Ethics approvals
This study was approved by the research ethics committee of Tokyo Metropolitan Institute of Medical Science (Permission number: 16-25). The research use of autopsied brains was conducted in compliance with Japan’s Postmortem Examination and Corpse Preservation Act and with informed consent from the next of kin of all subjects. All methods were carried out in accordance with relevant guidelines and regulations (Declaration of Helsinki).
Gold enhancement of the QDs on nickel grids
To monitor the influence of gold enhancer on QD particles, 1.5 μl of QD565, QD655, or QD705 solutions (1:30 dilution in distilled water, Invitrogen, Thermo Fisher Scientific, MA, USA) were placed onto the formvar coated nickel grids, and the grids were air-dried. GoldEnhance EM plus (Nanoprobes, NY, USA) is an autometallographic enhancer which catalytically deposits gold ions in solution onto the nanoparticle as metallic gold. 40 μl droplets of the reagent mixture of GoldEnhance EM plus were prepared according to the manufacturer’s protocol, and placed on the Parafilm (Bemis Company, Inc, WI, USA) strip. The grids were immersed in the GoldEnhance EM plus droplets for various time (1 s, 2 min, or 5 min) (Fig. 1Aa). Alternatively, the grids were immersed in 1% gold (III) chloride solution or 0.5% silver nitrate solution diluted in distilled water for 3 min (Fig. 1Aa). The grids were subsequently rinsed with distilled water, air-dried, and observed under EM.
Gold enhancement of QD particles intensifies their electron density by progressive accumulation of gold. (A) (a) QD655 and QD565 particles before and after gold enhancement for 1 s, 2 min, and 5 min using GoldEnhance EM plus® are shown. Alternatively, Gold enhancement using 1% gold (III) chloride solution for 3 min, or silver enhancement using 0.5% silver nitrate solution for 3 min was also performed on QD655 particles. Enhanced EM contrast of the whole particles of QD565 and the tips of rod-shaped QD655 was obtained after gold enhancement. Bars: 20 nm. (b) A brain section labeled for GFAP/QD565 was gold enhanced using 1% gold (III) chloride solution for 3 min. Bar: 100 nm. (B) High-angle annular dark field scanning transmission electron microscope (HAADF-STEM) image and EDX spot analysis of gold enhanced QD565 (a,d,e), native QD705 (b,f), and gold enhanced QD705 particles (c,g,h,i) on a nickel grid. Numbered rectangles in the EM images correspond to the areas for EDX spot analysis. Gold enhanced QD705 consisted of two different crystalloid structures with different interplanar spacing. GE: gold enhancement. Bars: 2 nm. (C) (a) Gradual increase in size of high contrast area during continuous STEM observation. EBI: electron beam irradiation. (b) Sizes of high density nanospheres on QD565, 655, and QD705 after GE was obtained 2 min and 7 min after continuous EBI. * P < 0.05 (Welch’s two sample t-test). Bars: 5 nm. (D) Change in color and absorbance (450 nm) of quantum dot (QD) solution by gold enhancement in the well. GE gold enhancement.
Gold enhancement of the QD in microtiter plates
To quantify bright-field coloration after gold enhancement of QDs, gold enhancement reaction was performed in the microtiter plates and absorbance was measured. 10 μL goat anti-mouse IgG antibody (1:500 / phosphate-buffered saline (PBS)) was immobilized in the wells of microtiter plates (Nunc-Immuno MicroWell 96-well Plates, Thermo Fisher Scientific, MA, USA) overnight at 4 °C. Subsequently, the wells were incubated with 100 μL normal mouse serum (1:500/PBS) for 1 h at room temperature (RT). Some of the wells were further incubated with biotinylated goat anti-mouse IgG (1:1000/PBS) for 1 h, and ABC complex (1:50/PBS, Vector Laboratories, CA, USA) for 1 h at RT, for the subsequent reaction with streptavidin. Wells were washed three times with PBS containing 0.05% Triton X-100 between each step. Subsequently, after washing three times with 0.1 M Tris HCl, QD565-conjugated goat anti-mouse IgG (1: 200/0.1 M Tris HCl, Invitrogen, CA, USA), QD655-conjugated streptavidin (1: 200/0.1 M Tris HCl, Invitrogen, CA, USA), QD705-conjugated goat anti-mouse IgG (1: 200/0.1 M Tris HCl, Invitrogen, CA, USA) were added to each well and allowed to stand for 1 h. After washing three times with 0.1 M Tris HCl, 100 μL GoldEnhance EM plus was added, and reaction was carried out for up to 30 min. Absorbance was measured at 450 nm using Biotrak II (GE healthcare, IL, USA) at 1, 8, 16, and 30 min.
