Pathways to clinical CLARITY: volumetric analysis of irregular, soft, and heterogeneous tissues in development and disease

Three-dimensional tissue-structural relationships are not well captured by typical thin-section histology, posing challenges for the study of tissue physiology and pathology. Moreover, while recent progress has been made with intact methods for clearing, labeling, and imaging whole organs such as the mature brain, these approaches are generally unsuitable for soft, irregular, and heterogeneous tissues that account for the vast majority of clinical samples and biopsies. Here we develop a biphasic hydrogel methodology, which along with automated analysis, provides for high-throughput quantitative volumetric interrogation of spatially-irregular and friable tissue structures. We validate and apply this approach in the examination of a variety of developing and diseased tissues, with specific focus on the dynamics of normal and pathological pancreatic innervation and development, including in clinical samples. Quantitative advantages of the intact-tissue approach were demonstrated compared to conventional thin-section histology, pointing to broad applications in both research and clinical settings.

(a) Clearing kinetics were determined using UV-spectrophotometer measurements taken over time, here illustrated using brain tissue fixed with A1B1P4 CLARITY gel and cleared at 37°C. Under all clearing conditions, transmission of longer wavelength light is more efficient than shorter wavelength light, with implications for optimal fluorophore selection. n = 6 2 mm sections. (b) When cleared under identical conditions, an array of different tissue types results in vastly different levels of optical transparency due to organ-organ heterogeneity. n = 6 2 mm sections.
(c) 2 mm slices of brain tissue fixed with different hydrogel formulations were cleared and measured for transparency at different time points. Clearing rates were computed by fitting clearing curves to the first order transformation reaction kinetics equation y = Y max (1 -e -kt ) and normalized to Brain A4B4P4. n = 6 2 mm sections. (d) Rates of tissue clearing were increased for all tissue types when clearing was performed at 60°C, illustrated here with several tissue types embedded in A1B1P4 CLARITY gel. n = 6 2 mm sections. (e) Summary of clearing rates for different tissue types when cleared at 60°C. A1B1P4 CLARITY tissues can be cleared 15-30x faster at 60°C, A4B4P0 CLARITY tissues 50x faster or more when compared to A4B4P4 cleared at 37°C. n = 6 2 mm sections. (a) When removed from clearing solution once the tissue was clear, all gel types retained protein effectively and were suitable for detection using antibody probes, illustrated here with parvalbumin (PV) antibody incubated for 48 hours. However, prolonged exposure to clearing solution can result in catastrophic protein loss at both 37°C and 60°C for tissues fixed with A4B4P0 gel. Images were acquired with identical microscope power settings and are maximum intensity projections over 1 mm. Scale bar = 50 µm. (b) Quantification of protein loss in tissues immediately after clearing. n = 6. (c) Quantification of protein loss in tissues after 2 weeks' excess in clearing solution at 37°C. A4B4P0 gel lost excessive protein, even compared to PFA fixed tissue in SDS. n = 6. (d-g) Quantification of protein loss in brain, muscle, liver, and kidney tissue over time while cleared at 37°C. Arrowheads indicate when brain tissue is optically cleared; those values are used in panel (b). n = 6. For each detected islet, a convex hull was fit to positive pixels. Objects was found to be fewer than 5 pixels in area (representing 124 µm 2 in area, less than a cell), were excluded from analysis. An ellipse was fit to each hull to determine the major and minor axes. The hull was offset by two pixels (10 µm) toward the outer dimension to generate an islet neighborhood. Scale bar: 250 µm. (c) Example raw signals from the islet and neural channel are shown with thresholded, positive neural pixels and fitted convex hull islet, followed by the islet neighborhood (green) overlayed on the thresholded neural channel (red    (Sst)). This 200 µm-thick frozen section from a healthy adult pancreas was obtained from a human tissue bank, embedded in A1B1P4 hydrogel, and cleared at 60°C for 2 days. Scale bar = 500 µm (low magnification), 100 µm (high magnification of different islets for each stain). (b) Unlike CLARITY samples processed following PFA fixation, we found that fresh frozen specimens require blocking with serum prior to antibody staining to reduce background binding. Scale bar = 100 µm. (c) Frozen specimens processed with CLARITY could be stained through multiple rounds of antibody labeling and removal. Arrows indicate identical islets. Staining was performed with 3-6 hour antibody incubations, and antibodies were removed through clearing at 60°C for 24 hours. Scale bar = 500 µm. Figure S6: CLARITY in human cancer specimens. (a) CLARITY enabled clearing, dye and antibody labeling, and imaging of pathological samples obtained from a tissue bank, such as this frozen melanoma biopsy, embedded in A1B1P4 hydrogel, cleared for 2 weeks at 37°C, labeled with DAPI, Col4, and melanA, and imaged on a confocal microscope. Scale bar = 500 µm (b) When compared with a 2D optical section, CLARITY enables visualization of complex tissue architecture that may be diagnostically relevant, such as vasculature labelled with Col4. Scale bar = 50 µm. (c, d) CLARITY enables visualization of complex 3-dimensional tumor boundaries, including structures that change drastically within serial sections, such as appendages or offshoots (arrows), which can be assessed as serial sections or through manually-annotated volumetric rendering with viewing software such as Imaris. Scale bar = 100 µm. (e-g) Multiple human tumor types can be cleared and visualized in 3D utilizing CLARITY, enabling visualization of glandular or tubular structures (e.g. in ovarian cancer), and tumor specific diagnostic markers such as Cytokeratin 7 (Ck7, in lung non-small cell carcinoma) and Cytokeratin 20 (Ck20, in colorectal carcinoma). Scale bars: 500 µm (e, low magnification), 100 µm (e, high magnification), 50 µm (f, g).  Fig. 4) was cleared, stained, and imaged using a biphasic CLARITY protocol, the sample was stored in PBST. The sample was then submitted to a clinical pathology lab for frozen section processing and hematoxylin/eosin staining. (b) Hematoxylin/eosin labeling of post-CLARITY tissue reliably labels intact structures such as islets (arrows). Images were taken with a light microscope under 10x magnification. Scale bar = 50 µm.

