- Email: [email protected]
Distribution of the surface antigen HAM-4 and cytoskeleton during reformation of bile-canalicular structures in rat primary cultured hepatocytes
EXPERIMENTAL
CELL
RESEARCH
199,
50-55 (19%)
Distribution of the Surface Antigen HAM-4 and Cytoskeleton during Reformation of Bile-Canalicular Structures in Rat Primary Cultured Hepatocytes TOMOO SAWADA, HIROSHI ITAI, YOSHIHISA FUJIKURA, HIROMICHI MASAKATSU TAMECHIKA, AND TETSUO FUKUMOTO’ Department
of Anatomy,
Yamaguchi
University
School
INTRODUCTION
Hepatocytes are of epithelial origin. As in other epithelia hepatocytes have cell surface polarity, with three different surface domains: the apical (bile-canalicular) surface, the lateral (contiguous) surface, and the basal (sinusoidal) surface. Surface polarity results in many molecules being differentially distributed in each surface domain [l-5]. Some monoclonal antibodies against cell surface molecules can detected these differences in molecular composition of the three surface domains (6, 71. HAM-4 is a monoclonal antibody which recognizes a wheat germ agglutinin-binding protein on the surface of
0014.4827/92
$3.00
Yamaguchi
MATERIALS Preparation
755, Japan
and Primary
Culture
AND
METHODS
of Dissociated
Hepatocytes
Hepatocytes were prepared by the method described by Berry and Friend [16] and Seglen [17]. An adult DA rat was perfused through the portal vein and out through the abdominal aorta for 30-45 min with Ca’+-free Krebs-Henseleit Ringer bicarbonate buffer containing 0.05% collagenase, pH 7.5, saturated with 95% O2 and CO,. The liver was teased with a spatula and filtered through four layers of gauze. The dissociated hepatocytes were suspended in 50 ml of Hanks’ buffer solution and allowed to settle for 15 min. The pellet was then resuspended. After three such washes, remaining bile canaliculi be-
requests should be addressed.
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
Ube-City,
rat hepatocytes. The antigen recognized by HAM-4 is distributed mainly on the surface of the bile canaliculus (BC) both in uiuo and in vitro in dissociated (free) hepatocytes [&lo]. In primary culture, hepatocytes reform structures (BC-structures) corresponding to the bile canaliculus in uiuo and this reconstruction (BC-reconstruction) provides a model in which the surface polarity can be investigated. Previous reports have described the fundamental processes of BC-reconstruction: formation of intercellular closed spaces which become the BC-structures [ 111, formation of cell-to-cell junctional structures [12], and recovery of some biochemical characteristics [13]. In relation to the cytoskeleton, Gautam et al. [14] reported that the redistribution of Mg’+-ATPase into bile canaliculi depended on actin filaments but not on microtubules. It has also been found that microtubule inhibitors do not inhibit BC-reconstruction [15]. Although there have been numerous studies of BCreconstruction in vitro and the reacquisition of cell surface polarity, the development of hepatocyte surface polarity is not yet fully understood. Hepatocytes have many different surface molecules and there may be distinct mechanisms controlling the distribution of each one. HAM-4 antigen represents a specific molecular marker of the apical hepatocyte surface. We therefore investigated the distribution of HAM-4 antigen molecules during BC-reconstruction and focused on the role of microtubules during this process.
The distribution of rat bile-canalicular surface antigen (HAM-4 antigen) and cytoskeletal elements (microtubules, actin filaments, and cytokeratin filaments) was examined during the reformation of bile-canalicular structures (BC-structures) in primary cultures of dissociated hepatocytes obtained following collagenase perfusion. HAM-4 antigen, which initially dispersed after cell dissociation, became focused into regions of cell-to-cell contact even before formation of BC-structures. Typical bile-canalicular microvilli also appeared in these regions before the intercellular spaces were completely closed. Finally, after in vitro reformation of BC, HAM-4 antigen was localized specifically at the BC-surface. The process of BC-reformation and the intracellular organization of actin and cytokeratin filaments were not significantly affected by microtubule inhibitors (nocodazole, colcemid, and colchicine). However, the localization of HAM-4 antigen molecules at the surface of BC was disrupted by these inhibitors, suggesting that the distribution of HAM-4 antigen, which represents a marker for the reconstruction of surface polarity, is dependent on microtubule function. 0 1992 Academic Press, Inc.
