Reagents
All reagents used in this study were listed in Supplementary Table 1.
Animal tissue collection and fixation
One porcine kidney was purchased from a local farmer's market, and immediately immersed into the fixation solution containing 4% (w/v) paraformaldehyde (PFA) and 20% (w/v) acrylamide (AA) at room temperature overnight. The renal cortex part of the tissue was cut into 3 mm sizes and stored in this fixation solution at 4°C indefinitely. The tissue piece to be imaged by transmission electron microscopy (TEM) was further fixed in 2.5% (w/v) glutaraldehyde at room temperature overnight.
The rat tissues, including heart, liver, intestine, and kidney were harvested from 8-week-old SD rats. The tissues were cut into 3 mm sizes and were either fixed in 4% paraformaldehyde at room temperature overnight for expansion microscopy study, or fixed in 2.5% (w/v) glutaraldehyde at room temperature overnight for TEM study. The mouse lung tissue was harvested from 4-week old C57BL/6 mice. The tissues were cut into 3 mm sizes and fixed in 4% paraformaldehyde at room temperature overnight. The animal study was approved by The Laboratory Animal Ethics Committee of Experiment Animal Center of the Soochow University.
ZOOM protocol
Following the protocol described in 6, the fixed porcine kidney tissue block was incubated in the ZOOM monomer solution containing 30% (w/v) AA, 0.01% (w/v) N,N'-methylenebisacrylamide (BIS), 4% (w/v) PFA, and 0.1% VA-044 (w/v) in phosphate buffered saline (PBS) at 4 ºC shaker for 16 hours. Then gelation proceeded for 4 hours at 50°C. The sample was then incubated in the denaturation buffer (200 mM SDS, 200 mM NaCl, and 50 mM Tris-HCl, pH 9) at 95°C for 30 min, followed by incubation at 80°C for 24 hours. The sample was washed in PBS and 0.1% (w/v) Triton X-100 (PBST) at room temperatures for 4 hours. Next, the sample was cryo-sectioned and stained as described below, followed by complete expansion in deionized water for confocal imaging.
MAP protocol
Following the protocol described in 3, the fixed porcine kidney tissue block was incubated in the MAP monomer solution containing 10% (w/v) SA, 30% (w/v) AA, 0.1% (w/v) BIS, and 0.1% VA-044 (w/v) in PBS at 4 ºC shaker for 16 hours. Then gelation proceeded for 4 hours at 50°C. The sample was then incubated in the denaturation buffer (200 mM SDS, 200 mM NaCl, and 50 mM Tris-HCl, pH 9) at 70°C for 24 hours and then 95°C for 24 hours. After denaturation, the sample was washed in PBST at room temperatures for 4 hours. Next, the sample was cryo-sectioned and stained as described below, followed by complete expansion in deionized water for confocal imaging.
Pan-ExM protocol
For the pan-ExM experiment, the protocol by M'Saad and Bewersdorf 9 was followed. Briefly, the fixed porcine kidney tissue block was incubated in post-fix solution (0.7% (w/v) PFA + 1% (w/v) AA in PBS) for 7 hours at 37 ºC. The sample was then incubated in the first expanding gel solution containing 19% (w/v) SA, 10% AA (w/v), 0.1% (w/v) N,N’-(1,2-dihydroxyethylene)bisacrylamide (DHEBA) + 0.1% (w/v) VA-044 in PBS for 16 hour at 4°C followed by gelation at 55°C for 3 hours. After heat denaturation in a solution of 200 mM NaCl, 200 mM SDS, 50 mM Tris, pH 6.8 at 73°C for 1 hour, the gel samples were incubated in deionized water for 4 hours. Next, the expanded gel was incubated in the second re-embedding neutral gel solution (10% (w/v) AA, 0.05% (w/v) DHEBA, 0.1% (w/v) VA-044 in PBS) at 4°C shaker for 16 hours and gelled at 55°C for 3 hours. After being washed in PBS, the sample was incubated in post-fix solution (0.7% PFA + 1% AA in PBS) for 8 hours at 37°C, followed by washing in PBS. Next, the sample was incubated in the third expanding gel (19% (w/v) SA, 10% AA (w/v), 0.1% (w/v) BIS, and 0.1% (w/v) VA-044 in PBS at 4°C shaker for 16 hours, and gelled at 55°C for 3 hours. The gel was then incubated in 0.2 M NaOH for 5 hours in order to dissolve DHEBA, followed by washing in PBS for 2 hours. Next, the sample was cryo-sectioned and stained as described below, followed by complete expansion in deionized water for confocal imaging.