Introduction:
Published CLARITY protocols 1-3 , designed for solid organs with fixed architectures such as the brain, result in the formation of a solid-phase gel on exterior surfaces of the tissue specimen which must subsequently be mechanically removed, potentially damaging fine structures. Furthermore, this rigid gel results in deleterious clumping, shearing, and expansion within tissue cavities, which is particularly harmful for soft and irregular tissue samples. We formulated a two-phase hydrogel consisting of a solid gel only within the tissue parenchyma itself but liquid elsewhere, including at irregular tissue and cavity boundaries. This enables fragile tissues to achieve mechanical and chemical stability -including clinical biopsies and mouse embryoswithout damage incurred from solid gel formation during the embedding process.
In addition to considerations made for soft and irregular tissues, this protocol is directly compatible with any fixed tissue sample, without the need for initial hydrogel-based perfusion or fixation, thereby making it particularly suitable for use with fixed clinical samples.

Nomenclature:
Hydrogel formulations (e.g. A4B4P4) are reported as proportions of key chemical components Acrylamide (A), Bis-Acrylamide (B), and Paraformaldehyde (P) relative to original concentrations published in Chung et al. 1 Acrylamide polymer mixtures can either form solid hydrogels or remain in liquid phase depending on the concentration of polymer components. Here we report the material phase of the hydrogel as solid or liquid when polymerized directly from solution. Liquid hydrogels will take on a solid phase within the tissue only.  Table 5: A4B0P4 Hydrogel Solution. This hydrogel forms a liquid phase when polymerized.
For animal specimens, this can be achieved via perfusion of ice-cold PBS followed by perfusion of ice-cold 4% PFA or by direct immersion fixation.
For fresh clinical specimens, this can be achieved via rinsing with ice-cold PBS followed by direct immersion fixation in ice-cold 4% PFA.
For frozen specimens, samples can be thawed either in ice-cold 4% PFA or directly in CLARITY hydrogel solution.
The volume of PFA should be in excess of sample size. For most applications, 50 mL in a falcon tube is sufficient.
After PFA fixation, samples can be stored (i.e. in PBS or cryoprotectant and frozen) indefinitely.
2. (Optional) Perform any post-fixation processing of tissue as with standard histology, for example, decalcification of bony tissue or blocking into smaller pieces.
Decalcification of mouse bones can be performed with incubation in 10-20% EDTA for 2-14 days depending on animal age and size of bone.
Antigen retrieval may be appropriately performed at this point for some epitopes. For example, recovery of BrdU labeling of nuclear DNA requires a short incubation in 0.1M HCl.
3. Post-fix tissue sample in CLARITY hydrogel at 4°C.
Gel formulation should be selected based on desired CLARITY properties (i.e. faster clearing speed vs. improved protein retention and multi-round staining).
The length of incubation period will depend on sample size. We have found a general rule to be approximately 12 hours of incubation per 1000um sample thickness at a minimum (e.g. 1 mm-thick sections incubated for 12 hours, 6 mmthick samples such as mouse brain incubated for 2-3 days).
The volume of hydrogel should be in excess of sample size. For most applications, 50 mL in a falcon tube is sufficient.
4. If using Solid Phase CLARITY hydrogel, transfer sample to Liquid Phase polymerization solution immediately prior to polymerization.