r To whom reprint
of Medicine,
KUNIKI,
50
SURFACE
ANTIGEN
OF CULTURED
tween pairs of hepatocytes were detected by fluorescent staining for actin filaments. About 90% of hepatocytes were solitary and virtually no remaining bile canaliculi were detected even between clumped hepatocytes. Dissociated hepatocytes were seeded in William’s medium E (Nissui) supplemented with 10% fetal bovine serum and gentamytin (50 pg/ml). Cells were seeded at 1 X lo5 cells/O.2 ml/cm* on glass coverslips. At the time of initiation of cultures, less than 1% of hepatocytes had BC-structures recognizable by anti-actin-staining, which represented bile canaliculi remaining after dissociation. Therefore, bilecanalicular spaces found at later stages in culture are referred to as reconstructed BC-structures, although we had no direct evidence for this regarding individual structures. Transmission
Electron Microscopy
(TEM)
Hepatocytes were fixed in 2% glutaraldehyde, 2% formaldehyde, and 0.025% CaCl,, modified from Karnovsky [ 181 and postfixed in 2% osmium tetroxide, in 0.1 M sodium cacodylate, pH 7.4. Fixed specimens were dehydrated through a graded acetone series and embedded in resin (Epon 812). Ultrathin sections were cut and examined in a transmission electron microscope JEM 200CX. Fluorescent
Staining
of HAM-4 and Cytoskeletons
The following antibodies were used for indirect immunofluorescence of HAM-4 and cytoskeletal elements: the culture supernatant of HAM-4 hybridoma, anti-u-tubulin monoclonal antibody (Pharmacia, Sweden) and anti-(40,45, and 52.5 kDa)-cytokeratin monoclonal antibody PKK-1 (Labsystems Oy, Finland) as primary antibodies. FITC (Auorescein isothiocyanate)-labeled goat F(ab’), anti-mouseIgG was used as secondary antibody. For the staining of actin filaments, NBD-phallacidin was used. To stain for HAM-4 and actin filaments, hepatocytes were fixed with cold methanol (10 min). For staining of microtubules and cytokeratin filaments, hepatocytes were first immersed in extraction buffer (30-60 min) before fixation in methanol. Extraction buffer [19] contained 20% glycerol, l-2% Triton X-100,10 mM KCl, 5 mM ethylene glycol bis(&aminoethyl ether)N,N’-tetraacetic acid, 1 mM PMSF, and 25 mM imidazole, pH 7.0. Inhibitors Colchicine and lumicolchicine. Colchicine and P-lumicolchicine were dissolved in washing or culture medium to give a final concentration of 100 PM. Colcemid. A stock solution (1 mM in distilled water) was diluted to give a final concentration of 10 PM in culture or washing medium. Nocodazole. A stock solution (1 n&f nocodazole in dimethyl sulfoxide (DMSO)) was diluted in washing or culture medium to give a final concentration of 10 PM. One percent DMSO affected the distribution of neither HAM-4 antigen nor cytoskeletal elements. RESULTS
Reformation
of BC and Distribution
of HAM-4
Antigen
More than 90% of hepatocytes were present as single cells immediately after dissociation and washing. Initially, HAM-4 stained the dissociated hepatocyte outline uniformly. After l-l.5 h, most cells had adhered to the glass coverslips and lo-30% of them were touching each other and formed pairs of cells. However, these pairs had rarely formed closed-intercellular-spaces and hepatocytes remained rounded in appearance. Within
HEPATOCYTES
51
