A Tissue Clearing Method for Neuronal Imaging from Mesoscopic to Microscopic Scales

Summary

The protocol provides a detailed method of neuronal imaging in brain slice using a tissue clearing method, ScaleSF. The protocol includes brain tissue preparation, tissue clarification, handling of cleared slices and confocal laser scanning microscopy imaging of neuronal structures from mesoscopic to microscopic levels.

Abstract

A detailed protocol is provided here to visualize neuronal structures from mesoscopic to microscopic levels in brain tissues. Neuronal structures ranging from neural circuits to subcellular neuronal structures are visualized in mouse brain slices optically cleared with ScaleSF. This clearing method is a modified version of ScaleS and is a hydrophilic tissue clearing method for tissue slices that achieves potent clearing capability as well as a high-level of preservation of fluorescence signals and structural integrity. A customizable three dimensional (3D)-printed imaging chamber is designed for reliable mounting of cleared brain tissues. Mouse brains injected with an adeno-associated virus vector carrying enhanced green fluorescent protein gene were fixed with 4% paraformaldehyde and cut into slices of 1-mm thickness with a vibrating tissue slicer. The brain slices were cleared by following the clearing protocol, which include sequential incubations in three solutions, namely, ScaleS0 solution, phosphate buffer saline (–), and ScaleS4 solution, for a total of 10.5–14.5 h. The cleared brain slices were mounted on the imaging chamber and embedded in 1.5% agarose gel dissolved in ScaleS4D25(0) solution. The 3D image acquisition of the slices was carried out using a confocal laser scanning microscope equipped with a multi-immersion objective lens of a long working distance. Beginning with mesoscopic neuronal imaging, we succeeded in visualizing fine subcellular neuronal structures, such as dendritic spines and axonal boutons, in the optically cleared brain slices. This protocol would facilitate understanding of neuronal structures from circuit to subcellular component scales.

Introduction

Tissue clearing methods have improved depth-independent imaging of biological and clinical samples with light microscopy, allowing for extraction of structural information on intact tissues^1,2. Optical clearing techniques could also potentially speed up, and reduce the cost for histological analysis. Currently, three major clearing approaches are available: hydrophilic, hydrophobic, and hydrogel-based methods^1,2. Hydrophilic approaches surpass in preserving fluorescence signals and tissue integrity and are less toxic compared to the other two approaches^3,4.

A hydrophilic clearing method, ScaleS, holds a distinctive position with its preservation of structural and molecular integrity as well as potent clearing capability (clearing-preservation spectrum)⁵. In a previous study, we developed a rapid and isometric clearing protocol, ScaleSF, for tissue slices (~1-mm thickness) by modifying the clearing procedure of ScaleS⁶. This clearing protocol requires sequential incubations of brain slices in three solutions for 10.5-14.5 h. The method is featured with a high clearing-preservation spectrum, which is compatible even with electron microscopy (EM) analysis (Supplementary Figure 1), allowing for multi-scale high-resolution three dimensional (3D) imaging with accurate signal reconstruction⁶. Thus, ScaleSF should be effective especially in the brain, where neuronal cells elaborate exuberant processes of tremendous length, and arrange specialized fine subcellular structures for transmitting and receiving information. Extracting structural information with scales from circuit to subcellular levels on neuronal cells is quite useful toward better understanding of brain functions.

Here, we provide a detailed protocol to visualize neuronal structures with scales from the mesoscopic/circuit to microscopic/subcellular level using ScaleSF. The protocol includes tissue preparation, tissue clarification, handling of cleared tissues, and confocal laser scanning microscopy (CLSM) imaging of cleared tissues. Our protocol focuses on interrogating neuronal structures from circuit to subcellular component scales. For a detailed procedure for preparation of the solutions and stereotaxic injection of adeno-associated virus (AAV) vectors into mouse brains, refer to Miyawaki et al. 2016⁷ and Okamoto et al. 2021⁸, respectively.

Protocol

All the experiments were approved by the Institutional Animal Care and Use Committees of Juntendo University (Approval No. 2021245, 2021246) and performed in accordance with Fundamental Guidelines for Proper Conduct of Animal Experiments by the Science Council of Japan (2006). Here, male C57BL/6J mice injected with AAV vector carrying enhanced green fluorescent protein (EGFP) gene and parvalbumin (PV)/myristoylation-EGFP-low-density lipoprotein receptor C-terminal bacterial artificial chromosome (BAC) transgenic mice (PV-FGL mice)⁹ were used. PV-FGL mice were maintained in C57BL/6J background. No sex-based differences were found with regard to this study.