Setting up a two-phase hydrogel during polymerization enables the formation of a rigid, solid hydrogel throughout the tissue parenchyma while keeping the surrounding solution in liquid phase for easy removal. Within the tissue, where there are high concentrations of proteins to serve as additional cross-linking agents, all CLARITY formulations will form solid structures.

Solid Phase CLARITY Formulations: A4B4P4, A4B4P0
Liquid Phase polymerization solution: A1B1P4, A4B0P4, A4B0P0, PBS As the Liquid Phase solution is only transient, we have found that the selection of the specific formulation used to be a matter of user preference, and Solid Phase polymerization is robust even when performed in PBS. We recommend maximizing similarity between Solid Phase and Liquid Phase for each individual sample: i.e. using A4B0P4 for A4B4P4 gels, and A4B0P0 for A4B4P0 gels.
5. If using Liquid Phase CLARITY hydrogel, keep sample in Liquid Phase CLARITY hydrogel.
As the tissue contains high concentrations of proteins that will serve as additional cross-linkers to the hydrogel, the polymerized gel will naturally form a solid within the tissue parenchyma and remain a liquid otherwise.
6. Polymerize CLARITY Hydrogel to embed tissue specimen in polyacrylamide (Option 1): As described previously 1,3-5 , polymerization can be performed by first degassing the solution under a vacuum, and flushing the solution with nitrogen gas in order to remove oxygen, an inhibitor of the polymerization reaction. Keep under vacuum conditions for 10 minutes before quickly recapping the tube.
(Option 2): Polymerization can also be achieved through direct inhibition of oxygen via a thick layer of hydrophobic oil -such as Castor Oil (Fisher AC404165000) or Sunflower Oil (Fisher NC9967447) -on top of the liquid phase solution in the tube 6 . While simpler due to not needing vacuum or nitrogen sources, this method retains trace dissolved oxygen within the liquid solution, which can be overcome with longer polymerization times.   Table 8: Alternative Clearing Solution. 4% SDS in Phosphate buffer is also an effective clearing solution.
As previously reported 2,5 , SDS concentration can be raised to 8%, which modestly improves clearing speed. We have found that an alternative clearing solution consisting of 4-8% SDS in 1x PBS is capable of clearing tissue at similar rates.
Clearing should be performed at the appropriate temperature. Prolonged exposure to temperatures higher than ~40°C will quench endogenous fluorescent proteins 7 so should not be used when preservation of endogenous fluors are desired. A4B4P0 gels will clear rapidly at elevated temperatures (60°C) but also lose protein quickly, so careful optimization of clearing temperature will be necessary depending on the molecular marker of interest. However, other gels including A1B1P4 and A4B4P4 should be cleared at elevated temperatures when possible for maximal speed.
For human tissues, clearing at elevated temperatures with A1B1P4 hydrogel is recommended for maximal clearing speed and protein retention.  1 week -*Note, clearing of A4B4P0 tissue is not recommended at 60°C due to accelerated loss of proteins Supplemental Table 9: Approximate Clearing Times. Empirical estimates for the duration for passive clearing in Clearing Solution of 2 mm tissue slices.
12. When tissue is clear, remove from SDS Clearing Solution and wash in PBST (1x PBS + 0.1% Triton-X) 3x, at least one hour each, and once overnight at room temperature.
Clarity is most easily judged by observing a uniform translucency throughout the tissue. Partially cleared tissue will frequently have spots of greater opacity toward the center.
Thoroughly washing the sample to remove trace amounts of SDS is critical. Residual SDS will precipitate when the sample is placed in refractive index matching solutions, causing opacity and tissue damage.

