The Basics of Expansion Microscopy

Abstract

Optical imaging techniques are often used in neuroscience to understand brain function and discern disease pathogenesis. However, the optical diffraction limit precludes conventional optical imaging approaches from resolving nanoscopic structures with feature sizes smaller than 300 nm. Expansion microscopy (ExM) circumvents this limit by physically expanding preserved tissues embedded in a swellable hydrogel. Biomolecules of interest are covalently linked to a polymer matrix, which then isotopically expands at least 100-fold in size in pure water after mechanical homogenization of the tissue-gel. The sample can then be investigated with nanoscale precision using a conventional diffraction-limited microscope. The protocol described here is a variant of ExM that uses regents and equipment found in a typical biology laboratory and has been optimized for imaging proteins in expanded brain tissues.

INTRODUCTION

Studying nanoscale configuration of biomolecules in the brain is critical for understanding normal brain function and disease development. Conventional optical imaging techniques have provided insight into the cellular and sub-cellular organization of brain tissues. However, these systems are unable to laterally resolve features smaller than 200 – 300 nm due to the fundamental physical limits imposed by diffraction, typically given as half the ratio of the wavelength of light to the numerical aperture of the imaging system. Super-resolution imaging techniques overcome this limit (Combs & Shroff, 2017; S. W. Hell, 2007; Schermelleh, Heintzmann, & Leonhardt, 2010) by various strategies, such as employing single-molecule localization methods (photo-activated localization microscopy (PALM)(Betzig et al., 2006) stochastic optical reconstruction microscopy (STORM)(Rust, Bates, & Zhuang, 2006), structured light to encode higher spatial frequencies (structured illumination microscopy (SIM)(Gustafsson, 2000), or engineering of the point spread function (stimulated emission depletion (STED) (Stefan W. Hell & Wichmann, 1994). Although these powerful tools have furthered understanding of the brain at the nanoscale, they are not optimized in terms of cost, speed, and quality for imaging brain tissues over extended volumes.

As an emerging super-resolution imaging technique, expansion microscopy (ExM)(F. Chen, Tillberg, & Boyden, 2015), relies on physical, rather than optical, magnification. The protocol described here demonstrates how a variant of protein retention ExM (ProExM)(Tillberg et al., 2016) can be applied to mouse brain tissue. Proteins are covalently anchored to a swellable hydrogel matrix using a commercially available molecule, Acryloyl-X SE (AcX). This hydrogel matrix is formed in situ, where the initiator ammonium persulfate (APS) and accelerator tetramethyl-ethylenediamine (TEMED) trigger the crosslinking of sodium acrylate with the N-N’-methylenebisacrylamide to form a dense polyelectrolyte hydrogel. The samples are then mechanically homogenized using a non-specific proteinase K (ProK) that digests away many of the biomolecules in the sample. Unlike the original ExM protocol, the monomer solution used here has an increased concentration of sodium acrylate (SA) and acrylamide, providing an increased expansion factor. When fully expanded in water, the sample can achieve a 5-fold linear expansion allowing nanoscale features (~50 nm) to be imaged using a conventional confocal microscope or other super-resolution imaging systems.

STRATEGIC PLANNING

An overview for performing ExM on brain tissue sections is shown in Figure 1. The protocol can be performed using reagents and imaging systems found in most biology labs. Although there is relatively little active time and the protocol can be paused at many points, the timing of two key steps must be followed closely: preparation of the gelling solution and homogenization of the gelled sample. Like other ExM variants, this protocol is not compatible with cyanine dyes (Cy3, Cy5, Alexa 647). It is suggested that other fluorescent labels such as Alexa 488, Alexa 546, ATTO 647N, or CF633 labels are used.

Figure 1:

Schematic of the Expansion Microscopy (ExM) workflow. After appropriate fixation of the brain tissue, it is frozen and sectioned. Samples are then stained using conventional staining protocols and pre-expansion images are obtained using a fluorescent microscope. After imaging, samples are treated with Acryloyl-X SE (AcX) to anchor proteins to the hydrogel. A monomer solution is then applied to the sample and after in situ polymerization the sample is mechanically homogenized using proteinase K (ProK). Samples can then be expanded in pure water prior to imaging.

