Localization of Actin and Cytokeratin in Hepatocytes and Biliary Epithelial Cells in Normal Human Livers Using a Low Temperature Resin and
Post-Embedding I mmunohistochemistry
NAOKO KONG and YASUNI NAKANUMA The Second Department of Pathology,
University School of Medicine, Kanazawa 920
KoNo, N. and NAKANUMA, Y. Localization of Actin and Cytokeratin in Hepatocytes and Biliary Epithelial Cells in Normal Human Livers Using a Low Temperature Resin and Post-Embedding Immunohistochemistry. Tohoku J. Exp.
Med., 1989, 157 (4), 325-335 Using a low temperature resin Lowicryl K4M and post-embedding immunogold labeling, localization of actin and cytokeratin was examined in hepatocytes and biliary epithelial cells in normal human livers.
Preservation of the liver tissues was satisfactory at the electron microscopic level.
In immunolabeling of actin, gold particles were mainly found on microvilli and around the junctional complexes and at cell borders in the hepatocytes as well as biliary epithelial cells. In immunolabeling of cytokeratin, the gold particles were rather preferentially found in the perinuclear cytoplasm and around the junctional complexes in the biliary epithelial cells and hepatocytes. These particles were mainly found to be located on the intermediate filaments at higher magnification, although a considerable number of intermediate filaments were not labeled in the hepatocytes and biliary epithelial cells. Double staining method using the gold particles of different sizes clearly confirmed the abovementioned distribution of actin and cytokeratin in the same cells. From these observations, it was suggested that post-embedding immunoelectron microscopy using Lowicryl K4M is a useful tool for analysis of cytoskeletal organization in the liver tissue. cyto- skeleton ; post-embedding ; low temperature resin ; liver
Ultrastructural localization of antigens in tissues and cells has been perfor- med on relatively thick slices of unfixed or lightly fixed tissues which are then post-fixed, dehydrated and embedded (pre-embedding method) (Nakane and Piece 1966; Nagura et al. 1979). This procedure in time consuming and require skillful techniques. Recently, it has been found that many tissue antigens survive fixation, dehydration and resin embedding, and the technique of post-embedding
Received February 20, 1989; revision accepted for publication March 7 , 1989.
Mailing address : Yasuni Nakanuma, M.D., The Second Department of Pathology , K anazawa University School of Medicine, Kanazawa 920, Japan.
immunoelectron cytochemistry has been developed (Both et al. 1978 ; Roth 1982).
Unfortunately, the usual resins must be polymerized at temperature of 50-6o°C, and this may result in the reduction in the antigenic activity of some proteins.
This problem may be overcome if tissues are processed and embedded in resins at low temperature. Lowicryl K4M which has found to be a suitable embedding medium at low temperature, is now generally accepted to be the resin of choice of post-embedding immunoelectron microscopy (Carlemal et al. 1980 ; McPhail et al.
1987: Okanoue et al. 1988a, b).
Three filamentous system, actin-containing microfilaments, microtubules, and intermediate filaments (IFs), represent the major cytoskeletal components of most eukaryotic cells including hepatobiliary system (Imanari et al. 1981; Katsuma et al. 1987, 1988; Holm et al. 1988). The cytokeratins are the largest group within the IF group and are characteristic of epithelial cells. An interacting cytoskeletal network is believed to be necessary for the regulation of cellular function and the maintenance of cell shape and mobility (Imanari et al. 1981; Katsuma et al. 1987, 1988; Holm et al. 1988). Several pathologic conditions, including necrodegener- ative processes and neoplastic transformation, are known to interfere with the structure and interaction of the cytoskeletal system (Denk and Lackinger 1986;
In this study, the distribution of actin and cytokeratin was immunoelectron microscopically examined in the hepatocytes and adjoining biliary epithelial cells in normal human livers using a low temperature resin and post-embedding im- munoelectron microscopy. Such an approach has been rarely done in the he- patobiliary system to our knowledge (Lemanski et al. 1985; Okanoue et al.
