The combination of antibody labeling, optical clearing, and advanced light microscopy allows three-dimensional analysis of complete structures or organs. Described here is a simple method to combine immunolabeling of thick kidney slices, optical clearing with ethyl cinnamate, and confocal imaging that enables visualization and quantification of three-dimensional kidney structures.
Optical clearing techniques render tissue transparent by equilibrating the refractive index throughout a sample for subsequent three-dimensional (3-D) imaging. They have received great attention in all research areas for the potential to analyze microscopic multicellular structures that extend over macroscopic distances. Given that kidney tubules, vasculature, nerves, and glomeruli extend in many directions, which have been only partially captured by traditional two-dimensional techniques so far, tissue clearing also opened up many new areas of kidney research. The list of optical clearing methods is rapidly growing, but it remains difficult for beginners in this field to choose the best method for a given research question. Provided here is a simple method that combines antibody labeling of thick mouse kidney slices; optical clearing with cheap, non-toxic and ready-to-use chemical ethyl cinnamate; and confocal imaging. This protocol describes how to perfuse kidneys and use an antigen-retrieval step to increase antibody- binding without requiring specialized equipment. Its application is presented in imaging different multicellular structures within the kidney, and how to troubleshoot poor antibody penetration into tissue is addressed. We also discuss the potential difficulties of imaging endogenous fluorophores and acquiring very large samples and how to overcome them. This simple protocol provides an easy-to-setup and comprehensive tool to study tissue in three dimensions.
The growing interest in studying entire organs or large multicellular structures have led to the development of optical clearing methods that involve imaging of transparent tissue in three dimensions. Until recently, the best methods to estimate cell number, length, or volume of whole structures have been stereology or exhaustive serial sectioning, which is based on the systemic sampling of tissue for subsequent analysis in two dimensions1,2,3. However, these methods are time-consuming and need a high level of training and expertise4. Optical clearing methods overcome these problems by equilibrating the refractive index throughout a sample to make tissue translucent for 3-D imaging5,6,7.
Several optical clearing methods have been developed which fall into two main categories: solvent-based and aqueous-based methods. Aqueous-based methods can be further divided into simple immersion8,9, hyperhydration10,11, and hydrogel embedding12,13. Solvent-based methods dehydrate the tissue, remove lipids and normalize the refractive index to a value around 1.55. Limitations of most solvent-based methods are quenching of endogenous fluorescence of commonly used reporter proteins such as GFP, solvent toxicity, capacity to dissolve glues used in some imaging chambers or objective lenses, and shrinkage of tissue during dehydration14,15,16,17,18,19,20,21. However, solvent-based methods are simple, time-efficient, and can work in a number of different tissue types.
Aqueous-based methods rely on the immersion of the tissue in aqueous solutions that have refractive indices in the range of 1.38-1.528,11,12,22,23,24. These methods were developed to preserve endogenous fluorescent reporter protein emission and prevent dehydration-induced shrinkage, but limitations of most aqueous-based clearing methods include a longer duration of the protocol, tissue expansion, and protein modification (i.e. partial denaturation of proteins by urea in hyperhydrating protocols such as ScaleA2)7,11,23,25. ScaleS addressed tissue expansion by combining urea with sorbitol, which counterbalances by dehydration the urea-induced tissue expansion, and preserved the tissue ultrastructure as evaluated by electron microscopy10. Tissue shrinkage or expansion affects the absolute sizes of structures, distances between objects, or cell density per volume; thus, the measurement of size changes upon clearing of the tissue may help interpret the obtained results7,26.
In general, a protocol for optical clearing consists of multiple steps, including pretreatment, permeabilization, immunolabeling (if required), refractive index matching, and imaging with advanced light microscopy (e.g., two-photon, confocal, or light-sheet fluorescence microscopy). Most of the clearing approaches have been developed to visualize neuronal tissue, and emerging studies have validated their application in other organs5. This comprehensive tool has been previously demonstrated to allow reliable and efficient analysis of kidney structures, including glomeruli27,28, immune infiltrates28, vasculature28, and tubule segments29, and it is an ideal approach to better the understanding of glomerular function and tubule remodeling in health and disease.
