Advertisement

Laser Scanning–Based Tissue Autofluorescence/Fluorescence Imaging (LS-TAFI), a New Technique for Analysis of Microanatomy in Whole-Mount Tissues

Open AccessPublished:April 30, 2012DOI:https://doi.org/10.1016/j.ajpath.2012.02.032
      Intact organ structure is essential in maintaining tissue specificity and cellular differentiation. Small physiological or genetic variations lead to changes in microanatomy that, if persistent, could have functional consequences and may easily be masked by the heterogeneity of tissue anatomy. Current imaging techniques rely on histological, two-dimensional sections requiring sample manipulation that are essentially two dimensional. We have developed a method for three-dimensional imaging of whole-mount, unsectioned mammalian tissues to elucidate subtle and detailed micro- and macroanatomies in adult organs and embryos. We analyzed intact or dissected organ whole mounts with laser scanning–based tissue autofluorescence/fluorescence imaging (LS-TAFI). We obtained clear visualization of microstructures within murine mammary glands and mammary tumors and other organs without the use of immunostaining and without probes or fluorescent reporter genes. Combining autofluorescence with reflected light signals from chromophore-stained tissues allowed identification of individual cells within three-dimensional structures of whole-mounted organs. This technique could be useful for rapid diagnosis of human clinical samples and possibly the effect of subtle variations such as low dose radiation.
      The investigation of the histoarchitecture of organs is imperative for the study of organismal development, postnatal morphogenesis, and pathogenesis. These investigations require not only pinpointing the spatial locations (and dislocations in disease) of parenchymatous cells of organs but also assaying the organization of the surrounding three-dimensional (3D) microanatomy. Conventional methods for imaging biological materials range from ultra–high-magnification modalities such as electron microscopy to relatively low-resolution compound optical microscopy (reviewed elsewhere
      • Walter T.
      • Shattuck D.W.
      • Baldock R.
      • Bastin M.E.
      • Carpenter A.E.
      • Duce S.
      • Ellenberg J.
      • Fraser A.
      • Hamilton N.
      • Pieper S.
      • Ragan M.A.
      • Schneider J.E.
      • Tomancak P.
      • Heriche J.K.
      Visualization of image data from cells to organisms.
      ). A major limitation of the former is that its resolving capabilities are so high that one is susceptible to miss the proverbial forest for the trees.
      • Walter T.
      • Shattuck D.W.
      • Baldock R.
      • Bastin M.E.
      • Carpenter A.E.
      • Duce S.
      • Ellenberg J.
      • Fraser A.
      • Hamilton N.
      • Pieper S.
      • Ragan M.A.
      • Schneider J.E.
      • Tomancak P.
      • Heriche J.K.
      Visualization of image data from cells to organisms.
      The use of optical microscopy combined with immunofluorescence can surmount this problem, ie, it can be used to observe tissue architecture on a larger scale while still maintaining subcellular resolution. However, this modality requires extensive processing of organs into thin sections, staining each with dyes or fluorescent antibodies, and imaging an array of slides. Moreover, to find a given microanatomic structure in tissue sections demands enormous time and effort (as well as expense), especially when the tissue is small and the target microanatomy is difficult to locate.
      Subsequent 3D reconstruction of the tissue sections to study organ microanatomy is cumbersome, and the quality of resulting images is proportional to the number and quality of the individual sections and re-registration of these sections. To overcome these limitations, we have developed a laser scanning–based tissue autofluorescence/fluorescence imaging (LS-TAFI), a laser scanning confocal microscopy (LSCM)–based technique to capture tissue autofluorescence and to image the histological architectures of whole-mounted organs. Using a number of examples from prenatal and postnatal mammary gland morphologies as well as friend virus B-type (FVB/n) transgenic polyoma virus middle T-antigen [Tg(PyMT)] mammary pathology, we demonstrate that LS-TAFI robustly records and images macropattern features of biological tissues and simultaneously penetrates their complexity down to the single-cell level. Combining this modality with fluorophoric or nonfluorophoric staining leads to simple and accurate determination of single-cell shape and identification and localization of specific cells within native tissue environments. LS-TAFI holds promise to advance the study of the role of architecture in organ development, physiology, and pathology, as we demonstrate here.