1. Tissue preparation

1. Perfusion fixation
   NOTE: Perform steps 1.1.1 through 1.1.3 in a fume hood to limit the exposure to paraformaldehyde (PFA).
   1. Anesthetize adult male mice (8–16 weeks old) by an intraperitoneal injection of overdose of sodium pentobarbital (200 mg/kg). Confirm adequacy of anesthesia by the absence of toe-pinch withdrawal and eye-blink reflexes.
   2. Open the thoracic cavity and cut the right atrial appendage with surgical scissors. Perfuse the mice with 20 mL of ice-cold phosphate buffer saline (PBS) using a 23 G needle attached to a 20 mL syringe, followed by perfusion of 20 mL of ice-cold 4% PFA in 0.1 M phosphate buffer (PB) using another 20 mL syringe.
      CAUTION: PFA is toxic and teratogenic. Avoid inhalation or contact with skin, eyes, and mucous membrane.
   3. Remove brain tissues from the skull with tweezers. Transfer the brain tissues to a 15 mL tube containing 4% PFA in 0.1 M PB, protect the samples from light, and gently rock overnight at 4 °C on a shaker at 50–100 rpm.
   4. NOTE: The harvested brain tissues can be stored for several weeks in 0.02% sodium azide (NaN₃) in PBS at 4 °C.
      CAUTION: NaN₃ is toxic. Avoid inhalation or contact with skin, eyes, and mucous membrane. Handle it inside a fume hood.
2. Brain slice preparation
   1. Prepare 4% agar in PBS by adding 2 g of agar to 50 mL of PBS. Microwave the mixture until the agar is fully dissolved. Let the solution cool to 40-45 °C.
      NOTE: The viscous property provided by agaropectin, a major component of agar, improves ease of cutting of tissue slices.
   2. Add 10 mL of the agar solution to a 6-well culture plate. Submerge the brain tissue in the agar solution using forceps. Let the agar solidify on ice.
   3. Remove the embedded brain tissue from the well and trim the agar with a razor blade. Secure the agar block onto the bottom of the vibratome bath with superglue and pour 0.1 M PB in the buffer tray.
   4. Clean another razor blade using a lint-free tissue paper soaked in ethanol and attach the blade to the blade holder of the vibrating tissue slicer.
   5. Set the sectioning speed to 0.14 mm/s with 1.4 mm amplitude and the frequency to 75–77 Hz. Cut the brain tissue into 1-mm-thick slices and collect the slices in a 6-well cell culture plate containing PBS.
      NOTE: The brain slices can be stored for several weeks in 0.02% NaN₃ in PBS at 4 °C.

2. Tissue clarification

NOTE: The compositions of ScaleS solutions used are listed in Table 1. Samples should be protected from light by covering with a foil. The clearing steps is shown in Figure 1A.

1. Add 8 mL of ScaleS0 solution to one well of a 6-well cell culture plate and add 8 mL of ScaleS4 solution to another well of the plate and pre-warm to 37 °C in an incubator.
2. Transfer the brain slices to the pre-warmed ScaleS0 solution with a spatula and incubate for 2 h at 37 °C in a shaking incubator at 90 rpm.
3. Transfer the permeabilized brain slices in 8 mL of PBS(–) in a 6-well cell culture plate with a spatula and wash for 15 min by keeping in an orbital shaker at 40–60 rpm. Repeat this twice.
4. Transfer the brain slices in the pre-warmed 8 mL of ScaleS4 solution with a spatula and clear them by incubating in a shaking incubator at 90 rpm for 8–12 h at 37 °C. A cleared brain slices can be seen in Figure 1.

3. Brain slice mounting

NOTE: A customizable imaging chamber is used for reliable mounting of cleared brain slices (Figure 2)⁶. The chamber consists of the chamber frame and bottom coverslip. The microscope stage adaptors are also designed to mount the imaging chamber on microscope stages directly (Figure 2A,B). The chamber frame and microscope stage adaptors can be 3D-printed using in-house or outsourced 3D-printing services. 3D computer-aided design (CAD) data of the imaging chamber are provided in Furuta et al. 2022⁶.