Tissue preparation from autopsied human brain
The autopsied brains were routinely immersion-fixed in 4% formalin for 4 weeks. Tissue blocks of the temporal lobes were washed in 0.1 mol/L phosphate buffer (PB) and were cryoprotected with 15% sucrose/PB overnight and then in 30% sucrose/PB overnight at 4 °C. The tissue was then immersed in Optimal Cutting Temperature compound (Sakura Finetek Japan, Tokyo, Japan) and rapidly frozen on the stage of a carbon dioxide freezer kept below −10 °C. 26 μm-thick floating sections were prepared by cutting the tissue on a sliding microtome equipped with a freezing stage below −10 °C for double labeling toward phospho-tau (Ser202, Thr205) and glial fibrillary acidic protein (GFAP). Protocols for floating sections are summarized in Table 1.
Dual immunolabelling with QDs or QDs/fluoronanogold
26 μm-thick floating sections were washed with PBS, blocked for 30 min in 5% normal goat serum/0.05% NaN3/PBS, and incubated with the mixture of anti-phospho-tau (Ser202, Thr205) antibody AT8 (1:1000, mouse monoclonal, Thermo Fisher Scientific, MA, USA), and anti-GFAP (1:100, rabbit polyclonal, DAKO, Agilent Technologies, CA, USA), diluted in the blocking buffer, for approximately 96 h at 4 °C. For single immunolabeling for ubiquitin, anti-ubiquitin (1:100, rabbit polyclonal, DAKO, Agilent Technologies, CA, USA) was used as primary antibody. The sections were washed with PBS, incubated with biotinylated secondary anti-rabbit antibody (1:200, BA-1000, Vector Laboratories, CA, USA) and AT8 (1:1000), diluted in PBS, for 3 days. After washing three times with PBS, sections were subsequently incubated with avidin biotinylated enzyme complex (1:200, Elite ABC, Vector Laboratories, CA, USA) diluted in PBS for 1 h, to enhance the interaction between biotinylated secondary antibody and streptavidin. Sections were washed with Tris-buffered saline (TBS). AT8 antibody was selectively labeled with QD565 conjugated with F(ab’)2 anti-mouse IgG (H + L) (1:50, Invitrogen, CA, USA), and anti-GFAP antibody or anti-ubiquitin antibody was selectively labeled with either QD655 conjugated with streptavidin (1:50, Invitrogen, CA, USA) (Fig. 2A) or Alexa488-1.4 nm nanogold conjugated with streptavidin (1:50, Nanoprobes, NY, USA) (Supplementary Figs. 3, 4), diluted in TBS for 90 min in the dark. After washing with TBS, the sections were put in a glass container filled with washing buffer, and scooped up using the face of the glass slide. The glass slide was tilted to remove droplets of buffer and glycerol was added gently. A cover glass was placed and the edge of the cover glass was sealed. We avoided the use of PAP pen, prolonged incubation in PBS at low concentration, or anti-fade reagent containing p-phenylenediamine, to avoid quenching of QD fluorescence.
EM distinction of QD565 for AT8 and QD655 for GFAP after gold enhancement and their LM counterparts on double immunolabeled section. (A) A section from the autopsied brain was double immunolabeled for AT8/QD565 (magenta) and GFAP/QD655 (green). (a) Extended focus image (EFI) of an NFT (open triangle) and astrocytes and its cross-sectional reconstruction (20 μm thick). Fluorescent signals from QD565 and QD655 were similarly intense throughout the entire depth of the section. (b) EFI of astrocytes on the same section at different site. An astrocyte enclosed in the rectangle was targeted for subsequent EM observation. (c,d) Bright-field LM counterpart of targets in (a,b) after gold enhancement and post-fixation. QD565 and QD655 labeled structures are darkened. (e–j) EM images corresponding to the LM images, after gold enhancement of the section. The areas enclosed by rectangles correspond to the images just below. While the filaments in NFT (e,g,i) was labeled with high contrast spherical QD565 (arrow), the glial fibers of the astrocyte (f,h,j) was labeled with rod-like QD655, which showed high contrast nanospheres deposited at the tips of the low contrast QD655 nanorods (filled triangle). The QD565 (in i, arrow) and the QD655 label (in j, filled triangle) are readily distinguishable from each other by their characteristic shapes after gold enhancement. Bars in a: 20 μm, b: 100 μm, c: 20 μm, d: 100 μm, e: 10 μm, f: 2 μm, g: 200 nm, h: 100 nm, i and j: 20 nm. (B) Another section of the same autopsied brain was labeled with ubiquitin/QD655 (magenta). A fluorescence image of a Lewy body (a, open triangle) were directly compared with bright-field LM image (b) and EM image (c–e) after gold enhancement. Loose and short ubiquitin-positive fibers of a Lewy body was labeled by rod-like QD655 (e, filled triangle), which showed high contrast nanosphere deposits at the tips of the nanorods. Bars in a and b: 20 μm, c: 5 μm, d: 100 nm, e: 20 nm.
