple phosphorylation peptide (pThr212/pSer214/pThr217) used in this study64. Additionally, it has also been reported that AT100 is immunoreactive for glial tau lesions such as tufted astrocytes in PSP and astrocytic plaques in CBD102,103. CBD and PSP tau neuropathology was immunolabelled with 7F2, AT180, and 1F10 in the striatum, a region prone to tau accumulation in both of these diseases36,37. For CBD, some cases showed similar levels of 1F10 immunoreactivity compared to that of 7F2 and AT180 while others were somewhat lower. Remarkably, 1F10 immunoreactivity in PSP cases was markedly lower in both the caudate and putamen, compared to 7F2. We next sought to immunostain PiD cases, as this tauopathy is primarily characterized by the aggregation of 3-repeat tau isoforms, to further understand if the immunoreactivity pattern is dependent on the isoforms prone to accumulate in these diseases. In our limited cohort of PiD cases, 1F10 sensitivity for PiD neuropathology was similar to two other phospho-tau monoclonal antibodies, 7F2 and AT180. Therefore, since differences in 1F10 sensitivity relative to 7F2 or AT180 were not observed in AD, this suggests that the presence of the 1F10 epitope is less abundant in PSP indicating phospho-epitope heterogeneity in tauopathies. In line with previous observations on the biomarker site pThr181, pThr217 is not elevated at comparable levels to that of AD in primary tauopathies such as CBD and PSP54,58,59,70,89,90,91,104. A previous study found that an in-house pThr217 monoclonal antibody labelled CBD, PSP, and PiD tau neuropathology comparable to that of AT8 despite distinguishing AD from CBD and PSP in a CSF immunoassay90. Interestingly, pThr217 is also elevated in patients with the FTDP-17 mutation R406W, Creutzfeldt-Jakob disease, and Niemann-Pick disease Type C indicating increased pThr217 is not specific for AD and that the presence of Aβ is not a required for increases in levels of tau pThr217105,106,107. Immunodetection using our novel pThr217-selective monoclonal antibody reveals clear differences in tau immunolabelling in PSP compared to 7F2. The paucity of pThr217 tau in CSF and plasma from several primary tauopathies may stem from this modification not being generated or, in contrast to AD, is present at level below the threshold for detection despite robust tau accumulation.

Furthermore, it is hypothesized that the source of glial tau aggregates in primary tauopathies is derived from neuronal tau through a prion-type mechanism. It is possible pThr217 may not be a potent seed-competent species, or kinases which phosphorylate tau at Thr217 are less active or not present in glial cells compared to neurons. A previous study found that pThr217 is turned over at a quicker rate compared to its non-phosphorylated counterpart108. Under pathological conditions where tau is heavily modified by PTMs35 hyperphosphorylation at Thr217 may confer a change in normal turnover leading to buildup of pathological species, which can’t be managed in the context of pathological environments.

Targeting pThr217 using a monoclonal antibody has already been accomplished in a phase 1 clinical trial. Following administration in patients with AD and healthy participants, CSF pThr217 was attenuated, but whether clinical benefits are derived from targeting this phospho-site remains to be determined109. Moreover, the pathological consequences of pThr217 still require further studies to determine the contributions of pThr217 towards disease heterogeneity. 1F10 can potentially be used to determine the role of tau pThr217 towards aggregate formation in preclinical models of tauopathy. Other applications may include the development of a 1F10 immunoassay in order to diagnose and monitor phospho-tau species in CSF or plasma in patients with prodromal AD, in addition to potential utility as an immunotherapy.

Materials and methods

Antibody generation

Peptides were synthesized and purified as a service by GenScript (Piscataway, NJ, USA). Synthetic peptide [CSRSR(pThr)P(pSer)LP(pThr)PPTREPKK] corresponds to residues 208–225 of full length 2N/4R human tau with phosphate groups on Thr212, Ser214, and Thr217. The immunization peptide contained an added Cysteine at the N-terminus for conjugation to Imject maleimide-activated mariculture keyhole limpet hemocyanin (mcKLH; Thermo Scientific, Waltham, MA, USA). Antibody generation was performed using standard techniques for hybridoma production110. Conjugated peptide was used to immunize female BALB/c mice (Jackson Laboratory, Bar Harbor, ME, USA), spleens from immunized mice were harvested, and white blood cells were fused with mouse myeloma cells (Sp2/O-Ag14; ATCC, Manassas, VA, USA). Hybridoma clones were selected using Hypoxanthine-Aminopterin-Thymidine (HAT) supplement (Sigma Aldrich, St. Louis, MO, USA) and the surviving clones were initially screened for reactivity by ELISA using the immunization peptide. Antibody isotype for 1F10 was determined to be IgM using a mouse monoclonal isotyping kit (Sigma-Aldrich, St. Louis, MO, USA).

