The pathologic examination of biopsied or surgically resected tissue is central to advancing our understanding of disease mechanisms, establishing diagnoses, determining prognoses, and guiding decisions for treatments. Pathology examination can take place at the gross, microscopic, and molecular levels. For microscopic examination, most tissues are routinely formalin-fixed and paraffin-embedded. These tissue blocks are then thinly sectioned, stained with hematoxylin and eosin (H&E), and mounted on glass slides for examination using a standard light microscope. Although these thin H&E sections are immensely efficient and powerful tools for discovery and patient care, we have forgotten that they are extraordinarily thin, and therefore incomplete, representations of significantly larger pieces of tissue. As a result, lesions with complex and informative 3-dimensional (3D) anatomy are incompletely conceptualized as flat 2D architectures.
This 2D histology does not completely represent important characteristics of 3D lesions and can be misleading. For example, in the cancer biology field the budding of neoplastic cells at the leading edges of infiltrating adenocarcinomas had been interpreted in 2D to be the migration of single detached cancer cells, while, in fact, 3D reconstruction of consecutive sections of slides immunohistochemically labeled for cytokeratin revealed that in many instances these cells that appeared isolated in 2D were, in fact, 3D branch-like projections of collective cancer cells.1 Perhaps more importantly, 3D relationships are lost in thin H&E sections. Cross sections of perineural invasion are easily appreciated in 2D sections in cancers, but in 2D one cannot define the point at which the cancer invaded the nerve. Histology viewed in 2D also leaves some questions unanswered. For example, immune cells are often appreciated in the neighborhood of neoplastic cells, but their distribution in 3D is unclear in 2D sections. Are they randomly scattered or is there a 3D architectural organization? As pathologists have viewed sections in 2 dimensions for so long, we have forgotten to think in 3 dimensions and to ask important architectural questions.
3D imaging has been adopted in other specialties, often with stunning results. 3D ultrasonography, 3D computed tomography, and 3D magnetic resonance imaging, have not only fundamentally changed clinical care, but they have also improved our understanding of diseases.2–4 For example, 3D computed tomography of the pancreas can reveal the relationship of cancers of the head of the gland to major vessels such as the superior mesenteric artery and vein.5,6 Similarly, 3D organoid cell culture systems of murine and human pancreata have recently been generated to recapitulate the characteristics of infiltrating pancreatic ductal adenocarcinomas better than flat 2D-cultured cell lines.7 3D visualization of 3D pancreatic lesions has great potential to expand our understanding of a broad variety of pathologies.
Recently, several groups have examined mouse and human pancreata in 3D using clearing and the results have been eye-opening.8–12 For example, as discussed in greater detail later, our group applied clearing with immunolabeling for cytokeratin 19 to a series of surgically resected human pancreatic cancers and showed that neoplastic glands of invasive cancer often track in the connective tissue parallel to muscular arteries.12
Clearing with 3D histopathologic evaluation of normal or pathologic lesions will provide new insights into the pathogenesis of diseases, and may provide an understanding of the 3D basis for diagnostic features appreciated in 2D.13 In this review, we will discuss tissue clearing, and 3D microscopic examination with emphasis on the application of these technologies to the pancreas.
Tissues are composed of a diverse mixture of cells, connective tissues, and fluids. Water is the most prominent fluid, and the structural components of cells are predominantly made of lipids and proteins. Each of these tissue components has a different refractive index (RI), and when light waves pass from a medium with one RI into a medium with a different RI, the light waves bounce (reflect) and bend (refract).14 It is this irregular scattering, as well of absorption and reflection, of light as it passes through tissues that makes tissues opaque.
More specifically, RI is defined by ratio of the speed of light when light passes through a substance relative to the speed of light when it passes through a vacuum. For example, the RI of air is 1, that of water is 1.33, and that of the oil is 1.52.15 As light passes from water rich tissues with a relatively low RI into lipid rich structures, such as cell membranes, with a higher RI, light will bend and scatter, creating optical nontransparency.16
Tissue clearing techniques work by creating tissues with a uniform RI which reduces the scattering light.16 For example, the hydrogel embedding method creates tissues with uniform RIs in the range of 1.33 to 1.47, simple emersion methods create tissues with uniform RIs in the range of 1.42 to 1.52, and solvent-based methods create tissues with uniform RIs of 1.56 to 1.57.17 The scattering of light is reduced in these tissues with uniform RIs, and when combined with the removal of pigments to reduce the absorption of light, the tissues become transparent.
