Fig 2: Human BIN1 isoform 9 favors the release of Tau via EVs in vitro as well as increasing the area of pTau staining in vivo. (A) Western blot of Tau and BIN1 in cell lysates (left panel) and EVs (right panel) from HEK293T cells transfected with Tau 2N4R alone or in presence of BIN1 isoform 1 or 9 (indicated by arrows). GAPDH (left panel) represents a cell lysate loading control. CLATHRIN (right panel) serves as an extracellular vesicle loading control. One representative experiment shown out of 4 independent replicates. WB images shown represent cropped sections of full blots shown in Supplementary Fig. 12. (B) Western blot of Tau and BIN1 in cell lysates (left panel) and EVs (right panel) from HEK293T cells transfected with Tau 2N4R P301L-YFP alone or in presence of BIN1 isoform 1 or 9 (indicated by arrows). GAPDH (left panel): cell lysate loading control. CLATHRIN (right panel): extracellular vesicle loading control. One representative experiment shown out of 4 independent replicates. WB images shown represent cropped sections of full blots shown in Supplementary Fig. 12. (C) Experimental scheme. Artwork by Channa Bao. (D) Representative image of Immunohistochemical staining of pTau (40E8) in hippocampi of PS19 mice injected with AAV expressing GFP or GFP-P2A-hBIN1 isoform 9. Scale: 600 µm. (E) Plot representing the percentage of pTau staining area in the ipsilateral hippocampus (area of pTau staining normalized to the size of the ipsilateral hippocampus for each animal). Each triangle represents one animal. Unpaired parametric Student’s t-test two tailed, 95% confidence: p-value: *p < 0.05; **<0.01; ***p < 0.001.

Fig 3: Tau propagation modulates mouse hippocampal CA1 response to optogenetic and pharmacologic stimulation.(A) Top row of photomicrographs show bright field images of wild-type (WT) or Wfs1-Cre mouse brain hippocampal slices 4 weeks after injection of a Cre-dependent channelrhodopsin-2 (Chr2) with or without Cre-dependent P301L mutant tau expression, positioned on a multielectrode array. Bottom row of photomicrographs show visualization of YFP in hippocampal slices from wild-type and Wfs1-Cre mouse brains 4 weeks after injection of a Cre-dependent Chr2 with or without Cre-dependent P301L mutant tau expression. Scale bars, 100 µm. (B) Shown is baseline neuronal activity in the hippocampal CA1 slices measured by multielectrode array 4 weeks after the injection of a Cre-dependent Chr2 with or without Cre-dependent P301L mutant tau expression in wild-type and Wfs1-Cre mouse brains [one-way ANOVA, F(2,18) = 0.33, Wt-Chr2 group: n = 8 animals; Wfs1-Chr2 group: n = 7 animals; Wfs-Chr2-tau group: n = 6 animals]. (C) Shown is hippocampal CA1 slice neuronal activity measured by multielectrode array after dentate gyrus electrical stimulation, 4 weeks after the injection of a Cre-dependent Chr2 with or without Cre-dependent P301L mutant tau expression in wild-type and Wfs1-Cre mouse brains [one-way ANOVA, tau effect F(2,18) = 0.08, Wt-Chr2 group: n = 8 animals; Wfs1-Chr2 group: n = 7 animals; Wfs-Chr2-tau group: n = 6 animals]. (D) Left: Shown is a heatmap representing the degree of hippocampal CA1 activation in response to light stimulation of YFP- and Chr2-expressing axons in hippocampal slices, 4 weeks after the injection of a Cre-dependent Chr2 with or without Cre-dependent P301L mutant tau expression in wild-type or Wfs1-Cre mouse brains. (D) Right: Quantification of CA1 neuronal activity after light activation of YFP-expressing channelrhodopsin axons in hippocampal slices during the total 20 s of recording after light stimulation [one-way ANOVA, tau effect F(2,18) = 8.30 **P < 0.01; Tukey’s post hoc, *P < 0.05 and **P < 0.01. Wt-Chr2 group: n = 8 animals; Wfs1-Chr2 group: n = 7 animals; Wfs-Chr2-tau group: n = 6 animals]. (E) Quantification of CA1 neuronal activity after light activation of YFP and Chr2-expressing axons in hippocampal slices during only the first 2 s after light stimulation [one-way ANOVA, tau effect F(2,18) = 7.19 **P < 0.01; Tukey’s post hoc, *P < 0.05 and **P < 0.01. Wt-Chr2 group: n = 8 animals; Wfs1-Chr2 group: n = 7 animals; Wfs-Chr2-tau group: n = 6 animals]. (F to H) Shown is a representative parasagittal section of Wfs1-Cre mouse brain injected unilaterally with AAV.PHP.eB-Flex-hM3D-mCherry (DREADDs activator) into the hippocampal CA1 region and AAV2/6-Flex-P301L mutant tau or AAV2/6-Flex-GFP injected unilaterally into the MEC. Four weeks after viral vector injection and 90 min after intraperitoneal injection of clozapine at 0.3 mg/kg, hippocampal CA1 sections were immunostained with (F) anti-mCherry antibody (red) and anti-GFP antibody (green); (G) anti-human tau HT7 antibody and (H) anti–c-fos antibody. Scale bars, 200 µm. (I) Quantification of c-fos staining in (H) for the contralateral (internal control for clozapine) and ipsilateral sides [two-way ANOVA, interaction effect F(1,22) = 19.86; Tukey’s post hoc, ****P < 0.0001; ns, not significant; control: n = 6 animals; P301L mutant tau: n = 7 animals].