      Materials and Methods

      Mouse and Tissue Samples

      For imaging of the whole-mount mammary gland, FVB mice were raised until 2 weeks of age and then sacrificed. Mammary glands from FVB/n Tg(PyMT) mice were isolated at 10 weeks. Whole-mount lung and pancreas were dissected from C57BL/6 mice after the mammary gland was isolated. Embryo in stage 21 was isolated from mid-pregnant C57BL/6 mice. For imaging, β-gal staining in mammary gland, Mmp14 (+/lacZ),
      • Yana I.
      • Sagara H.
      • Takaki S.
      • Takatsu K.
      • Nakamura K.
      • Nakao K.
      • Katsuki M.
      • Taniguchi S.
      • Aoki T.
      • Sato H.
      • Weiss S.J.
      • Seiki M.
      Crosstalk between neovessels and mural cells directs the site-specific expression of MT1-MMP to endothelial tip cells.
      and Mmp14 (wild type, +/+) C57BL/6 mice at the age of 3 weeks were used. Experimental animal protocols were followed in accordance with guidelines set by the Lawrence Berkeley National Laboratory's Animal Welfare and Research Committee (AWRC) or the University of California Davis. The mammary glands from FVB mice and the lung and pancreas from C57BL/6 mice were fixed with 4% paraformaldehyde for 30 minutes and then stained with carmine alum. Each tissue was dehydrated with ethanol and xylene after the fixation. β-Gal staining was performed as described by Yana et al
      • Yana I.
      • Sagara H.
      • Takaki S.
      • Takatsu K.
      • Nakamura K.
      • Nakao K.
      • Katsuki M.
      • Taniguchi S.
      • Aoki T.
      • Sato H.
      • Weiss S.J.
      • Seiki M.
      Crosstalk between neovessels and mural cells directs the site-specific expression of MT1-MMP to endothelial tip cells.
      Tissue sections were counterstained with H&E for cytological visualization.

      Tissue Preparation Procedures for LS-TAFI

      Mammary glands were spread on a glass slide, whereas other tissues were put in Tissue-Tek cassettes. Isolated tissues were fixed with 4% paraformaldehyde in phosphate-buffered saline solution for 15 minutes. If tissues needed to be stained with carmine, they were fixed with Carnoy's solution (75% ethanol and 25% glacial acetic acid) overnight. If needed, tissues were stained with carmine alum or hematoxylin. Carmine-stained mammary tissues were destained with destaining solution (2% HCl in 70% ethanol) for more than 16 hours. Tissues were then dehydrated in increasing concentrations (70%, 80%, 90%, 95%, and 100%) of ethanol for 30 minutes each. Fat was dissolved by treating with Toluene or Xylene for 48 hours. Tissues were mounted with Permount mounting medium (Fisher Scientific, Pittsburgh, PA).

      Mammary Glands

      Whole-mount mammary gland was allowed to attach to Superfrost Plus white glass slides on glass slide. The surface of whole-mount tissue was covered with cover glass (no. 1; Thermo Scientific, Walthlam, MA) and was kept at room temperature for at least 2 days until Permount solidified. The surface of cover glass needed to be even while Permount solidified.

      Other Tissues

      Whole-mount tissues were prepared in 8-well chamber cover-glass (no. 1; Thermo Scientific) and mounted with Permount. Lung and embryo could float in Permount, so they needed to be softly pushed down to the bottom of the chamber cover-glass with a glass pipette while Permount solidified.

      β-Gal–Stained Tissues

      β-Gal–stained tissue
      • Yana I.
      • Sagara H.
      • Takaki S.
      • Takatsu K.
      • Nakamura K.
      • Nakao K.
      • Katsuki M.
      • Taniguchi S.
      • Aoki T.
      • Sato H.
      • Weiss S.J.
      • Seiki M.
      Crosstalk between neovessels and mural cells directs the site-specific expression of MT1-MMP to endothelial tip cells.
      was fixed with 4% paraformaldehyde in phosphate-buffered saline solution for 15 minutes and then dehydrated as described above.

      Confocal Microscopy

      A laser scanning confocal microscope (LSM710; Carl Zeiss AG, Oberkochen, Germany) with lasers producing light (25 mW Argon laser: 458,488 and 514 nm, 15 mW DPSS-561 laser: 561 nm, 30 mW Diode 405-30 laser, 405 nm) was used for the entire work. The setting for each tissue sample was optimized before scanning because of the different characteristic intensity of autofluorescence. To observe tissue samples, a ×10, 0.3 NA lens or a ×40, 1.1 NA water immersion lens was used. For imaging tissues, the following settings for main dichroic beam splitter (MBS) and detection band/range (DB) were used for imaging whole mounts: i) MBS, 561 nm: DB, 630 to 740 nm; ii) MBS, 458/561 nm: DB 568 to 712 nm; iii) MBS 458/561 nm: DB 611 to 728 nm and MBS, 405 nm: DB 371 to 456 nm; iv) MBS, 458/561 nm: DB 568 to 740 nm (autofluorescence for embryo); v) MBS, 488 nm: DB 495 to 553 nm (autofluorescence for embryo) and MBS, 561 nm: DB568 to 758 nm (autofluorescence); and vi) MBS, 405 nm: DB 371 to 495 nm (β-gal, reflection MBS T80/R20) and MBS, 458/561 nm: DB, 568 to 712 nm (autofluorescence). Setting i was used for carmine-stained mammary gland; settings i and iii were used for hematoxylin-stained mammary gland from MMTV-PyMT mouse; settings iv and v were used for carmine-stained lung and pancreas and unstained embryo; and setting vi was used for Mmp14 (+/lacZ) with β-gal staining. Each scan was performed on starting at the whole-mount surface and imaging to a depth up to ∼800 μm. 3D image reconstitution from Z-stack images was performed with IMARIS software version 7.11 (Bitplane, Zurich, Switzerland) and ZEN (Carl Zeiss AG).

