In situ seeding assay: A novel technique for direct tissue localization of bioactive tau

Abstract

Proteins exhibiting prion-like properties are implicated in tauopathies. The prion-like traits of tau influence disease progression and correlate with severity. Techniques to measure tau bioactivity such as RT-QuIC and biosensor cells lack spatial specificity. Therefore, we developed a histological probe aimed at detecting and localizing bioactive tau in situ. We first induced the recruitment of a tagged probe by bioactive Tau in human brain tissue slices using biosensor cell lysates containing a fluorescent probe. We then enhanced sensitivity and flexibility by designing a recombinant probe with a myc tag. The probe design aimed to replicate the recruitment process seen in prion-like mechanisms based on the cryo-EM structure of tau aggregates in Alzheimer disease (AD). Using this novel probe, we observed selective staining of misfolded tau in pre- and post-synaptic structures within neurofibrillary tangles and neurites, whether or not associated with neuritic plaques. The probe specifically targeted AD-associated bioactive tau and did not recognize bioactive tau from other neurodegenerative diseases. Electron microscopy and immunolabeling further confirmed the identification of fibrillar and non-fibrillar tau. Finally, we established a correlation between quantifying bioactive tau using this technique and gold standard biosensor cells. This technique presents a robust approach for detecting bioactive tau in AD tissues and has potential applications for deciphering mechanisms of tau propagation and degradation pathways.

Keywords: Alzheimer disease; astrocyte; histopathology; prion-like; synapse; tau; tauopathies.

Figure 1.

Tau fragments are recruited by bioactive Tau and visualized in situ. (A) Schematic of the different constructs used in this study compared to full length Tau and the sequence included in the central core of AD fibrils. Full-length Tau protein (2N4R) is a 441aa protein with a proline-rich domain and a microtubule binding domain. The central core of tau fibrils in AD is constituted by aa. 306-378. The CRL-3275 cell line expresses the segment 243-366 of tau with a P301S mutation and tagged with CFP (243-366-CFP) or YFP. The HEK 293T clone 18 expresses a chimeric protein composed of Tau fragment 244-379 with P301L, K311Q, K317Q, K321Q mutations (3 pseudo-acetylation), and the Neon Green fluorophore attached at the C-terminus with an (EAAK)₃ linker. Recombinant probe myc-297-391 is composed of Tau fragment 297-391 with a myc-tag attached at the N-terminus. Myc-297-391-2xKQ is composed of Tau fragment 297-391 with K343Q, K353Q mutations and a myc-tag attached at the N-terminus. (B) AD and control brain slices (BA8) are incubated with 243-366-CFP or 244-379-3xKQ-NG for 24 h. Maximum intensity projections of confocal images demonstrate the staining of tangles and neuritic threads by the recruited 244-379-3xKQ-NG (green) and the bound Tau (AT8) antibody (purple) on AD brain, but not on control tissue. No signal is visualized with 243-366-CFP on AD tissue; however, the presence of Tau deposits is confirmed by the Tau (AT8) signal. (C) AD brain slices (BA8) are incubated with Myc-297-391 or Myc-297-391-2xKQ for 24 h. Then, immunofluorescence staining is performed with myc (green), and pTau181 (purple) conjugated antibodies. Maximum intensity projections of confocal images show the recruitment of Myc-297-391 and Myc-297-391-2xKQ probe in tangles and neuritic threads. The probe strongly colocalizes with the pTau181 antibody. (D) Quantification of myc signal on virtual slides showed more stained area in AD (BA8) compared to control (BA8) for both myc-297-391 (0.64 ± 0.26 for AD vs 0.010 ± 0.00; P = .005) and myc-297-391-2xKQ (0.29 ± 0.099 vs 0.01 ± 0.058; P = .023). There is no statistical difference between the 2 probes (P = .99). Each point corresponds to one patient (3 controls and 3 AD). Error bars at SEM. Kruskal-Wallis test with multiple comparisons and Bonferroni corrected alpha value. *P < .05; **P < .01.