Although this all sounds very new, tissue clearing has actually been around for almost a century. Tissue clearing was first introduced by the German anatomist Werner Spalteholz in 1911.18 In order to study the coronary arteries, Spalteholz created transparent muscle tissue by sequentially soaking the tissues in alcohol, clover oil or xylene, and Canada balsam.18 He finally established clearing solutions that included methyl salicylate, benzyl benzoate, and even wintergreen oil.18 Over the ensuing decades, many different methods have been developed to clear tissues. The techniques can be divided into solvent-based clearing and aqueous-based clearing. Prototypes of solvent-based clearing techniques are benzyl alcohol, benzyl benzoate (BABB), an improved version of the clearing method used by Spalteholz, 3D imaging of solvent-cleared organs (3DISCO), immunolabeling-enabled 3DISCO and ultimate 3DISCO.19–22 The aqueous-based clearing techniques can be subdivided in techniques that use only simple immersion and techniques that use a hydrogel. The prototypes of the simple immersion technique are FocusClear, 2, 2′-thiodiethanol, sucrose, ClearT, and see deep brain (SeeDB).23–27 These techniques do not remove lipids, but replace the liquid in and around the tissue with a high RI, water-based solution. Although Scale and clear unobstructed brain imaging cocktail and computational analysis (CUBIC) are also aqueous-based clearing techniques, they also exploit the removal of lipids from the tissue and the use of urea to penetrate, partially denature and thus hydrate the hydrophobic regions of high RI proteins, to decrease the overall RI of the tissue.28,29 Hydrogel-based clearing techniques immobilize the proteins in an acrylamide-based matrix and clear the tissue through the complete removal of all lipids from the tissue. Prototypes of this method are clear lipid-exchanged acrylamide-hybridized rigid imaging/immunostaining/in situ hybridization–compatible tissue hydrogel (CLARITY), passive clarity techniques and perfusion-assisted agent release in situ.30,31 Each of these methods have their own strengths and weaknesses which have been well summarized elsewhere.16,17,32,33 Although many different clearing methods have been developed, only a few of these methods have been applied to the pancreas, and even fewer to the dense fibrotic tissues that characterize human pancreatic cancer.10–12,34–40 Table 1 summarizes clearing methods that have been reported to clear pancreatic parenchyma.
Solvent-based clearing methods typically use methanol, with or without hexane, or tetrahydrofurane as extra dehydrant.16,20 These reagents remove water and lipids from the tissues, resulting in homogenously dehydrated protein-rich dense samples with an average RI, higher (>1.5) than those of water and lipid.16 More of the remaining lipids can be removed and the tissues further RI matched using a second step that includes BABB, dibenzyl ether, or methyl salicylate.16–18 There are several short comings to these solvent-based clearing approaches: first, since water is essential for maintaining emissions from fluorescent protein chromophores, these solvent-based methods, which remove water, result in the rapid quenching of fluorescent signals.16 Second, these solvent-based methods, again because they remove water, can cause significant tissue shrinkage.30
Aqueous-based methods have also been widely used to clear pancreatic parenchyma, especially normal tissues that are less fibrotic than are cancers.10,11,35–40 These aqueous-based methods include SeeDB, CUBIC, and CLARITY, and they have the advantage of greater preservation of the fluorescent signals.27–29 These methods use 1 of 3 approaches, (1) simple immersion of specimens into a water-based solution with a higher RI (1.42 to 1.56), similar to the RI of the lipids and proteins, (2) partial removal of lipids in combination with the hydration of the hydrophobic regions of proteins (hyperhydration) to lower the RI, or (3) hydrogel embedding in combination with the complete removal of lipids.16 Aqueous-based methods have some limitations including the requirement of electrophoresis devices, constant perfusion, or high temperatures,42–45 which may prevent complete tissue clearing of dense fibrotic pancreatic parenchyma (desmoplastic reaction).
Although aqueous-based methods may be ideal to clear normal tissues, including normal pancreas tissue, these techniques may not be effective with abnormal tissues with dense fibrosis. Therefore, solvent-based methods may be preferred with dense fibrotic tissues. For example, as discussed in detail below, we recently reported a modified iDISCO method to clear surgically resected human pancreas tissue and showed that this method could be used to visualize a number of pancreas lesions, including pancreatic intraepithelial neoplasias (PanINs) and intraductal papillary mucinous neoplasms (IPMNs), and even infiltrating pancreatic ductal adenocarcinomas.12
Irrespective of which specific approach is used, tissue clearing techniques, as noted above, can be thought of as having 4 basic steps: (a) pretreatment of specimens, (b) permeabilization and/or removal of lipid (delipidation), (c) immunolabeling (antibody penetration), and (d) clearing (RI matching) steps.17 There are some exceptions, for example the hydrogel-based techniques, where the clearing step is performed before the immunolabeling step.
Pretreatment steps are used for various reasons in the different protocols. In general, during this step, pigments are removed from the specimens as pigments may absorb light. Hydrogen peroxide (H2O2) is often used, because it reduces the internal pigments of the tissue and reduces autofluoresence of other proteins by oxidizing them.21,41 Recently, N-butyldiethanolamine has also been shown to be a potent agent for decolorization.46
In the hydrogel-based methods, pretreatment is used to introduce acrylamide as a monomer into the tissue. This is carried out in combination with formaldehyde, to crosslink proteins, nucleic acids, and other small molecules to the monomer. Eventually, the monomer is polymerized at 37°C to form the hydrogel that keeps the bound molecules in place during delipidation.