      Results

      Application of LS-TAFI to Three-Dimensionally Visualize Mouse Mammary Gland Whole Mounts

      The mammary gland is among the most widely studied organs in postnatal developmental biology. Its ability to undergo radical architectural changes with hormonal cycles and its intricate arrangement of epithelial and stromal cells (along with the vasculature) allow morphologists to study various aspects of differentiation and morphogenesis. This made it the ideal system in which to test the power and applicability of our imaging method. Typically, mammary glands are either stained with carmine alum to visualize their morphology in whole mount or are sectioned and stained with H&E to identify tissue and cellular characteristics. The abdominal mammary gland from a 2-week-old FVB mouse was isolated and stained with carmine alum to visualize the spatial arrangement of its ductal tree (Figure 1, A and B). A higher-magnification image taken with an optical microscope facilitates more detailed visualization of individual end buds (Figure 1C), and conventional histological sectioning of the whole mount with H&E staining details the internal structure of end buds and their surrounding cells (Figure 1D).
      Figure thumbnail gr1
      Figure 1The application of LS-TAFI to three-dimensionally visualize mouse mammary gland whole mount. A–D: Mammary gland whole-mount imaging with optical microscope. A: Carmine alum-stained mammary gland whole mount. Mammary gland was isolated from female 2-week-old FVB mouse. Scale bar = 1 cm. B: Higher magnification of inset in A showing the mammary ductal tree. Scale bar = 1 mm. C: Higher magnification of inset in B showing end buds. Scale bar = 25 μm. D: H&E-stained mammary gland tissue section. Scale bar = 25 μm. E–H: Three-dimensional (3D) reconstitution of mouse mammary gland ductal tree from confocal analysis. E: Surface rendering of reconstituted 3D image from confocal scanning of mouse mammary gland. Whole-mount sample was analyzed directly with laser scanning confocal microscope (LSM710; Zeiss). To detect fluorescence and autofluorescence signal effectively, MBS 561 nm and DB 630-740 nm were used. Magnification, ×10. Confocal images were reconstituted into 3D image using IMARIS (Bitplane). Pseudo-coloring (green) was done for better visualization. Z-stack of confocal scanned images is available as (available at http://ajp.amjpathol.org). F: Image of surface reconstitution of result obtained in E. To decipher the 3D depth on image, surface reconstitution was performed with IMARIS. Actual fluorescence intensity of E was used for image processing. Scale bar = 200 μm. G: Surface rendering of reconstituted images of end buds (box in E). Magnification, ×40. H: Image of surface reconstitution of G. Scale bar = 20 μm. Sections of confocal scanned images and 3D movie of ductal tree are available as (available at http://ajp.amjpathol.org). I, K, and M: Images of mammary end buds. I: Surface rendering of reconstituted 3D image from confocal scanning of mouse mammary end buds. Magnification, ×40. Scale bar = 20 μm. Z-stack of confocal scanned images is available as (available at http://ajp.amjpathol.org). K: Z-stack section of confocal image from I. Scale bar = 100 μm. M: Image of mammary ductal epithelial cell layers of end bud (box in K). Lumen (asterisk), Luminal epithelial cell (white arrowhead), myoepithelial cell (blue arrowhead), and stromal cell (yellow arrowhead) are highlighted. Scale bar = 50 μm. J, L, and N: Images of blood vessels. J: Surface rendering of reconstituted 3D image from confocal scanning of blood vessels in mouse mammary gland. Magnification, ×40. Pseudo-color (green) is shown for better visualization. Scale bar = 20 μm. Z-stack of confocal scanned images is available as (available at http://ajp.amjpathol.org). L: Section of confocal image from J. Scale bar = 100 μm. N: Higher-magnification image of blood vessel in mammary gland. Red blood cells (white arrowhead) and smooth muscle cells (blue arrowhead) are highlighted (box in L). Scale bar = 50 μm. Arrowheads indicate blood cell in vessel lumen. Pseudo-color (red) is shown for better visualization.
      To address the limitations of conventional imaging methodologies previously elaborated, and motivated by prior studies of Schistosoma mansoni that demonstrated that organ structure of the entire organism could be visualized by LSCM,
      • Machado-Silva J.R.
      • Pelajo-Machado M.
      • Lenzi H.L.
      • Gomes D.C.
      Morphological study of adult male worms of Schistosoma mansoni Sambon, 1907 by confocal laser scanning microscopy.
      we used LS-TAFI to capture the microstructure of whole-mounted, carmine-stained tissue before sectioning. Using extensive iterations, we scanned the entire mammary ductal tree by LSCM using specially tuned settings (MBS and DB; MBS 561 nm and DB 630-740 nm) on a Zeiss LSM710 to elicit tissue fluorescence/autofluorescence. High-resolution images were scanned and reconstituted by image processing software (IMARIS, version 7.1.1), allowing visualization of the 3D architecture of the entire gland (compare Figure 1B and 1E; see also Supplemental Movie S1 at http://ajp.amjpathol.org). A surface reconstruction image from Figure 1E is shown in Figure 1F.