Figure 2.

The probe is specific for Tau AD conformation. (A, B) AD brain slices (BA8) are incubated with recombinant myc-297-391 for 24 h after pre-treatment with PBS for 24 h (PBS), 297-391 for 24 h, myc-297-391 for 24 h, or formic acid for 10 min followed by PBS for 23 h and 50 min. Subsequently, immunofluorescence staining is performed with myc and pTau181 antibodies. (A) Maximum intensity projections of representative confocal images. (B) Quantification of the signal in the cortex of 5 different cases on virtual slides shows that, compared with the PBS pre-treatment, pretreatment with 297-391 decreases the myc signal, and formic acid dramatically decreases the myc signal; however, the pTau181 signal remains similar. Pre-treatment with the probe containing the myc-tag does not alter the myc signal. Error bars at standard error of the mean. One-way ANOVA for paired values with multiple comparisons and Bonferroni corrected alpha value, performed on the logarithm of the values *P < .05; **P < .01; ***P < .001. (C, D) Brain slices from (BA8) cortex of 3 controls, 3 AD, 3 PSP, 3 CBD, and 3 PiD are incubated with recombinant myc-297-391 for 24 h. Subsequently, immunofluorescence staining is performed with myc and pTau181 antibodies. Maximum intensity projections of representative confocal images show the presence of neurites and tangles stained by myc-297-391 and pTau181 in AD (compare to Figure 1C), tufted astrocytes (inset) and neuronal inclusion only stained by pTau181 in PSP, astrocytic plaques (inset) and neuronal inclusion only stained by pTau181 in CBD and Pick bodies (inset) stained by pTau181 in PiD. (D) Quantification of the signal in the cortex with virtual slides shows that compared to control, AD exhibits a higher area covered. There are no differences between control and PSP, CBD, or PiD. Error bars at SEM.

Figure 3.

(A) AD brain slices (BA8) are incubated with myc-297-391 for 24 h. Then, immunofluorescence staining is performed with myc, Aβ, and pTau181 conjugated antibodies. Maximum intensity projections of confocal images show the recruitment of the probe in tangles (arrow) and neuritic threads, sometimes included in neuritic plaques (arrowhead). The probe strongly colocalizes with the pTau181 antibody. (B) AD brain slices (BA8) were incubated with myc-297-391 and subsequently probed with primary antibodies against GFAP, pTau181, and myc, all conjugated with fluorophores. The slices were then imaged using super-resolution microscopy, and the resulting images were reconstructed and analyzed using Imaris software. The representative super-resolution image displays DAPI (blue) myc-297-391 (green), pTau (red), GFAP (purple), and the colocalization between myc and pTau181 (yellow). The first row corresponds to maximum intensity projections of the super-resolution microscopy images. The second row shows the 3D reconstructions for each channel. The third row, from left to right, displays the 3D reconstruction of myc colocalized with GFAP (in green), myc and GFAP channels separately, pTau and GFAP channels separately, and finally the GFAP channel with myc colocalized with pTau181 (in yellow). Rare astrocytes exhibiting colocalization between GFAP and myc are observed. (C) AD brain slices (BA8) were incubated with myc-297-391 and subsequently probed with primary antibodies against PSD95, synapsin1, and myc, all conjugated with fluorophores. The slices were then imaged using super-resolution microscopy, and the resulting images were reconstructed and analyzed using Imaris software. The representative super-resolution image displays DAPI (blue), myc-297-391 (green), PSD95 (red), synapsin1 (purple), and the colocalization between PSD95 and synapsin1 (yellow). The first row corresponds to maximum intensity projections of the super-resolution microscopy images. The second row, from left to right, displays the 3D reconstruction of: PSD95, synapsin1 and myc channels; the myc channel; synapsin1 channel (purple) and myc colocalized with synapsin1; PSD95 channel and myc colocalized with PSD95. The third row, from left to right, displays an inset of the 3D reconstruction of: PSD95, synapsin1 and myc-297-391 channels; the colocalization between PSD95 and synapsin (yellow) and the colocalization between PSD95, synapsin1 and myc-297-391 (green); synapsin1 channel (purple) and myc-297-391 colocalized with synapsin1; PSD95 channel and myc-297-391 colocalized with PSD95. Circles show examples of synapses containing myc-297-391. Tau is observed in pre, post, and complete synapses. (D) Colocalization between myc-297-391 and synapsin1 (pre-synapses), PSD95 (post-synapses), and both synapsin1 + PSD95 (mature synapses) were quantified. To ensure that colocalization is not random, the same quantifications were performed after rotating the myc-297-391 channel by 180 degrees. Results show more myc-297-391 colocalization with pre-, post-, and mature synapses before flipping the myc channel compared to after flipping. This confirms the specific colocalization of bioactive Tau. The proportion of synapses containing bioactive Tau does not differ significantly between pre-, post-, and mature synapses. Error bars at standard error of the mean. Kruskal-Wallis test with multiple comparisons and Bonferroni corrected alpha value *P < .05; **P < .01; ***P < .001.