Summarized here is a solvent-based method that combines immunostaining of kidney tubules; optical clearing with cheap, non-toxic, and ready-to-use chemical ethyl cinnamate (ECi); and confocal microscopy imaging that allows complete tubule visualization and quantification. This method is simple, combines antigen-retrieval of kidney slices with staining of commercial antibodies, and does not require specialized equipment, which makes it accessible to most laboratories.
NOTE: All experimental procedures described here were approved by the Institutional Animal Care and Use Committee (IACUC) of Oregon Health and Science University, Portland, Oregon, USA, and relevant local authorities in Aachen, Germany.
1. Retrograde Abdominal Aortic Perfusion and Fixation of Mouse Kidneys
2. Tissue Preparation and Immunostaining
3. Tissue Clearing
4. ConfocalImaging and Image Analysis
NOTE: For imaging, other microscopy techniques can be used as long as the refractive index matching solution is compatible with the objective lens. This protocol uses an inverted confocal microscope.
Kidneys are complex organs comprised of 43 different cell types31. Most of these cells form large multicellular structures such as glomeruli and tubules, and their function is highly dependent on interactions with each other. Classical 2-D histological techniques partially capture these large structures and may miss focal changes within intact structures31. Thus, 3-D analysis using optical clearing techniques helps to understand how they function in health and disease.
Most solvent-based optical clearing techniques will at least partially quench the endogenous fluorescent reporter protein signal. Quenching of YFP-tagged Parvalbumin was observed upon dehydration and optical clearing with ECi (Figure 1). However, other strong fluorescent proteins may resist signal-quenching, and adjustment of pH to basic levels (pH 8-11) may stabilize endogenous fluorescent proteins14,20,28,30. In addition, aqueous-based methods should be considered if fluorescent protein emission preservation is desired. In this current report, it is demonstrated that this solvent-based clearing protocol is compatible with antibody labeling; although, it remains challenging to achieve deep immunolabeling of tissue, especially in a dense cell-rich organ such as the kidney.
Retrograde abdominal aortic perfusion of the kidneys was then performed to remove blood cells and open up tubules (Figure 2A,B). This approach decreases autofluorescence and improve antibody diffusion. Antibody concentration depends on the size of tissue and abundance of the antigen. Therefore, the superficial signal can relate to either poor antibody penetration or insufficient amount of antibody (Figure 3A,B, see also Movie 1). For large samples or extremely abundant markers, higher concentrations of antibody and replenishment of antibody after 1-2 days may be required. If the antigen of interest is expressed in kidney vasculature, intravascular delivery of the antibody should be considered (Figure 3C). However, antibodies targeting proteins in the apical membrane of tubule epithelial cells do not cross the glomerular filtration barrier due to molecular size, thus causing unspecific signals in blood vessels and glomeruli (Figure 3D).
This protocol can be applied to kidney slices or whole kidneys (Figure 4A,B). To test antibody-specificity, only controls were subjected to secondary antibody (Figure 4C). The tissue can be labeled with antibodies to detect a specific cell population (e.g., proliferating cells, Figure 4D) or to visualize whole tubule segments using segment-specific antibodies (Figure 4E). Moreover, the combination of multiple antibodies provides the opportunity to colocalize different proteins in 3-D (e.g., to detect segment-specific tubule hyperplasia as it occurs in tubule remodeling upon different stimuli) (Figure 4F,G; also see Movie 2).
Figure 1: Quenching of endogenous fluorescent reporter protein signal. (A) The endogenous fluorescent reporter protein derived from ParvalbuminYFP+transgenic mouse is detectable in 5 μm thin PFA-fixed frozen kidney sections. The arrows mark some fluorescent-labeled parvalbumin+ cells located in the early part of the distal convoluted tubule. (B) No evident fluorescent signal after the optical clearing of kidney tissue is observed. Please click here to view a larger version of this figure.