BASIC PROTOCOL: ExM for Intact Brain Tissue

Materials

Phosphate-buffered saline (PBS)

Blocking buffer: MAXblock™ Blocking Medium (Active Motif) or prepared blocking buffer containing 5% (v/v) normal donkey serum in 0.1% (w/v) Triton X-100 in PBS. Serum should be from the same host as the secondary antibodies.

Staining buffer: MAXbind™ Staining Medium (Active Motif) or PBST containing 1% (w/v) Triton X-100 in PBS.

Washing buffer: MAXwash™ Washing Medium (Active Motif) or PBS.

Primary and secondary antibodies

10% (w/v) paraformaldehyde solution in PBS

30% (w/v) sucrose solution in PBS

Optimal cutting temperature compound (OCT)

Glycerol solution: 30% glycerol (v/v), 30% ethylene glycol (v/v) in PBS

secondary antibodies for immunostaining

DAPI (optional)

AcX stock solution (see recipe)

Digestion buffer (see recipe)

Low-melt agarose (optional)

Freezing microtome

Petri dishes or 6-well plastic-bottom plates

6-well glass bottom plates with No. 1.5 cover glass

Paintbrush

15 mL conical tubes

50 mL conical tubes

Razor blade

Forceps

No. 1.5 coverslips

Microscope slides

1.5 mL Eppendorf tube

2 mL Eppendorf tube

Diamond tipped pen

Hydrophobic pen

Tissue Preparation

The steps described here are given as an example; specimens can be prepared using a preferred protocol if desired. The final thickness of the tissue affects practical factors including the diffusion time of reagents and antibody penetration. Typically, 30 μm slices are used, however slices up to 100 μm thick have been used.

1. Perfuse tissue with 10% PFA. Post-fix the brain in 10% PFA overnight at 4 °C.

2. Cryoprotect the tissue by incubating the brain in 30% (w/v) sucrose in PBS until the brain sinks (usually 24 hours).

3. Section the brain in 30 μm slices using a freezing microtome and store the specimens at 4 °C in glycerol solution.

   Note: Tissue slices can be stored for several weeks at 4 °C before proceeding to the next steps.

Immunostaining/immunohistochemistry staining (optional)

Immunostaining can be performed using a preferred staining protocol. The provided steps follow a typical immunofluorescence (IF)/immunohistochemistry (IHC) protocol for PFA-fixed, free-floating sections and can be replaced with a preferred staining protocol. Timing and concentrations are given as suggestions, and parameters should be optimized for specific reagents and experimental conditions.

1. In a 6-well plastic-bottom plate, wash the samples three times in PBS for 10 minutes each at room temperature. Transfer samples between well plates with a soft paint brush.

2. Permeabilize the tissues by incubating in blocking buffer for 1 hour at 37 °C or 6 hours at room temperature.

3. Incubate the samples with primary antibodies at manufacturer-recommended or experimentally-determined concentrations in preferred staining buffer overnight (at least 16 hours) at room temperature. Samples can also be placed on a shaker at low speed.

   Note: The blocking buffer given in the Reagents and Solutions section or 1x PBS can be used in place of the staining buffer.

4. Remove the antibody solution and wash the samples three times in 1x PBS for 10 minutes each at room temperature.

5. Incubate the samples with fluorescently-conjugated secondary antibodies at concentrations of approximately 10 μg/mL (optionally with 300 nM DAPI) in staining buffer for 3 hours at room temperature. Samples can be placed on a shaker at low speed.

6. Remove the antibody solution and wash the samples three times in 1x PBS for 10 minutes each at room temperature.

   Note: Samples can be stored in PBS at 4 °C for several weeks before gelation.

7. Obtain pre-expansion images using a fluorescent imaging system of choice, if desired, by gently placing the samples on a glass slide with a paint brush. Add a small amount of 1x PBS to the bottom of the slide and apply coverslip.

   Note: Pre-expansion images can be used to determine the expansion factor and thus the biological length of the expanded samples. If possible, fields of view should contain fiducial markers that can be easily identified in post-expansion samples.

8. Remove the coverslip with a razor blade after images have been taken and transfer samples into 1x PBS with a paint brush.