      Capturing tissue fluorescence via LS-TAFI can also be used to reveal cellular architecture within a given tissue. For instance, within the mammary gland, terminal end buds are typically difficult to orient and locate within tissue sections. By scanning at high resolution, we were able to resolve microstructural details and visualize cells occupying distinct locales within the gland. Figure 1G demonstrates the use of this technique to locate and resolve premature terminal end buds at the end of a ductal branch in specific locations within the mammary ductal tree (see Supplemental Movie S2 at http://ajp.amjpathol.org). By comparing a reconstructed-fluorescent image (Figure 1G) with its 3D surface reconstruction image (Figure 1H), epithelial, myoepithelial, adipocyte-like, and fibroblast-like cells can be identified, along with microvasculature. The spatial density of these different cell types can be analyzed with this method as well. For example, in Figure 1, G and H, stromal cell types are concentrated around budding ductal edges, in agreement with prior studies by Pollard et al
      • Ingman W.V.
      • Wyckoff J.
      • Gouon-Evans V.
      • Condeelis J.
      • Pollard J.W.
      Macrophages promote collagen fibrillogenesis around terminal end buds of the developing mammary gland.
      In addition, resolution of epithelial ducts within the mammary gland was possible: the ducts are typically a bi-layered tubular structure with the stromal cells, fibroblasts, and adipocytes surrounding the ductal tree.
      • Wiseman B.S.
      • Werb Z.
      Stromal effects on mammary gland development and breast cancer.
      Sequential magnification of a reconstituted, high-resolution 3D region of the ductal tree allowed visualization of the ductal lumen where luminal epithelial cells; myoepithelial and stromal cells can be seen near a terminal end bud (Figure 1, I, K, and M; see also Supplemental Movie S3 at http://ajp.amjpathol.org). Elsewhere in the gland, the microvasculature could be visualized (Figure 1, J, L, and N; see also Supplemental Movie S4 at http://ajp.amjpathol.org). Interestingly, cells could be observed clearly within the blood vessel lumen, as could perivascular cells aligned with their long axes parallel to that of the vessel (Figure 1N). These results demonstrate the use and advantages of capturing tissue fluorescence of whole-mounted mammary gland via LS-TAFI to visualize epithelial and extra parenchymal microstructures.
      Part of the rationale for developing the technique was to detect details of a very small piece of anatomy or lesion in a much larger sample. Another part of the rationale was to develop a technique that could lend itself to discerning subtle consequences of an insult or treatment condition. As such, we asked whether fine differences in mammary gland development could be assessed using LS-TAFI. We examined samples from the mammary glands of two mouse strains that we are currently using to understand how low-dose radiation may affect architecture of the mammary gland: one resistant (Mus spretus; SPRET/EiJ) and one susceptible (Mus musculus; BALB/c) to mammary cancer.
      • Nagase H.
      • Bryson S.
      • Cordell H.
      • Kemp C.J.
      • Fee F.
      • Balmain A.
      Distinct genetic loci control development of benign and malignant skin tumours in mice.
      • Quigley D.A.
      • To M.D.
      • Perez-Losada J.
      • Pelorosso F.G.
      • Mao J.H.
      • Nagase H.
      • Ginzinger D.G.
      • Balmain A.
      Genetic architecture of mouse skin inflammation and tumour susceptibility.
      As shown in Supplemental Figure S1 (available at http://ajp.amjpathol.org), distinct differences in the terminal end bud microanatomy of the two mouse strains could be easily appreciated even without radiation. Although the size of the mammary glands in the whole mounts of SPRET/EiJ is smaller than those of the BALB/c (data not shown), the overall size and structure of end buds appear to be similar between strains in terms of cap cells and body cells and orientation when imaged using conventional microscopy. However, using LS-TAFI reveals a clear difference in the stromal cells (fibroblasts and myofibroblasts), as well as the structure of the adipose tissue. The SPRET/EiJ has more abundant periductal stroma, and the adipocytes are characterized by smaller amounts of lipid and by higher cell density. Although some of these same features can be appreciated in traditional histological sectioning, the 3D approach allows z-sections to be compared more easily and preserves the samples for further immunohistological analysis if needed.

