JOURNAL OF ELECTRON MICROSCOPY TECHNIQUE 18:223-230 (1991)
Tridimensional Ultrastructure of Perfusion Fixed Gastrointestinal Epithelial Cells by High Resolution Scanning Electron Microscopy CYRIL W. KENDALL, A. VENKETESHWER RAO, SUSAN A. JANEZIC, ROBERT J. TEMKIN, MARTIN J. HOLLENBERG, AND PETER J. LEA Departments of Nutritional Sciences (C.W.K., A.V.R., S.A.J.) and Anatomy (R.J.T., M.J.H., P.J.L.), Faculty of Toronto, Toronto, Ontario, Canada M5S 1A8
KEY WORDS
of
Medicine, University
SEM, Intestinal morphology, Intracellular structure, Mitochondria, Cell membrane
ABSTRACT Improvements in the design of modern scanning electron microscopes (SEM) and new methods of specimen preparation incorporating chemical removal of the cytosol and cytoskeleton, now make it possible to view cells and their organelles in three dimensions (3D) at high magnification. In this experiment, high resolution SEM (HRSEM) utilizing new methods of tissue preparation was used to study the intracellular structures of the mouse ileum. In addition, in vivo intestinal perfusion was used to further enhance cellular preservation. Using these modifications it was possible to visualize, in 3D, the fine structure of intestinal epithelial cells and intracellular organelles such as the nucleus, mitochondria, endoplasmic reticulum, and Golgi complex, as well as microvilli and cell membrane. Whole mitochondria appeared as irregularly shaped organelles which contained tubular cristae. Plate-like cristae were not observed. The brush border was found to be a closely packed array of cylindrical projections. The extensive folding and structural intricacy of lateral cell membranes between absorptive cells could only be appreciated by viewing this tissue with 3D HRSEM. The use of HRSEM to study 3D ultrastructure of cells and their organelles will improve our understanding of the structure-function relationships in both the healthy and diseased gastrointestinal tract. INTRODUCTION
microscopy (TEM) of serial thin sections in combination with computer assisted 3D reconstruction (Lea and Our current understanding of the normal surface Pawlowski, 1986). However, these methods are technistructure of the intestinal tract is largely based on ob- cally demanding and time consuming and are exservations from standard scanning electron microscopy tremely selective given the fact that only small areas of (SEM). SEM has also been extensively used to charac- the cell can be readily surveyed. Due to improvements in design of the modern SEM terize surface morphological changes associated with such intestinal disorders as ulcerative colitis (Siew, and better methods of specimen preparation incorpo19831, Crohn’s disease (Rickert and Carter, 19771, and rating chemical removal of the cytosol and the cyintestinal neoplasms (Riddell et al., 1977; Weese et al., toskeleton, a fast, comprehensive and reliable method 1987). These studies have provided some insight into of viewing cellular components in 3D now exists. High the relationship between intestinal structure and func- resolution scanning electron microscopy (HRSEM) has tion and have partially illustrated the pathological pro- been successfully used to study intracellular organelle cesses involved in these diseased states. Furthermore, structure in such varied organs as the liver, retina, SEM has been used to study the surface structural kidney, thyroid, and brown adipose tissue (Hollenberg changes associated with differing dietary habits in hu- and Lea, 1989; Lea and Hollenberg, 1988). The resolumans (Cooke et al., 1968; Owen and Brandborg, 1977) tion of the SEM image produced by the HRSEM techand the consumption of various dietary components in nique is comparable to a conventional TEM in the biexperimental animals (Cassidy et al., 1981; Paulini et ological mode and also enables cell and organelle structure to be viewed in stereo 3D. al., 1987). While standard SEM techniques have been helpful In this study HRSEM was used to investigate the in showing the surface structural alterations associ- intracellular structure of cells found in the mouse ated with the above conditions, they have not allowed the detailed observation of intracellular ultrastructure in three-dimensions (3D). In order to study the spatial organization of cells and their organelles, researchers have used techniques including freeze-fractureifreeze Received November 23, 1989; accepted in revised form January 15, 1990. etching (Caldwell et al., 1987; Heuser and KirchAddress reprint requests to A.V. Rao, Department of Nutritional Sciences, hausen, 1985), or conventional transmission electron Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada M5S 1A8.
0 1991 WILEY-LISS. INC.
Figs. 1-4
HRSEM OF GASTROINTESTINAL EPITHELIUM
small intestine. In addition, a unique method of tissue fixation utilizing in vivo intestinal perfusion with a fixative was used to further enhance tissue preservation (Kendall e t al., 1987). Using these techniques it was possible to study, a t comparatively high magnification, aspects of 3D cell architecture and organelle ultrastructure of epithelial cells from mouse ileum.
MATERIALS AND METHODS Female C57 BL/6J and Balb cJ mice, approximately 16 weeks of age and 20-25 grams, were maintained on a standard AIN 76A diet in a n accredited facility and treated in compliance with the Ontario Animals for Research Act. Tissue Preparation for HRSEM Mice were anaesthetized with intraperitoneal sodium pentobarbitol injections. The abdominal cavity of the animals was opened by a midline incision to expose the intestinal tract and a needle inserted for delivery of the perfusate. The intestinal lumen was initially perfused for 5 minutes with a control buffer to clear it of debris. The control buffer (NaC1 120 mmol/l, KC1 5 mmol/l and NaHCO, 23 mmol/l) was isotonic with mouse ileostomy fluid at 296 mOsm/kg, pH 8.0 (Chadwick et al., 1979). This was followed by a primary fixative (0.5% glutaraldehyde and 0.5% paraformaldehyde in 0.07 M phosphate buffer (ph 7.2)) which was perfused for 10-15 minutes. Sections of the ileum (3-5 mm2) were excised after primary fixation (Kendall et al., 1987) and processed for HRSEM (Lea and Hollenberg, 1988). The tissue samples were washed in phosphate buffer (20-30 minutes) and post-fixed in 1.0% osmium tetroxide (OsO,) in phosphate buffer for 1.5 hours. Samples were again washed in buffer, placed in 25% dimethylsulfoxide (DMSO) in distilled water for 30 minutes, then in 50% DMSO for 30 minutes. Samples were freeze cleaved by placing them in liquid freon 22, transferring to a brass well immersed in
Fig. 1. HRSEM micrograph of a n intestinal villus. The epithelium consists of a single layer of cells, primarily columnar absorptive cells (AC) lined up along the border of the villus with perpendicular orientation to the microvilli (MV). A central lacteal (L) occupies the core of the villus. X 790. Fig. 2. HRSEM micrograph of columnar absorptive cells. Oval nuclei (N) located in the basal zone of the cell, occupy much of the cell's total volume. Cell membrane (CM) and many mitochondria (M) are observed. x 6,600. Fig. 3. HRSEM micrograph of columnar absorptive cell nucleus (N). Cleavage plane through a nucleus reveals the form and distribution of chromatin. The nucleus is fairly homogeneous, but contains denser regions of nucleoli (arrows) which appear to be distributed throughout its volume. x 17,700. Fig. 4. HRSEM micrograph of lateral cell membranes (CM) reveals their highly complex folding. Tightness in packing of adjacent columnar absorptive cells is achieved by intertwining of cell membranes. The cell membranes of the lower two cells appear to be entwined into a rope-like structure (arrow). A number of subcellular organelles can also be observed including Golgi complex (G), microvilli (MV), mitochondria (MI, and nucleus (N). x 6,900.