MATERIALS AND METHODS
Surgical specimens of human livers (n = 2) were received fresh, and were dissected into two parts. One part was fixed in 10% buffered formalin and embedded in paraffin for light microscopic examination. The other was further dissected into cubes with sizes of approxi- mately 0.5 mm and fixed by immersion in a solution containing 1% glutaraldehyde in 0.1 M phosphate buffered solution (PBS) (pH 7.4) for 2 hr at room temperature. After an overnight buffer wash at 4°C, the tissue was dehydrated through graded ethanol and embedded in Lowicryl K4M according to the method reported by Carlemal et al. (1980), Roth et al. (1978, 1982) and Lemanski et al. (1985). That is, the fixed tissues in glutaralde- hyde were rinsed with 0.1 M PBS, and they were dehydrated in 30% ethanol for 30 min (0°C), 50% ethanol for 60 min (-20°C), 70% ethanol for 60 min (-35°C), 90% ethanol for 60 min (-35°C), and 100% ethanol for 60 min (-35°C). The dehydrated tissues were passed through by 100% Lowicryl K4M diluted 1:1 with 100% ethanol for 1 hr, 100% K4M diluted 3: 1 with 100% ethanol for 1 hr, 100% K4M for 1 hr and then in the second changes of 100% K4M overnight at -35°C. The tissues were polymerized for 24-48 hr in an ultraviolet radiation chamber at -35°C. Semithin sections (1.0 m) were stained with toluidine blue for selection of an appropriate area. For immunoelectron microscopy, ultra-thin sections (80-100 nm) were mounted on carbon-coated nickel grids and were immersed in 0.01 M PBS containing 1% bovine serum albumin for 20 min. The sections
were preincubated in a drop of mouse monoclonal antibodies to actin or cytokeratin for 1 hr, rinsed in PBS, incubated in a drop of secondary antibodies for 30 min and then reacted with 15 nm protein A-gold particles for 30 min at room temperature. The sections were then rinsed in distilled water, fixed by 1% osmiun tetraoxide for 30 min, and stained with uranyl acetate and lead citrate and were examined with a Hitachi H-300 electron micro- scope.
Double immunostaining method using protein A-gold of different particle sizes was used to demonstrate the distribution of actin and cytokeratin on the same tissue section.
Grid-mounted sections were subjected to the sequential double protein A-gold staining method, described by Gueze and associates (1981) and Roth (1978, 1982). The sections were incubated in 1% bovine serum albumin for 20 min. The sections were then incubated in the primary antibody to actin and incubated by secondary antibodies and then a protein A-gold labeled with a smaller gold particle (5 nm). Later, the reaction sites of the second primary monoclonal antibody to cytokeratin were identified by secondary antibodies and then protein A-gold labeled with a larger gold particle (15 nm). The sections were then rinsed in distilled water, fixed by 1% osmium tetraoxide for 30 min, and stained with uranyl acetate and lead citrate, and then examined with a Hitachi H-300 electron microscope.
A list of primary and secondary antibodies used in this study and their optimal dilution are shown in Table 1. Protein A-gold of different particle sizes (5 nm and 15 nm) were purchased from E.Y. Lab. Inc., (San Mateo, CA, USA). Monoclonal antibodies (KL-1 and CAM) against cytokeratin are known to stain hepatocytes as well as biliary epithelial cells in man (Van Eyken et al. 1987).
No positive stain was obtained when either of primary monoclonal antibodies, secon- dary antibodies or protein A-gold had been omitted or replaced by PBS in the staining procedures, respectively.
Liver tissues were generally well preserved and identification of cellular organelles were possible at the ultrastructural level, though the pictures were not so cripsy compared to those of the liver tissues which were fixed in 2% glutaralde- hyde, postfixed in 1% osmium tetraoxide, and then embedded in epoxy resin (Ray 1987). Immunolabeling using monoclonal and polyclonal antibodies produced a consistent specific labeling with negligible background, and individual gold particles were clearly identified in the sections stained with uranyl acetate and
TABLE 1. List of primary and secondary anti bod ies and their optimal dilution
Fig 1. Immunocytochemical localization of actin in hepatocytes. Gold particles are mainly present at the bile canalicular region, microvilli, along the lateral cell border and at the sinusoidal microvilli. Gold particles are also seen on the sinusoidal endothelial cells. Arrows denote bile canaliculi. X 5,600.