      Application of LS-TAFI to Three-Dimensionally Visualize Pathological Tissue

      To extend this method to the diagnosis of pathological tissues, mammary glands from 10-week-old FVB/n Tg (PyMT) mice (Figure 2A) were stained with hematoxylin and then scanned and visualized by LS-TAFI (3D reconstituted image shown in Figure 2B). We coupled two unique excitation settings (described in Materials and Methods) to elicit fluorescence specifically from the mounted tissue in addition to that of red blood cells, thus allowing resolution of the blood microvasculature surrounding PyMT tumors (Figure 2C). Scanning through the high-resolution 3D reconstruction of the tumor-laden tissue allowed us to examine the microanatomy within early hyperplastic nodules, revealing hyperproliferative structures lacking a central lumen (Figure 2, D and E; see also Supplemental Movie S5 at http://ajp.amjpathol.org). However, some nodules retained a more normal structure (Figure 2E). Using LS-TAFI, we were thus able to resolve microarchitecture on a single-cell level within tumor-laden mammary tissue.
      Figure thumbnail gr2
      Figure 2The application of LS-TAFI to three-dimensionally visualize pathological tissue. A: Hematoxylin-stained mammary gland whole-mount imaging with optical microscopy. Mammary gland is isolated from 10-week-old MMTV-PyMT FVB mouse. Magnification, ×10. Scale bar = 1 cm. B: Surface rendering of reconstituted 3D image from confocal scanning of MMTV-PyMT mouse mammary gland. Whole-mount sample was analyzed directly with LSCM. To detect fluorescence and autofluorescence signal effectively, MBS 458/561 nm, DB 568–712 nm, and 40× lens were used. Scale bar = 500 μm. C: Surface rendering of reconstituted 3D image of PyMT tumor nodules. To detect fluorescence and autofluorescence signal effectively, MBS 458/561 nm and DB 568–712 nm were used (red). Red blood cells were highlighted with image setting of MBS 405 nm and DB 371–456 (green). Arrowheads indicate blood microvasculature. Scale bar = 40 μm. Z-stack of confocal scanned images is available as (available at http://ajp.amjpathol.org). D: Z-stack section of confocal image from C. Tumor nodules are highlighted with arrowheads. Scale bar = 30 μm. E: Z-stack section of confocal image from C. Tumor nodules are highlighted with white arrowhead and phenotypically normal nodule is indicated with blue arrowhead. Scale bar = 30 μm.

      Application of LS-TAFI to Visualize Whole-Mount Lung, Pancreas, and Embryo

      To explore the application of this methodology beyond the mammary gland, we identified settings that facilitated the visualization of nuclei and cellular structures within whole-mounted lung and pancreas isolated from 3-week-old C57BL/6 mice using LS-TAFI. The epithelial network as a whole (Figure 3A; see also Supplemental Movie S6 at http://ajp.amjpathol.org) in addition to alveolar geometry (Figure 3B) could be visualized in the lung. Imaging the surface of the pancreas revealed a more cell-dense tissue (Figure 3C; see also Supplemental Movie S7 at http://ajp.amjpathol.org); imaging through the pancreas at higher resolution allowed the identification of pancreatic islets and pancreatic alveoli (Figure 3D).
      Figure thumbnail gr3
      Figure 3The application of LS-TAFI to visualize whole-mount lung, pancreas, and embryo. A and B: Images of lung surface from 3-week-old C57BL/6 mice. A: Surface rendering of reconstituted 3D image from confocal scanning of lower lobe of the lung. Scale bar = 150 μm. MBS 561 nm with DB 568-758 nm captured the intense fluorescent signals from the nuclei (red), and MBS 488 nm with DB 495-583 nm outlined the shape of alveoli (green). Magnification, ×10. Z-stack of confocal scanned images is available as (available at http://ajp.amjpathol.org). B: Section of confocal image from A. Scale bar = 30 μm. C and D: Images of the surface of pancreas. C: Surface rendering of reconstituted 3D image from confocal scanning of pancreas. Same image setting as lung was used. Scale bar = 150 μm. Z-stack of confocal scanned images is available as (available at http://ajp.amjpathol.org). D: Section of confocal image from C. Pancreatic islets (white arrowheads) and pancreatic alveoli (blue arrowhead) are highlighted. Scale bar = 30 μm. E and H: Images of C57BL/6 mouse embryo in Theiler stage 21. E: Surface rendering of reconstituted 3D image from confocal scanning of embryo. Magnification, ×10. Z-stack of confocal scanned images is available as (available at http://ajp.amjpathol.org). Scale bar = 500 μm. F: One image of Z-sections shows a part of brain (inset) and mammary placodes (white arrowheads). Scale bar = 500 μm. G: Higher-magnification image of boxed area of F. Scale bar = 200 μm. H: Higher-magnification image of boxed area of G. Individual cells are indicated as arrowheads in G and H. Magnification, ×10. Scale bar = 50 μm.
      We were further able to extend this technique to unstained mouse embryos. An embryo at Theiler stage 21 was scanned at high resolution via LS-TAFI and reconstituted as a 3D image (Figure 3E; see also Supplemental Movie S8 at http://ajp.amjpathol.org). Individual mammary placodes could be identified (Figure 3F), and we were even able to discern individual cells within the skull and brain (Figure 3, G and H). The LSCM-based tissue scanning technique described here can simultaneously visualize macrostructural details of different types of biological tissues and the microanatomical positions of the cellular constituents of the same tissues and organs.

      Application of LS-TAFI to Visualize Mammary Gland Whole Mount with β-Gal Staining