225
liquid nitrogen and cleaved with a liquid nitrogen chilled razor blade which was driven by a sharp blow from a hammer. Cleaved samples were thawed in 50% DMSO a t room temperature, washed in buffer, and post-fixed with 1.0% OsO, in phosphate buffer for 1hour. The samples were once again washed in buffer and the cytosol was extracted by immersing the samples in 0.1% OsO, in phosphate buffer for up to 6 days. The specimens were washed in buffer, placed in 1.0% tannic acid in distilled water for 1 hour, washed, placed in 1.0% OsO, for 1 hour, and returned to buffer. Subsequently, tissue samples were dehydrated in increasing concentrations of ethanol and then dried with a liquid COz critical point dryer (Tousimis 810 Sample Dryer, Rockville, MD). Specimens were mounted on aluminum stubs using carbon conductive paint.
Metal Coating of HRSEM Samples Samples were sputter coated with gold-palladium using argon gas as the ionizing plasma (Polaron SEM E5100 Coating System, United Kingdom). The average thickness of gold-palladium metal film that was applied to the samples was 10 nm. This film thickness resulted in optimal signal to noise ratios for the high resolution mode of imaging used in the SEM (Lea and Hollenberg, 1989a). SEM A Hitachi Model S-570 SEM, with a tungsten filament set a t a n accelerating voltage of 20 KV and a working distance of -2 mm was used to examine the specimens. A negative working distance means that the specimen was introduced into the bore of the objective lens to a maximum of 2 mm. The upper secondary electron detector, situated in the SEM lens above the objective lens, was used to collect electrons. Micrographs were recorded by photographing the surface of a high resolution cathode ray tube using Polaroid positivehegative film. RESULTS HRSEM made it possible to analyse aspects of both cell organization and organelle ultrastructure. By cleaving longitudinally through the finger-shaped villi of the mouse ileum it was possible to clearly identify absorptive, goblet, and Paneth cells. The epithelium consisted of a single layer of cells, primarily columnar absorptive cells, running perpendicular to the microvillar brush border (Fig. 1).These were interspersed with mucous secreting goblet cells. Secretory Paneth cells were observed in the base of the crypts. The nuclei were oval in shape and occupied the basal zone of the enterocyte (Fig. 2). In specimens where the cleavage plane passed through the nucleus, it was possible to observe the form and distribution of the strands of chromatin (Fig. 3). Between cells, the lateral cell membranes showed extensive infolding and it appeared that in some cases the cell membranes of adjacent cells were joined into a complex structure (Fig. 4). The brush border of the columnar absorptive cells was made up of a great number of very fine, closely packed microvilli, approximately
Figs. 5-8.
HRSEM OF GASTROINTESTINAL EPITHELIUM
227
Fig. 9. HRSEM of rough endoplasmic reticulum consisting of parallel arrays of flattened cisternae (C) and numerous ribosomes (arrows) found in a Paneth cell. Large secretory granules (V) and a mitochondrion (M) can also be observed. x 28,730.
Fig. 10. HRSEM micrograph of a goblet cell. The apical region of the cell (A) is distended and contains many large mucous droplets (D). The nucleus (N) and other cytoplasmic organelles are displaced towards the basal region of the cell (B). x 5,575.
1.2-1.8 ym in length and 0.12 ym in diameter (Fig. 5). Microvilli appeared as hollow tubes. Beneath them a space approximately 0.3 ym in width, was observed
between the base of the microvilli and other cell organelles. The fine network of actin filaments that normally occupies the microvilli had been solubilized and extracted from the cells along with the fine-textured cytoplasmic matrix. The terminal web which usually occupies the region just below the microvilli was also extracted, creating a space. Many small pinocytotic transport vesicles were observed throughout the cell but were found to be concentrated in the region just below the microvilli. The enterocytes contained large numbers of mitochondria, indicating the high energy requirements and metabolic activity of these cells specifically and the intestinal tract in general (Fig. 6). The mitochondria were distributed above and below the nucleus but were more numerous in the basal portion of the cell. Both the external and internal structure of the mitochondria could be observed (Fig. 7). These mitochondria clearly had tube-like cristae,t approximately 30 nm in diameter. The Golgi were generally located in the supranuhear region and were closely associated with the nucleus (Fig. 8). The major components of the Golgi, the appa-
Fig. 5. HRSEM micrograph of brush border reveals it to be composed of closely packed microvilli (MV). The proteinaceous coat normally found on the surface of the microvilli, the actin filaments which occupy the core of microvilli, and the filaments of the terminal web have been solubilized during tissue preparation leaving a space. Pinocytotic transport vesicles (arrows) are observed below the microvilli. x 16,200. Fig. 6. HRSEM micrograph shows that mitochondria (arrows) are distributed above and below the nucleus in the enterocyte but are much more numerous in the basal portion of the cell (B). x 6,590. Fig. 7. HRSEM micrograph of mitochondria (M) reveals they are irregularly shaped organelles containing tubular cristae (arrows) approximately 30 nm in diameter. x 25,000. Fig. 8. HRSEM micrograph of intracellular organelles of a n absorptive cell including Golgi complex (GI, mitochondria (M), nucleus (N), small vesicles (arrows), and a n intricately structured cell membrane (CM). x 19,500.
228
C.W. KENDALL ET AL.
ture was that much of the cytoplasm around the nucleus was occupied by rough endoplasmic reticulum cisternae in parallel arrays.