Other agents can also be applied during pretreatment, depending on the tissue type. Bones can be demineralized with EDTA.47
Permeabilization and/or Removal of Lipid (Delipidation)
A permeabilization step can help enhance the penetration of the clearing solution and the penetration of antibodies into tissue specimens. Lipid removal (delipidation) also contributes to tissue clearing, because lipids usually contribute to the inhomogeneity of RIs in biospecimens. However, lipid removal, when extreme, will also result in the removal of membrane proteins. Hydrogel-based methods use the hydrogel to keep the membrane proteins in place allowing for more extreme delipidation and tissue clearing.30 Permeabilization and delipidation typically are accomplished using solvents (dichloromethane, methanol, tertiary butanol, tetrahydrofuran, dimethyl sulfoxide), sorbitol, urea, or detergents (sodium dodecyl sulfate, saponin, or triton X-100).16
Immunolabeling (Antibody Penetration)
Cell membranes are significantly more porous after lipids have been partially removed, facilitating the delivery of antibodies inside of the cells. Traditional immunolabeling of 5-µm sections depends on the simple diffusion of antibodies. However, antibody diffusion into the thick slabs of tissue used in clearing is a slow process and can take several weeks.30 Several methods have therefore been developed to improve antibody penetration in thick slabs of tissue.
Kim et al42 devised a stochastic electrotransport system which enhances antibody penetration without affecting the low-electromobility molecules and which can be used together with CLARITY. They further used a rotational electric field to selectively deliver antibodies throughout porous samples.42
Centrifugal forces applied with a bench-top centrifuge at a speed of 600 rcf have also been shown to improve the delivery of antibodies deep into tissue.43 A similar effect can be achieved with the application of convectional flow forced by placing tissues and antibodies into the chamber of a syringe and then pumping the syringe under pressue.43
An alternative method, system-wide control of interaction time and kinetics of chemicals (SWITCH), has been introduced to improve antibody distribution.44 In SWITCH the binding of antibodies is suppressed until antibodies reach even distribution throughout the tissue, then antibodies are “SWITCHED ON” to bind their target epitope using pH-dependent reactivity.44
Clearing (Refractive Index Matching)
As discussed earlier, in this step the RI of the specimens is homogenized using a variety of solutions, producing transparent specimens. The RI achieved with the hydrogel embedding method ranges from 1.33 to 1.47, while of the RIs achieved with simple emersion methods range from 1.42 to 1.52. The RIs with solvent-based methods range from 1.56 to 1.57.17
Although conventional light microscope is good for the evaluation of thin 2D H&E slides, it is not suitable for evaluation of 3D architecture in thick slabs of cleared tissue. Therefore, light sheet microscopy, confocal microscopy, or 2-photon microscopy is used to evaluate the 3D histology of cleared samples. Familiarity with a few terms is needed before we discuss these different types of microscopes. The working distance of microscope is defined by the distance from the front edge of the microscope objective to the plane at which an area of the sample to be studied is in focus.15 The numerical aperture of a microscope defines the ability of the objective to capture sample details and how much light enters the objective.15 In general, long working distance (>5 mm) and high numerical aperture (>0.9) is recommended for evaluating the 3D histology of cleared tissues.15,16
Light Sheet Microscopy
Light sheet microscopy was developed to enhance the 3D visualization of large specimens.48 Light sheet microscopy can be used to visualize volumetric specimens ranging in size all the way from <1 to up to 15 mm.49 Light sheet microscopy works by optical sectioning with 2 thin counter-propagating sheets of LASER beams that are perpendicular to the focal plane of the objective.50 The objective collects fluorescent light which is emitted only from this thin optical section. 49 3D reconstruction of the volumetric specimens can then be obtained by serial stacking of the images in the axial dimension.50 Schematic illustration of how light sheet microscopy works to visualize 3D pancreatic tissues is depicted in Fig. 1.
Light sheet microscopes tend to cause less photobleaching effect and can produce 3D images that cover larger volumes at higher speeds than can confocal microscopes.15 Another strength of light sheet microscopy is that it can be applied to large specimens as long working distances can be obtained when the objective is immersed into the specimen chamber medium.49 However, lenses with longer working distances tend to have lower resolution.15 As described below, confocal LASER microscopy can be used to obtain higher resolution images when needed.