1. For chamber preparation, attach the chamber frame to a coverslip using a pressure-sensitive adhesive.
2. Prepare 1.5% agarose in ScaleS4D25(0) solution (ScaleS4 gel) by adding 1.5 g of agarose to 100 mL of the solution in a bottle. Mix the solution well by stirring and microwave the solution until the agarose is fully dissolved. Once done, allow the solution to cool to 37 °C.
3. Mount the cleared brain slice onto the bottom coverslip of the imaging chamber with a spatula. Wipe away the excess solution from the cleared slice using a clean lint-free tissue paper.
4. Add the ScaleS4 gel on the brain slice using a micropipette to fill the imaging chamber. Place another coverslip on top with forceps and place a piece of lint-free tissue paper and a glass slide on the coverslip in this order.
5. Transfer the imaging chamber to a refrigerator at 4 °C. Place metal weights on the glass slide and leave them for 30 min.
6. Remove the metal weights, glass slide, lint-free tissue paper, and coverslip from the imaging chamber, and wipe the excess gel away (Figure 2A,B).
7. Place the imaging chamber in a 60 mm glass Petri dish and attach the rim of the imaging chamber to the dish with a putty-like pressure sensitive adhesive. Attach the chamber at multiple points to the Petri dish.
8. Pour ScaleS4 solution in the dish and shake gently for 1 h at 20-25 °C on an orbital shaker at 40-60 rpm. Substitute with fresh solution and remove air bubbles on the gel surface by gently scraping the surface using a 200 µL pipette tip. Mount the immersed imaging chamber on a microscope stage (Figure 2C).

4. CLSM imaging

1. Acquire images using a CLSM equipped with a multi-immersion objective lens of a long working distance (WD) (16x/0.60 numerical aperture [NA], WD = 2.5 mm).
   NOTE: High NA objective lenses can provide high diffraction-limited resolution.
2. Turn on all the relevant imaging equipment (workstation, microscope, scanner, lasers, and mercury lamp) and launch a CLSM imaging software.
3. Set the correction collar of the multi-immersion objective lens to 1.47. ScaleS4 solution has a refractive index (RI) of around 1.47^5,7. RI mismatch-induced aberrations can disturb the image formation (Figure 3).
4. Immerse the objective lens in the solution, and let it approach the slice slowly. Remove any air bubbles trapped on the tip of the objective lens. Find regions of interest (ROIs) in the cleared tissues using epifluorescence.
5. Set image acquisition parameters by testing appropriate settings.
   1. Determine the bit depth for image acquisition. The data size of the image increases with the bit depth.
   2. Set the detection wavelength. Adjust the appropriate gate for the detector according to the emission spectrum. Ensure that the detection wavelength does not cover any laser lines.
   3. Set the xy resolution. Larger formats provide better xy resolutions, but it takes longer time to collect the images.
   4. Set the scan speed. A slower scan speed provides a high signal-to-noise ratio. However, it also increases the pixel dwell time and the risk of photobleaching. Choose accordingly.
   5. Adjust the pinhole size. The pinhole size controls optical section thickness. A smaller pinhole size creates a thinner optical section, and thus better z resolution, but reduces fluorescence signal. Making the pinhole size larger provides a thicker optical section with stronger fluorescence signal.
   6. Set the laser power, the detector/amplifier gain, and offset. Gradually increase the laser power and detector/amplifier gain until a suitable image is obtained. A high laser power carries the risk of photobleaching. Adjust the offset (contrast) appropriately to obtain a high signal-to-noise ratio.
   7. Determine the tilling area needed based on the size of ROI. Ensure that the entire length and width of the ROI is captured.
   8. Navigate the cleared tissues in all planes, and set the start and end points of the stack. Set the z-step size according to the desired z-resolution.
6. Collect images when satisfied with the image acquisition settings, and record captured images. Process the images using an image analysis software.