l-3 h of seeding, HAM-4 staining became localized to the area of attachment to neighboring cells, while staining diminished on other surfaces (Fig. 1A). After 3-6 h in culture, hepatocytes began to flatten onto the glass substratum and contacted one another more frequently. At this time, many pairs of cells already showed closed-intercellular-spaces (BC-structures) and various different stages of BC-reconstruction were revealed by TEM. These included pairs of hepatocytes just touching each other, pairs of hepatocytes which were in contact and showed slight thickening of adjacent surface membranes but which retained a defined space between them (LC; linear cell contact according to Wanson et al. [ ll]), and pairs of hepatocytes with closed-intercellular-spaces, where the typical microvilli of bile canaliculi in uiuo were observed. Such microvilli were also observed on free surface membranes adjacent to LC in some cell pairs and on either side of LC, before intercellular space closed completely (Fig. 2). During the increase in cell-to-cell contact between 3 and 6 h in culture, HAM-4 staining became more localized and intense in the adherent region between hepatocytes (Fig. 1B). At later times, strong fluorescence of HAM-4 staining was clearly restricted to the closed-intercellular-space between two hepatocytes (BC-structure), showing segregation of HAM-4 antigen onto the BC-surface. No staining was observed in the region where LC was thought to develop (Fig. lC, shown by an arrow). After 12-24 or 74 h in culture, the hepatocytes formed a flat monolayer in which each cell was in contact with several others forming a cord-line. BC-structures fused and developed into long tubular canals within the cell sheet. At these stages, HAM-4 stained the outline of the tubular BC-structures (Fig. 1C) and their branches (Fig. 1D).
Cytoskeleton In addition to HAM-4 antigen, cytoskeletal structures (microtubules, cytokeratin filaments, and actin filwere studied. Microtubule distribution aments) changed in the late stages of culture after BC-reconstruction while cytokeratin and actin filament distribution did not change. Immediately after dissociation, microtubules were abundant near the periphery of the dissociated hepatocytes (Fig. 3A). Cytokeratin and actin filaments were also distributed uniformly at the periphery of the cells (data not shown). Microtubules in solitary adherent hepatocytes were evenly distributed, exhibiting radial rays extending in all directions (data not shown). In contrast, in paired hepatocytes microtubules showed uneven distribution,
52
SAWADA
ET AL
FIG. 1. HAM-4 indirect immunofluorescence on cultured hepatocytes. X1200. (A) Two hepatocytes forming a couplet after 1.5 h of culture, show strong Buorescence at the region of cell-to-cell contact and the open surface of the contact region. (B) Strong fluorescence is restricted to the region of cell-to-cell contact in a couplet after 1.5 h of culture. (C) Fluorescence is restricted to the surface of a BC-structure in a couplet after 12 h of culture. Linear contact is thought to occur between the BC-surface and the outer surface of the couplet where no fluorescence is detected (arrow). (D) Fluorescence reveals the contour of a BC-structure formed in a hepatocyte sheet after 74 h ofculture. The BC-structure is expanded and branched.
being particularly abundant and sometimes forming networks in the cytoplasm near cell-to-cell contact regions (Fig. 3B). This dense distribution remained even
in cell pairs which had already reconstructed small BCstructures (Fig. 3C). Cytokeratin filaments remained underlying the whole cell surface both during and
FIG. 2. Transmission electron microscopy of cultured hepatocytes during the formation of BC (6 h). Bar, 1 Wm. (A) A couplet in which surface differentiation has occurred in the area adjacent to a linear contact (arrow). Microvilli are more concentrated in this area, and the area is still open to the outside at this stage. (B) Surface differentiation in the area adjacent to a linear contact. Microvilli show the typical shape of a bile canaliculus.