Confocal (LASER Scanning) Microscopy
Confocal scanning microscopy has been used to characterize structural and cytologic details based on excitation of endogenous or exogenous fluorophores with or without the use of a LASER. Confocal LASER scanning microscopy works as a computer-assisted epifluorescent microscope, which rejects much of the out-of-focus fluorescent light. A LASER light is precisely focused to illuminate a highly delineated spot in the specimen. This reduces the fluorescent light emitted by tissues outside the focal plane of the objective lens.50 At each illuminated spot, fluorescent light emitted by the tissue is separated from the incident exciting LASER light using a dichroic mirror. In order to be detected the light also has to pass through a confocal pinhole, which rejects fluorescent light generated outside of the focal plane of the objective, and therefore most of the unwanted light scattered by the specimen is not seen.50 Images are created by scanning diffraction limited spots in a raster pattern across the specimen with 2 mirrors mounted on coordinating galvanometer motors.50 The light passing through the confocal pinhole is detected with a photomultiplier tube, and the signal is processed by a computer to create a digital image.50 Schematic illustration of how confocal LASER microscopy works to visualize 3D pancreatic tissues is depicted in Fig. 1.
Confocal microscopy can obtain near-real time images with histologic resolution.51 Historically the imaging depth of confocal LASER scanning microscopy has been limited by the working distance of the objectives of the microscope, which typically do not exceed 4 mm.52 However, microscope makers have recently developed customized objectives for use with specific clearing reagents. For example, Leica sells customized microscope specific for CLARITY, Olympus provides objectives for CLARITY, Scale, and SeeDB, and Zeiss makes objectives for several clearing reagents. The working distance for these specialized objectives is now to 5.6 to 8 mm,15 theoretically allowing for the evaluation of 3D histology of tissues with up to 8 mm in thickness.
Two-photon (LASER Scanning) Microscopy
Two-photon LASER scanning microscopy uses long wavelength near-infrared LASERs, which can reduce light scattering and increase depth of penetration.53 High-power pulsed LASERs are restricted to the focal plane resulting in excitation of 2 lower energy photons, which cause optical sectioning.54 The lower energy 2 photons reduce phototoxicity and bleaching compared with confocal LASER scanning microscopy.54 Imaging depth reach up to 1 mm.54 Therefore, 2-photon microscopy can only be used for the evaluation of 3D histology of small sized tissues that are 1 to 2 mm in thickness.
Recently the successful clearing of pancreatic tissue has been reported by several groups.10,12,34,35,37–39,41 Most of the previous studies have focused on clearing normal (nonfibrotic) pancreas and on imaging the endocrine components of the gland.8,11,34,35,37–39 There have been only a few studies of the exocrine pancreas.12 Previous studies that have applied clearing to the pancreas are summarized in Table 1.
Most of the previous studies that applied tissue clearing to the pancreas used the mouse pancreata, since the mouse pancreas is small and therefore more easily cleared and imaged.34 Kim et al34 characterized green fluorescent protein (GFP)-tagged beta-cell distribution in mouse pancreas after tissue clearing with sucrose. Fu et al35 reported islet microstructures and the vasculature of the mouse pancreas cleared with FocusClear. The same group later reported streptozosin-treated injured islets had Schwann cell and pericyte plasticity.38 Tang et al37,39 cleared mouse and human pancreata and identified a pancreatic neurovascular network and the coupling of ganglionic and islets via the network.
Wong et al11 applied a simplified CLARITY method that used acrylamide-free sodium dodecyl sulfate–based tissue delipidation to mouse pancreas tissue. With this modified technique, they demonstrated differences in the distribution of beta-cells between wild type and mouse insulin 1 promoter GFP mice, the latter of which express enhanced green fluorescent protein (EGFP) under the control of the mouse insulin 1 promoter.11
Tainaka et al41 cleared the whole body of a mouse, including the pancreas, using CUBIC-perfusion techniques, and demonstrated the superiority of CUBIC-perfusion over SeeDB methods in clearing techniques. Lee et al40 also cleared the whole body of a mouse, but they used the CLARITY technique. They observed that the optimal electrical current conditions for the clearing of different mouse organs vary based on the consistency of the organs, and found that 250 mA was the best condition for clearing the mouse pancreas with the CLARITY technique.40
There have only been a handful of reports of clearing applied to the human pancreas.8,12,37 These studies have provided novel insights into human diseases not appreciated in conventional 2D histology sections. Tang et al37 cut surgically resected human pancreas to 350-μm thick sections using a vibratome, cleared the sections with RapiClear clearing solution, and observed the endocrine pancreas in these sections. They were able to demonstrate subtle 3D relationships between the vasculature and the endocrine pancreas.37 They observed α-cell distributions in the core and mantle layer of islets of Langerhans, and showed that as arterioles entered the core of the islets they break into capillaries. They also mapped the distribution of sympathetic nerve fibers along the islet microvasculature.37 They found that condensed nerve fiber networks encircle the islets of Langerhans, and showed connections of nerve fibers from islets with microvasculature and ganglia.37 In a separate study Fowler et al sectioned pancreatic tissue at up to 5 mm of thickness, and cleared the sections by immersing them in fructose and thioglycerol.8 Their observations were similar to those of Tang and colleagues. They demonstrated large islet vessels branching off into small capillaries.8 Using 3D histologic evaluation, they were able to capture larger numbers of islets, which increased the statistical power of their observations.8
We applied a modified iDISCO method to clear thick slabs of grossly normal human pancreas parenchyma (up to 1.5-mm thick) and to slabs of human pancreatic ductal adenocarcinoma (0.6 mm in thickness) obtained from surgically resected specimens and the results were dramatic (Fig. 2A).12 Normal pancreatic tissue, intraductal precursor lesions of invasive cancer, including PanINs and IPMNs, and pancreatic ductal adenocarcinomas were all successfully cleared and visualized in 3D.12 This allowed us to visualize features in 3D which were difficult to appreciate in previous thin 2D sections.