Representative Results

Optical clearing of a mouse brain slice of 1-mm thickness was achieved using this protocol. Figure 1B represents transmission images of a mouse brain slice before and after the clearing treatment. The tissue clearing method rendered a 1-mm-thick mouse brain slice transparent. A slight expansion in final sizes of brain slices was found after the incubation in the clearing solution for 12 h (linear expansion: 102.5% ± 1.3%). The preservation of fluorescence and structural integrity of the tissues was assessed with targeted EGFP expression in the plasma membrane in PV-FGL mice (Figure 1C). In these mice, somatodendritic membrane-targeted EGFP is expressed in PV-positive neurons⁹. EGFP expression targeted to the plasma membrane in the somatodendritic region was maintained after the treatment (Figure 1C). Additionally, the previous EM study shows well-preserved structural integrity in brain tissues cleared with ScaleSF (Supplementary Figure 1)⁶.

RI mismatch-induced aberrations caused a noticeable loss of image brightness and resolution (Figure 3). A 1-mm-thick brain slice of PV-FGL mouse was cleared and imaged under a CLSM equipped with a multi-immersion objective lens of a long WD. Adjustment of the correction collar of the objective lens to the water position (RI 1.33) hampered clear visualization of EGFP-positive neurons located at the depths of 400 µm and 800 µm due to low brightness and low contrast. (Figure 3A,C). These neurons were clearly visualized with the same CLSM, when the correction collar was adjusted to match the ScaleS4 solution (RI 1.47; Figure 3B,D). RI-matching between an immersion fluid and objective lens is critical for accurate 3D imaging in optically cleared tissues.

Lastly, mouse neocortical neurons were utilized to demonstrate the feasibility of the protocol. A mouse brain injected with AAV2/1-SynTetOff-EGFP vector¹⁰ in the primary somatosensory cortex (S1) was fixed with 4% PFA in 0.1 M PB. Coronal slices of 1-mm thickness were prepared from the brain with a vibrating tissue slicer. After clearing and mounting on the imaging chamber, neuronal imaging targeted to neocortical neurons was conducted (Figure 4). A 3D reconstruction of EGFP-labeled neurons in the 1-mm-thick brain slice is represented in Figure 4A. A higher magnification image shows individual dendritic arbors decorated with dendritic spines (Figure 4B). We further showed axon terminal arborizations and axonal boutons in the contralateral cortex (Figure 4C).

Figure 1: Optical clearing of mouse brain slices of 1-mm thickness. (A) The schedule for ScaleSF tissue clearing. (B) Transmission images of a 1-mm-thick brain slices before (left) and after (right) treatment. (C) A 3D volume rendering of the cerebral cortex of a PV-FGL mouse cleared using the tissue clearing method. (D,E) xy images in (C) at the depths of 250 µm (D) and 750 µm (E). (F,G) Enlarged view of the rectangles outlined in (D) and (E). Images appearing in (C-G) are deconvoluted before the rendering process. Abbreviations: pia = pia mater, WM = white matter. Scale bar: 2 mm in (B), 500 µm in (C), 200 µm in (D,E), and 40 µm in (F,G). This figure has been modified from Furata et al. 2022⁶. Please click here to view a larger version of this figure.

Figure 2: A customizable 3D-printed imaging chamber for tissue slice visualization. (A,B) A schema drawing (A) and picture (B) of a customizable 3D-printed imaging chamber. The imaging chamber is composed of a chamber frame, a bottom coverslip and microscope stage adaptors. Cleared tissue slices are placed onto the bottom coverslip and embedded in ScaleS4 gel. The chamber frame, bottom coverslip and microscope stage adaptors are customizable according to the size and thickness of tissue slices. (C) An imaging setup with the imaging chamber. The imaging chamber is immersed in ScaleS4 solution in a Petri dish and mounted on a stage of an upright CLSM. This figure has been modified from Furata et al. 2022⁶. Please click here to view a larger version of this figure.

Figure 3: Compromised deep imaging caused by an RI mismatch between an objective lens and ScaleS4 solution. (A-D) xy images of the cerebral cortex of a PV-FGL mouse at depths of 400 µm (A,B) and 800 µm (C,D). The correction of collar of a multi-immersion objective lens is adjusted to 1.33 in (A,C) and 1.47 in (B,D). Images are acquired with the same parameters except for RIs of the objective lens. Scale bar: 50 µm. Please click here to view a larger version of this figure.