SURFACE
ANTIGEN
OF CULTURED
53
HEPATOCYTES
FIG. 3. Anti-tubulin indirect immunofluorescence during BC formation. X1200 (A) Hepatocytes 1.5 b after dissociation. Local concentration is not seen. (B) A couplet of hepatocytes after 12 h of culture. The fluorescence is concentrated under a BC-structure. (C) A couplet observed after 24 h of culture, extended widely over the coverslip. Microtubules are concentrated under the BC-structure. (D) A sheet of hepatocytes after 68 h of culture. Microtubules have extended, forming a network in the cells. Microtubules are not concentrated near B&structures (*). n, nucleus.
after BC-reconstruction although they were slightly more abundant at the edges of BC-structures in some cases (Fig. 4A). Some cytoplasmic filaments were also observed. In contrast, actin filaments were specifically localized around BC-structures (data not shown). In the late stages of culture, when BC-structures had expanded and developed into tubular canals, the dense arrays of microtubules beneath the BC-structures disappeared and a relatively uniform microtubular network was observed throughout the cytoplasm (Fig. 3D). The distribution of actin and cytokeratin filaments did not change after BC-structures had reformed (data not shown).
The Effect of Inhibitors
on HAM-4
and Cytoskeletons
In the presence of colchicine, ,&lumicolchicine, colcemid, and nocodazole, hepatocytes were slow to attach and spread on the glass surface and did not form a flat monolayer. However, BC-structures formed even in the presence of these inhibitors (Fig. 5A) but with l-3 h delay compared with control cultures. No microtubules were detected by indirect immunofluorescence after cells were washed and cultured in the presence of inhibitors. However, actin filaments (Fig. 5C) and cytokeratin filaments (Fig. 4B) showed similar distribution in cultures both with and without inhibitors; both types of
SAWADA
ET AL
tribute to the unequal distribution of surface molecules in epithelial cells. In previous studies, the nonuniform distribution of some surface molecules was associated with cell adhesion molecules [20,21] or cell-to-cell junctional structures [22,23], and an example in which tight junctions were not involved has also been reported [24]. With regard to the cytoskeleton, epithelial cell surface molecules have previously been divided into two groups: the group of molecules whose redistribution was affected by microtubule inhibitors [25] and the other which was not affected by such inhibitors [14, 25, 261. In this study, HAM-4 antigen redistribution did not require tight junctions and junctional complexes, since redistribution of this antigen was observed before for-
FIG. 4. Indirect immunofluorescence of cytokeratin filaments. ~1200. (A) Four hepatocytes with BC-structures (arrows) cultured for 12 h. The fluorescence is especially concentrated at the BC-structures and the region of cell-to-cell contact. The fluorescence is also evident along the whole periphery of the cells. (B) A couplet with a BC-structure (arrow) after 12 h of culture in the presence of 100 pi%4 colchicine. The cells are outlined by fluorescence, which is especially strong around the BC-structure.
filament were observed beneath the surface of reconstructed BC-structures after 12 h of culture. HAM-4 staining decreased significantly during washing, disappeared during culture, and was not detected even on the surface of BC-structures (Fig. 5B) in the presence of inhibitors. However, P-lumicolchicine (100 &Z) affected neither BC-reconstruction nor distribution of HAM-4-staining, which was concentrated at the BC-surface (Figs. 5D and 5E). DISCUSSION The redistribution of HAM-4 antigen started at a very early stage in BC-reformation and finally localized at the BC-surface. This process appeared to be associated with the presence during BC-reconstruction of densely packed microtubules at points of cell-to-cell contact since microtubule inhibitors distrupted HAM-4 antigen distribution. Intercellular junctional structures and cytoskeleton are two possible major factors which may directly con-
FIG. 5. Hepatocytes cultured in 100 FM colchicine or fl-lumicolchicine. x1200. (A, B, C) Hepatocytes cultured for 12 h in colchicine. BC-structure (*) is evident by phase-contrast microscopy (A), and actin filaments (C, staining with NBD-phallacidin) are concentrated just beneath the surface of the BC-structure. However, no specific accumulation of HAM-4 antigen (B, HAM-4 staining) is detected at the B&structure. (D, E) A couplet observed after 12 h of culture in @lumicolchicine. The BC-structure (s) has expanded considerably as shown by phase-contrast microscopy (D), and a specifically rich distribution of HAM-4 antigen is detected (E, HAM-4 staining). n, nucleus.