In normal pancreatic parenchyma, we observed the stunning 3D branching of the pancreatic duct system, as well as regularly distributed small periductal glands extending off of larger normal ducts.12 We observed that some ductules, rather than separating as they branched as branches of a tree would, instead looped back to apparently rejoin the duct system (Supplemental Video 1, Supplemental Digital Content 1, http://links.lww.com/PAP/A23). Secondary changes of focal ductal obstruction were observed, including focal dilation of the lumina and acini in a lobule producing a ball of grapes appearance.12 These cleared human pancreata could also be used to visualize, in 3D, the architecture of both PanINs and IPMNs (Supplemental Video 2, Supplemental Digital Content 2, http://links.lww.com/PAP/A24). The full size and shape of these lesions could be determined when the lesions were fully contained in the slab of tissue studied, and the intraluminal papillary projections were beautifully visualized.12 When invasive cancers were cleared, the cancer cells were haphazardly arranged in 3D, forming, in some cases, sheets of mesenchymal appearing cells, and in others irregular-shaped tubular structures with numerous blunt-ended projections (Fig. 2B). Invasive cancer growing into preexisting ducts (cancerization of ducts) was visualized in 3D.12 In contrast to the nice uniform papillae of PanIN lesions, the invasive cancer cells in cancerization of ducts irregularly grew across the duct lumina, forming jagged bridges that spanned the lumina. The use of several different antibodies and fluorophores allows for the visualization of the expression of multiple proteins in 3D (Figs. 2C, D).
The 3D visualization of cleared human pancreatic cancer highlighted invasive carcinoma growing within blood vessels and a unique finding not fully appreciated in 2D sections. Long tubules of neoplastic cells were visualized growing in the surrounding connective tissue parallel to blood vessels. Although we labeled pancreatic tissues with cytokeratin 19, the blood vessels could also be detected by the distinct autofluorescent signal generated by their tortuous elastic lamina (not all autofluoresence is bad!). The finding of long tubes of invasive cancer growing parallel to blood vessels helps explain the 2D clinical finding that a gland next to a muscular vessel supports the diagnosis of cancer.55 It also suggests a mechanism for this—there may be something structural in the perivascular connective tissue, perhaps a weak tissue plane, or small channels in the connective tissue, that leads to the preferential growth of neoplastic cells in this location.12,56,57
The confluence of advances in microscopy and tissue clearing has opened the door to the 3D visualization of human anatomy and diseases, and provides an extraordinary opportunity to advance our understanding of the true complexity of human diseases.33,36,58 Several questions remain to be asked, and several opportunities are yet to be explored.
What is the maximal thickness of pancreatic parenchyma that can be visualized with current technologies? Using human exocrine pancreatic tissues, we evaluated up to 1.5-mm thick slabs of normal pancreas parenchyma and 0.6-mm thick slabs of pancreatic ductal adenocarcinoma sections. 12 Fowler et al8 reported that they were able to evaluate up to 5-mm thick sections of normal pancreas parenchyma. Considering the working distance for the specialized objectives used is currently as high as 8 mm, one could theoretically visualize sections as thick as 16 mm (flipping the piece over one can visualize 8 mm from each surface) with confocal LASER scanning microscopy.15 With a light sheet microscope, one could go even deeper, but the cost would be that only lower power magnification would be available.49 Future technical advances with development of advanced microscopic techniques may extend the limit of lesions that can be visualized in 3D.
Can antibody penetration be improved? Antibody penetration in thick pancreatic tissues, especially dense fibrotic desmoplastic stroma of pancreatic ductal adenocarcinomas, is a significant hurdle in the 3D study of these tissues. Several strategies can be applied to improve antibody penetration. Single-domain antibodies (Fab′)2 or nucleic acid aptamers, which are smaller than standard antibodies, can penetrate thick pancreatic tissue may reduce time for labeling.16 In addition, forces, such as centrifugal, pressure, or electrophoretic fields, can facilitate antibody penetrations.40,42,43 We applied single-domain antibodies (Fab′)2 and also used centrifugal forces to enhance antibody penetration and to reduce needed incubation times.12 New technologies that improve antibody penetration will allow even more to be seen.