Cyc-ExM protocol
Fixed tissues were incubated at 30 ºC in gel solution (20% (w/v) acrylamide, 3% (w/v) paraformaldehyde, 0.1% (w/v) V50, pH 7) for several hours, depending on the size of the sample, with 1 hour for a 3 mm sized sample. Polymerization of the hydrogel was then carried out at 80°C for 2 hours. The sample was then transferred to the denaturation solution (200 mM SDS, 200 mM NaCl, 50 mM Tris-HCl, pH 9.0) at 80°C for 2 hours. This completes one cycle of expansion. To continue with more cycles, simply repeat the above incubation, polymerization, and denaturation step, except that the incubation time and denaturation time needs to be increased if the sample size was increased after each cycle of expansion (Fig. 1a), in order to allow sufficient diffusion and denaturation within the sample. Once the desired number of cycles are completed, the sample was cryo-sectioned and stained as described below, followed by incubation in 60% (w/v) sucrose solution for confocal imaging.
Cryo-sectioning
600 µm thick sections of the samples (either unexpanded or expanded according to various protocols) were obtained by cryo-sectioning. Briefly, the tissue was embedded in optimal cutting temperature (OCT) compound, frozen at -28 ºC and sectioned at the same temperature in a cryostat microtome (Hisure HS3090A). The sections were then collected in PBS.
NHS ester staining
For unexpanded samples, the tissue section was incubated 1 mM 5-Carboxy-tetramethylrhodamine N-succinimidyl ester (TAMRA NHS ester) in PBS for 1 hour in a room temperature shaker. The sample was subsequently washed three times in PBS for 30 min each.
For expanded samples, the tissue section was incubated 1 mM TAMRA NHS ester in 0.1 M Na2CO3, 2% (w/v) SDS, 0.2% (w/v) Triton X-100, pH 11.6 for 1 hour in a room temperature shaker. The sample was subsequently washed three times in the same buffer without dyes for 10 min each. For comparison, several expanded samples were stained in 1 mM TAMRA NHS ester in PBS (pH 7.2), which resulted in much higher fluorescence background.
To amplify the NHS ester signal in expanded samples (especially for samples expanded for more than 7 cycles), biotin-streptavidin based signal amplification methods were used. For 8-cycle expanded porcine renal tissue, the tissue section was stained in 1 mM biotin NHS ester in 0.1 M Na2CO3, 2% (w/v) SDS, 0.2% (w/v) Triton X-100, pH 11.6 for 1 hour in a room temperature shaker. After washing in the same buffer for 1 hour at room temperature, the tissue was stained by 1 mg/mL streptavidin conjugated to TAMRA in PBS for 1 hour in a room temperature shaker, followed by washing in PBST. A slightly different signal amplification strategy was used for imaging of Cyc-ExM expanded rat heart, liver, and kidney tissues. the tissue blocks after the 1st cycle of expansion were stained with 1 mM biotin NHS ester in PBST for 2 hours in a room temperature shaker, followed by washing in PBST and subsequent cycles of gelation and expansion. After the 10th cycle of expansion, the tissue was stained by 1 mg/mL streptavidin conjugated to TAMRA in PBS for 1 hour in a room temperature shaker, followed by washing in PBST.
Maleimide staining
The sections of Cyc-ExM expanded tissue was incubated in 50 mM tris(2-carboxyethyl)phosphine (TCEP) in PBST for 30 min, followed by incubation in 0.2 mM CF568 maleimide in PBST for 1 hr. The sample was then washed in PBST for 1 hour.
Nuclear staining
For nuclear staining, the unexpanded samples were incubated in 10 µg/mL 4',6-diamidino-2-phenylindole (DAPI) in PBS for 30 min at room temperature. The expanded samples were incubated in 10 µM SYTOX Green in PBS for 30 min at room temperature.
Antibody labeling
To immunolabel proteins in the tissue before expansion, the fixed porcine kidney tissue block was further cut into 1 mm3 sizes. The tissue sample was incubated in primary antibodies diluted in PBST for 16 hours at a room temperature shaker, followed by washing in PBST for 8 hours. Then the sample was incubated in secondary antibodies diluted in PBST for 16 hours, followed by washing in PBST for 8 hours. The primary antibodies used were rabbit anti-podocin (P0372, Sigma-Aldrich; 1:20 dilution), and rabbit anti-actinin-4 (HPA001873, Sigma-Aldrich; 1:20 dilution), The secondary antibody was goat anti-rabbit Fab fragment conjugated to biotin (111-007-003, Jackson ImmunoResearch; 1:40 dilution). The sample was then processed by Cyc-ExM as described above. To fluorescently label the bound antibody after expansion, the expanded tissue was sectioned, first stained by 1 mM NHS-Fluorescein in 0.1 M Na2CO3, 2% (w/v) SDS, 0.2% (w/v) Triton X-100, pH 11.6 for NHS ester staining, and then stained by 1 mg/mL streptavidin conjugated to TAMRA in PBST for 1 hour in a room temperature shaker, followed by washing in PBST.
Confocal microscopy of samples
The sectioned and stained samples were placed flat on glass-bottom dishes with excess solution wicked by Kimwipes. Images were acquired using a Zeiss 800 confocal system. Unless stated otherwise, images were taken from a 60× objective, with a pinhole diameter of 33 µm, and a voxel size of 0.0821 µm×0.0821 µm×0.32 µm (before being divided by the expansion factor).