Fig 2. Immunohistochemical localization of actin in hepatocytes. Gold parti- cles are found on microvilli and pericanalicular cytoplasm and around the junctional complexes. A few microvilli are devoid of these particles.
Fig . 3. Immunocytochemical localizatin of actin in biliary epithelial cells. Gold particles are mainly present at the microvilli and around the junctional complexes. There are also gold particles in periductal fibroblast (F) and endothelial cells of the portal vein (E). x 6,600.
Fig 4. Immunocytochemical localization of actin in Gold particles are mainly found on microvilli and complexes (arrow). X 13,000.
lead citrate, when optimal dilution of these staining reagents was applied. The best results with these antibodies were obtained on the section surfaces of Lowicryl K4M embedded liver tissue. Reactivity with these antibodies became reduced in the sections obtained from the inner portions of the blocks.
In immunolabeling of actin, the gold particles were localized on the bile canalicular and sinusoidal microvilli, around bile canaliculi and the lateral cell borders of the hepatocyyes (Figs. 1 and 2). Gold particles were virtually absent in nucleus, mitochondria, rough and smooth endoplasmic reticulum, and in bile canalicular lumina. In the biliary epithelial cells the particles were localized on the lumina,l microvilli, and around the junctional complexes, especially around the tight junction, and also at the cell borders (Figs. 3 and 4). However, some of microvilli of the hepatocytes and biliary epithelial cells were devoid of gold particles, and gold particles were rather discontinuously distributed along the cell borders of some of the hepatocytes and biliary epithelial cells. Gold particles were also found in the cytoplasm of endothelial cells and other mesenchymal cells in the portal tracts and sinusoidal lining cells (Figs. 1 and 3).
In immunostaining of cytokeratin using monoclonal antibodies, the gold particles were scattered in the cytoplasm of the hepatocytes and biliary epithelial cells (Figs. 5, 6 and 7), especially in the perinuclear cytoplasm and around the junctional complexes, especially desmosomes. The gold particles were found in almost all biliary cells while they were inconstant in hepatocytes ; That is, gold
Fig.5. Immunocytochemical localization of cytokeratin in hepatocyte. Gold particles are mainly present at the pericanalicular cytoplasm and around the junctional complexes. x 36,000.
particles were found in some hepatocytes but rare or absent in others in the same section. At higher magnification, the gold particles were mainly attached on some bundles of IFs (Fig. 7), though a considerable number of IFs were not decorated by the gold particles. The gold particles were rarely found within the nucleus, mitochondria, rough and smooth endoplasmic reticulum, and bile canalicular or biliary lumina. The endothelial cells and other mesenchymal cells were virtually negative for these gold particles. There was no distinct difference between the staining pattern of cytokeratin in the hepatocytes or biliary epithelial
cells by CAM 5.2 and that by KL-l.
Double staining procedure disclosed simultaneously larger gold particles corressponding to cytokeratin and smaller ones to actin in the same hepatocytes as well as biliary cells (Fig. 8). Distribution of small particles was identical to that of actin and distribution of larger ones to that of cytokeratin, respectively, mentioned above.
The present study showed that preservation of liver tissues which were lightly Fig . 6. Immunocytochemical localization of cytokeratin in biliary epithelial cells.
Gold particles are mainly present at the perinuclear cytoplasm and around the junctional complexes. N, nucleus ; L, ductular lumen. X 26,000.
fixed in glutaraldehyde and embedded in Lowicry K4M at low temperature was satisfactory at the electron microscope level, consistent with the previous reports (Carlemal et al. 1980; Lemanski et al. 1985; McPhail et al. 1987; Okanoue et al.
1988a). Furthermore, the immunocytochemical labeling using protein A-gold as the electron dense markers, also showed that actin and cytokeratin were positively stained in the hepatocytes and biliary epithelial cells and background staining was negligible in Lowicry K4M embedded liver tissue (Lemanski et al. 1985) when the optimum dilution of primary or secondary antibodies and protein A-gold was applied. Double staining also showed heterogeneous distribution of actin and cytokeratin in the same hepatocytes and biliary epithelial cells, respectively.