Molecular Labeling
In general, CLARITY is compatible with all existing forms of fluorescence-based molecular labeling, including antibody staining, in situ hybridization*, and small molecule dyes such as DAPI or Propidium Iodide for labeling nuclei. Autofluoresence of the tissue, particularly in short wavelength channels (i.e. 488nm), combined with nuclear labeling is sufficient to provide gross structural information analogous to Hematoxylin/Eosin staining.
Due to the nature of imaging modalities used with CLARITY, it is not compatible with nonfluorescent, colorimetric stains sometimes used in thin-section histopathology with bright field imaging, such as Crestyl Violet or Diaminobenzidine-Peroxidase, as these can create imaging artifacts while casting shadows through large image volumes. However, should users desire such stains, this can be overcome with the use of well-established fluorescent analogs, including NeuroTrace (ThermoFisher) and Tyramide Signal Amplification (TSA) (Perkin-Elmer) reagents.

Antibody Staining
As with all antibody-based protocols, specific staining conditions will need to be optimized for individual probes, as each antibody has unique properties related to binding affinity, epitope specificity, and other biophysical properties that determine optimal staining conditions. We describe general best practices that have been effective for a large number of antibodies tested by our lab and others 1,3,5,10-12 .
Staining can be performed in PBST buffer at room temperature with shaking, in volumes sufficient to fully immerse the sample (typically 1-2 mL for tissue sections 2 mm or smaller), with thorough washing (at least 3x 1 hour each with one overnight wash) with PBST following both primary and secondary antibody incubations.
Antibody concentration depends on the size of the tissue and abundance of the epitope. We have found that 1:200 dilution is sufficient for even moderately abundant markers (i.e. insulin in the pancreas or parvalbumin in the brain) but higher concentrations of 1:50 or more may be required for extremely abundant markers (such as neurofilament in the brain). For large samples, replenishment of additional antibody after several days of incubation may be beneficial.
Incubation time depends on gel type and size of sample.
Recommended minimum antibody incubation time (per 1 mm thickness): • A4B4P4: 48 hours • A1B1P4: 12 hours • A4B4P0: 48 hours Unlike with thin section histology, we have found there is no additional benefit to treating tissues with a blocking reagent prior to antibody staining, with the exception of when immunolabeling CLARITY samples derived from frozen clinical specimens.

Antibody Removal
After imaging, antibodies can be disassociated from their epitopes by incubation in Clearing Solution at 60°C 1, 10 . As the antibodies are not fixed to the hydrogel backbone, the Clearing Solution will "clear" the antibody probes similar to other cleared macromolecules. While the precise incubation time required will depend heavily on antibody binding strength, a general guide is about 1 day of clearing per 500um of stained tissue. It is recommended that the user confirm the removal of antibodies by checking the tissue under an epifluorescent or confocal microscope for signs of fluorescence. If desired, the user can verify that primary antibodies are also removed by staining again with only secondary antibodies.