The application of multicolor immunolabeling to cleared pancreatic tissues8,39 provides unique opportunities to evaluate the 3D relationships of normal ducts, precursor lesions, invasive carcinoma, immune cells, islets of Langerhans, vessels, and nerve fibers (Supplemental Video 3, Supplemental Digital Content 3, http://links.lww.com/PAP/A25). Fowler et al8 demonstrated immunolabeling with 8 different primary antibodies using conjugated primary antibodies with a combination of fluorophores. By using direct conjugation of fluorophores with probes may reduce the incubation time for secondary antibodies.16 In addition, chances of nonspecific antibody bindings between multiple primary and secondary antibodies may be reduced. However, longer exposure times may be needed to visualize the weaker signals generated by conjugated primary antibodies. In contrast, Tang et al39 showed immunolabeling with 5 different nonconjugated primary antibodies from 2 different species, including guinea pig and rabbit. Although both groups used different methods, they wonderfully demonstrated multiple immunolabeling in endocrine pancreas. The potential here is clearly great.
The 3D visualization of cleared human pancreas will provide an extraordinary opportunity to understand the true complexity of human pancreatic lesions in many aspects. First, some genetic mutations are associated with changes in protein expression that can be visualized in 3D, potentially allowing for the 3D visualization of the patterns of driver gene mutation in precursor lesions and in invasive cancers. This 3D visualization may provide novel insight into pancreatic ductal adenocarcinoma tumorigenesis.
Recently developed 3D organoid systems of pancreatic ductal adenocarcinomas can be used as powerful research tools for drug screening and evaluation of new biomarkers for pancreatic ductal adenocarcinomas.7,59 The parallels between 3D organoids and 3D cleared tissues, potentially will allow researchers to validate hypothesis generated studying organoids using cleared tissue and vice-versa.
Third, recent advances in therapies that modulate the immune system to treat patients with cancers of a number of organs, including lung, head and neck, and stomach, suggest great potential in defining the 3D relationships of immune and cancer cells.60 3D evaluation of microenvironment of pancreatic ductal adenocarcinomas, including special associations among cancer cells, stromal cells, and immune cells, may provide insights for immune-suppressive environments for pancreatic ductal adenocarcinomas and overcoming ways for immunotherapy-resistant mechanisms of cancers.
Fourth, vascular invasion is one of the most important mechanisms underlying distant metastases of pancreatic ductal adenocarcinomas.61 Identifying mechanisms of vascular invasions with cleared pancreatic ductal adenocarcinoma tissues will be helpful for making therapeutic regimens for preventing vascular invasion of pancreatic ductal adenocarcinomas.
Thanks to recent technological advances of tissue clearing and advanced microscopy, the remarkable world of 3D anatomy and pathology to pancreatic pathologists are newly opened to pathologists. The application of labeling and clearing to human pancreatic parenchyma can provide detailed visualization of normal pancreatic anatomy and it can be used to characterize the 3D architecture of various disease processes, including precursor lesions, such as PanINs and IPMNs, to infiltrating pancreatic ductal adenocarcinomas.
1. Bronsert P, Enderle-Ammour K, Bader M, et al. Cancer
cell invasion and EMT marker expression: a three-dimensional study of the human cancer
-host interface. J Pathol. 2014;234:410–422.
2. Vannier MW, Hildebolt CF, Marsh JL, et al. Craniosynostosis: diagnostic value of three-dimensional CT reconstruction. Radiology. 1989;173:669–673.
3. Yoffie RL, Vannier MW, Gutierrez FR, et al. Three-dimensional magnetic resonance imaging of the heart. Radiol Technol. 1989;60:305–309.
4. Levine RA, Weyman AE, Handschumacher MD. Three-dimensional echocardiography: techniques and applications. Am J Cardiol. 1992;69:121H–130HH; discussion 31H–34H.
5. Wen Z, Yao F, Wang Y. 64-Slice spiral computed tomography and three-dimensional reconstruction in the diagnosis of cystic pancreatic tumors. Exp Ther Med. 2016;11:1506–1512.
6. Roth HR, Lu L, Lay N, et al. Spatial aggregation of holistically-nested convolutional neural networks for automated pancreas
localization and segmentation. Med Image Anal. 2018;45:94–107.
7. Boj SF, Hwang CI, Baker LA, et al. Organoid models of human and mouse ductal pancreatic cancer
. Cell. 2015;160:324–338.
8. Fowler JL, Lee SS, Wesner ZC, et al. Three-dimensional analysis of the human pancreas
. Endocrinology. 2018;159:1393–1400.
9. Chien HJ, Peng SJ, Hua TE, et al. 3-D imaging of islets in obesity: formation of the islet-duct complex and neurovascular remodeling in young hyperphagic mice. Int J Obes (Lond). 2016;40:685–697.
