Box 1. Tissue preparation, microscopy methods and imaging techniques

The following sections describe methods used to label and image the spiny dendritic branches and amyloid beta plaques shown in Figures 4 and 5 of this protocol, and to reconstruct the apical and basal dendritic trees of pyramidal neurons from multiple-tiled image stacks using confocal and multiphoton laser scanning microscopy (CLSM and MPLSM), also described by Rodriguez et al.³⁰ and Wearne et al.¹⁹.

Tissue preparation and microscopy methods

Animals: The brains of long-tailed macaque monkeys (Macaca fascicularis), Tg2576 mice older than 12 months, and wild-type littermates, were used for cell loading and 3D reconstruction of spiny dendritic branches and entire pyramidal neurons shown in this protocol and in our previous studies that have been used for 3D neuronal reconstructions^{19, 30, 35, 36}. All animal protocols were approved by the Animal Care and Use Committee at Mount Sinai School of Medicine. For reconstruction of amyloid plaques from human tissue, blocks from the prefrontal cortex (Brodmann's area 9) and from the hippocampus were obtained at autopsy, after verification of consent, from patients with Alzheimer disease (78-85 years old, post-mortem interval 3-8 h) through the Mount Sinai School of Medicine Alzheimer's Disease Research Center Brain Bank. One hemisphere was fixed in 4% (wt/vol) paraformaldehyde for 3-4 weeks and subsequently sampled for quantitative histopathology³⁷.

Tissue preparation. For imaging and reconstruction of single neurons, spiny dendrites and multineuron networks, mice were perfused transcardially with a mixture of 4% paraformaldehyde and 0.125% glutaraldehyde in phosphate buffer (phosphate-buffered saline (PBS); pH 7.4) for 10 min. Monkeys were perfused intracardially as described previously^{36, 38, 39}. In brief, the monkeys were deeply anesthetized with ketamine hydrochloride (25 mg kg^-1) and pentobarbital sodium (20–35 mg kg^-1, i.v.), intubated and mechanically ventilated. The chest was opened to expose the heart, and 1.5 ml of 0.1% sodium nitrite was injected into the left ventricle. The descending aorta was clamped and the monkeys were perfused transcardially with cold 1% paraformaldehyde in PBS for 1 min and then for 12 min with cold 4% paraformaldehyde and 0.125% glutaraldehyde in PBS. Brains were carefully removed from the skull, hemisected under a dissecting microscope and sectioned on a vibratome (Leica) at 200 m for mice¹⁹ and at 400 m for monkeys^{35, 36}. For imaging of amyloid beta plaques in transgenic mice, senile plaques were first visualized in thick sections with 0.01% thioflavine S, and sections were coverslipped with SlowFade (Molecular Probes). This permits visualization of the loaded neurons in the context of amyloid plaque distribution in Tg2576 mice and reveals the possible effects of amyloid on the integrity of dendrites and spines. Materials from human brains were cut at 300 m on a vibratome and the sections were stained with thioflavine S for senile plaque visualization as above. For monkeys, after hemisection, 4–5 mm thick blocks of tissue from prefrontal cortex area 46 were taken perpendicular to the axis of the principal sulcus and processed as described above and by Duan et al.^{35, 36}.

Autofluorescence quenching. The autofluorescent component of an image due to lipofuscin may be removed by spectral imaging and linear unmixing, as described in "Imaging techniques," below. Moreover, the autofluorescence of human sections due to lipofuscin can be quenched by treatment of the mounted sections in 1% Sudan Black B in 70% alcohol, rehydrated and coverslipped⁴⁰. This provides an efficient removal of bleed-through autofluorescence in case high levels of lipofuscin are present in human materials and in the oldest mice.