Fig. 11. HRSEM micrograph of a Paneth cell found in the base of a crypt. These cells are characterized by large, heterogeneous, membrane bound spherical granules (V) and extensive, rough endoplasmic reticulum (ER), and Golgi complex (G). x 11,670.
ratus, the transtubular network, and small transport vesicles could all be observed in 3D. Also associated with the nuclear region of the cell was the rough endoplasmic reticulum which consisted of cisternae with many ribosomes on their surface (Fig. 9). Goblet cells were found to be dispersed among the columnar absorptive cells. They were easily identified by large mucigen droplets (Fig. 10).The mucigen droplets were located primarily in the apical portion of the cell, which resulted in a distention of this region and gave these cells their typical goblet shape. A membrane enveloped each mucigen droplet. The nucleus and the rest of the cytoplasmic organelles were displaced toward the basal region of the cell. Sparse microvilli were observed on the cell surface. Paneth cells were located at or near the base of the crypts. They were distinguishable from the other cell types by the presence of large spherical secretory granules in the supranuclear region (Fig. 11). The granules appeared heterogeneous, having a dense core and a lighter peripheral zone. In humans and a number of other species these granules appear to be homogeneous (Erlandsen et al., 1974). Another distinguishing fea-
DISCUSSION The use of HRSEM in conjunction with a n improved method for tissue fixation utilizing in vivo lumenal perfusion, allowed for detailed morphological analysis of cell organization and organelle ultrastructure of the mouse ileum. Standard SEM techniques reveal the small intestine to be characterized by finger-shaped villi with a surface epithelium composed of tightly packed microvilli. Unfortunately they provide little information on the organization and structure of cellular components underlying the surface epithelium. With HRSEM a s applied in this study, it was possible to readily identify various cell types and to visualize, in 3D, their cellular organization and the ultrastructure of cellular organelles such as the nucleus, microvilli, cell membrane, mitochondria, endoplasmic reticulum, and Golgi complex. This method of tissue preparation also prevented ice crystal formation in the infranuclear region, a problem reported by Osatake e t al. (1985) when using rapid freeze fixation in a SEM study of the rat intestine. The HRSEM methodology as applied in our study proved especially useful for observing the 3D structure of mitochondria. Externally the mitochondria appeared as irregularly shaped, elongated rods while their internal structure was found to contain tubular cristae, approximately 30 nm in diameter. Mitochondria were generally thought to contain plate-like cristae. Our finding corroborates recent work by Lea and Hollenberg (198913) which has revealed that the majority of cell types examined contain tubular cristae, approximately 30 nm in diameter and not plate-like cristae. They proposed that the tubular cristae were continuous with the inner mitochondrial membrane and spanned the entire matrix of the organelle, thereby functionally increasing the surface area of the inner mitochondrial membrane. This interpretation also fits the morphology of the mitochondria observed in mouse ileum cells. The lateral cell membranes of the columnar absorptive cells appeared to be structurally complex, having a very high degree of folding. The amount of folding was so great that no gaps were found between adjacent cells and it appeared as if the cells were locked together. Such a n arrangement would be very important for maintaining the correct orientation and proper functioning of epithelial cells. The extent of folding and the closeness of packing of epithelial cells was much greater than that which has been described by conventional SEM or TEM methods (Creamer, 1974; Henry, 1982; Pfeiffer et al., 1974). The complexity of the true structure could only be appreciated by viewing the membranes in 3D. The brush border was observed to be a closely packed array of apparently hollow cylindrical projections. Normally, the lumenal surface of the plasmalemma of the microvilli is covered with a continuous glycocalyx (fuzzy coat) structure. The supporting cytoskeleton consists of actin filaments that root in the apex of the
HRSEM OF GASTROINTESTINAL EPITHELIUM
microvilli and attach to microtubules situated at their base in a region known as the terminal web (Creamer, 1974). It would appear that this protein coat, the fine filaments which occupy the core of each microvillus, as well as the filaments which make up the terminal web were solubilized during tissue preparation. Lea et al. (1990) have observed that intracellular, lipid associated membranes are well preserved during tissue preparation techniques for HRSEM whereas the cytoplasmic matrices and proteinaceous components such as actin, intermediate filaments, and microtubules are washed out. As a result, the proteins that make up the fuzzy coat and the filaments which occupy the central core of microvilli and the terminal web of columnar absorptive cells are solubilized. It has been suggested that some cellular proteins may be decomposed and eventually leached out of the cracked cell by the dilute osmium maceration process, while membrane proteins are protected from oxygen radical dependent denaturation by lipid molecules which preferentially absorb the oxygen radicals. The HRSEM technique employed in our study has several advantages over conventional SEM and TEM methods. The Hitachi 53-570 SEM obtains a resolution of approximately 3 nm, as compared with 15 nm using conventional SEM techniques and about 2 nm for the examination of biological specimens using TEM. In this regard, the HRSEM specimen preparation techniques allow direct comparisons to be made between TEM micrographs of cells and organelles from thin sections and our SEM micrographs. HRSEM also permits visualization of intracellular structures well below the fracture plane. In conventional freeze fracture preparations viewed by TEM, only that part of the sample which lies directly at the fracture surface can be visualized. HRSEM permits visualization of cell architecture to a depth of field equivalent to that contained in 15-30 conventional TEM thin sections. This depth of field also enables researchers to examine cell organization and organelle structure in 3D. This adds a new perspective and provides a more accurate picture of the true cell organization and organelle ultrastructure. Two major technological advances have made possible the examination of cells and organelles in 3D by HRSEM. The first is due to improvements in equipment design, which has enhanced the resolution capabilities of the Hitachi S-570 SEM. The second involves improvements in tissue preparation. In order to observe intracellular structures by SEM it is necessary to remove the cytoplasmic matrix of the cell. The chemical methods pioneered by Tanaka and co-workers have only partly achieved this (Osatake et al., 1985; Tanaka and Mitsushima, 1984). The method involves initially fixing the tissue, quick freezing, cleaving and thawing of the cell, then immersion for several days in dilute OsO,. While the exact action of OsO, is unknown (Lea et al., 1990),it is believed t o produce an initial gelation of cytosolic proteins which makes them more soluble and capable of leaching out of the cleaved cell. As reported in the literature, most researchers have relied upon in vitro immersion fixation to study intestinal morphology. Intraluminal injection of fixatives
229
has also been utilized. However, perfusion fixation has several advantages over immersion fixation. Further development and adaptation of perfusion fixation for SEM with very good ultrastructural preservation of mouse intestine has resulted in a more accurate description of intestinal morphology (Kendall et al., 1987). A comparison of the two methods has been discussed extensively by Glauert (1974)and demonstrated by Gertz et al. (1975). Vascular or in vivo lumenal perfusion allows the fixation of tissues before the death of the animal. This maximizes tissue preservation, minimizes alterations in cell structure, and allows for more uniform tissue preservation. The potential for mechanical damage is also reduced as the tissue hardens during fixation, before direct manipulation. Thus, in vivo lumenal perfusion as used in this study helped to prevent mechanical damage and reduced the number of artifacts while uniformly preserving comparatively normal intestinal epithelium. This methodology will enable researchers to study and better understand the structure-function relationships of cells and intracellular organelles of the intestinal epithelium. Studies are currently in progress in our laboratory to determine the structural-functional alterations associated with the progression of various gastrointestinal diseases and the morphological responses of the intestinal epithelium to dietary components.