10. Butterworth E, Dickerson W, Vijay V, et al. High resolution 3D imaging of the human pancreas
neuro-insular network. J Vis Exp. 2018. Doi: 10.3791/56859.
11. Wong HS, Yeung PKK, Lai HM, et al. Simple and rapid tissue clearing
method for three-dimensional histology of the pancreas
. Curr Protoc Cell Biol. 2017;77:19.20.1–19.20.10.
12. Noe M, Rezaee N, Asrani K, et al. Immunolabeling of cleared human pancreata provides insights into three-dimensional pancreatic anatomy and pathology. Am J Pathol. 2018;188:1530–1535.
13. Roberts N, Magee D, Song Y, et al. Toward routine use of 3D histopathology as a research tool. Am J Pathol. 2012;180:1835–1842.
14. Ariel P. A beginner’s guide to tissue clearing
. Int J Biochem Cell Biol. 2017;84:35–39.
15. Marx V. Microscopy
: seeing through tissue. Nat Methods. 2014;11:1209–1214.
16. Richardson DS, Lichtman JW. Clarifying tissue clearing
. Cell. 2015;162:246–257.
17. Richardson DS, Lichtman JW. SnapShot: tissue clearing
. Cell. 2017;171:496–e1.
18. Spalteholz W. Über das Durchsichtigmachen von Menschlichen und Tierischen Präparaten: Nebst Anhang: Über Knochenfärbung. Leipzig: Hirzel; 1911.
19. Dodt HU, Leischner U, Schierloh A, et al. Ultramicroscopy: three-dimensional visualization of neuronal networks in the whole mouse brain. Nat Methods. 2007;4:331–336.
20. Erturk A, Becker K, Jahrling N, et al. Three-dimensional imaging of solvent-cleared organs using 3DISCO. Nat Protoc. 2012;7:1983–1995.
21. Renier N, Wu Z, Simon DJ, et al. iDISCO: a simple, rapid method to immunolabel large tissue samples for volume imaging. Cell. 2014;159:896–910.
22. Pan C, Cai R, Quacquarelli FP, et al. Shrinkage-mediated imaging of entire organs and organisms using uDISCO. Nat Methods. 2016;13:859–867.
23. Chiang AS, Lin WY, Liu HP, et al. Insect NMDA receptors mediate juvenile hormone biosynthesis. Proc Natl Acad Sci USA. 2002;99:37–42.
24. Staudt T, Lang MC, Medda R, et al. 2,2’-thiodiethanol: a new water soluble mounting medium for high resolution optical microscopy
. Microsc Res Tech. 2007;70:1–9.
25. Tsai PS, Kaufhold JP, Blinder P, et al. Correlations of neuronal and microvascular densities in murine cortex revealed by direct counting and colocalization of nuclei and vessels. J Neurosci. 2009;29:14553–14570.
26. Kuwajima T, Sitko AA, Bhansali P, et al. ClearT: a detergent- and solvent-free clearing method for neuronal and non-neuronal tissue. Development. 2013;140:1364–1368.
27. Ke MT, Fujimoto S, Imai T. SeeDB: a simple and morphology-preserving optical clearing agent for neuronal circuit reconstruction. Nat Neurosci. 2013;16:1154–1161.
28. Hama H, Kurokawa H, Kawano H, et al. Scale: a chemical approach for fluorescence imaging and reconstruction of transparent mouse brain. Nat Neurosci. 2011;14:1481–1488.
29. Susaki EA, Tainaka K, Perrin D, et al. Whole-brain imaging with single-cell resolution using chemical cocktails and computational analysis. Cell. 2014;157:726–739.
30. Chung K, Wallace J, Kim SY, et al. Structural and molecular interrogation of intact biological systems. Nature. 2013;497:332–337.
31. Yang B, Treweek JB, Kulkarni RP, et al. Single-cell phenotyping within transparent intact tissue through whole-body clearing. Cell. 2014;158:945–958.
32. Azaripour A, Lagerweij T, Scharfbillig C, et al. A survey of clearing techniques for 3D imaging of tissues with special reference to connective tissue. Prog Histochem Cytochem. 2016;51:9–23.
33. Seo J, Choe M, Kim SY. Clearing and labeling techniques for large-scale biological tissues. Mol Cells. 2016;39:439–446.
34. Kim A, Kilimnik G, Hara M. In situ quantification of pancreatic beta-cell mass in mice. J Vis Exp. 2010;40:1970.
35. Fu YY, Lu CH, Lin CW, et al. Three-dimensional optical method for integrated visualization of mouse islet microstructure and vascular network with subcellular-level resolution. J Biomed Opt. 2010;15:046018.
36. Tanaka N, Kaczynska D, Kanatani S, et al. Mapping of the three-dimensional lymphatic microvasculature in bladder tumours using light-sheet microscopy
. Br J Cancer
37. Tang SC, Baeyens L, Shen CN, et al. Human pancreatic neuro-insular network in health and fatty infiltration. Diabetologia. 2018;61:168–181.