Cell loading in fixed slices with Lucifer Yellow. For intracellular injection, thick sections were incubated with 4,6-diamidino-2-phenylindole (DAPI; Sigma), a fluorescent nucleic acid stain, for at least 5 min, to reveal the cytoarchitecture under UV excitation. Sections were mounted on nitrocellulose filter paper, immersed in phosphate-buffered saline (pH 7.4) and visualized under epifluorescence. Using DAPI staining as a guide, neurons are impaled with sharp micropipettes and loaded with 5% Lucifer Yellow (Molecular Probes) in distilled H₂O under a direct current of 3–8 nA for 10–12 min or until the dye fills distal processes and no further loading is observed^{19, 35, 36}. Alternatively, to prevent artifacts due to artificial sectioning of dendrites at either surface of the tissue slab, which occurs in most cases for average-sized pyramidal neurons loaded within the first 100 m of the surface, cells can be loaded blindly deep in the tissue, aiming with the micromanipulator at depths of about 200 m. This permits a fuller loading of all dendritic branches. After neuronal labeling, sections were fixed again in 4% paraformaldehyde and 0.125% glutaraldehyde in PBS for 4 h at 4 °C, washed and stored in PBS. Sections were then mounted on uncoated slides, coverslipped in PermaFluor and loaded neurons were visualized by CLSM. Neurons loaded deeply in the tissue must be imaged using a long working distance lens and MPLSM, using techniques described by Rodriguez et al.³⁰

Imaging techniques

Laser scanning microscopy (confocal and multiphoton). Tissue sections were examined using a BioRad Radiance 2000 multiphoton laser-scanning microscopy (MPLSM) system (Hercules) equipped with a Coherent Mira 900F Ti:sapphire laser and a Kr/Ar (488, 568 nm) laser. Tissue sections were also examined using a Zeiss LSM 510 META confocal laser scanning microscope (Jena) equipped with Ar (488 nm) and HeNe (543, 633 nm) lasers and/or a Leica TCS-SP (UV) confocal laser scanning microscope, equipped with Ar (488 nm), Kr (568 nm) and HeNe (633 nm) lasers. For MPLSM, fluorescence from labeled neurons was imaged by using the Ti:sapphire laser (tuned to 860 nm for Lucifer Yellow) and a UplanApo/IR 60X (1.2 n.a., 0.25 mm WD) water immersion objective lens. Serial optical sections (1,024 1,024 pixels; pixel dimensions = 0.2 0.2 m) were collected through focus at 0.2 m intervals. For higher resolution imaging, a UPlanApo 100/1.35 n.a. objective (WD 0.1 mm) was used on the BioRad system in the multi-photon mode (i.e., with the Ti:sapphire laser) and/or the single-photon (confocal) mode (with the 488 nm line of the Kr/Ar laser). Single photon (confocal) imaging was also made on a Leica TCS-SP (UV) CLSM using a 100, 1.4 n.a. PlanApo lens and the 488 nm line of an Ar laser or on a Zeiss LSM 510 META confocal laser scanning using a 63, 1.4 n.a. PlanApo lens and the 488 nm line of an Ar laser. Serial optical sections (1,024 1,024 pixels with pixel dimensions = 0.18 0.18 m (BioRad) or 0.1 0.1 m (Zeiss or Leica)) were collected through focus at 0.1 m intervals and correction for signal attenuation will be performed as described above. Collected data are saved to disk as BioRad pic files or stacks of TIFF images (Leica and Zeiss) and processed independently of each other before volume integration and reconstruction with the VIAS-NeuronStudio system³⁰.

Spectral imaging and linear unmixing. For the amyloid plaque reconstructed in Figure 5c,d of this protocol, thioflavine S-labeled slides were imaged on a Zeiss LSM 510 META confocal laser scanning microscope equipped with a 63, 1.4 n.a. oil immersion PlanApo objective lens. The META detector was used to generate spectral signatures (over a 240 nm range) for thioflavine S-labeled plaques and for lipofuscin autofluorescence in unlabeled sections. Lambda scanning was performed for thioflavine S-labeled sections and advanced linear unmixing (using the LSM 510 software) was used to separate the spectrally overlapping fluorescence of the thioflavine S-labeled senile plaques and the lipofuscin autofluorescence. The lipofuscin component of the image was removed, leaving the thioflavine S components for 3D reconstruction and analysis using the Rayburst sampling algorithm (see Wearne et al.¹⁹, Fig. 12 and p. 676, for further details).