Fig 4: Trem2 KO exosomes exhibit enhanced tau seeding activity. A Western blot of tau and Alix in lysates and exosomal preparations from tau-loaded WT microglia with or without LPS/ATP stimulation. B Comparison of tau and Alix in lysates and exosomes from tau-loaded WT and Trem2 KO microglia following LPS/ATP stimulation by Western blot analysis. C ELISA measurement of tau levels in WT and Trem2 KO microglial exosomes from 3 independent cultures, 2 replicate ELISA runs per culture. D Experimental design schematic, WT and Trem2 KO microglia exposed to tau oligomers were treated (or untreated) with exosome inhibitor (GW4869), followed by exosomal induction with LPS/ATP. Purified exosomes were then assayed for tau seeding using the tau-RD FRET biosensor cell line, or injected into mouse brain for histopathological analysis. E Exosomes were purified from 1 × 107 WT or Trem2 KO microglia in the absence of tau (no tau), pre-loaded with tau oligomers (+ tau) and induced with LPS/ATP, or Trem2 KO microglia pre-loaded with tau and pre-treated with 10 µM GW4869 prior to LPS/ATP release (+ tau, + exosome inhib), and applied to a tau RD reporter cell line. Tau fret signals were then visualized by fluorescence microscopy. Bar = 10 µm. F Quantification of the area of FRET-positive signals detected (µm2) per cell from representative images in (E). G Characterizing pathological effects of exosomal tau preparations in vivo. Exosomal preparations from tau-loaded or control (“no tau”) WT or Trem2 KO microglia were stereotactically injected into 5-month old WT mouse hippocampus, and stained for AT8 pathology (green) or nuclei (DAPI, blue) 21 days after injection. Bar = 50 µm. H Number of AT8-positive cells (per 2.5 × 105 µm2 area) were scored from 6 independently injected animals per treatment/genotype, graph represent mean ± SE. Graphs in (C), (F) and (H) depict mean ± SE. Statistical significance was determined by Student’s t-test (C), One-way ANOVA with Tukey’s multiple comparison (F, H), **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig 5: BIN1-associated, Tau-containing EVs are seeding-competent. (A) BIN1 immunoprecipitated EVs show the presence of various isoforms of Tau (long exposure shown) as well as BIN1 with molecular weights compatible with that of isoform 6 and 9 (as shown in lane containing protein lysate from 293T cells transfected with BIN1 isoform 6 and 9 expressing plasmids and indicated by arrows). TST101 represents exosome marker. True-Blot secondary antibodies (Rockland) had been used to avoid detection of non-specific bands. One representative experiment shown out of 3 independent replicates. WB images shown represent cropped sections of full blots shown in Supplementary Fig. 11. (B) Tau immunoprecipitated material shows the presence of various truncated and full-length Tau isoforms (long exposure shown) with molecular weights compatible with control aggregated Tau fibrils (short exposure shown). No BIN1 or TST101 was detected. True-Blot secondary antibodies (Rockland) had been used to avoid detection of non-specific bands. One representative experiment shown out of 3 independent replicates. WB images shown represent cropped sections of full blots shown in Supplementary Fig. 11. (C) Nanoparticle tracking analysis of extracellular vesicles purified by affinity immunopurification with anti-BIN1 from human CSF. One representative experiment out of 3 independent replicates. (D) BIN1 immunoprecipitated EVs (black bars) as well as BIN1 immunodepleted CSFs (gray bars) were prepared from samples obtained from 10 different individuals and tested in the Tau-FRET assay for Tau oligomerization. Results are plotted as Integrated FRET Density values for each sample (reference codes reported). Samples containing pTau and/or oligoTau by previous western blot analysis are highlighted in red. Samples devoid of pTau and/or oligoTau by previous western blot analysis are highlighted in blue. Elution Buffer (EB) serves as a negative control, whereas Tau fibrils (3 µM) serve as a positive control.