Anchoring and Gelation

1. Prepare 1.5 mL of the anchoring solution in a 1.5 mL Eppendorf tube by diluting the Acryloyl-X SE (AcX) stock to 0.1 mg/mL in PBS.

   Note: Samples fixed with non-aldehyde fixatives can be anchored with a 0.03 mg/mL AcX solution. Aldehyde-fixed samples require a higher AcX concentration (0.1 mg/mL) as the fixation results in fewer free amines available to react with AcX.

2. Place the sample in the tube with the anchoring solution with a paint brush and incubate the sample for 3 hours at room temperature.

   Note: Alternatively, the samples can be incubated in the anchoring solution overnight at 4 °C.

3. While the sample is incubating, prepare the monomer, 4-Hydroxy-TEMPO (4HT) inhibitor, tetramethylethylenediamine (TEMED) accelerator, and ammonium persulfate (APS) initiator stock solutions according to the instructions in the Reagents and Solutions section.

   Note: The monomer stock solution can be prepared in advanced and stored at 4 °C for up to 3 months or at −20 °C for long-term storage. The 4HT and TEMED stock solutions can be prepared in 1 mL aliquots and stored at −20 °C for up to 6 months. APS can lose efficacy after long-term storage, so it is suggested to prepare small quantities (<0.1 mL) immediately before mixing the gelling solution.

4. Construct a gelling chamber (Figure 2A) by cutting two spacers from a glass coverslip with a diamond-tipped pen and placing them on a glass slide approximately 2 cm apart. Draw a small circle with the hydrophobic pen between the glass spacers to keep solution on the sample.

5. At the end of the anchoring treatment, prepare the gelling solution by combining 200 μL of the monomer solution with 2.67 μL TEMED stock (1:75 dilution), 2 μL 4-HT stock (1:100 dilution), and 4 μL APS stock solution in a 1.5 mL Eppendorf tube.

   Note: The gelling solution should be kept cold and the APS should be added immediately before using in order to prevent premature gelling of the solution.

6. Pipette the gelling solution into the gelling chamber, taking care to keep the solution within the circle drawn with the hydrophobic pen. Carefully transfer the sample out of the AcX solution with a paint brush and place the tissue directly into the gelling solution. Flatten the tissue onto the glass. (Figure 2B).

7. Place the glass slide into a 100 mm Petri dish and incubate the tissue in the gelling solution for 30 minutes at 4 °C.

8. After incubation, gently place a coverslip over the sample chamber, taking care to keep the tissue flat on the glass and avoiding air bubble from forming over the tissue.

9. Place a damp paper wipe in the Petri dish next to the gelling chamber, put the lid on the Petri dish, and incubate the sample in a humidified incubator overnight at 37 °C.

   Note: Alternatively, samples can be gelled by incubating at 37 °C for 2 hours followed by incubation at 60 °C for 30 minutes.

Figure 2:

Samples are gelled in a chamber constructed from a glass slide with two cover glass spacers (A). After anchoring, cold gelling solution is pipetted on to the slide and the tissue is gently placed in the gelling solution (B). The samples are then gelled (C, left) and homogenized. After homogenization, the samples can be fully expanded using deionized water (C, right) Scale bar is 10 mm.

Digestion and Expansion

1. After the sample has gelled, remove the sample from the incubator and carefully remove the coverslip using a razor blade. Trim the blank gel surrounding the tissue and use the razor blade to peel away the blank gel. Gently remove the sample from the glass slide using the razor blade and place the sample in a 2 mL Eppendorf tube using the razor blade or a soft brush.

   Note: The sample should be removed carefully from the glass slide. If the sample does not transfer easily, wet a brush or razor blade with 1X PBS.

   If desired, the sample can also be cut into smaller pieces prior to removal from the glass slide.

2. Add 2 mL of freshly prepared digestion solution containing 8 units/mL Proteinase K to the Eppendorf tube containing the sample. Incubate the tissue in the solution overnight (approximately 16 hours) at room temperature. The samples should appear clear after complete homogenization.

3. Transfer the homogenized sample to a dish compatible with your fluorescence imaging system, such as a 6-well glass-bottom plate for an epifluorescence system. Wash the sample 3 times in 1x PBS for 10 minutes each at room temperature. If desired, the sample can be re-stained with 300 nM DAPI, which washes away during the digestion process.