Other antibodies

Total tau rabbit polyclonal antibody 3026 and mouse monoclonal antibody 7F2 specific for tau phosphorylated at Thr205 were previously described63. Conformation-dependent mouse monoclonal tau antibodies 5E2 and 2F12 were previously described64. Tau pThr217 rabbit monoclonal antibody (Clone E9Y4S, Catalog # 51625S) was purchased from Cell Signaling (Danvers, MA, USA). Tau pThr217 rabbit polyclonal antibody (Catalog # 44–744, Lot # 2973333) was purchased from Invitrogen (Waltham, MA, USA). AT180 is a mouse monoclonal antibody specific for tau phosphorylated at Thr23166,71 was purchased from Thermo Fisher Scientific (Waltham, MA, USA). AT100 is a mouse monoclonal antibody specific for tau phosphorylated at Thr212/Ser21464,71,72,73 was purchased from Thermo Fisher Scientific (Waltham, MA, USA). Anti-HA tag clone 12CA5 mouse monoclonal antibody was purchased from Roche (Indianapolis, IN, USA). Anti-actin clone C4 mouse monoclonal antibody111 was purchased from Thermo Fisher Scientific (Waltham, MA, USA).

Generation of Recombinant Tau protein

cDNA encoding full-length human 2N/4R tau isoform was cloned into bacterial expression vector pRK172112, expressed in Escherichia coli, and purified by mono S chromatography as previously described113.

Enzyme-linked immunosorbent assay (ELISA)

96-well ELISA plates (MaxiSorp; Thermo Fisher Scientific, Waltham, MA, USA) were coated with 100 ng of peptides in 100 µL phosphate buffered saline (PBS; 11.9 mM phosphates, pH 7.4, 137 mM NaCl, 2.7 mM KCl) per well and incubated overnight at 4 °C. Wells were washed with PBS and blocked with 5% fetal bovine serum (FBS) in PBS for approximately 1 h at room temperature (RT). Primary antibodies were added to wells and incubated at RT. After PBS washes, plates were incubated with horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG + IgM secondary antibody or goat anti-rabbit IgG secondary antibody (Jackson Immuno Research Labs, West Grove, PA, USA) diluted in blocking solution. Plates were washed with PBS and 3,3’,5,5’-tetramethylbenzidine (TMB substrate, Thermo Fisher Scientific, Waltham, MA, USA) was used as a chromogenic substrate for HRP and added to each well. The reactions were stopped by adding 1 N HCl and the optical density was measured at 450 nm with a plate reader.

Cell culture and cell harvest

HEK293T cells (ATCC, Manassas, VA) were cultured in Dulbecco’s Modified Eagle Medium with L-Glutamine and high glucose (4.5 g/L) supplemented with 10% FBS and antibiotics (100 units/mL penicillin, 100 µg/mL streptomycin) at 37 °C in 5% CO2. Cells were co-transfected using calcium phosphate precipitation as previously described with some modifications114. 1.5 µg of total DNA was mixed with 18.75 µl of 0.25 M CaCl2 and mixed. An equal amount of 2X BES buffer (50 mM N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid (BES), 280 mM NaCl, 1.5 mM Na2HPO4, pH 6.96) was added to the DNA/CaCl2 mixture 1/5th at a time stepwise and vortexed after each addition, incubated at 15–20 min at room temperature, and then added dropwise to 1 mL of media. Plasmids were co-transfected at a 1:1 ratio, with final total DNA used being the same. The following day, media was replaced with media containing 3% FBS supplemented with antibiotics and cells were harvested 48 h thereafter. Transfected cells were washed with PBS and harvested in SDS-sample buffer (10 mM Tris, pH 6.8, 1 mM EDTA, 40 mM DTT, 0.005% Bromophenol Blue, 0.0025% Pyronin Yellow, 1% SDS, 10% Sucrose) and incubated at 95 °C for 10 min and stored at -80 °C till further use.