Figure 4: Neuronal imaging in a 1-mm-thick mouse brain slice cleared with ScaleSF. (A) 3D volume rendering of EGFP-labeled mouse neocortical neurons in the S1. Neocortical neurons are labeled with the AAV2/1 SynTetOff-EGFP vector. (B) Dendritic arbors of EGFP-labeled neocortical neurons. A maximum intensity projection (MIP) image from a depth of 39 µm to 48 µm is represented. Arrowheads indicate dendritic spines. (C) EGFP-labeled axon terminals in the contralateral cortex. A MIP image from a depth of 481.5 µm to 513 µm is represented. Arrowheads indicate axonal boutons. Images appearing in (B) and (C) are deconvoluted. Scale bars: 300 µm in (A) and 10 µm in (C). The bar in (C) applies to (B) as well. Please click here to view a larger version of this figure.

Supplementary Figure 1: Ultrastructure in brain slices cleared with ScaleSF, CUBIC, and PACT. (A-C) Transmission EM images of mouse cerebral cortex cleared with ScaleSF (A), CUBIC (B), and PACT (C). Mouse brains are fixed with 4% PFA containing 1% glutaraldehyde. Ultrathin sections are prepared from cleared brain slices. Membrane structures are severely damaged in brain slices cleared with CUBIC (B) and PACT (C). Arrowheads indicate postsynaptic membranes. Scale bar: 500 nm. This figure has been modified from Furata et al. 2022⁶. Please click here to download this File.

Recipies for ScaleS solutions
ScaleS0 solution
Reagent	Final concentration
D-sorbitol	20% (w/v)
Glycerol	5% (w/v)
Methyl-β-cyclodextrin	1 mM
γ-cyclodextrin	1 mM
Dimethyl sulfoxide	3% (v/v)
10x PBS(–)	1x
ScaleS4 solution
Reagent	Final concentration
Urea	4 M
D-sorbitol	40% (w/v)
Glycerol	10% (w/v)
Triton X-100	0.2% (w/v)
Dimethyl sulfoxide	25% (v/v)
ScaleS4D25(0) solution
Reagent	Final concentration
Urea	4 M
D-sorbitol	40% (w/v)
Glycerol	10% (w/v)
Dimethyl sulfoxide	25% (v/v)

Table 1: Composition of the three ScaleS solutions. The compositions of ScaleS0, ScaleS4, and ScaleS4D25(0) solutions are listed. For a detailed procedure for preparation of these solutions, refer to Miyawaki et al. 2016⁷.

Discussion

Critical steps within the protocol
There are a few critical steps in the protocol that should be conducted with utmost caution to obtain meaningful results. Uniform fixation of samples is imperative for 3D imaging within large-scale tissues. The objective lens, sample, and immersion fluid should have matching RI. RI-mismatch among them will lead to highly disturbed imaging of EGFP-expressing cells within the cleared brain slices (Figure 3). The correction collar adjustment of the objective lens to the immersion fluid minimizes depth-induced spherical aberrations to maximize the signal, contrast and spatial resolution in 3D imaging. The RIs of the prepared solutions can be measured using a refractometer.

Troubleshooting of the technique
Longer storage of the solutions can affect clearing capability, and its capacity for preservation of fluorescence signals and structural integrity. Freshly prepared solutions should be used. These solutions can be stored up to 1 month at 4 °C. Isometricity is critical for effective and efficient neuronal imaging with accurate signal reconstruction. Although a slight expansion in sample sizes was observed after incubation for 12 h (Figure 1B), the expansion can be controlled by decreasing the incubation period between 8-12 h. For more accurate 3D depth imaging, the correction collar might need to be adjusted at a given plane due to depth-induced spherical aberration.

Modifications of the technique
In the present study, we utilized mouse brain tissues to demonstrate the feasibility of the protocol. Yet the protocol described here can be also used for large-brained animals, such as primates. Indeed, this protocol has been used in common marmoset (Callithrix jacchus) brain tissues, and succeeded in simultaneous visualization of neural circuit and subcellular structures of its corticostriatal circuits⁶. The microscope stage adaptors are designed to mount the imaging chamber on microscope stages directly⁶ (Figure 2A,B). Cleared tissue slices are observable using an inverted microscope through the bottom coverslip of the imaging chamber. Following restoration of cleared brain tissues with PBS(-) (deScaling)^5,11, we can prepare tissue sections of 20-µm to 50-µm thickness from brain tissues that were cleared with ScaleSF (re-sectioning)⁶. Subcellular structures captured within cleared tissues can be imaged again on re-sections with a high NA objective lens of a short WD. Clearing of brain slices perfused with fixatives containing glutaraldehyde has been achieved using this protocol, providing superior ultrastructure preservation⁶. ScaleSF achieves a high-level of ultrastructure preservation that allows for EM analysis in optically cleared tissues⁶ (Supplementary Figure 1). The EM compatibility of this method is particularly useful for imaging structures with the scales from macroscopic to nanoscopic level.