   Note: The samples should be contained in a vessel large enough to accommodate the fully expanded sample. A 6-well plate can accommodate samples with a pre-expansion diameter of 0.6 cm or smaller.

4. Expand the sample by washing in an excess of ddH₂O in repeated rounds for 10 minutes each at room temperature until the sample no longer increases in size, typically 3–5 washes.

   Note: Handling of the samples after expansion should be avoided as the gels are more fragile in their expanded state.

5. Image the fully expanded samples on a preferred fluorescent imaging system. Optional: the sample can be immobilized in 1.5–2% low melt agarose as described in Supplementary Protocol 1.

Supplementary Protocol 1: Immobilization in Low Melt Agarose

1. Prepare 1.5–2% (w/v) low-melt agarose in ddH₂O.

2. Melt the solution in a water bath or microwave and pipette the agarose solution around the edges of the gel.

3. After the agarose has solidified, add additional ddH₂O to the sample to prevent dehydration.

REAGENTS AND SOLUTIONS

Deionized, distilled water should be used in all recipes and protocol steps. All percent solutions are given in terms of weight/volume percent unless otherwise noted.

Acryloyl-X, SE Stock Solution

Combine 5 mg Acryloyl-X, SE (AcX; Invitrogen cat. no. A20770) with 500 μL anhydrous dimethylsulfoxide (DMSO) to make a 10 mg/mL solution. After the AcX is dissolved, the stock solution can be aliquoted in 20 μL aliquots and stored at −20 °C with a desiccant to prevent loss of activity.

10% Ammonium Persulfate (APS) Stock Solution

0.01g APS

0.099 mL ddH₂O

APS stock solution should be prepared immediately before adding to the gelling solution as it has been found to lose reactivity after long-term storage.

Digestion Solution (25 mM ETDA, 0.8 M NaCl, 50 mM Tris pH 8.0, 0.5% Triton X-100)

25 mL 0.5M EDTA pH 8

23.38g NaCl

25 mL 1 M Tris pH 8 (3.03 g Tris Base in 25 mL ddH₂O)

2.25 g Triton X-100

Add ddH₂O for a total volume of 500 mL

The solution can be scaled up or down as needed and stored at 4 °C.

0.5% 4-Hydroxy-TEMPO (4HT) Stock Solution

0.05 g 4HT

10 mL ddH₂O

4HT stock solution can be prepared in 1 mL aliquots and stored at −20 °C for at least 6 months.

Monomer Solution (5% AA, 0.1% Bis, 15% SA, 11.7% NaCl)

1 mL 50% acrylamide (AA)

0.137 mL ddH₂O

0.5 mL 2% N,N′-methylenebisacrylamide (Bis)

1 mL 10x PBS

3.95 mL 38% sodium acrylate (SA)

3.34 mL 35% sodium chloride (NaCl)

The monomer stock solution can be stored at 4 °C for up to 3 months or at −20 °C for longer.

10% (v/v) tetramethylethylenediamine (TEMED) Stock Solution

1 mL TEMED

9 mL ddH₂O

TEMED stock solution can be prepared in 1 mL aliquots and stored at −20 °C for at least 6 months.

COMMENTARY

Background Information

The ability to resolve nanoscale features in brain tissue can provide insight into how biomolecules affect brain function. Although super-resolution imaging techniques have allowed researchers to explore neural circuitry at the biomolecular scale, these techniques do not readily lend themselves to imaging extended three-dimensional samples. ExM was developed as an alternative to super-resolution techniques where chemical processes are used to physically, rather than optically, magnify the sample.

ExM and its derivatives, reviewed elsewhere (Wassie, Zhao, & Boyden, 2019), can be used to anchor fluorescent labels (F. Chen et al., 2015), proteins (Asano et al., 2018; Chozinski et al., 2016; Ku et al., 2016; Tillberg et al., 2016; Truckenbrodt et al., 2018; Zhao et al., 2017), RNA (Asano et al., 2018; Fei Chen et al., 2016; Tsanov et al., 2016; Wang, Moffitt, & Zhuang, 2018) and other key biomolecules to a polymer network in situ. Following chemical processing, the tissue hydrogels can then be isotopically expanded in water. Because expansion physically separates molecules of interest, nanoscale features (<70 nm) can be resolved using conventional diffraction limited imaging systems. Even further magnification can be achieved through iterative expansion (Chang et al., 2017) or new gel chemistry (Cipriano et al., 2014; Truckenbrodt et al., 2018; Truckenbrodt, Sommer, Rizzoli, & Danzl, 2019).