ACKNOWLEDGMENTS The authors wish to thank B. Calvieri and D. Friday for their excellent technical assistance. REFERENCES Caldwell, R.B., Slapnick, S.M., and McLaughlin, B.J. (1987) Quantitative freeze fracture and filipin binding study of retinal pigment epithelial cell basement membranes in diabetic rats. Exp. Eye Res., 44:245-252. Cassidy, M.M., Lightfoot, F.G., Grau, L.E., Story, J.A., Kritchevsky. D., and Vahouney, G.V. (1981) Effect of chronic intake of dietary fiber on the ultrastructural topography of rat jejunum and colon: a scanning electron microscopy study. Am. J. Clin. Nutr., 34:218228. Chadwick, V.S., Gaginella, T.S., Carlson, G.L., Debongnie, J.C., Phillips, S.F., and Hoffman, A.F. (1979) Effect of molecular structure on bile acid-induced alterations in absorptive function, permeability and morphology in the perfused rabbit colon. J. Lab. Clin. Med., 94:661-669. Cooke, G.C., Kajubi, S.K., and Lu, F.D. (1968) Jejunal morphology of the African in Ueanda. J . Pathol.. 98:157-167. Creamer, B. (1974yMainly about structure. In: The Small Intestine. B. Creamer, ed. William Heinemann Medical Books Limited, London, pp. 1-23. Erlandsen, S.L., Parsons, J.A., and Taylor, T.D. (1974) Ultrastructural immunocytochemical localization of lysozyme in the Paneth Cells of man. J . Histochem. Cytochem., 22:401-413. Gertz, S.D., Rennels, M.L., Forbes, M.S., and Nelson, E. (1975) Preparation of vascular endothelium for scanning electron microscopy: a comparison of the effects of perfusion and immersion fixation. J . Microsc. 105:309-313. Glauert, A.M. (1974) Fixation, dehydration and embedding of biological specimens. In: Practical Methods in Electron Microscopy, vol. 3. A.M. Glauert, ed. American Elsevier Publishing Co. Ltd., New York, pp. 1-207. Henry, K. (1982) Ultrastructure of the small intestine. In: Small Intestine. V.S. Chadwick and S.F. Phillips, eds. Buttersworth Scientific, Toronto, pp. 19-39. Heuser, J.E., and Kirchhausen, T. (1985) Deep etch views of clathrin assemblies. J. Ultrastruct. Res., 92:l-27. Hollenberg, M.J., and Lea, P.J. (1989) Advantages of visualization of ~
230
C.W. KENDALL ET AL.
organelle structure by high resolution scanning electron microscopy. Scanning, 11:157-168. Kendall, C.W., Rao, A.V., and Lea, P.J. (1987) An in vivo intestinal perfusion model for studying the effects of dietary components on intestinal ultrastructure: a method utilizing scanning electron microscopy. Nutr. Rep. Int., 36:1221-1235. Lea, P.J., and Hollenberg, M.J. (1988) High resolution scanning electron microscopy of mitochondria from various rat tissues. Proc. Electron Microsc. SOC. Am., 46:14-15. Lea, P.J., and Hollenberg, M.J. (1989a) Mitochondrial structure revealed by high resolution scanning electron microscopy. Am. J . Anat., 184:245-257. Lea, P.J., and Hollenberg, M.J. (198913) Mitochondrial structure revealed by scanning electron microscopy (SEMI. In: Cells and Tissues: A Three Dimensional Approach by Modern Techniques in Microscopy. P.M. Motta, ed. Plenum Press, New York, pp. 63-70. Lea, P.J., Hollenberg, M.J., Temkin, R.J., and Khan, P.A. (1990) Chemical extraction of the cytosol using osmium tetroxide for high resolution scanning electron microscopy. J . Electron Microsc. Tech. (in press). Lea, P.J., and Pawlowski, A. (1986) Human melanocytic nevi. I. Electron microscopy and three dimensional computer reconstruction of nevi and basement membrane zone from ultrathin serial sections. Acta Dermatol. Venereol., 127:5-15. Osatake, J., Tanaka, K., and Inoue, T. (1985) An application of rapid
freezing, freeze substitution-fixation for scanning electron microscopy. J . Electron Microsc. Tech., 2201-208. Owen, R.L., and Brandborg, L.L. (1977) Jejunal morphological consequences of vegetarian diets in humans. Gastroenterology, 721111. Paulini, I., Mehta, T., and Hargis, A. (1987) Intestinal structural changes in African green monkeys after long term psyllium or cellulose feeding. J. Nutr., 117:253-266. Pfeiffer, C.J., Rowden, G., and Weibel, J . (1974) Gastrointestinal U1trastructure: An Overview of Scanning and Transmission Electron Micrographs. Academic Press Inc., San Diego, pp. 1-294. Rickert, R.R., and Carter, H.W. (1977)The gross light microscopic and scanning electron microscopic appearance of the early lesions of Crohn’s disease. Scanning Electron Microsc., II:179-186. Riddell, R.H., Path, M.R.C., and Levin, R. (1977) Ultrastructure ofthe ‘transitional’ mucosa adjacent to large bowel carcinoma. Cancer 40:2509-2522. Siew, S. (1983) The application of electron microscopy in the clinical investigation of the human colon. Scanning Electron Microsc. IV: 191 1-1929. Tanaka, K., and Mitsushima, A. (1984) A preparation method for observing intracellular structures by scanning electron microscopy. J . Microsc., 1 3 3 2 1 3 4 2 2 , Weese, J.L., Gilchrist, K.W., and Albrecht, R.M. (1987) Scanning microscopy of colonic mucin during carcinogenesis: Is i t clinically applicable? Ultrastruct. Pathol. 11:429-433.