Figure 2: Evaluation of the kidney perfusion quality. (A) Periodic Acid Schiff (PAS)-stained paraffin embedded tissue (5 μm thin sections) shows few tubules with a slightly opened lumen (typically proximal tubules which possess brush border; see arrows). (B) All tubules are opened up after good kidney perfusion. Please click here to view a larger version of this figure.
Figure 3: Troubleshooting whole-mount immunolabeling. (A,B) A thick kidney slice was stained with an antibody against sodium-chloride cotransporter-2 (NKCC2), which is expressed in the thick ascending limb of the loop of Henle. The signal is detectable at the surface of the tissue (arrows in (B)). However, the antibody did not penetrate into the tissue [see arrowhead in (B)]. (C) Intravascular antibody injection targeting proteins expressed in the vessels (e.g., CD31) allows fast and homogenous blood vessel staining. The round structures with a strong signal are glomeruli. (D) However, antibodies targeting proteins expressed in the apical membrane of tubule epithelia (e.g., phosphorylated sodium-chloride cotransporter (phospho-NCC) which is expressed in the distal convoluted tubule) will not cross the glomerular filtration barrier, thus causing unspecific signals in the vasculature (see arrowhead pointing at afferent arterioles and other vessels) and glomeruli (see arrows). Please click here to view a larger version of this figure.
Figure 4: Representative results of immunolabeled optically cleared kidney tissue. (A,B) This protocol allows optical clearing of the whole kidney. (C) A representative z-stack to show absence of non-specific binding of secondary antibody as control without prior incubation. (D) 3-D visualization of a representative z-stack shows proliferating bromodeoxyuridine (BrdU)+ cells in kidney medulla. (E) Medullary collecting ducts are visualized using an antibody against aquaporin-2 (AQP2). (F,G) Analysis and quantification of BrdU+ cells within AQP2+ medullary collecting ducts. For (F,G), see also Movie 2. This figure has been modified from Saritas et al.29. Please click here to view a larger version of this figure.
Movie 1 (Related to Figure 3): 3-D visualization of NKCC2+ thick ascending limb of the loop of Henle. The movie demonstrates poor antibody penetration in an optically cleared kidney slice. Please click here to download the video.
Movie 2 (Related to Figure 4): 3-D visualization of BrdU+ cells and AQP2+ medullary collecting ducts in an optically cleared kidney slice. BrdU+ cells (in red) and AQP2+ medullary collecting ducts (in green) are shown. Imaris spot and surface algorithms were used to determine BrdU+ cells outside (turquoise spots) or inside (pink spots) AQP2+tubules (green surfaces). This movie originally appeared in Saritas et al.29. Please click here to download the video.
Optical clearing techniques have received wide attention for 3-D visualization and quantification of microanatomy in various organs. Here, solvent-based clearing method (ECi) was combined with immunolabeling for 3-D imaging of whole tubules in kidney slices. This method is simple, inexpensive, and quick. However, other research questions may be best answered with other clearing protocols5. It is also important to keep in mind that solvent-based methods cause tissue-shrinkage at variable degrees, mainly due to the dehydration step14,18. Most solvent-based methods (e.g., ECi14) also at least partially quench endogenous fluorescence of reporters such as GFP or tdTomato; thus, FDISCO32, CLARITY12, or CUBIC11 may serve as alternative protocols. However, the quenching effect of ECi is variable and depends on individual constructs of each reporter mouse28. In addition, use of 1) a fluorescent antibody against GFP and other reporters or 2) modified solvent-based protocols that preserve endogenous fluorescence signals30,32 can also be an alternative approach in this context. It is worth mentioning that several groups have tested the compatibility of aqueous-based protocols24,33,34 to visualize RNA in 3-D, while solvent-based clearing methods have not yet been tested. Thus, aqueous-based clearing methods such as the modified version of CLARITY33 should be preferred when RNA analysis is considered.