Critical Parameters

The most critical parameter for successfully carrying out the protocol is the timing of both the gelation and homogenization steps. Because the polymerization of the gelling solution is temperature dependent, the preparation of the gelling solution and application of the solution to the tissue should be done while the mixture is still cold. Adding the APS stock last also helps prevent premature gelation. If the gelling solution polymerizes too quickly, the sample will not be anchored to the gel matrix, resulting in loss of target molecules during homogenization. Timing of sample homogenization depends on the properties of the tissue (thickness, type) as well as the temperature the homogenization is carried out. Too short of digestion can lead to tissue cracking and incomplete digestion, while too long of digestion can result in loss of probes.

Troubleshooting

Critical problems that can arise when following this protocol are most likely to occur during the gelation and homogenization steps. Issues with gel quality and inadequate homogenization can lead to distortions, lack of expansion, and/or loss of target biomolecules.

Premature gelation can result in insufficient anchoring of the biomolecules to the gel matrix, causing distortions, limiting expansion, and loss of target molecules. Because the polymerization of the monomer solution is temperature dependent, premature gelation can occur if the mixed gelling solution is too warm. To prevent this, the mixed gelling solution should be applied as quickly as possible after the addition of the initiator (APS). If necessary, the mixed gelling solution can be kept on ice prior to application to the sample. If the mixed solution continues to gel prematurely, fresh initiator (4HT) stock should be prepared and if necessary, the concentration of the 4HT stock can be increased.

If the gel fails to polymerize, the SA in the monomer solution may be of poor quality. SA should be stored at −20 °C in a desiccated environment for up to 3 months. If the 38% SA stock solution is tinted yellow, then the SA is not pure and should be discarded. If too little of the initiator (APS) is used or if the APS stock is too old, the gel may not fully polymerize. As APS is not stable in water long-term, it is suggested that the APS stock be made immediately before mixing with the monomer solution. To test the quality of reagents and insure the gelling solution polymerizes appropriately, a blank gel can be made by following steps 14–20.

Air bubbles trapped under the lid of the gelling chamber can cause distortions in the tissue after homogenization. These bubbles can be moved away from the sample by adding more gelling solution through the side of the gelling chamber, gently lifting and replacing the chamber lid, or a small drop of gelling solution can be added to the lid before placing over the sample. Bubbles not in contact with the tissue will not cause distortion issues.

Distortions and ruptures can be due to inadequate anchoring, the flatness of the tissue prior to gelling, and/or insufficient homogenization. Inadequate anchoring can be the result of premature gelation (discussed above) or reduced reactivity of AcX, the anchoring compound. AcX loses reactivity after long-term storage or if it is exposed to water. Therefore, it is suggested to be stored for up to 6 months in a desiccated environment at −20 °C. Distortions can also be caused by air bubbles being trapped under the cover glass of the gelation chamber.

If the tissue is not completely flat when gelled, the unevenness will be magnified by the expansion factor. Thus, care should be taken when placing the sample in the gelling chamber to ensure the sample is as flat as possible. If necessary, the spacers can be omitted from the gelling chamber and the chamber lid can be placed directly on top of the tissue submerged in the gelling solution. However, this will result in a thinner gel which may be less sturdy and more prone to breaking during handling.

If distortions and ruptures are not believed to be caused by problems during gelation or if the expansion factor is smaller than expected, the sample may be insufficiently digested. Sample digestion is dependent on time, temperature, as well as the properties (thickness, density) of the tissue. After adequate digestion, the tissue gel should be clear. Under digestion may be due to too short of an incubation time in the digestion buffer. If increasing the digestion time or increasing the amount of ProK does not solve the problem, the ProK may need to be replaced. To preserve the activity of ProK, it should be stored at −20 °C and aliquoted into smaller volumes in order to avoid too many free-thaw cycles.

