Nanoscale imaging of clinical tissue samples can improve understanding of disease pathogenesis. Expansion pathology (ExPath) is a version of expansion microscopy (ExM), modified for compatibility with standard clinical tissue samples, to explore the nanoscale configuration of biomolecules using conventional diffraction limited microscopes.
In modern pathology, optical microscopy plays an important role in disease diagnosis by revealing microscopic structures of clinical specimens. However, the fundamental physical diffraction limit prevents interrogation of nanoscale anatomy and subtle pathological changes when using conventional optical imaging approaches. Here, we describe a simple and inexpensive protocol, called expansion pathology (ExPath), for nanoscale optical imaging of common types of clinical primary tissue specimens, including both fixed-frozen or formalin-fixed paraffin embedded (FFPE) tissue sections. This method circumvents the optical diffraction limit by chemically transforming the tissue samples into tissue-hydrogel hybrid and physically expanding them isotropically across multiple scales in pure water. Due to expansion, previously unresolvable molecules are separated and thus can be observed using a conventional optical microscope.
Investigating the molecular organization of tissues in a three-dimensional (3D) context can provide new understanding of biological functions and disease development. However, these nanoscale environments are beyond the resolution capabilities of conventional diffraction limited microscopes (200−300 nm), where the minimal resolvable distance, d is defined by d α<!––> λ/NA. Here λ is the wavelength of light and NA is the numerical aperture (NA) of the imaging system. Recently, direct visualization of fluorescently labeled molecules has been made possible by newly developed super-resolution imaging techniques1,2,3, including stimulated emission depletion (STED), photo-activated localization microscopy (PALM), stochastic optical reconstruction microscopy (STORM), and structured illumination microscopy (SIM). Although these imaging techniques have revolutionized understanding of biological function at the nanoscale, in practice, they often rely on expensive and/or specialized equipment and image processing steps, can have slower acquisition time comparing to conventional optical imaging, require fluorophores with specific characteristics (such as photo-switching capability and/or high photostability). In addition, it remains a challenge to perform 3D super-resolution imaging on tissue specimens.
Expansion microscopy (ExM), first introduced in 20154, provides an alternative means of imaging nanoscale features (<70 nm) by physically expanding preserved samples embedded in a swellable polyelectrolyte hydrogel. Here, key biomolecules and/or labels are anchored in situ to a polymer network that can be isotopically expanded after chemical processing. Because the physical expansion increases the total effective resolution, molecules of interest can then be resolved using conventional diffraction-limited imaging systems. Since the publication of the original protocol, where custom synthesized fluorescent labels were anchored to the polymer network4, new strategies have been used to directly anchor proteins (protein retention ExM, or proExM)5,6,7,8,9 and RNA9,10,11,12 to the hydrogel, and increase physical magnification through iterative expansion13 or adapting gel chemistry8,14,15.
Here we present an adapted version of proExM, called expansion pathology (ExPath)16, which has been optimized for clinical pathology formats. The protocol converts clinical samples, including formalin-fixed paraffin-embedded (FFPE), hematoxylin and eosin (H&E) stained, and fresh-frozen human tissue specimens mounted on glass slides, into a state compatible with ExM. Proteins are then anchored to the hydrogel and mechanical homogenization is performed (Figure 1)16. With a 4-fold linear expansion of the samples, multicolor super-resolution (~70 nm) images can be obtained using a conventional confocal microscope having only a ~300 nm resolution and can also be combined with other super-resolution imaging techniques.