      To explore the potential of LS-TAFI for other experimental applications, we used a transgenic model that carries a nuclear localization signal-tagged lacZ as the trans-gene to indicate promoter activity, here specifically of the matrix metalloproteinase 14 (Mmp14) gene.
      • Yana I.
      • Sagara H.
      • Takaki S.
      • Takatsu K.
      • Nakamura K.
      • Nakao K.
      • Katsuki M.
      • Taniguchi S.
      • Aoki T.
      • Sato H.
      • Weiss S.J.
      • Seiki M.
      Crosstalk between neovessels and mural cells directs the site-specific expression of MT1-MMP to endothelial tip cells.
      The mammary glands from a 3-week-old Mmp14 (+/lacZ) C57BL/6 mouse and its sibling (wild type, +/+) were isolated and stained with β-gal. The gland from Mmp14 (+/lacZ) stained positively with β-gal (Figure 4B), whereas the wild-type mammary gland, which lacks the lacZ transgene, did not stain with β-gal as expected (Figure 4A). An eosin-stained tissue section of mammary gland from Mmp14 (+/lacZ) mouse revealed a number of cells throughout the gland (both epithelial and stromal) with Mmp14 promoter activity on β-gal staining (Figure 4C).
      Figure thumbnail gr4
      Figure 4The application of LS-TAFI to visualize mammary gland whole mount with β-gal staining. β-gal-stained mammary gland whole mount and a tissue section from Mmp14 (+/lacZ) mouse. A: β-gal-stained mammary gland whole mount from control mouse (wild type). B: β-Gal–stained mammary gland from Mmp14 (+/lacZ) mouse. C: Tissue section of β-gal-stained mammary gland from Mmp14 (+/lacZ) mouse stained with eosin. Mammary glands were isolated from female 3-week-old Mmp14 (+/lacZ) mouse. D: Surface rendering of 3D imaging of mammary gland whole mount from Mmp14 (+/lacZ) mouse with β-gal staining. To detect fluorescence signal effectively, MBS 458/561 nm with DB 568–712 nm and 40× objective lens were used. β-Gal staining was visualized from confocal scan with reflection light from MBS 405 nm (reflection: MBS T80/R20) with DB 371–495 nm. Pseudo-color (blue) is shown for better visualization. Z-stack of confocal scanned images is available as (available at http://ajp.amjpathol.org). E: Section of confocal scan from D, highlighting surface of terminal end bud. F: A section of confocal scan, highlighting inside of terminal end bud.
      To assay for cell shape and to locate β-gal staining in whole mounts, we determined specific settings to capture autofluorescent signals from the whole-mounted tissue itself in addition to light reflected from the β-gal staining (precise settings described in Materials and Methods). The reconstituted 3D image shown in Figure 4D and Supplemental Movie S9 (available at http://ajp.amjpathol.org) reveals β-gal–positive cells stained blue primarily within the mammary epithelial of terminal end buds (Figure 4, E and F). The above examples demonstrate the utility of LS-TAFI to detect specific cellular subtypes throughout an unadulterated organ and to spatially localize the activity of a given promoter to show cell position within the tissue. Clearly, the technique could be extended to any lacZ-containing transgene.