Western blotting

Equal volumes of HEK293T cell lysates were loaded onto 10% SDS-polyacrylamide gels, resolved by SDS-PAGE, and electrophoretically transferred onto nitrocellulose membranes. Membranes were blocked in 5% nonfat dry milk or 5% bovine serum albumin in Tris-buffered saline (TBS; 50 mM Tris, pH 7.5, 150 mM NaCl). Antibodies were added to membranes either diluted in blocking solution or undiluted for hybridoma-secreted antibodies which were then incubated overnight at 4°C. Membranes were washed in TBS three times and then incubated in HRP-conjugated goat anti-mouse or anti-rabbit secondary antibody diluted in blocking solution for 1 h at RT. Membranes were washed three times in TBS and subsequently imaged using Western Lightning Plus ECL reagents (PerkinElmer, Waltham, MA, USA) followed by chemiluminescence imaging (GeneGnome XRQ, Syngene).

Immunohistochemistry of mouse and human tissue

Formalin fixed human brain was obtained from the University of Florida Neuromedicine Brain and Tissue Bank (UF HBTB) (Supplementary Tables S1–S5). Mice were euthanized with CO2 and perfused with a heparin/PBS solution. Mouse tissue was harvested and fixed in 70% EtOH/150 mM NaCl. Human and mouse brain tissue was embedded in paraffin and sectioned. Slides were deparaffinized in xylenes and rehydrated in alcohols (100%, 100%, 90%, 70%) and then washed in water. For antibodies 1F10 and AT180, heat induced epitope retrieval was performed with H2O/0.05% Tween 20 in a steamer for 1 h. For antibodies 7F2 and AT100, a Target Retrieval Solution, Citrate pH 6 (Agilent Technologies, Santa Clara, CA, USA) was used for 1 hour. Slides were then allowed to cool in running water. Endogenous peroxidases were quenched using PBS/1.5% H2O2 with 0.005% Triton X-100 for 15–20 minutes. Slides were washed in water followed with 0.1 M Tris, pH 7.6. 5% nonfat dry milk/0.1 M Tris, pH 7.6 was used as a blocking solution for 1F10 and 2% FBS/0.1 M Tris, pH 7.6 was used as a blocking solution for 7F2, AT180 and AT100. Primary antibodies were applied undiluted or diluted in blocking solution to slides which were then incubated overnight at 4˚C. The following day, slides were washed 3 times in 0.1 M Tris, pH 7.6 and tissue sections were incubated in blocking solutions. Biotinylated goat anti-mouse IgM, Mu chain specific secondary antibody (Vector Laboratories, Newark, CA, USA) diluted 1:500 in blocking solution (1F10), or biotinylated goat anti-mouse IgG diluted 1:1000 (AT100) or 1:3000 (7F2 and AT180) was added to the sections for 1 h at RT. ImmPRESS HRP horse anti-mouse IgG polymer (Vector Laboratories, Newark, CA, USA) diluted 1:10 was also added to the secondary antibody solution, except for mouse tissue sections immunostained with 7F2. Slides were washed 3 times in 0.1 M Tris, pH 7.6 and incubated in blocking solution. Avidin-biotin complex (ABC) solution (Vectastain ABC kit; Vector Laboratories, Newark, CA, USA) was diluted 1:500 in blocking solution for 1F10, 1:1000 for AT100, or 1:3000 for 7F2 and AT180 and applied to slides for 1 h at RT. Slides were washed 3 times in 0.1 M Tris, pH 7.6 and tissue sections were developed using chromogen 3,3’-diaminobenzidine (DAB kit, KPL, SeraCare, Gaithersburg, MD, USA) and counterstained with hematoxylin (Sigma Aldrich, St. Louis, MO, USA). Sections were washed in tap water and dehydrated in alcohols (70%, 90%, 100%, 100%) followed by xylenes and coverslipped using cytoseal.

Semi-quantification of Tau neuropathology

Slides were scanned using an Aperio AT2 slide scanner at 40x magnification (Aperio Technologies, Vista, CA, USA). Monoclonal antibody 7F2 was used to establish a baseline. 3 fields of view at 10x magnification were captured for each stained section which were then scored by three raters (S.Q., S.P., B.I.G.) using a 0 (no pathology), 1 (low pathology), 2 (moderate pathology), or 3 (high pathology) scale. The scores were average and graphed using GraphPad Prism software (San Diego, CA, USA).