1. Preparation of Stock Reagents and Solutions
2. Preparation of Archived and Freshly Prepared Clinical Tissue Slides for ExPath
3. In Situ Polymerization of Specimens
4. Sample Digestion
5. Sample Expansion and Imaging
If the protocol has been successfully carried out (Figure 1), samples will appear as a flat and transparent gel after mechanical homogenization (Figure 3A) and can expand by a factor of 3−4.5x in water (Figure 3B), providing an effective resolution of ~70 nm depending on the final expansion factor and imaging system used5,16. Figure 4 shows example images of a 5 µm thick FFPE kidney sample processed using the ExPath protocol. Full expansion of the gelled samples resulted in a 4.5x expansion factor in water, giving an effective resolution of ~63 nm when imaged using a 0.95 NA objective. The tissue was first pre-processed with xylene to remove paraffin and rehydrated (section 2.1.1), followed by antigen-retrieval with the citrate buffer (section 2.2). The recovered tissue was then stained with antibodies for alpha-actinin 4 (ACTN4) and vimentin, along with DAPI to visualize nuclear DNA and wheat germ agglutinin (WGA), to label carbohydrates. The specimen was then imaged using a spinning disk confocal microscope (Figure 4A,B,D). Tissues were treated following the above protocol and fully expanded in water (Figure 4C,E). Similarly, Figure 5 shows example images of H&E stained normal breast tissue (Figure 5A) that was treated with xylene (section 2.1.2) to remove the cover glass and then treated as an FFPE sample (section 2.1.1). During homogenization, the H&E stain is eliminated from the tissue. Post-expansion images of the DAPI stained sample obtained on a spinning disk confocal microscope (Figure 5B) compared with pre-expansion images indicate an expansion factor of 5.1, giving an effective resolution of ~43 nm when imaged with an NA 1.15 lens.
Example data for unfixed frozen kidney slices processed using the ExPath protocol can be seen in Figure 6. The tissue was first fixed in cold acetone (section 2.1.3) and stained with ACTN4, vimentin, DAPI, and WGA. Comparison of pre-expansion (not shown) and post-expansion images indicate an expansion factor of 4.5. Comparison data from the acetone-fixed frozen kidney samples to that of the FFPE samples (Figure 4C,E) shows that there is a decrease in quality of the ACTN4 staining in the expanded FFPE sample, which may be due to a degradation of the antigenicity due to the fixation method16.
Figure 1: Schematic of the expansion pathology (ExPath) workflow. Pre-processing of clinically archived tissue slides is first performed based on the storage format. Samples are then stained using conventional immunostaining protocols and pre-expansion images are obtained. Samples are then treated with acryloyl-X SE (AcX) to anchor proteins to the hydrogel. In situ polymerization is preformed prior to mechanical homogenization using proteinase K (ProK). Samples can then be expanded in ddH2O prior to imaging. Please click here to view a larger version of this figure.
Figure 2: Gelation chamber for pathology samples. (A) Two spacers, such as two pieces of #1.5 cover glass cut with a diamond knife, are placed on either side of the tissue after incubating the sample in gelling solution at 4 °C. The spacers should be thicker than the tissue slices, to prevent compression of the sample. (B) A lid, such as a piece of #1.5 cover glass, is used to cover the sample prior to incubation at 37 °C. Please click here to view a larger version of this figure.
Figure 3: Example of sample expansion. A 5 µm thick kidney sample post homogenization, (A) pre- and (B) post-expansion in ddH2O is shown. Grid squares are 5 mm x 5 mm. Please click here to view a larger version of this figure.
Figure 4: Representative results from 5 µm thick FFPE kidney samples. (A) Pre-expanded image. (B) Magnified pre-expansion image of the outlined region of interest in panel A and (C) the corresponding post-expansion image of the same region of interest after the sample was fully expanded in water. (D) Magnified image of the outlined region of interest in panel B and (E) the corresponding post-expansion image of the same region of interest after the sample was fully expanded in water. Cracking, distortions, and loss of labeled targets can be the result of inadequate anchoring and/or homogenization (F,G). All images were obtained using a spinning disk confocal microscope with a 20x (NA 0.95; water immersion) objective (A-E,G) or 10x (NA 0.5) objective (F). Blue, DAPI; green, vimentin; red, alpha-actinin 4 (ACTN4); magenta, WGA. Scale bars = 100 µm (A,F,G, yellow scale bars indicate post-expansion images); 50 µm (B), 50 µm (C, physical size post expansion 225 µm; expansion factor 4.5); 25 µm (D); 25 µm (E, physical size post expansion 112.5 µm; expansion factor 4.5). Please click here to view a larger version of this figure.
Figure 5: Representative results from H&E stained normal breast tissue. (A) Brightfield image of pre-expanded H&E stained tissue taken with a 40x (0.95 NA) objective. (B) Post-expansion DAPI image of the same region of interest after the sample was fully expanded in water. The image was obtained on a spinning disk confocal microscope using a 40x (NA 1.15; water immersion) objective. Scale bars = 10 µm (A, the yellow scale bar indicates post-expansion images) and 2 µm (B, physical size post expansion 50.1 µm; expansion factor 5.1). Please click here to view a larger version of this figure.