Cells or gene products of interest can be visualized using transgenic mice with endogenous fluorescent protein expression, but the generation of genetically engineered mouse lines is time-consuming and expensive. Therefore, antibody labeling is more practical and provides more flexibility, although the immunolabeling of large tissue is challenging. In contrast to the original ECi protocol14, our approach combines a modified ECi optical clearing method and immunolabeling, which uses an antigen-retrieval step. Heat-induced antigen-retrieval will denature proteins, but helps to recover loss of antigenicity during PFA-fixation35 and improves antibody-binding.
There are a few critical steps in this protocol. First, it is important to perform good perfusion of the kidney. Hemoglobin containing red blood cells limits the imaging depth7 and expanded tubules with open lumen enhance antibody penetration. The opening of tubules also allows better distinguishing of proteins expressed in the epithelial apical membrane from other apically expressed proteins on the contralateral side. However, artificially expanded tubules may negatively influence tubule structure and mask specific pathophysiological response of the kidney (e.g. dilation of injured proximal tubules,) but not of healthy tubules36. Second, antibody incubation at 37 °C rather than 4 °C or RT improves antibody penetration27. However, each antibody has unique properties; thus, temperature during antibody incubation needs to be optimized for individual probes. Third, the use of Alexa Fluor-647 (far-red spectrum) secondary antibodies helps increase the signal-to-noise ratio, especially since kidneys emit a large amount of autofluorescence in the blue-green spectrum37.
Retro-orbital injection14 or perfusion of kidney38 with fluorophore-conjugated antibodies against proteins expressed in blood vessels is an elegant and fast way to label kidney vasculature. However, proteins in the apical membrane of tubules remain inaccessible by intravascular injection since antibodies cannot cross the glomerular filtration barrier with its cut-off molecular weight for filtration of proteins in the range of 60-65 kDa. Therefore, the use of small antibody fragments and engineered variants such as Fab fragments (~55 kDa), diabodies (~50 kDa), tandem scFv (~28 kDa) or nucleic-acid aptamers (~6-30 kDa) with preserved molecular recognition properties of antibodies may provide an opportunity to access the apical membrane of tubules38,39. In addition, the combination of small antibody fragments with electric fields40 or pressure13 or the use of sodium dodecyl sulfate (SDS)-based clearing protocols to remove lipids24,41,42 may enable fast and efficient antibody penetration of tissue.
Several groups demonstrated that perfusion with clearing reagent not only reduces the protocol time, but also increases tissue transparency19,24,43,44. Therefore, cannulation of the abdominal aorta or the renal artery with subsequent perfusion of the kidney with dehydrating and refractive index matching solutions should be considered to achieve better and faster tissue clearing. However, we did not perfuse the whole kidney with clearing reagents and used sliced kidneys for several reasons. First, different antigens can be visualized by antibody labeling of multiple kidney slices from one kidney. Second, less time and antibody are needed to perform immunolabeling of a kidney slice compared to a whole kidney. Third, 3-D imaging of large samples generates data sets up to several terabytes, which can be challenging to manage for most workstations. Therefore, data from tissue slices or subsets of bigger files are more convenient to perform complex operations such as automated counting or distance measurements.
In this protocol, confocal microscopy is used to perform imaging with single-cell resolution, which is particularly relevant in colocalization studies. However, confocal imaging requires laser scanning, since its imaging speed is only practical for small pieces of cleared tissue. To perform 3-D morphometric analysis of multicellular structures such as long tubule segments or even whole organs, faster microscope techniques such as light sheet fluorescent microscopes (LSFM) are necessary. LSFM allow fast imaging when the highest cellular resolution is not essential. For example, the lengths of distal convoluted tubules were recently assessed by combining whole-mount immunolabeling, optical clearing based on CLARITY12, and LSFM29. Unfortunately, commercial LSFM is expensive and not always compatible with solvent-based clearing protocols. In fact, CLARITY was chosen for this study, since our particular LSFM with an objective customized for CLARITY was not compatible with ECi29. However, Klingberg et al. demonstrated that ECi is in principle compatible with LSFM14.