      Discussion

      The elucidation of the macro- and microanatomy within organs is an important component of understanding the development of physiological and pathological states. Under current conventional microscopy techniques, tissues and organs need to be sectioned and stained with dyes conjugated to different markers or specific antibodies to localize specific cell types and to reveal tissue architecture. Alternatively, one can use genetically engineered fluorescent proteins and reflected confocal imaging.
      • Tilli M.T.
      • Parrish A.R.
      • Cotarla I.
      • Jones L.P.
      • Johnson M.D.
      • Furth P.A.
      Comparison of mouse mammary gland imaging techniques and applications: reflectance confocal microscopy GFP imaging, and ultrasound.
      These methods, albeit informative, come with a number of limitations. For example, in studying mammary gland development, a large number of sequential sections are required to elaborate the 3D organization of ducts and terminal end buds and the exact spatial arrangement of surrounding stromal constituents. To capture the one near-perfect image that contains all of the different cell types is cumbersome. Aside from the fact that drawing a meaningful conclusion from one image is not prudent, the process of sectioning the tissues itself may damage samples or irreversibly alter their original structure. This could lead to faulty and inaccurate rendering and co-registration of 3D reconstruction, which could distort our anatomical understanding. The application of confocal microscopy to whole mounts could solve many of these issues because of the enhanced Z-resolution inherent to this technique. However this modality has not been applied to whole tissues, apparently because of the assumption that without traditional immunostaining to label microstructures, the results would not prove fruitful.
      Here, we have shown that LS-TAFI can capture meaningful macro- and microanatomy from precisely such samples. Furthermore, we demonstrate that staining tissues is not always necessary for ascertaining the position and shape of individual cells within whole mounts. Thus LS-TAFI will allow simultaneous resolution of micro- and macrostructures within whole tissues and will yield accurate 3D spatial data. Although we have not pinpointed the exact source of tissue autofluorescence that enables LS-TAFI to capture such images, the putative candidates that generally fluoresce within biological tissues include NADPH, aromatic amino acids, lipo-pigments, flavin, porphyrin, and macromolecular components of the extracellular matrices.
      • Monici M.
      Cell and tissue autofluorescence research and diagnostic applications.
      • Richards-Kortum R.
      • Sevick-Muraca E.
      Quantitative optical spectroscopy for tissue diagnosis.
      In our current work, lipo-pigments are the most probable sources because their wavelength kinetics fit the spectrum within which we imaged our tissues. In contrast, NADPH lies outside our observation spectrum and is an unlikely source of autofluorescence in our specimens. The fluorescing entity may vary from one tissue to another based on their chemical constitution and may be difficult to identify. Identifying the origin(s) of autofluorescence remains a veritable future endeavor.
      LS-TAFI can be used to detect not only fluorescent and autofluorescent signals but also to detect light that is reflected by color reaction–based stains. For detecting lacZ activity in tissues, β-D-galactopyranoside (X-gal) is used as the substrate. X-gal is digested by β-galactosidase, which generates galactose and 5-bromo-4-chloro-3-hydroxyindole. The oxidized product of 5-bromo-4-chloro-3-hydroxyindole is a blue compound (5,5′-dibromo-4,4′-dichloro-indigo) that can be visualized as reflected light after excitation with 405 nm laser. This result further indicates that LS-TAFI could be used to ascertain with great accuracy the location and cellular targets of drugs with chromophoric properties or with pharmacological reactions that result in products detectable by their color. In addition to staining the lacZ reporter, β-gal staining can also be used to identify senescent cells by staining at lower pH.
      • Dimri G.P.
      • Lee X.
      • Basile G.
      • Acosta M.
      • Scott G.
      • Roskelley C.
      • Medrano E.E.
      • Linskens M.
      • Rubelj I.
      • Pereira-Smith O.
      • Peacocke M.
      • Campisi J.
      A biomarker that identifies senescent human cells in culture and in aging skin in vivo.
      When we assayed for such cells within mammary glands from older mice (32 weeks of age, BALB/c), we were able to locate β-gal–positive cells (ie, senescent cells with lower intracellular pH; data not shown). There are now many engineered mice that carry lacZ or fluorescent proteins under the control of tissue-specific or ubiquitous gene promoters.
      • Spergel D.J.
      • Kruth U.
      • Shimshek D.R.
      • Sprengel R.
      • Seeburg P.H.
      Using reporter genes to label selected neuronal populations in transgenic mice for gene promoter, anatomical, and physiological studies.
      Visualizing these signals in transgenic mice with LS-TAFI would allow the evaluation of promoter or protein activities as a function of time (four dimensional, 4D) along with cell distributions in different tissues. Furthermore, LS-TAFI might be useful in diagnosis of pathologies from biopsy specimens by reducing the time required for diagnosis as well as by enhancing the sensitivity for visualizing tissues in the Z-plane. Additionally it allows the quantification of morphogenetic parameters relevant to 3D microanatomies of organ whole mounts down to a fine level of detail. Finally, it may be possible to extract or microdissect areas of interest identified in LS-TAFI for molecular analysis similar to the currently available 2D laser capture microdissection technique.
      There are clear advantages of our methodology compared with conventional diagnostic imaging techniques; however, there are also potential limitations that necessitate further development to advance the technique. One important technical constraint is the imaging depth for analysis, which is dependent on the laser wavelength, type of confocal microscopy, numerical aperture of the objective lens, and thickness of the tissue sample. The first three are hardware issues and can be addressed with other commercially available microscopes. The latter could be solved by tuning laser intensity or preparing thinner samples. In the data presented, ∼800 μm could be captured with depth-associated signal attenuation but without loss of signal-to-noise ratio. It must be mentioned that multiphoton laser scanning microscopy
      • Denk W.
      • Strickler J.H.
      • Webb W.W.
      Two-photon laser scanning fluorescence microscopy.
      • Helmchen F.
      • Denk W.
      Deep tissue two-photon microscopy.
      is able to penetrate deeper into tissues and resolves the issues mentioned above to some extent, although these instruments are exceedingly expensive and not within the reach of many investigators.
      Combining LS-TAFI with existing methods
      • Walter T.
      • Shattuck D.W.
      • Baldock R.
      • Bastin M.E.
      • Carpenter A.E.
      • Duce S.
      • Ellenberg J.
      • Fraser A.
      • Hamilton N.
      • Pieper S.
      • Ragan M.A.
      • Schneider J.E.
      • Tomancak P.
      • Heriche J.K.
      Visualization of image data from cells to organisms.
      such as electron microscopy or magnetic resonance imaging may provide a combined image with greater information at minimal additional effort. Clinically, it might be possible to use MRI and LS-TAFI in concert to target organs for analysis. An outstanding advance by Xie et al uses stimulated Raman spectroscopy to image lipid bodies
      • Freudiger C.W.
      • Min W.
      • Saar B.G.
      • Lu S.
      • Holtom G.R.
      • He C.
      • Tsai J.C.
      • Kang J.X.
      • Xie X.S.
      Label-free biomedical imaging with high sensitivity by stimulated Raman scattering microscopy.
      and chromophores
      • Min W.
      • Lu S.
      • Chong S.
      • Roy R.
      • Holtom G.R.
      • Xie X.S.
      Imaging chromophores with undetectable fluorescence by stimulated emission microscopy.
      without labeling them. Combining their techniques with LS-TAFI would allow identification of the spatial localization of subcellular organelles or pathologies within 3D organ environments. LS-TAFI can extract richer visual information on tissue- and cell-level microanatomy than all other contemporary techniques except two-photon microscopy. Once hardware is upgraded, much better imaging resolution and larger image size (both in area and depth) may be captured to translate the use of LS-TAFI from research laboratories to clinical settings.

      Acknowledgments

      We thank Xuefei Tian, Jamie L. Inman, and Alvin T. Lo for helping with mouse maintenance and tissue sampling and Damir Sudar for critical reading of the manuscript.