Figure 6: Representative results of fresh frozen kidney samples after expansion, obtained on a spinning disk confocal microscope with a 40x (NA 1.15; water immersion) objective. Blue = vimentin; green = alpha-actinin 4 (ACTN4); red = collagen IV; grey = DAPI. Scale bar = 10 µm (physical size post expansion 45 µm; expansion factor 4.5). Please click here to view a larger version of this figure.
Component | Stock concentration | Stock volume (mL) | Final concentration | Final amount per 10 mL |
Sodium acrylate | 0.380 g/mL | 2.25 | 0.086 g/mL | 0.86 g |
Acrylamide | 0.500 g/mL | 0.50 | 0.025 g/mL | 0.25 g |
N,N′-Methylenebisacrylamide | 0.020 g/mL | 0.50 | 0.001 g/mL | 0.10 g |
Sodium chloride | 0.292 g/mL | 4.00 | 0.117 g/mL | 1.17 g |
PBS | 10x | 1.00 | 1x | 1x |
Water | 1.15 | |||
Total Volume | 9.40 |
Table 1: Components of monomer solution. All concentrations are given in terms of g/mL (w/v) except PBS.
Here, we present the ExPath protocol16, a variant of proExM5 that can be applied to the most common types of clinical biopsy samples used in pathology, including FFPE, H&E stained, and fresh-frozen specimens on glass slides. Format conversion, antigen retrieval, and immunostaining of the specimens follow commonly used protocols that are not specific to ExPath. Unlike the original proExM protocol9, ExPath relies on a higher concentration of EDTA in the digestion buffer, which improves the expansion of formalin-fixed tissues, as demonstrated in the original ExPath study16. The protocol has been validated in 5−10 µm thick clinical specimens16, but could also be applied to thicker tissue samples with some modification. The most critical steps in this protocol are: 1) the timing of the gelation steps; 2) setup of the gelation chamber; 3) parameters for sample homogenization; and 4) handling of the gel.
The most critical parameter for this protocol is the timing of the gelation steps. If the gelling solution prematurely polymerizes, the sample will not be sufficiently anchored to the gel matrix. Inadequate anchoring and premature gelation can cause distortions, limit expansion, and result in the loss of target molecules (Figure 4F,G). The initiation of gelation is temperature dependent, therefore it is important to keep the mixed gelling solution at 4 °C before placing it on the target specimen. The initiator, APS, should be freshly prepared and added immediately before applying the gelling solution. APS is not stable at RT and can lose efficacy after undergoing freeze-thaw cycles. Insufficient anchoring can also be the result of reduced reactivity of the anchoring compound, AcX, which can lose activity after long term storage or after contact with water. AcX has been found to retain activity after 6 months of storage in a desiccated environment at -20 °C. If premature gelation is not the suspected cause of distortion, a new stock solution of AcX can be prepared.
During the setup of the gelation chamber, air bubbles can become trapped under the cover glass lid. If these bubbles are on top of or directly touching the sample, they can cause distortions. They can be moved away from the specimen by adding more gelling solution through the side of the chamber. To help prevent air bubbles, a small drop of gelling solution can be deposited on the lid before placing it over the tissue.
Sample digestion is dependent on time, temperature, as well as tissue properties, such as thickness and tissue type. Inadequate digestion can also cause distortions and result in an expansion factor that is smaller than expected. If incomplete digestion is suspected, the digestion time and/or ProK concentration can be increased, particularly in the case of thicker tissues. ProK can also lose activity over time. To preserve its activity, ProK should be stored at -20 °C and can be aliquoted into smaller volumes to avoid free-thaw cycles.
Care should be taken when handling homogenized samples, particularly when fully expanded. If the specimen is hard to locate when submerged in liquid, illuminating the container from different angles can scatter incident light to allow visualization of the gel. Fully expanded gels are fragile and can break during handling. Use of soft brushes and plastic spatulas is recommended to transfer the expanded gels.
The ExPath protocol provides a cost-effective alternative to current super-resolution imaging and electron microscopy techniques to interrogate nanoscale structures in clinical biopsy specimens. Although ProK provides even expansion of the sample after homogenization, the loss of proteins prevents interrogation of other targets of interest post-expansion. However, the protocol can be easily performed in any regular wet lab. Most importantly, imaging of nanoscale features can be carried out on conventional wide-field or confocal microscopes that are commonly found in biology laboratories and imaging core facilities.