38. Tang SC, Chiu YC, Hsu CT, et al. Plasticity of Schwann cells and pericytes in response to islet injury in mice. Diabetologia. 2013;56:2424–2434.
39. Tang SC, Shen CN, Lin PY, et al. Pancreatic neuro-insular network in young mice revealed by 3D panoramic histology. Diabetologia. 2018;61:158–167.
40. Lee H, Park JH, Seo I, et al. Improved application of the electrophoretic tissue clearing
technology, CLARITY, to intact solid organs including brain, pancreas
, liver, kidney, lung, and intestine. BMC Dev Biol. 2014;14:48.
41. Tainaka K, Kubota SI, Suyama TQ, et al. Whole-body imaging with single-cell resolution by tissue decolorization. Cell. 2014;159:911–924.
42. Kim SY, Cho JH, Murray E, et al. Stochastic electrotransport selectively enhances the transport of highly electromobile molecules. Proc Natl Acad Sci USA. 2015;112:E6274–E6283.
43. Lee E, Choi J, Jo Y, et al. ACT-PRESTO: Rapid and consistent tissue clearing
and labeling method for 3-dimensional (3D) imaging. Sci Rep. 2016;6:18631.
44. Murray E, Cho JH, Goodwin D, et al. Simple, scalable proteomic imaging for high-dimensional profiling of intact systems. Cell. 2015;163:1500–1514.
45. Treweek JB, Chan KY, Flytzanis NC, et al. Whole-body tissue stabilization and selective extractions via tissue-hydrogel hybrids for high-resolution intact circuit mapping and phenotyping. Nat Protoc. 2015;10:1860–1896.
46. Kubota SI, Takahashi K, Nishida J, et al. Whole-body profiling of cancer
metastasis with single-cell resolution. Cell Rep. 2017;20:236–250.
47. Greenbaum A, Chan KY, Dobreva T, et al. Bone CLARITY: clearing, imaging, and computational analysis of osteoprogenitors within intact bone marrow. Sci Transl Med. 2017;9:eaah6518.
48. Helmchen F, Denk W. Deep tissue two-photon microscopy
. Nat Methods. 2005;2:932–940.
49. Becker K, Jahrling N, Saghafi S, et al. Ultramicroscopy: light-sheet-based microscopy
for imaging centimeter-sized objects with micrometer resolution. Cold Spring Harb Protoc. 2013;2013:704–713.
50. Fritzky L, Lagunoff D. Advanced methods in fluorescence microscopy
. Anal Cell Pathol (Amst). 2013;36:5–17.
51. Ragazzi M, Longo C, Piana S. Ex vivo (fluorescence) confocal microscopy
in surgical pathology: state of the art. Adv Anat Pathol. 2016;23:159–169.
52. Yushchenko DA, Schultz C. Tissue clearing
for optical anatomy. Angew Chem Int Ed Engl. 2013;52:10949–10951.
53. Tanaka K, Toiyama Y, Okugawa Y, et al. In vivo optical imaging of cancer
metastasis using multiphoton microscopy
: a short review. Am J Transl Res. 2014;6:179–187.
54. Bousso P, Moreau HD. Functional immunoimaging: the revolution continues. Nat Rev Immunol. 2012;12:858–864.
55. Sharma S, Green KB. The pancreatic duct and its arteriovenous relationship: an underutilized aid in the diagnosis and distinction of pancreatic adenocarcinoma from pancreatic intraepithelial neoplasia. A study of 126 pancreatectomy specimens. Am J Surg Pathol. 2004;28:613–620.
56. Friedl P, Alexander S. Cancer
invasion and the microenvironment: plasticity and reciprocity. Cell. 2011;147:992–1009.
57. Benias PC, Wells RG, Sackey-Aboagye B, et al. Structure and distribution of an unrecognized interstitium in human tissues. Sci Rep. 2018;8:4947.
58. Qiu H, Wild AT, Wang H, et al. Comparison of conventional and 3-dimensional computed tomography against histopathologic examination in determining pancreatic adenocarcinoma tumor size: implications for radiation therapy planning. Radiother Oncol. 2012;104:167–172.
59. Huang L, Holtzinger A, Jagan I, et al. Ductal pancreatic cancer
modeling and drug screening using human pluripotent stem cell- and patient-derived tumor organoids. Nat Med. 2015;21:1364–1371.
60. Johnson BA III, Yarchoan M, Lee V, et al. Strategies for increasing pancreatic tumor immunogenicity. Clin Cancer
61. Hong SM, Goggins M, Wolfgang CL, et al. Vascular invasion in infiltrating ductal adenocarcinoma of the pancreas
can mimic pancreatic intraepithelial neoplasia: a histopathologic study of 209 cases. Am J Surg Pathol. 2012;36:235–241.