Light microscopic detection with a virtual slide system
The immunolabeled free-floating sections were mounted on glass slides and observed using a virtual slide system VS120 (Olympus, Tokyo, Japan) equipped with UPLSAPO objective lens (10, 20, and 40 times dry objectives, and a 60 times water-immersion objective, Olympus, Tokyo, Japan), a 130 W U-HGLGPS light guide-coupled illumination system (Olympus, Tokyo, Japan), a triple-band dichroic mirror unit U-DM3-DA/FI/TX (Olympus, Tokyo, Japan, 480–500 nm band-pass excitation filter, 505 nm dichroic mirror, 515–540 nm band-pass emission filter) (for Alexa 488), a filter set AT-Qdot565 (Chroma Japan Corp., Yokohama, Japan, 400–450 nm band-pass excitation filter, 485 nm dichroic mirror, 550–580 nm band-pass emission filter) installed in the filter cube U-MF2 (Chroma Japan Corp., Yokohama, Japan) (for QD565), a mirror unit U-DM3-QD655 (Olympus, Tokyo, Japan, 410–460 nm band-pass excitation filter, 505 nm dichroic mirror, 640–670 nm band-pass emission filter) (for QD655), a sensitive cooled charge-coupled device camera (ORCA-R2 C10600-10B, Hamamatsu Photonics, Shizuoka, Japan), and VS-ASW software (Olympus, Tokyo, Japan). Serial snapshots of immunofluorolabeled sections on motorized stage were captured by VS120 through 10 times objective lens in separate fluorescence channels, and put together to make a seamless broad image, covering the whole section. Areas of particular interest were also captured with 60 times water-immersion objective lens with multiple z-stacks at 1 μm intervals, or with 20 times objective lens at 2 μm intervals. The image size of each snapshot was 1376 pixels (horizontal) × 1038 pixels (vertical), at 0.645 μm/pixel with 10 times objective lens, 0.323 μm/pixel with 20 times objective lens, and 0.107 μm/pixel with 60 times objective lens. Cross sectional views were obtained by the ImageJ volume viewer plugin (Barthel, K. U., version 2.01, https://imagej.nih.gov/ij/plugins/volume-viewer.html).
Gold enhancement of QD labels on tissue sections and electron microscopy (EM) preparation
After the fluorescent virtual slide images were taken, the floating sections were gold enhanced and prepared for EM as follows. First, the cover glass of the floating sections was detached in TBS, and the sections were washed with TBS and post-fixed with 1% glutaraldehyde in TBS for 15 minutes at room temperature. After washing in TBS with 50 mM glycine, 1% bovine serum albumin in TBS, and distilled water, floating sections were immersed for 5 min in a 40 μl droplet of GoldEnhance EM plus reagent mixture on Parafilm, prepared according to manufacturer’s protocol. Sufficient bright-field coloration was obtained. On the other hand, gold enhancement was also attempted on another section by immersing the section in a 40 μl droplet of 1% gold (III) chloride solution for 5 min (Fig. 1Ab). However, 1% gold (III) chloride solution did not sufficiently alter the color of the labeling in bright field. Therefore, GoldEnhance EM plus was selected for further observation in tissue sections. After rinsing with distilled water, the sections were mounted with glycerol and examined with a virtual slide system VS120 equipped with a VC50 camera (Olympus, Tokyo, Japan) through a UPLSAPO 10× dry objective. The image size of each snapshot was 1376 pixels (width) × 1038 pixels (height), at 0.691 μm/pixel. The sections were then postfixed in 1% osmium tetroxide in 0.1M PB for 1.5 h at 4 °C, followed by dehydration in graded ethanol series (50%, 70%, 80%, and 90%) each for 5 min, and 100% ethanol for 30 min, respectively, three times. The sections were immersed in 100% propylene oxide and gently shaken for 15 minutes at room temperature. The 100% propylene oxide was then replaced with new 100% propylene oxide, and the same operation was performed a total of three times for 15 minutes each. The sections were infiltrated first with a 1:1 mixture of propylene oxide and epon (Epon 812, Taab, Berks, England) for 1 h, and then with pure epon for 1 h at room temperature. The floating sections were then placed between aclar films (Honeywell, NC, USA) with a sufficient amount of pure epon and heated in an oven at 65 °C for 3 days, allowing the resin to polymerize while the sections are stretched, as described previously20. Using VS120, LM images of the sections infiltrated with epon between the aclar films were taken to obtain a bright-field tiled image around the target. The target was then located under a stereomicroscope and this area was cut out along with the aclar film using scissors or a safety razor blade. The aclar film can be easily peeled off. On a glass slide, a gelatin capsule filled with epon was invertedly placed on top of the resin section and polymerized at 65°C for at least 48 hours. The polymerized capsule block was detached from the glass slide while heated on a hot plate20.