In conclusion, a simple ECi-based optical clearing method is demonstrated, which can be applied to any research project using fixed tissue slices ranging from ~100 μm to several millimeters in thickness. It also allows feasible analyses that previously required almost exhaustive efforts to complete, and eliminates the required assumptions and inferences associated with two-dimensional analysis of morphology. The combination of whole-mount immunolabeling, optical clearing, and advanced light microscopy will help advance the understanding of cellular function in health and disease.
The authors have nothing to disclose.
T. S. is supported by grants from the DFG German Research Foundation (332853055), Else Kröner-Fresenius-Stiftung (2015_A197), and the Medical Faculty of the RWTH Aachen (RWTH Returner Program). V. G. P. is supported by research fellowships from Deutsche Gesellschaft fur Nephrologie, the Alexander von Humboldt Foundation, and the National Health and Medical Research Council of Australia. D. H. E is supported by Fondation LeDucq. R. K. is supported by grants from the DFG (KR-4073/3-1, SCHN1188/5-1, SFB/TRR57, SFB/TRR219), the State of Northrhinewestfalia (MIWF-NRW) and the Interdisciplinary Centre for Clinical Research at RWTH Aachen University (O3-11).
0.22 µm filter | Fisher Scientific | 09-761-112 | |
15 mL conical tube | Fisher Scientific | 339650 | |
21 gauge butterfly needle | Braun | Venofix | |
3-way stopcock | Fisher Scientific | K420163-4503 | |
3D analyis software | Bitplane AG | IMARIS | |
3D analyis software | Cellprofiler | free open-source software | |
5-0 silk suture | Fine Science Tools | 18020-50 | |
50 ml plastic syringes | Fisher Scientific | 14-817-57 | |
Anti-BrdU monoclonal antibody | Roche | 11296736001 | |
Antibody diluent | Dako | S0809 | |
CD31-647 | BioLegend | 102516 | |
Citrate-based antigen retrieval solution | Vector Laboratories | H-3300 | |
curved hemostat | Fisher Scientific | 13-812-14 | |
Dako Wash Buffer | Agilent | S3006 | |
dissecting microscope | Motic | DSK-500 | |
Embedding cassettes | Carl Roth | E478.1 | |
Ethanol | Merck | 100983 | |
Ethyl cinnamate | Sigma-Aldrich | 112372 | |
Flexible film/Parafilm M | Sigma-Aldrich | P7793 | |
Goat anti-AQP2 | Santa Cruz Biotechnology | sc-9882 | |
Guinea pig anti-NKCC2 | N/A | N/A | DOI: 10.1681/ASN.2012040404 |
HCl | Carl Roth | P074.1 | |
Heparin | Sagent Pharmaceuticals | 401-02 | |
hemostat | Agnthos | 312-471-140 | |
horizontal rocker | Labnet | S2035-E | |
Imaging dish | Ibidi | 81218 | |
Ketamine | MWI Animal Health | 501090 | |
Micro serrefine | Fine Science Tools | 18052-03 | |
NaOH | Fisher Scientific | S318-500 | |
Operating scissors | Merit | 97-272 | |
Paraformaldehyde | Thermo Fischer Scientific | O4042-500 | |
Rabbit anti-phoshoThr53-NCC | PhosphoSolutions | p1311-53 | |
Silicone elastomer | World Precision Instruments Kwik-Sil | KWIK-SIL | |
Sodium azide | Sigma-Aldrich | S2002 | |
Tissue slicer | Zivic Instruments | HSRA001-1 | |
Triton X-100 | Acros Organics | AC215682500 | |
Vannas scissors | Fine Science Tools | 15000-00 | |
Vibratome | Lancer | Series 1000 | |
Xylazine | MWI Animal Health | AnaSed Inj SA (Xylazine) |