SURFACE
ANTIGEN
OF CULTURED
mation of tight junctions, which in turn has been reported to occur several hours after BC-reformation [ll, 121. The appearance of typical bile-canalicular microvilli on open surface areas suggested that not only HAM-4 redistribution but also the reconstruction of some other characteristic BC features might start before the BC-surface became enclosed and bordered by intercellular junctional structures. The role of LC, an intercellular contact established at an early stage, remains obscure. In this study, the redistribution of HAM-4 antigen exhibited a close relationship to microtubule distribution. Actin and cytokeratin filaments were not sufficient to support HAM-4 redistribution when microtubule function had been disrupted, There was no evidence for a direct association between HAM-4 antigen and microtubules. The shedding of HAM-4 antigen did not increase significantly in the presence of colchicine (data not shown). However, a possible hypothesis is that the lack of HAM-4 antigen redistribution and the decrease in HAM-4 staining may relate to the disruption of intracellular-molecular-transport, which may be associated with microtubules in hepatocytes as in some other cells [27-291. The distribution of microtubules during BC-reconstruction appears to support this hypothesis. Consequently, we speculate that the redistribution of HAM-4 antigen may occur by the shedding of the antigen from the whole of the cell surface and the specific targetting of new molecules to BC-surface under the control of the microtubule system. However, the possibility that superficial movement of HAM-4 antigen molecules occurs during this redistribution has not been eliminated. Recently, Kobayashi proved that HAM-4 recognizes dipeptidyl peptidase IV (personal communication). This enzyme is found abundantly on the bile-canalicular surface [30]. Thus, the data which we present here may identify the distribution of this enzyme. REFERENCES 1. 2. 3.
Toda, G., Oka, H., Oda, T., and Ikeda, Y. (1975) Bioc~m. Biophys. Actu 413,52-64. Wisher, M. H., and Evans, W. H. (1975) Biochem. J. 146,375388. Blitzer, B. L., and Boyer, J. L. (1978) J. Clin. Znuest. 62, 11041108.
Received March 15, 1991 Revised version received November
8, 1991
4. 5. 6. 7. 8. 9. 10. 11. 12.
55
HEPATOCYTES
Wall, D. A., and Hubbard, A. L. (1981) J. Cell Biol. SO, 687-696. Hubbard, A. L., Wall, D. A., and Ma, A. (1983) J. Cell Biol. 96, 217-229. Hubbard, A. L., Bartles, J. R., and Braiterman, L. T. (1985) J. Cell Biol. 100, 1115-1125. Fukumoto, T., Kimura, H., Naito, M., Miyamoto, M., Yamashita, A., and Sugiyama, H. (1984) Mol. Immunol. 21,285-291. Tamakoshi, K., Fukumoto, T., Kanai, K., and Yamashita, A. (1985) Clin. Exp. Zmmunol. 60, 373-380. Yamaguchi, K., Fujikura, Y., Sawada, T., and Fukumoto, (1987) J. Electron. Microsc. 36, 300. [abstract]
T.
Yamaguchi, K., Fujikura, Y., Kuniki, H., Itoh, K., Tamakoshi, K., and Fukumoto, T. (1991) Cell Struct. Funct. 16, 303-313. Wanson, J. C., Drochmans, P., Mosselmans, R., and Ronveaux, M. F. (1977) J. Cell Biol. 77,858-877. Sakisaka, S., Yoshida, H., Matsumoto, H., Ikejigi, N., Eguchi, T., Kawaguchi, M., and Tanikawa, K. (1981) J. Clin. Electron Mi-
crosc. 14, 404-405. 13.
Tanaka, K., Sato, M., Tomita, Biochem. 84,937-946.
14.
