Preferential uptake and higher toxicity of microvesicular/exosomal TDP-43. (A) TEM pictures of ultrathin sections of MVEs. A representative micrograph of intact vesicles with a bilayered lipid membrane is shown. (B–D) Anti–TDP-43 immunogold labeling was performed on ultrathin sections of MVEs (10-nm gold particles). Intraluminal (B and C) and membrane-associated (D) TDP-43 is shown. Bars, 50 nm. (E) Immunoblotting of HEK-293 cells transfected with myc–TDP-43 and their corresponding CM MVE-free supernatant after ultracentrifugation (sup) and MVEs. Overexpressed myc–TDP-43 was detected using a Myc antibody, and endogenous TDP-43 was detected in MVEs using a TDP-43 antibody. Flotillin was used as an MVE marker. The closed arrowhead indicates endogenous TDP-43, and the open arrowhead indicates myc–TDP-43. (F) MVEs were prepared from HEK-293 cells expressing either full-length luciferase alone (Luc), TDP-L1 + TDP-L2, or TDP-43 fused to full-length luciferase (TDP-Luc). Luciferase activity of MVE fractions was normalized to the activity measured in the respective host cells. n = 12 per group. (G and H) MVE fractions (MVEs) and MVE-free supernatants (sup) from HEK-293 cells transfected with control vector (Mock) or TDP-Luc were applied to naive HEK-293 cells (G) or primary neurons (H) and incubated for 3 d. Uptake of TDP-Luc was detected by luciferase activity measurement. n = 14. (I) Caspase-3/7 toxicity assay of naive primary neurons treated with MVE fraction (MVEs) or MVE-free supernatant (sup) containing TDP-Luc. n = 9 per group. Mean ± SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001. CPS, counts per second.