      Supplementary data

      • Supplemental Figure S1

        Images of terminal end bud from two different mouse strains at age 5 weeks. Carmine-stained mammary gland whole mount of SPRET/EiJ and Balb/cJ are shown. To detect fluorescence signal effectively, MBS 561 nm, DB 630–740 nm, and 40× objective lens were used. Scale bar = 20 μm. Lumen (asterisk), body cells (white arrowhead), cap cells (blue arrowhead), stromal cells (yellow arrowhead), and adipocytes (green arrowhead) are indicated.

      References

        • Walter T.
        • Shattuck D.W.
        • Baldock R.
        • Bastin M.E.
        • Carpenter A.E.
        • Duce S.
        • Ellenberg J.
        • Fraser A.
        • Hamilton N.
        • Pieper S.
        • Ragan M.A.
        • Schneider J.E.
        • Tomancak P.
        • Heriche J.K.
        Visualization of image data from cells to organisms.
        Nat Methods. 2010; 7: S26-S41
        • Yana I.
        • Sagara H.
        • Takaki S.
        • Takatsu K.
        • Nakamura K.
        • Nakao K.
        • Katsuki M.
        • Taniguchi S.
        • Aoki T.
        • Sato H.
        • Weiss S.J.
        • Seiki M.
        Crosstalk between neovessels and mural cells directs the site-specific expression of MT1-MMP to endothelial tip cells.
        J Cell Sci. 2007; 120: 1607-1614
        • Machado-Silva J.R.
        • Pelajo-Machado M.
        • Lenzi H.L.
        • Gomes D.C.
        Morphological study of adult male worms of Schistosoma mansoni Sambon, 1907 by confocal laser scanning microscopy.
        Mem Inst Oswaldo Cruz. 1998; 93: 303-307
        • Ingman W.V.
        • Wyckoff J.
        • Gouon-Evans V.
        • Condeelis J.
        • Pollard J.W.
        Macrophages promote collagen fibrillogenesis around terminal end buds of the developing mammary gland.
        Dev Dyn. 2006; 235: 3222-3229
        • Wiseman B.S.
        • Werb Z.
        Stromal effects on mammary gland development and breast cancer.
        Science. 2002; 296: 1046-1049
        • Nagase H.
        • Bryson S.
        • Cordell H.
        • Kemp C.J.
        • Fee F.
        • Balmain A.
        Distinct genetic loci control development of benign and malignant skin tumours in mice.
        Nat Genet. 1995; 10: 424-429
        • Quigley D.A.
        • To M.D.
        • Perez-Losada J.
        • Pelorosso F.G.
        • Mao J.H.
        • Nagase H.
        • Ginzinger D.G.
        • Balmain A.
        Genetic architecture of mouse skin inflammation and tumour susceptibility.
        Nature. 2009; 458: 505-508
        • Tilli M.T.
        • Parrish A.R.
        • Cotarla I.
        • Jones L.P.
        • Johnson M.D.
        • Furth P.A.
        Comparison of mouse mammary gland imaging techniques and applications: reflectance confocal microscopy.
        BMC Cancer. 2008; 8: 21
        • Monici M.
        Cell and tissue autofluorescence research and diagnostic applications.
        Biotechnol Annu Rev. 2005; 11: 227-256
        • Richards-Kortum R.
        • Sevick-Muraca E.
        Quantitative optical spectroscopy for tissue diagnosis.
        Annu Rev Phys Chem. 1996; 47: 555-606
        • Dimri G.P.
        • Lee X.
        • Basile G.
        • Acosta M.
        • Scott G.
        • Roskelley C.
        • Medrano E.E.
        • Linskens M.
        • Rubelj I.
        • Pereira-Smith O.
        • Peacocke M.
        • Campisi J.
        A biomarker that identifies senescent human cells in culture and in aging skin in vivo.
        Proc Natl Acad Sci USA. 1995; 92: 9363-9367
        • Spergel D.J.
        • Kruth U.
        • Shimshek D.R.
        • Sprengel R.
        • Seeburg P.H.
        Using reporter genes to label selected neuronal populations in transgenic mice for gene promoter, anatomical, and physiological studies.
        Prog Neurobiol. 2001; 63: 673-686
        • Denk W.
        • Strickler J.H.
        • Webb W.W.
        Two-photon laser scanning fluorescence microscopy.
        Science. 1990; 248: 73-76
        • Helmchen F.
        • Denk W.
        Deep tissue two-photon microscopy.
        Nat Methods. 2005; 2: 932-940
        • Freudiger C.W.
        • Min W.
        • Saar B.G.
        • Lu S.
        • Holtom G.R.
        • He C.
        • Tsai J.C.
        • Kang J.X.
        • Xie X.S.
        Label-free biomedical imaging with high sensitivity by stimulated Raman scattering microscopy.
        Science. 2008; 322: 1857-1861
        • Min W.
        • Lu S.
        • Chong S.
        • Roy R.
        • Holtom G.R.
        • Xie X.S.
        Imaging chromophores with undetectable fluorescence by stimulated emission microscopy.
        Nature. 2009; 461: 1105-1109