Gautam, A., Oi-Cheng, Ng., and Boyer, J. L. (1987) Hepatology 7,216-223. Brabander, M. D., Wanson, J. C., Mosselmans, R., Geuens, G., and Drochmans, P. (1978) Biol. Cell. 31, 127-140. Berry, M. N., and Friend, D. S. (1969) J. Cell Biol. 43, 506-520. Seglen, P. 0. (1976) Methods Cell Biol. 13, 29-83. Karnovsky, M. J. (1967) J. Cell Biol. 35, 213-236. Sawada, T., and Schatten, G. (1988) Cell Motil. 9,219-230.
15. 16. 17. 18. 19.
Y., and Ichihara,
A. (1978) J.
20.
Imhof, B. A., Vollmers, H. P., Goodman, W. (1983) Cell 35,667-675.
21.
Behrens, J., Birchmeier, W., Goodman, S. L., and Imhof, B. A. (1987) J. Cell Biol. 101, 1307-1313. Herzlinger, D. A., and Ojakian, G. K. (1984) J. Cell Biol. 98, 1777-1787. Hoi-Sang, U., Saier, M. N., Jr., and Ellisman, M. H. (1979) Exp. Cell Biol. 122, 384-391. Vega-Salas, D. E., Salas, P. J. I., Gundersen, D., and RodriguesBoulan, E. (1987) J. Cell Biol. 104,905-916. Eilers, U. K., Klumperman, J., and Hauri, H-P. (1989) J. Cell Biol. 108, 13-22. Salas, P. J. I., Misel, D. E., VegaSalas, D. E., Gundersen, D., Cereijido, M., and Rodrigues-Boulan, E. (1986) J. Cell Biol. 102, 1853-1867. Allen, R. D., Weiss, D. G., Hayden, J. H., Brown, D. T., Fujiwake, H., and Simpson, M. (1985) J. Cell Biol. 100,1736-1752. Vale, R. D., Schnapp, B. J., Reese, T. S., and Sheetz, M. P. (1985) Cell 40, 559-569. Burgess, T. L., and Kelly, R. B. (1987) Annu. Reu. Cell Biol. 3, 243-293. Fukui, Y., Yamamoto, A., Kyoden, T., Kato, K., and Tashiro, Y. (1990) Cell Struct. Funct. 15, 117-125.
22. 23. 24. 25. 26.
27. 28. 29. 30.
S. L., and Birchmeier,
CELL
RESEARCH
199,
50-55 (19%)
Distribution of the Surface Antigen HAM-4 and Cytoskeleton during Reformation of Bile-Canalicular Structures in Rat Primary Cultured Hepatocytes TOMOO SAWADA, HIROSHI ITAI, YOSHIHISA FUJIKURA, HIROMICHI MASAKATSU TAMECHIKA, AND TETSUO FUKUMOTO’ Department
of Anatomy,
Yamaguchi
University
School
INTRODUCTION
Hepatocytes are of epithelial origin. As in other epithelia hepatocytes have cell surface polarity, with three different surface domains: the apical (bile-canalicular) surface, the lateral (contiguous) surface, and the basal (sinusoidal) surface. Surface polarity results in many molecules being differentially distributed in each surface domain [l-5]. Some monoclonal antibodies against cell surface molecules can detected these differences in molecular composition of the three surface domains (6, 71. HAM-4 is a monoclonal antibody which recognizes a wheat germ agglutinin-binding protein on the surface of
0014.4827/92
$3.00
Yamaguchi
MATERIALS Preparation
755, Japan
and Primary
Culture
AND
METHODS
of Dissociated
Hepatocytes
Hepatocytes were prepared by the method described by Berry and Friend [16] and Seglen [17]. An adult DA rat was perfused through the portal vein and out through the abdominal aorta for 30-45 min with Ca’+-free Krebs-Henseleit Ringer bicarbonate buffer containing 0.05% collagenase, pH 7.5, saturated with 95% O2 and CO,. The liver was teased with a spatula and filtered through four layers of gauze. The dissociated hepatocytes were suspended in 50 ml of Hanks’ buffer solution and allowed to settle for 15 min. The pellet was then resuspended. After three such washes, remaining bile canaliculi be-
requests should be addressed.