Tridimensional Ultrastructure of Perfusion Fixed Gastrointestinal Epithelial Cells by High Resolution Scanning Electron Microscopy CYRIL W. KENDALL, A. VENKETESHWER RAO, SUSAN A. JANEZIC, ROBERT J. TEMKIN, MARTIN J. HOLLENBERG, AND PETER J. LEA Departments of Nutritional Sciences (C.W.K., A.V.R., S.A.J.) and Anatomy (R.J.T., M.J.H., P.J.L.), Faculty of Toronto, Toronto, Ontario, Canada M5S 1A8
KEY WORDS
of
Medicine, University
SEM, Intestinal morphology, Intracellular structure, Mitochondria, Cell membrane
ABSTRACT Improvements in the design of modern scanning electron microscopes (SEM) and new methods of specimen preparation incorporating chemical removal of the cytosol and cytoskeleton, now make it possible to view cells and their organelles in three dimensions (3D) at high magnification. In this experiment, high resolution SEM (HRSEM) utilizing new methods of tissue preparation was used to study the intracellular structures of the mouse ileum. In addition, in vivo intestinal perfusion was used to further enhance cellular preservation. Using these modifications it was possible to visualize, in 3D, the fine structure of intestinal epithelial cells and intracellular organelles such as the nucleus, mitochondria, endoplasmic reticulum, and Golgi complex, as well as microvilli and cell membrane. Whole mitochondria appeared as irregularly shaped organelles which contained tubular cristae. Plate-like cristae were not observed. The brush border was found to be a closely packed array of cylindrical projections. The extensive folding and structural intricacy of lateral cell membranes between absorptive cells could only be appreciated by viewing this tissue with 3D HRSEM. The use of HRSEM to study 3D ultrastructure of cells and their organelles will improve our understanding of the structure-function relationships in both the healthy and diseased gastrointestinal tract. INTRODUCTION
microscopy (TEM) of serial thin sections in combination with computer assisted 3D reconstruction (Lea and Our current understanding of the normal surface Pawlowski, 1986). However, these methods are technistructure of the intestinal tract is largely based on ob- cally demanding and time consuming and are exservations from standard scanning electron microscopy tremely selective given the fact that only small areas of (SEM). SEM has also been extensively used to charac- the cell can be readily surveyed. Due to improvements in design of the modern SEM terize surface morphological changes associated with such intestinal disorders as ulcerative colitis (Siew, and better methods of specimen preparation incorpo19831, Crohn’s disease (Rickert and Carter, 19771, and rating chemical removal of the cytosol and the cyintestinal neoplasms (Riddell et al., 1977; Weese et al., toskeleton, a fast, comprehensive and reliable method 1987). These studies have provided some insight into of viewing cellular components in 3D now exists. High the relationship between intestinal structure and func- resolution scanning electron microscopy (HRSEM) has tion and have partially illustrated the pathological pro- been successfully used to study intracellular organelle cesses involved in these diseased states. Furthermore, structure in such varied organs as the liver, retina, SEM has been used to study the surface structural kidney, thyroid, and brown adipose tissue (Hollenberg changes associated with differing dietary habits in hu- and Lea, 1989; Lea and Hollenberg, 1988). The resolumans (Cooke et al., 1968; Owen and Brandborg, 1977) tion of the SEM image produced by the HRSEM techand the consumption of various dietary components in nique is comparable to a conventional TEM in the biexperimental animals (Cassidy et al., 1981; Paulini et ological mode and also enables cell and organelle structure to be viewed in stereo 3D. al., 1987). While standard SEM techniques have been helpful In this study HRSEM was used to investigate the in showing the surface structural alterations associ- intracellular structure of cells found in the mouse ated with the above conditions, they have not allowed the detailed observation of intracellular ultrastructure in three-dimensions (3D). In order to study the spatial organization of cells and their organelles, researchers have used techniques including freeze-fractureifreeze Received November 23, 1989; accepted in revised form January 15, 1990. etching (Caldwell et al., 1987; Heuser and KirchAddress reprint requests to A.V. Rao, Department of Nutritional Sciences, hausen, 1985), or conventional transmission electron Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada M5S 1A8.
0 1991 WILEY-LISS. INC.
Figs. 1-4
HRSEM OF GASTROINTESTINAL EPITHELIUM
small intestine. In addition, a unique method of tissue fixation utilizing in vivo intestinal perfusion with a fixative was used to further enhance tissue preservation (Kendall e t al., 1987). Using these techniques it was possible to study, a t comparatively high magnification, aspects of 3D cell architecture and organelle ultrastructure of epithelial cells from mouse ileum.
MATERIALS AND METHODS Female C57 BL/6J and Balb cJ mice, approximately 16 weeks of age and 20-25 grams, were maintained on a standard AIN 76A diet in a n accredited facility and treated in compliance with the Ontario Animals for Research Act. Tissue Preparation for HRSEM Mice were anaesthetized with intraperitoneal sodium pentobarbitol injections. The abdominal cavity of the animals was opened by a midline incision to expose the intestinal tract and a needle inserted for delivery of the perfusate. The intestinal lumen was initially perfused for 5 minutes with a control buffer to clear it of debris. The control buffer (NaC1 120 mmol/l, KC1 5 mmol/l and NaHCO, 23 mmol/l) was isotonic with mouse ileostomy fluid at 296 mOsm/kg, pH 8.0 (Chadwick et al., 1979). This was followed by a primary fixative (0.5% glutaraldehyde and 0.5% paraformaldehyde in 0.07 M phosphate buffer (ph 7.2)) which was perfused for 10-15 minutes. Sections of the ileum (3-5 mm2) were excised after primary fixation (Kendall et al., 1987) and processed for HRSEM (Lea and Hollenberg, 1988). The tissue samples were washed in phosphate buffer (20-30 minutes) and post-fixed in 1.0% osmium tetroxide (OsO,) in phosphate buffer for 1.5 hours. Samples were again washed in buffer, placed in 25% dimethylsulfoxide (DMSO) in distilled water for 30 minutes, then in 50% DMSO for 30 minutes. Samples were freeze cleaved by placing them in liquid freon 22, transferring to a brass well immersed in
Fig. 1. HRSEM micrograph of a n intestinal villus. The epithelium consists of a single layer of cells, primarily columnar absorptive cells (AC) lined up along the border of the villus with perpendicular orientation to the microvilli (MV). A central lacteal (L) occupies the core of the villus. X 790. Fig. 2. HRSEM micrograph of columnar absorptive cells. Oval nuclei (N) located in the basal zone of the cell, occupy much of the cell's total volume. Cell membrane (CM) and many mitochondria (M) are observed. x 6,600. Fig. 3. HRSEM micrograph of columnar absorptive cell nucleus (N). Cleavage plane through a nucleus reveals the form and distribution of chromatin. The nucleus is fairly homogeneous, but contains denser regions of nucleoli (arrows) which appear to be distributed throughout its volume. x 17,700. Fig. 4. HRSEM micrograph of lateral cell membranes (CM) reveals their highly complex folding. Tightness in packing of adjacent columnar absorptive cells is achieved by intertwining of cell membranes. The cell membranes of the lower two cells appear to be entwined into a rope-like structure (arrow). A number of subcellular organelles can also be observed including Golgi complex (G), microvilli (MV), mitochondria (MI, and nucleus (N). x 6,900.