Upon removal, the tissue was attached to the resin cylinder. The final trimming of the resin cylinder around the target was guided by enhanced QD labeling visible under a stereomicroscope. Ultrathin sections of 70 nm were made and placed on formvar coated grids. Ultrathin sections were post-stained with uranyl acetate for 25 min and lead for 7 min and observed by EM.
EM observation and energy-dispersive X-ray (EDX) analysis
In order to confirm the presence of gold on the QD labelling on the sections after the gold enhancement, an energy spectrum of the metals was obtained on the EM sections. Ultrathin sections on EM grids were observed under a TEM(JEM-1400, JEOL, Tokyo, Japan) or a STEM , (Hitachi HD-2700, Hitachi High Technologies Corporation, Tokyo, Japan). Bright-field and high-angle annular dark-field (HAADF) STEM images were obtained using HD-2700, a 200 kV spherical aberration (Cs)-corrected STEM with secondary electron (SE) detector (Hitachi High Technologies Corporation, Tokyo, Japan). The STEM was also equipped with solid angle 100 mm2 silicon drift detector (SDD) Octane T Ultra (EDAX, AMETEK Inc., PA, USA), which is an elemental analysis instrument that uses characteristic X-rays from a specimen and distinguish different elements based on their energy spectra. Bright-field and HADDF-STEM images with 1280 × 960 pixels were acquired with an incident beam of 0.3 nm and a current of 0.5 nA for a frame time of 18 s per image (15 μs/pixel). EDX spot analyses were performed with an incident beam size of 0.3 nm and a current of 0.5 nA. The acquisition time of the EDX single point analysis was up to a few tens of seconds and varied depending on the specimen. In the EDX mapping, the EDX analysis was performed with the image resolution of 128 × 100 pixels, at 30,000 to 1,600,000 times magnification, and the total acquisition time was up to 12 min. Pixels containing peaks of different elements were displayed separately in different pseudo-color channels.
Image analysis
EM observations of gold-enhanced QDs showed that high contrast nanospheres were gradually generated and became larger on the QD particles as electron beam irradiation (EBI) for EM was continued, as described in the later sections. Therefore, the size change of high contrast nanospheres during EM was evaluated by image analysis. ADF-STEM images of a fixed range were obtained 2 min and 7 min after the start of continuous EBI. ImageJ (National Institutes of Health, Maryland, USA) was used to binarize these ADF-STEM images to highlight the high-contrast nanospheres. The images of QD705 and QD655 were binarized with a grayscale value of 175 as the threshold, and the images of QD565 were binarized by IsoData algorithm. The sizes of the high contrast areas as well as their X–Y coordinate in the images were enumerated automatically by particle analysis program of ImageJ. On images 7 min after EBI, we defined particles with more than 3.40 nm2 area as nanospheres. Corresponding QDs were identified in the image after 2 min from the start of EBI by their XY coordinates, and the area of the high contrast portions were obtained. The differences of the nanosphere sizes between 2 and 7 min after EBI were statistically analyzed as below. Next, to highlight both the high contrast nanospheres and surrounding minute particles on the ADF-STEM images, the images of QDs were binarized with the gray scale value 150, and high contrast portions were highlighted with pseudocolored yellow (Supplementary Fig. 1B).
Statistical analysis
All statistical analyses were performed using software R (version 3.6.0, R Foundation for Statistical Computing, Vienna, Austria.). To find the differences in the means of sizes of high-contrast nanospheres between samples, Welch’s two sample t-test was performed (Fig. 1C). P < 0.05 was considered significant.