Light sheet microscopy of Cyc-ExM expanded testis tissue
A mouse testis tissue expanded by 4 cycles of Cyc-ExM was trimmed to a size of 3 mm and stained by TAMRA NHS ester and Sytox Green, according to the similar protocols described above. The sample was then fully immersed in 60% sucrose solution for matching of refractive index. The sample and 60% sucrose solution were then mounted on the imaging chamber and imaged on a QLS-SCOPE (Planelight, S.L, acquired in Quantum Design China, Beijing office). The sample was imaged at 5× magnification, with xy pixel size of 1.37 µm×1.37 µm and a step size of 4.18 µm in the z direction. All images were acquired and stitched by QLS software (Planelight, S.L) and saved as 16-bit grayscale TIFF images for each channel. The 3D images were visualized using Zeiss Zen 3.2 software.
Transmission electron microscopy (TEM) of porcine and rat tissues
For comparison of resolution, TEM image was also collected. The 3 mm-sized glutaraldehyde-fixed tissue blocks from porcine kidney, rat heart and rat kidney were dehydrated and embedded in epoxy resin. Ultrathin sections of these blocks were mounted on copper grids, counterstained with 2% uranyl acetate and observed on a transmission electron microscope (Hitachi HT7800). The TEM samples were prepared and imaged at Servicebio Inc.
Measurement and calculation of expansion factor
The details of expansion factor calculations were shown in Supplementary Table 2. In most situations, to calculate the expansion factor, the nuclear cross-sectional area of the same types of cells were measured using FIJI/ImageJ software. The average nuclear cross-sectional area in expanded samples was divided by that of non-expanded samples. The square root of this ratio represents an estimate of the linear expansion factor, i.e.
$$EF=\sqrt{\frac{{A}_{post}}{{A}_{pre}}}$$
where Apre and Apost represent the average cross-sectional areas before and after expansion, respectively. Correspondingly, according to error propagation, the standard deviation of the expansion factor EF can be obtained as follows
$${\sigma }_{EF}=\frac{EF}{2}\sqrt{{\left(\frac{{\sigma }_{pre}}{{A}_{pre}}\right)}^{2}+{\left(\frac{{\sigma }_{post}}{{A}_{post}}\right)}^{2}}$$
where σpre and σpost represent the standard deviation of the cross-sectional area before and after expansion, respectively.
To calculate expansion factors for 10-cycle expanded tissue, smaller structures were used as references. For heart tissue, the sarcomere length in cardiac muscle was used as the reference structure, and for kidney and liver tissues, the width of mitochondria was used.
Assessing the isotropic expansion of glomerular ultrastructure.
From the pan-ExM and Cyc-ExM confocal datasets, we assessed the isotropy of expansion by comparing the ratio of FP width to SP width of porcine renal glomeruli to that of TEM images. We used FIJI/ImageJ to measure the widths of FP and SP. The sample size is n = 30 for all conditions.
Assessing the isotropic expansion of isolated glomeruli.
Individual glomeruli were dissected out from porcine kidney tissue blocks fixed by 4% (w/v) PFA and 20% (w/v) AA as above using two pairs of fine point tweezers under a stereo microscope (Sunny Optical Technology Co., Ltd). The glomeruli were then collected by a micropipette into a 1.5 mL microcentrifuge tube in PBS. To fluorescently label their anatomical structures, the glomeruli were then treated with 50 mM TCEP in PBS for 30 min, followed by incubation in 0.2 mM CF568 maleimide in PBS for 1 hr. The unbound dyes were washed away by PBS for 1 hour. Next, the glomeruli were incubated in gel monomer solutions following different expansion protocols (pan-ExM or Cyc-ExM). Immediately after the first gelation, the sample was imaged by confocal microscopy (60× objective lens), before proceeding to the subsequent denaturation and expansion procedures as described above. During expansion, the location and orientation of the glomeruli embedded in the hydrogel were carefully tracked with excess gel trimmed. After expansion, the sample was again treated with 50 mM TCEP in PBS for 30 min and stained with 0.2 mM CF568 maleimide in PBS for 1 hour, followed by washing in PBST. Finally, the same region of interest as in the pre-expansion image was relocated under the confocal microscope (20× objective lens) and imaged.
Distortion due to expansion was then quantified using existing methods for distortion vector field and root-mean-square (RMS) error calculation1, 30. Briefly, 2D images of the same region of interest before and after expansion were selected from 3D confocal z-stacks and aligned by similarity transformation using rotation, translation, and uniform scaling, which also yields the expansion factor. Next, non-rigid B-spline transformation was used to “warp” the post-expansion image to optimally fit the pre-expansion one. Finally, distances between feature points in the rigid-transformed image was compared with those in the non-rigid transformed image to calculate the root mean square measurement error as a function of distances for characterization of distortions. Three different regions were analyzed for each expansion protocol.