225
liquid nitrogen and cleaved with a liquid nitrogen chilled razor blade which was driven by a sharp blow from a hammer. Cleaved samples were thawed in 50% DMSO a t room temperature, washed in buffer, and post-fixed with 1.0% OsO, in phosphate buffer for 1hour. The samples were once again washed in buffer and the cytosol was extracted by immersing the samples in 0.1% OsO, in phosphate buffer for up to 6 days. The specimens were washed in buffer, placed in 1.0% tannic acid in distilled water for 1 hour, washed, placed in 1.0% OsO, for 1 hour, and returned to buffer. Subsequently, tissue samples were dehydrated in increasing concentrations of ethanol and then dried with a liquid COz critical point dryer (Tousimis 810 Sample Dryer, Rockville, MD). Specimens were mounted on aluminum stubs using carbon conductive paint.
Metal Coating of HRSEM Samples Samples were sputter coated with gold-palladium using argon gas as the ionizing plasma (Polaron SEM E5100 Coating System, United Kingdom). The average thickness of gold-palladium metal film that was applied to the samples was 10 nm. This film thickness resulted in optimal signal to noise ratios for the high resolution mode of imaging used in the SEM (Lea and Hollenberg, 1989a). SEM A Hitachi Model S-570 SEM, with a tungsten filament set a t a n accelerating voltage of 20 KV and a working distance of -2 mm was used to examine the specimens. A negative working distance means that the specimen was introduced into the bore of the objective lens to a maximum of 2 mm. The upper secondary electron detector, situated in the SEM lens above the objective lens, was used to collect electrons. Micrographs were recorded by photographing the surface of a high resolution cathode ray tube using Polaroid positivehegative film. RESULTS HRSEM made it possible to analyse aspects of both cell organization and organelle ultrastructure. By cleaving longitudinally through the finger-shaped villi of the mouse ileum it was possible to clearly identify absorptive, goblet, and Paneth cells. The epithelium consisted of a single layer of cells, primarily columnar absorptive cells, running perpendicular to the microvillar brush border (Fig. 1).These were interspersed with mucous secreting goblet cells. Secretory Paneth cells were observed in the base of the crypts. The nuclei were oval in shape and occupied the basal zone of the enterocyte (Fig. 2). In specimens where the cleavage plane passed through the nucleus, it was possible to observe the form and distribution of the strands of chromatin (Fig. 3). Between cells, the lateral cell membranes showed extensive infolding and it appeared that in some cases the cell membranes of adjacent cells were joined into a complex structure (Fig. 4). The brush border of the columnar absorptive cells was made up of a great number of very fine, closely packed microvilli, approximately
Figs. 5-8.
HRSEM OF GASTROINTESTINAL EPITHELIUM
227
Fig. 9. HRSEM of rough endoplasmic reticulum consisting of parallel arrays of flattened cisternae (C) and numerous ribosomes (arrows) found in a Paneth cell. Large secretory granules (V) and a mitochondrion (M) can also be observed. x 28,730.
Fig. 10. HRSEM micrograph of a goblet cell. The apical region of the cell (A) is distended and contains many large mucous droplets (D). The nucleus (N) and other cytoplasmic organelles are displaced towards the basal region of the cell (B). x 5,575.
1.2-1.8 ym in length and 0.12 ym in diameter (Fig. 5). Microvilli appeared as hollow tubes. Beneath them a space approximately 0.3 ym in width, was observed
between the base of the microvilli and other cell organelles. The fine network of actin filaments that normally occupies the microvilli had been solubilized and extracted from the cells along with the fine-textured cytoplasmic matrix. The terminal web which usually occupies the region just below the microvilli was also extracted, creating a space. Many small pinocytotic transport vesicles were observed throughout the cell but were found to be concentrated in the region just below the microvilli. The enterocytes contained large numbers of mitochondria, indicating the high energy requirements and metabolic activity of these cells specifically and the intestinal tract in general (Fig. 6). The mitochondria were distributed above and below the nucleus but were more numerous in the basal portion of the cell. Both the external and internal structure of the mitochondria could be observed (Fig. 7). These mitochondria clearly had tube-like cristae,t approximately 30 nm in diameter. The Golgi were generally located in the supranuhear region and were closely associated with the nucleus (Fig. 8). The major components of the Golgi, the appa-
Fig. 5. HRSEM micrograph of brush border reveals it to be composed of closely packed microvilli (MV). The proteinaceous coat normally found on the surface of the microvilli, the actin filaments which occupy the core of microvilli, and the filaments of the terminal web have been solubilized during tissue preparation leaving a space. Pinocytotic transport vesicles (arrows) are observed below the microvilli. x 16,200. Fig. 6. HRSEM micrograph shows that mitochondria (arrows) are distributed above and below the nucleus in the enterocyte but are much more numerous in the basal portion of the cell (B). x 6,590. Fig. 7. HRSEM micrograph of mitochondria (M) reveals they are irregularly shaped organelles containing tubular cristae (arrows) approximately 30 nm in diameter. x 25,000. Fig. 8. HRSEM micrograph of intracellular organelles of a n absorptive cell including Golgi complex (GI, mitochondria (M), nucleus (N), small vesicles (arrows), and a n intricately structured cell membrane (CM). x 19,500.
228
C.W. KENDALL ET AL.
ture was that much of the cytoplasm around the nucleus was occupied by rough endoplasmic reticulum cisternae in parallel arrays.