Limitations of the technique
The protocol described here allows us to visualize neuronal structures from circuit to subcellular scales in brain slices of 1-mm thickness. However, three limitations remain in the protocol. The first is the clearing capability of the clearing protocol. ScaleSF is a clearing protocol for brain slices, not for the whole brain. Although brain slices of 1-mm thickness can provide good knowledge of dendritic and local axonal arbors¹², information about axonal projections spanning the entire brain is fragmentary and incomplete in the slices^13,14,15,16. The second is the imaging resolution. Using the protocol described here, we succeeded in visualizing subcellular neuronal structures, such as dendritic spines and axonal boutons, in an optically cleared brain slice (Figure 4). However, the resolution of the objective lens used in this study, xy resolution of 400-750 nm, is not sufficient to resolve more fine structures of neuronal cells. Given high NA objective lenses are typically designed for oil-immersion (RI 1.52), RI-mismatch with the solutions (RI 1.47) might prevent high-resolution imaging with these objective lenses. The third is fluorescent protein labeling of neuronal cells. The labeling method limits broad applications of our imaging technique. Histochemical and/or immunohistochemical techniques that label large-scale tissues while maintaining tissue integrity would significantly advance the protocol provided here.

Significance with respect to existing methods and future applications of the technique
In the present study, we describe a detail protocol for neuronal imaging from mesoscopic to microscopic structures using ScaleSF tissue clearing. The protocol described here makes it possible to visualize neuronal structures from circuit to subcellular levels in a reasonable amount of time without specialized equipment, facilitating understanding of neuronal structures from circuit to component scales. Neurons elaborate exuberant processes of tremendous length and arrange specialized fine structures for transmitting and receiving information. Thus, neuronal imaging requires a tissue clearing method that exerts potent clearing capability as well as a high-level of tissue preservation for simultaneous visualization of both large and small-scale structures. However, tissue clearing methods featured with high clearing capabilities aggressively remove lipids and pigments for extensive tissue clarification^3,4, compromising tissue integrity^5,6,17 (Supplementary Figure 1). This is in stark contrast to the clearing protocol used here that achieves a high-level of structure preservation⁶ (Supplementary Figure 1). Hence, ScaleSF tissue clearing allows for effective and efficient neuronal imaging that requires multi-scale high-resolution 3D imaging with accurate signal reconstruction.

Materials

Name	Company	Catalog Number	Comments
16x multi-immersion objective lens	Leica Microsystems	HC FLUOTAR 16x/0.60 IMM CORR VISIR
Agar	Nacalai Tesque	01028-85
Agarose	TaKaRa Bio	L03
Dimethyl sulfoxide	Nacalai Tesque	13407-45
D-Sorbitol	Nacalai Tesque	06286-55
γ-cyclodextrin	Wako Pure Chemical Industries	037-10643
Glycerol	Sigma-Aldrich	G9012
Huygens Essential	Scientific Volume Imaging	ver. 18.10.0p8/21.10.1p0 64b
Imaris	Bitplane	ver. 9.0.0
Leica Application Suite X	Leica Microsystems	LAS X, ver. 3.5.5.19976
Methyl-β-cyclodextrin	Tokyo Chemical Industry	M1356
Paraformaldehyde	Merck Millipore	1.04005.1000
Phosphate Buffered Saline (10x; pH 7.4)	Nacalai Tesque	27575-31	10x PBS(–)
Sodium azide	Nacalai Tesque	31233-55
Sodium pentobarbital	Kyoritsu Seiyaku	N/A
TCS SP8	Leica Microsystems	N/A
Triton X-100	Nacalai Tesque	35501-15
Urea	Nacalai Tesque	35940-65
Vibrating tissue slicer	Dosaka EM	PRO7N