Although the fluorescence signal is expected to be dimmer when the sample is fully expanded (less fluorophores per unit volume), loss of target molecules and/or little or no fluorescence signal may be due to insufficient anchoring of the tissue (discussed above) or over-digestion of the sample. If over-digestion is suspected, the homogenization time should be adjusted accordingly.

After digestion, the samples can be fragile, so care should be taken when handling the expanded gels. Handling of the samples in their fully expanded states should be avoided. As the digestion process clears away components of the tissue that are not anchored to the polymer matrix, the sample may be hard to locate when submerged in liquid. Illuminating the container at different angles and gently shaking the container can scatter light to make the gel more readily visible.

Understanding Results

Samples treated with the described protocol will result in an isotropically expandable tissue-hydrogel. When fully expanded by ~100 folds in volume, it will have a refractive index of water (~1.33) and will appear clear. Successfully homogenized samples will have no cracks or distortions and will retain the fluorescent labels applied pre-expansion. Example outcomes can be seen in Figure 3. After fixation, a sagittal section of mouse striatum was stained with DAPI and labeled for tyrosine hydroxylase, synaptophysin, and alpha-internexin and then imaged using a spinning disk confocal microscope (Figure 3A,E). After gelation and mechanical homogenization, the sample (Figure 3B) was expanded 5.27-fold in water, giving an effective resolution of ~45 nm when imaged with a 40x (1.1NA) objective. With expansion, individual puncta of synaptophysin in presynaptic terminals can be readily resolved and quantified (Figure 3C,D). Likewise, individual white matter tracts indicated with alpha-internexin can be traced through the striatum.

Figure 3:

Representative results from sagittal sections of mouse striatum. Pre-expanded samples (A,E) were stained with DAPI (blue) and labeled for tyrosine hydroxylase (green), synaptophysin (red), and alpha-internexin (magenta). Successful completion of the protocol will result in a 5.27-fold expansion of the tissue (B). A magnified image of the outline region of interest in (A,B) are shown in (C,D), respectively. Distortions and loss of florescent signals can occur if samples are over-homogenized (F). All images were taken on a spinning disk confocal microscope using either a 1.1 NA 40x (A-D) or 20× 0.95 NA (E,F) water immersion objective. Scale bars (yellow scale bars indicate post-expansion images): 10 μm (A,B; post-expansion physical size 52.7 μm); 5 μm (C,D; post-expansion physical size 26.4 μm); (C,D) 100 μm.

Distortions can occur when the sample has not completely homogenized, while over homogenization can result in the loss of the target molecules of interest. Brain tissue can be particularly sensitive to homogenization conditions. For example, performing homogenization in for 3 hours at 60 °C rather than overnight at room temperature can completely remove the tyrosine hydroxylase and synaptophysin signals (Figure 3F) while keeping the DAPI and alpha-internexin intact.

The maximum achievable resolution will be dependent on the specifics of the imaging system used, as the effective resolution depends on both the expansion factor and the NA of the microscope objective. It should also be noted that expansion effectively dilutes fluorescent signal as fluorophores become more physically separated resulting in less fluorophores imaged per pixel compared to its unexpanded state. Thus, samples may appear dimmer in their fully expanded states and may require higher excitation intensity and/or longer acquisition times.

Time Considerations

The timing of the protocol is primarily dependent on the thickness of the sample, whether immunofluorescence staining is performed, and the temperature at which various incubation steps are carried out. For pre-fixed and pre-labeled tissues, the protocol can be carried out in 1 day while fresh, un-labeled tissues can be processed in as little as 3 days. Overall, the protocol has minimal active time, with most of the time consumed by incubating or washing the sample, thus in principle allowing multiple samples to be processed in parallel.

Significance Statement.

Interrogation of the brain at the nanoscale level can provide insight into how nanoscopic configuration and misconfiguration of biomolecules gives rise to complex brain function and disease. However, conventional optical microscopes are diffraction limited and therefore incapable of pinpointing the spatial organization of molecules with nanoscale precision. Although super-resolution microscopy techniques circumvent optical diffraction limit, most require delicate, expensive setups and lengthy acquisition times. Rather than optical magnification, expansion microscopy (ExM) offers diffraction unlimited image resolution through imaging physically magnified tissue specimens using conventional diffraction limited microscopes.