Figure 4.

Schematic of the experimental design: (A) High molecular weight (HMW) proteins and sarkosyl-insoluble proteins extracted from AD (BA8) and control brain (BA8) are incubated with recombinant myc-297-391 for 24 h. Subsequently, HMW proteins and free probes are separated using SEC. The fractions containing HMW proteins and the sarkosyl-insoluble fraction are then applied to EM grids and immunostained with myc antibody and pTau181 antibody, followed by secondaries conjugated with gold beads (6 nm for myc secondary antibody and 12 nm for pTau181 secondary antibody). The grids are observed by transmission electron microscopy, and random pictures are taken. The number of colocalized 6 nm and 12 nm beads ± fibers are counted. (B) Representative transmission electron microscopy images show colocalization between 12 nm and 6 nm beads on amorphous material in HMW and sarkosyl-insoluble fractions extracted (arrow) from AD brain and 12 nm beads on fibers (arrowhead) on sarkosyl-insoluble fraction extracted from AD brain. (C) In AD HMW proteins, 12 nm beads are more frequently colocalized with 6 nm beads than in control cases. In sarkosyl-insoluble fractions, colocalized 6 nm and 12 nm beads are more frequently found on amorphous structures than on fibers. No difference is observed in the number of beads counted on amorphous structures between HMW and sarkosyl-insoluble fractions. There is no colocalization between 12 nm and 6 nm beads in HMW extracted from control cases. Error bars at standard error of the mean. Kruskal-Wallis test with multiple comparisons and Bonferroni corrected alpha value *P < .05; **P < .01.

Figure 5.

AD brain from 24 AD cases (BA8) has been stained with myc-297-391 and Tau AT8, and the signal in the cortex is quantified on virtual slides. From the other half of each brain sample, we extracted PBS-soluble proteins and sarkosyl-insoluble proteins. HMW proteins are extracted from the PBS-soluble fractions through SEC. Tau concentration is measured with dot-blot and bioactivity using in-vitro seeding assay and shown as integrated FRET density (IFD). (A, B, and C) myc-297-391 staining is positively correlated with FRET in the PBS fraction (r = 0.66 [0.33; 0.85]; P = .0006), in the HMW fraction (r = 0.53 [0.13; 0.78]; P = .0010) but not in the sarkosyl-insoluble fraction (r = −0.18 [−0.44; 0.41]; P = .94). (D, E, and F) Area covered by Myc-297-391 staining is positively correlated with total tau concentration in the HMW fraction (r = 0.63 [0.28; 0.83]; P = .0012), but not in PBS soluble fraction (r = 0.2925 [−0.15; 0.63]; P = .18), sarkosyl-insoluble fraction (r = −0.0029 [−0.42; 0.42]; P = .98). A two-tailed Spearman’s rank non-parametric correlation test was used, and r and P values are indicated on the plots. *P < .05; **P < .01; ***P < .001.