The authors have nothing to disclose.
This work was supported by the Faculty Start-up fund from the Carnegie Mellon University (YZ) and NIH Director’s New Innovator Award (DP2 OD025926-01 to YZ).
4-hydroxy-TEMPO (4HT) | Sigma Aldrich | 176141 | Inhibitor |
6-well glass-bottom plate (#1.5 coverglass) | Cellvis | P06-1.5H-N | |
Acetone | Fischer Scientifc | A18-500 | |
Acrylamide | Sigma Aldrich | A8887 | |
Acryloyl-X, SE (AcX) | Invitrogen | A20770 | |
Agarose | Fischer Scientifc | BP160-100 | |
Ammonium persulfate (APS) | Sigma Aldrich | A3678 | Initiatior |
Anti-ACTN4 antibody produced in rabbit | Sigma Aldrich | HPA001873 | |
Anti-Collagen IV antibody produced in mouse | Santa Cruz Biotech | sc-59814 | |
Anti-Vimentin antibody produced in chicken | Abcam | ab24525 | |
Aqua Hold II hydrophobic pen | Scientific Device | 980402 | |
Breast Common Disease Tissue Array | Abcam | ab178113 | |
DAPI (1 mg/mL) | Thermo Scientific | 62248 | Nuclear stain |
Diamond knife No. 88 CM | General Tools | 31116 | |
Ethanol | Pharmco | 111000200 | |
Ethylenediaminetetraacetic acid (EDTA) 0.5 M |
VWR | BDH7830-1 | |
FFPE Kidney Sample | USBiomax | HuFPT072 | |
Forceps | |||
Goat Anti-Chicken IgY (H+L), Highly Cross-Adsorbed CF488A | Biotium | 20020 | |
Goat Anti-Chicken IgY (H+L), Highly Cross-Adsorbed CF633 | Biotium | 20121 | |
Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Alexa Fluor 546 | Invitrogen | A11010 | |
MAXbind Staining Medium | Active Motif | 15253 | Can be substituted with non-commercial staning buffer of choice. |
MAXblock Blocking Medium | Active Motif | 15252 | Can be substituted with non-commercial blocking buffer of choice. |
MAXwash Washing Medium | Active Motif | 15254 | Can be substituted with non-commercial washing buffer of choice. |
Micro cover Glass #1 (24x60mm) | VWR | 48393 106 | |
Micro cover Glass #1.5 (24x60mm) | VWR | 48393 251 | |
N,N,N′,N′- Tetramethylethylenediamine (TEMED) |
Sigma Aldrich | T9281 | Accelerator |
N,N′-Methylenebisacrylamide | Sigma Aldrich | M7279 | |
Normal goat serum | Jackson Immunoresearch | 005-000-121 | For preparing blocking buffer. Dependent on animal host of secondary antibodies. |
Nunclon 4-Well x 5 mL MultiDish Cell Culture Dish | Thermo Fisher | 167063 | Multi-well plastic culture dish |
Nunclon 6-Well Cell Culture Dish | Thermo Fisher | 140675 | |
Nunc 15mL Conical | Thermo Fisher | 339651 | |
Nunc 50mL Conical | Thermo Fisher | 339653 | |
Orbital Shaker | |||
Paint brush | |||
pH Meter | |||
Phosphate Buffered Saline (PBS), 10x Solution | Fischer Scientifc | BP399-1 | |
Plastic Petri Dish (100 mm) | Fischer Scientifc | FB0875713 | |
Proteinase K (Molecular Biology Grade) | Thermo Scientific | EO0491 | |
Razor blade | Fischer Scientifc | 12640 | |
Safelock Microcentrifuge Tubes 1.5 mL | Thermo Fisher | 3457 | |
Safelock Microcentrifuge Tubes 2.0 mL | Thermo Fisher | 3459 | |
Sodium acrylate | Sigma Aldrich | 408220 | |
Sodium chloride | Sigma Aldrich | S6191 | |
Sodium citrate tribasic dihydrate | Sigma Aldrich | C8532-1KG | |
Tris Base | Fischer Scientifc | BP152-1 | |
Triton X-100 | Sigma Aldrich | T8787 | |
Wheat germ agglutinin labeled with CF640R | Biotium | 29026 | |
Xylenes | Sigma Aldrich | 214736 |