Subsequent rounds of Staining
Once antibodies are removed, the next round of staining can proceed exactly as indicated above (Step 1). Note that while every effort is made to minimize protein loss through overclearing, some is unavoidable; therefore it is recommended to prioritize detection of low-abundance signals in early rounds of staining, and high-abundance signals in later rounds.

Conjugated Antibodies
Our lab and others 1,3,13 have found success with the use of dye-conjugated antibodies, either commercially available (i.e. Alexa-conjugated anti-GFP, ThermoFisher A-31852) or conjugated in the lab using commercially available kits following the manufacturer's instructions (Alexa Fluor Antibody Labeling Kit -ThermoFisher A-20186).
Advantages of using conjugated antibodies include a reduced experiment duration (due to the need for only one antibody incubation rather than two) and a greater ability to multiplex (due both to the ability to simultaneously use primary antibodies from the same species without signal cross-contamination, and the ability to perform multi-round staining by quenching fluorescent signals from previous rounds rather than complete antibody removal.
Disadvantages of conjugated antibodies include the loss of signal amplification that occurs between primary antibody detection and secondary antibody detection, which can be crucial for the visualization of some low abundance markers.

Imaging
High-resolution and rapid imaging of CLARITY-processed tissue is discussed extensively in previous reports 1, 3,14 . As all images in this report were generated with a commercially available confocal microscope (Olympus FV-1100) with a standard objective lens (10x water immersion, 0.6NA, 3 mm WD, Olympus), we summarize our procedure below. b. Sample swelling Several reagents, such as Glycerol, are known to significantly swell or shrink the tissue from its original dimensions, which could be advantageous (for example, for resolving fine structural features 16 ) or disadvantageous (increased imaging time), depending on specific applications. The reagent combinations used in this study (A1B1P4 gel with RapiClear) do not significantly change tissue dimensions.
c. Ability to titrate Refractive Index Commercial reagents are optimized for tissues with refractive index of 1.45. However, the refractive index of a sample may be higher or lower depending on properties such as macromolecular content. For example, calcified bones have a refractive index of approximately 1.53 [17][18][19] . Homemade reagents are easily tuned to have optimal refractive indices for any tissue specimen, as described previously for Histodenz 5 , TDE 6 , and diatrozic acid/iodixanol 10 . It is recommended that users check the final refractive index of their solutions using a commercially available refractometer.
2. Mount sample in imaging chamber constructed from a glass slide, clay, and a coverslip or glass dish, as previously described 1,3,4 .
3. Fill imaging chamber with refractive index solution, remove bubbles and seal with Kwik-Sil (World Precision Instruments, #KWIK-SIL).
5. Analyze images using appropriate tools such as free analysis software (ImageJ), commercial 3D image analysis software (Imaris, Bitplane), or custom image processing software, depending on application.

CLARITY Extensions
Biphasic CLARITY is a modular approach that may be extended and combined with innovations from other volumetric histology methods with minimal re-engineering. Decolorization reagents from CUBIC may be used to remove heme elements from CLARITY tissues 9 , higher temperatures may be used to increase reaction kinetics 7 , antibody penetration may be improved with the use of binding and non-binding buffers 10 , and refractive index matching may be accomplished using a wide variety of affordable, commercially available reagents and home-made mixtures, including FocusClear, RapiClear, glycerol, Histodenz, 2,2'-thiodiethanol, CUBIC-2, or diatrizoic acid/iodixanol 1, 5,8,10,13,20 . Additionally, end-users can select advanced components, such as custom perfusion chambers and light sheet microscopes 2,3 , to add onto a simple CLARITY framework in a modular manner depending on specific application.