The previous studies using the heavy meromyosin and ultrastructures at higher resolution (Oda et al. 1974; Imanari et al. 1981) disclosed that actin- microfilaments were localized at the canalicular and sinusoidal microvilli, around the bile canaliculi and at the lateral cell border of the hepatocytes. Biliary epithelial cells showed preferential distribution of actin-microfilaments on the luminal microvilli, around the desmosomes and also on the cell border (Okanoue et al. 1988h). The present study showed the similar distribution pattern of gold
Fig . 7. Immunocytochemical localization of cytokeratin particles are located on the bundles of intermediate clear cytoplasm. N, nucleus. X 23,000.
in biliary filaments in
the Gold perinu-
particles reflecting the presence of actin. These actin-microfilaments may be related to the tonus of canaliculus or ductal lumen, formation of microvilli, perstalsis and secretion of the hepatocytes as well as biliary epithelial cells
(Imanari et al. 1981; Denk and Lackinger 1986; Van Eyken et al. 1988).
It is recently pointed out that a very dense filamentous network of IFs was present throughout the cytoplasm of the hepatocytes as well as biliary epithelial cells, especially dense around the bile canaliculi, at the cell borders and in the perinuclear cytoplasm in the hepatocytes as well as biliary epithelial cells (Ka- tsuma et al. 1987, ; Okanoue et al. 1988a, b). The present study showed that gold particles reflecting the presence of cytokeratin were scattered in the cyto- plasm, but rather dense in the perinuclear cytoplasm, around the desmosomes, and
at the cell borders. Gold particles in the hepatocytes outnumbered those in the biliary epithelial cells supports the finding that IFs in biliary epithelial cells were more numerous compared with hepatocytes (Nakanuma et al. 1982; Okanoue et al. 1988b).
Van Eyken et al. (1987, 1988) reported using light histochemistry and a panel of commercially available anticytokeratin antibodies, reported that monoclonal antibodies such as CAM 5.2 and KL-1 stained intensely biliary epithelial cells and, to a lesser degree in the staining intensity and distribution, hepatocytes.
And they postulated that antigenecity of cytokeratin of the hepatocytes is
Fig.8. Biliary epithelial cell. Sequential double method. Gold particles of 5 nm (small arrows) arrows) corresponding to actin and cytokeratin, lumen. x66,000.
protein A-gold staining and those of 15 nm (large
respectively. L, ductular
different from that of bile duct epithelial cells (Burton et al. 1984; Van Eyken et al. 1987, 1988). Katsuma et al. (1987, 1988) recently reported that antigenecity of IFs were also different in the different parts of hepatocytes. The data obtained in this study that reaction products of cytokeratin were rather constantly found in a majority of biliary epithelial cells while they were inconstant in hepatocytes, support the heterogeneous antigenecity of cytokeratin in hepatocytes and also the difference of antigenecity of cytokeratin between hepatocytes and biliary epith- elial cells, above mentioned.
In conclusion, the method used in this study seems useful for the detection of actin microfilaments as well as cytokeratin IFs in the hepatocytes or biliary epithelial cells. However, some of microvilli were negative for actin and a considerable number of IFs were negative for cytokeratin in the hepatocytes as well as biliary epithelial cells, suggesting that the tissue processing in this study might have masked or destroyed part of the immunoreactivity of actin and cytokeratin. Because a majority of IFs were negative in the biliary epithelial cells and hepatocytes, a majority of epitopes of cytokeratin were masked in IFs in these cells and some devices are necessary to remove this masking in the im- munolabeling of cytokeratin. Further studies seem mandatory to resolve this problem.
1) Burton Vila, MM., Haot, J. & Desmet, V.J. (1984) Cholestatic features in focal nodular hyperplasia of the liver. Liver, 4, 387-395.
2) Carlemal, R., Villiger, W. & Acertarin, J.D. (1980) Advanced in specimen prepara- tion for electron microscopy. I. Novel-low-temperature embedding resin and a refor- mulated Vestal Experimentia, 36, 740.
3) Denk, H. & Lackinger, E. (1986) Cytoskeleton in liver diseases. Semin. Liver Dis., 6, 199-211.
4) Geuze, H.J., Slot, J.W., Van Cer Ley, P.A., Scheffer, C.T. & Griffith, J.M. (1981) Use of colloidal gold particles in doublelabeling immunoelectron microscopy of ultrathin frozen sections. J. Cell Biol., 89, 653-665.