Fig. 11. HRSEM micrograph of a Paneth cell found in the base of a crypt. These cells are characterized by large, heterogeneous, membrane bound spherical granules (V) and extensive, rough endoplasmic reticulum (ER), and Golgi complex (G). x 11,670.
ratus, the transtubular network, and small transport vesicles could all be observed in 3D. Also associated with the nuclear region of the cell was the rough endoplasmic reticulum which consisted of cisternae with many ribosomes on their surface (Fig. 9). Goblet cells were found to be dispersed among the columnar absorptive cells. They were easily identified by large mucigen droplets (Fig. 10).The mucigen droplets were located primarily in the apical portion of the cell, which resulted in a distention of this region and gave these cells their typical goblet shape. A membrane enveloped each mucigen droplet. The nucleus and the rest of the cytoplasmic organelles were displaced toward the basal region of the cell. Sparse microvilli were observed on the cell surface. Paneth cells were located at or near the base of the crypts. They were distinguishable from the other cell types by the presence of large spherical secretory granules in the supranuclear region (Fig. 11). The granules appeared heterogeneous, having a dense core and a lighter peripheral zone. In humans and a number of other species these granules appear to be homogeneous (Erlandsen et al., 1974). Another distinguishing fea-
DISCUSSION The use of HRSEM in conjunction with a n improved method for tissue fixation utilizing in vivo lumenal perfusion, allowed for detailed morphological analysis of cell organization and organelle ultrastructure of the mouse ileum. Standard SEM techniques reveal the small intestine to be characterized by finger-shaped villi with a surface epithelium composed of tightly packed microvilli. Unfortunately they provide little information on the organization and structure of cellular components underlying the surface epithelium. With HRSEM a s applied in this study, it was possible to readily identify various cell types and to visualize, in 3D, their cellular organization and the ultrastructure of cellular organelles such as the nucleus, microvilli, cell membrane, mitochondria, endoplasmic reticulum, and Golgi complex. This method of tissue preparation also prevented ice crystal formation in the infranuclear region, a problem reported by Osatake e t al. (1985) when using rapid freeze fixation in a SEM study of the rat intestine. The HRSEM methodology as applied in our study proved especially useful for observing the 3D structure of mitochondria. Externally the mitochondria appeared as irregularly shaped, elongated rods while their internal structure was found to contain tubular cristae, approximately 30 nm in diameter. Mitochondria were generally thought to contain plate-like cristae. Our finding corroborates recent work by Lea and Hollenberg (198913) which has revealed that the majority of cell types examined contain tubular cristae, approximately 30 nm in diameter and not plate-like cristae. They proposed that the tubular cristae were continuous with the inner mitochondrial membrane and spanned the entire matrix of the organelle, thereby functionally increasing the surface area of the inner mitochondrial membrane. This interpretation also fits the morphology of the mitochondria observed in mouse ileum cells. The lateral cell membranes of the columnar absorptive cells appeared to be structurally complex, having a very high degree of folding. The amount of folding was so great that no gaps were found between adjacent cells and it appeared as if the cells were locked together. Such a n arrangement would be very important for maintaining the correct orientation and proper functioning of epithelial cells. The extent of folding and the closeness of packing of epithelial cells was much greater than that which has been described by conventional SEM or TEM methods (Creamer, 1974; Henry, 1982; Pfeiffer et al., 1974). The complexity of the true structure could only be appreciated by viewing the membranes in 3D. The brush border was observed to be a closely packed array of apparently hollow cylindrical projections. Normally, the lumenal surface of the plasmalemma of the microvilli is covered with a continuous glycocalyx (fuzzy coat) structure. The supporting cytoskeleton consists of actin filaments that root in the apex of the
HRSEM OF GASTROINTESTINAL EPITHELIUM
microvilli and attach to microtubules situated at their base in a region known as the terminal web (Creamer, 1974). It would appear that this protein coat, the fine filaments which occupy the core of each microvillus, as well as the filaments which make up the terminal web were solubilized during tissue preparation. Lea et al. (1990) have observed that intracellular, lipid associated membranes are well preserved during tissue preparation techniques for HRSEM whereas the cytoplasmic matrices and proteinaceous components such as actin, intermediate filaments, and microtubules are washed out. As a result, the proteins that make up the fuzzy coat and the filaments which occupy the central core of microvilli and the terminal web of columnar absorptive cells are solubilized. It has been suggested that some cellular proteins may be decomposed and eventually leached out of the cracked cell by the dilute osmium maceration process, while membrane proteins are protected from oxygen radical dependent denaturation by lipid molecules which preferentially absorb the oxygen radicals. The HRSEM technique employed in our study has several advantages over conventional SEM and TEM methods. The Hitachi 53-570 SEM obtains a resolution of approximately 3 nm, as compared with 15 nm using conventional SEM techniques and about 2 nm for the examination of biological specimens using TEM. In this regard, the HRSEM specimen preparation techniques allow direct comparisons to be made between TEM micrographs of cells and organelles from thin sections and our SEM micrographs. HRSEM also permits visualization of intracellular structures well below the fracture plane. In conventional freeze fracture preparations viewed by TEM, only that part of the sample which lies directly at the fracture surface can be visualized. HRSEM permits visualization of cell architecture to a depth of field equivalent to that contained in 15-30 conventional TEM thin sections. This depth of field also enables researchers to examine cell organization and organelle structure in 3D. This adds a new perspective and provides a more accurate picture of the true cell organization and organelle ultrastructure. Two major technological advances have made possible the examination of cells and organelles in 3D by HRSEM. The first is due to improvements in equipment design, which has enhanced the resolution capabilities of the Hitachi S-570 SEM. The second involves improvements in tissue preparation. In order to observe intracellular structures by SEM it is necessary to remove the cytoplasmic matrix of the cell. The chemical methods pioneered by Tanaka and co-workers have only partly achieved this (Osatake et al., 1985; Tanaka and Mitsushima, 1984). The method involves initially fixing the tissue, quick freezing, cleaving and thawing of the cell, then immersion for several days in dilute OsO,. While the exact action of OsO, is unknown (Lea et al., 1990),it is believed t o produce an initial gelation of cytosolic proteins which makes them more soluble and capable of leaching out of the cleaved cell. As reported in the literature, most researchers have relied upon in vitro immersion fixation to study intestinal morphology. Intraluminal injection of fixatives
229
has also been utilized. However, perfusion fixation has several advantages over immersion fixation. Further development and adaptation of perfusion fixation for SEM with very good ultrastructural preservation of mouse intestine has resulted in a more accurate description of intestinal morphology (Kendall et al., 1987). A comparison of the two methods has been discussed extensively by Glauert (1974)and demonstrated by Gertz et al. (1975). Vascular or in vivo lumenal perfusion allows the fixation of tissues before the death of the animal. This maximizes tissue preservation, minimizes alterations in cell structure, and allows for more uniform tissue preservation. The potential for mechanical damage is also reduced as the tissue hardens during fixation, before direct manipulation. Thus, in vivo lumenal perfusion as used in this study helped to prevent mechanical damage and reduced the number of artifacts while uniformly preserving comparatively normal intestinal epithelium. This methodology will enable researchers to study and better understand the structure-function relationships of cells and intracellular organelles of the intestinal epithelium. Studies are currently in progress in our laboratory to determine the structural-functional alterations associated with the progression of various gastrointestinal diseases and the morphological responses of the intestinal epithelium to dietary components.