5) Holm, R., Nesland, J.M., Attramadal, A. & Johannessen, J.V. (1988) Double- staining methods at the ultrastructural level applying colloidal gold conjugates.
Ultrastruct. Pathol., 12, 279-290.
6) Imanari, H., Kuroda, H. & Tamura, K. (1981) Microfilaments around the bile canaliculi in patients with intrahepatic cholestasis. Gastroenterol. Jpn., 16, 168-173.
7) Katsuma, Y., Swierenga, S.H.H., Khettry, U., Marceau, N. & French, SW. (1987) Changes in the cytokeratin intermediate filament cytoskeleton associated with Mallory body formation in mouse and human liver. Hepatology, 7, 1215-1223.
8) Katsuma, Y., Marceau, N., Ohta, M. & French, SW. (1988) Cytokeratin intermediate filaments of rat hepatocytes : Different cytoskeletal domains and their three- dimentional structure. Hepatology, 8, 559-568.
9) Lemanski, L.F., Paulson, D.J., Hill, C.S., Davis, L.A., Riles, L.C. & Lim, S.S. (1985) Immunoelectron microscopic localization of a-actin on Lowicryl embedded thin- sectioned tissues. J. Histochem. Cytochem., 33, 515-522.
10) McPhail, G.D., Finn, T. & Isaacson, P.G. (1987) A useful low temperature method for post-embedding electron immunocytochemistry in routine histopathology. J.
Pathol., 151, 231-238.
11) Nagura, H., Brantzaly, P., Nakane, P. & Brown, W.R. (1979) Ultrastructural locali- zation of J chain in human intestinal mucosa. J. Immunol., 123, 1044-1050.
12) Nakane, P. & Piece, G.G. (1966) Enzyme-labeled antibodies : Preparation for locali- zation of antigens. J. Histochem. Cytochem., 14, 929-931.
13) Nakanuma, Y., Kono, N., Ohta, G., Kobayashi, K. & Kato, Y. (1982) Ultrastructural changes of bile duct epithelium in primary biliary cirrhosis in relation to progression of bile duct loss. Virchow Arch.(A), 398, 149-161.
14) Oda, M., Price, V.M., Fisher, M.M. & Phillips, M.J. (1974) Ultrastructure of bile canaliculi, with special reference to the surface coat and the pericanalicular web.
Lab. Invest., 31, 314-323.
15) Okanoue, T., Ohta, M., Kachi, K., Ohta, Y., Kanaoka, H., Sawa, Y., Kagawa, K., Takino, T. & French, SW. (1988a) Scanning electron microscopy of the hepatocyte
cytoskeleton in human liver tissue. J. Hepatol., 6, 291-298.
16) Okanoue, T., Ohta, M., Kachi, K., Ohta, Y., Sawa, Y., Kanaoka, H., Kagawa, K., Takino, T. & French, SW. (1988b) Intermediate filaments of hepatocytes and biliary epithelial cells in bile duct obstruction : Transmission and scanning electron microscopy study. Gastroenterol. Jpn., 23, 428-434.
17) Ray, MB. (1987) Distribution patterns of cytokeratin antigen determinants in alcoholic and nonalcoholic liver diseases. Hum. Pathol., 18, 61-66.
18) Roth, J. (1982) The preparation of protein A-gold complexes with 3 nm and 15 nm gold particles and their use in labeling multiple antigens on ultra-thin section.
Histochem. J., 14, 791-801.
19) Roth, J., Bendayan, M. & Orci, L. (1978) Ultrastructural localization of intracellular antigens by the use of protein A-gold complex. J. Histochem. Cytochem., 26, 1074-
20) Van Eyken, P., Sciot, R., Van Damme, B., Wolf-Peters, C.D. & Desmet, V.J. (1987) Keratin immunohistochemistry in normal human liver. Cytokeratin pattern of
hepatocytes, bile ducts and acinar gradient. Virchow Archiv. CAS, 412, 63-72.
21) Van Eyken, P., Sciot, R. & Desmet, V. (1988) Intrahepatic bile duct development in the rat : A cytokeratin-immunohistochemical study. Lab. Invest., 59, 52-59.