ACKNOWLEDGMENTS The authors wish to thank B. Calvieri and D. Friday for their excellent technical assistance. REFERENCES Caldwell, R.B., Slapnick, S.M., and McLaughlin, B.J. (1987) Quantitative freeze fracture and filipin binding study of retinal pigment epithelial cell basement membranes in diabetic rats. Exp. Eye Res., 44:245-252. Cassidy, M.M., Lightfoot, F.G., Grau, L.E., Story, J.A., Kritchevsky. D., and Vahouney, G.V. (1981) Effect of chronic intake of dietary fiber on the ultrastructural topography of rat jejunum and colon: a scanning electron microscopy study. Am. J. Clin. Nutr., 34:218228. Chadwick, V.S., Gaginella, T.S., Carlson, G.L., Debongnie, J.C., Phillips, S.F., and Hoffman, A.F. (1979) Effect of molecular structure on bile acid-induced alterations in absorptive function, permeability and morphology in the perfused rabbit colon. J. Lab. Clin. Med., 94:661-669. Cooke, G.C., Kajubi, S.K., and Lu, F.D. (1968) Jejunal morphology of the African in Ueanda. J . Pathol.. 98:157-167. Creamer, B. (1974yMainly about structure. In: The Small Intestine. B. Creamer, ed. William Heinemann Medical Books Limited, London, pp. 1-23. Erlandsen, S.L., Parsons, J.A., and Taylor, T.D. (1974) Ultrastructural immunocytochemical localization of lysozyme in the Paneth Cells of man. J . Histochem. Cytochem., 22:401-413. Gertz, S.D., Rennels, M.L., Forbes, M.S., and Nelson, E. (1975) Preparation of vascular endothelium for scanning electron microscopy: a comparison of the effects of perfusion and immersion fixation. J . Microsc. 105:309-313. Glauert, A.M. (1974) Fixation, dehydration and embedding of biological specimens. In: Practical Methods in Electron Microscopy, vol. 3. A.M. Glauert, ed. American Elsevier Publishing Co. Ltd., New York, pp. 1-207. Henry, K. (1982) Ultrastructure of the small intestine. In: Small Intestine. V.S. Chadwick and S.F. Phillips, eds. Buttersworth Scientific, Toronto, pp. 19-39. Heuser, J.E., and Kirchhausen, T. (1985) Deep etch views of clathrin assemblies. J. Ultrastruct. Res., 92:l-27. Hollenberg, M.J., and Lea, P.J. (1989) Advantages of visualization of ~
230
C.W. KENDALL ET AL.
organelle structure by high resolution scanning electron microscopy. Scanning, 11:157-168. Kendall, C.W., Rao, A.V., and Lea, P.J. (1987) An in vivo intestinal perfusion model for studying the effects of dietary components on intestinal ultrastructure: a method utilizing scanning electron microscopy. Nutr. Rep. Int., 36:1221-1235. Lea, P.J., and Hollenberg, M.J. (1988) High resolution scanning electron microscopy of mitochondria from various rat tissues. Proc. Electron Microsc. SOC. Am., 46:14-15. Lea, P.J., and Hollenberg, M.J. (1989a) Mitochondrial structure revealed by high resolution scanning electron microscopy. Am. J . Anat., 184:245-257. Lea, P.J., and Hollenberg, M.J. (198913) Mitochondrial structure revealed by scanning electron microscopy (SEMI. In: Cells and Tissues: A Three Dimensional Approach by Modern Techniques in Microscopy. P.M. Motta, ed. Plenum Press, New York, pp. 63-70. Lea, P.J., Hollenberg, M.J., Temkin, R.J., and Khan, P.A. (1990) Chemical extraction of the cytosol using osmium tetroxide for high resolution scanning electron microscopy. J . Electron Microsc. Tech. (in press). Lea, P.J., and Pawlowski, A. (1986) Human melanocytic nevi. I. Electron microscopy and three dimensional computer reconstruction of nevi and basement membrane zone from ultrathin serial sections. Acta Dermatol. Venereol., 127:5-15. Osatake, J., Tanaka, K., and Inoue, T. (1985) An application of rapid
freezing, freeze substitution-fixation for scanning electron microscopy. J . Electron Microsc. Tech., 2201-208. Owen, R.L., and Brandborg, L.L. (1977) Jejunal morphological consequences of vegetarian diets in humans. Gastroenterology, 721111. Paulini, I., Mehta, T., and Hargis, A. (1987) Intestinal structural changes in African green monkeys after long term psyllium or cellulose feeding. J. Nutr., 117:253-266. Pfeiffer, C.J., Rowden, G., and Weibel, J . (1974) Gastrointestinal U1trastructure: An Overview of Scanning and Transmission Electron Micrographs. Academic Press Inc., San Diego, pp. 1-294. Rickert, R.R., and Carter, H.W. (1977)The gross light microscopic and scanning electron microscopic appearance of the early lesions of Crohn’s disease. Scanning Electron Microsc., II:179-186. Riddell, R.H., Path, M.R.C., and Levin, R. (1977) Ultrastructure ofthe ‘transitional’ mucosa adjacent to large bowel carcinoma. Cancer 40:2509-2522. Siew, S. (1983) The application of electron microscopy in the clinical investigation of the human colon. Scanning Electron Microsc. IV: 191 1-1929. Tanaka, K., and Mitsushima, A. (1984) A preparation method for observing intracellular structures by scanning electron microscopy. J . Microsc., 1 3 3 2 1 3 4 2 2 , Weese, J.L., Gilchrist, K.W., and Albrecht, R.M. (1987) Scanning microscopy of colonic mucin during carcinogenesis: Is i t clinically applicable? Ultrastruct. Pathol. 11:429-433.