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
Ube-City,
rat hepatocytes. The antigen recognized by HAM-4 is distributed mainly on the surface of the bile canaliculus (BC) both in uiuo and in vitro in dissociated (free) hepatocytes [&lo]. In primary culture, hepatocytes reform structures (BC-structures) corresponding to the bile canaliculus in uiuo and this reconstruction (BC-reconstruction) provides a model in which the surface polarity can be investigated. Previous reports have described the fundamental processes of BC-reconstruction: formation of intercellular closed spaces which become the BC-structures [ 111, formation of cell-to-cell junctional structures [12], and recovery of some biochemical characteristics [13]. In relation to the cytoskeleton, Gautam et al. [14] reported that the redistribution of Mg’+-ATPase into bile canaliculi depended on actin filaments but not on microtubules. It has also been found that microtubule inhibitors do not inhibit BC-reconstruction [15]. Although there have been numerous studies of BCreconstruction in vitro and the reacquisition of cell surface polarity, the development of hepatocyte surface polarity is not yet fully understood. Hepatocytes have many different surface molecules and there may be distinct mechanisms controlling the distribution of each one. HAM-4 antigen represents a specific molecular marker of the apical hepatocyte surface. We therefore investigated the distribution of HAM-4 antigen molecules during BC-reconstruction and focused on the role of microtubules during this process.
The distribution of rat bile-canalicular surface antigen (HAM-4 antigen) and cytoskeletal elements (microtubules, actin filaments, and cytokeratin filaments) was examined during the reformation of bile-canalicular structures (BC-structures) in primary cultures of dissociated hepatocytes obtained following collagenase perfusion. HAM-4 antigen, which initially dispersed after cell dissociation, became focused into regions of cell-to-cell contact even before formation of BC-structures. Typical bile-canalicular microvilli also appeared in these regions before the intercellular spaces were completely closed. Finally, after in vitro reformation of BC, HAM-4 antigen was localized specifically at the BC-surface. The process of BC-reformation and the intracellular organization of actin and cytokeratin filaments were not significantly affected by microtubule inhibitors (nocodazole, colcemid, and colchicine). However, the localization of HAM-4 antigen molecules at the surface of BC was disrupted by these inhibitors, suggesting that the distribution of HAM-4 antigen, which represents a marker for the reconstruction of surface polarity, is dependent on microtubule function. 0 1992 Academic Press, Inc.
r To whom reprint
of Medicine,
KUNIKI,
50
SURFACE
ANTIGEN
OF CULTURED
tween pairs of hepatocytes were detected by fluorescent staining for actin filaments. About 90% of hepatocytes were solitary and virtually no remaining bile canaliculi were detected even between clumped hepatocytes. Dissociated hepatocytes were seeded in William’s medium E (Nissui) supplemented with 10% fetal bovine serum and gentamytin (50 pg/ml). Cells were seeded at 1 X lo5 cells/O.2 ml/cm* on glass coverslips. At the time of initiation of cultures, less than 1% of hepatocytes had BC-structures recognizable by anti-actin-staining, which represented bile canaliculi remaining after dissociation. Therefore, bilecanalicular spaces found at later stages in culture are referred to as reconstructed BC-structures, although we had no direct evidence for this regarding individual structures. Transmission
Electron Microscopy
(TEM)
Hepatocytes were fixed in 2% glutaraldehyde, 2% formaldehyde, and 0.025% CaCl,, modified from Karnovsky [ 181 and postfixed in 2% osmium tetroxide, in 0.1 M sodium cacodylate, pH 7.4. Fixed specimens were dehydrated through a graded acetone series and embedded in resin (Epon 812). Ultrathin sections were cut and examined in a transmission electron microscope JEM 200CX. Fluorescent
Staining
of HAM-4 and Cytoskeletons
The following antibodies were used for indirect immunofluorescence of HAM-4 and cytoskeletal elements: the culture supernatant of HAM-4 hybridoma, anti-u-tubulin monoclonal antibody (Pharmacia, Sweden) and anti-(40,45, and 52.5 kDa)-cytokeratin monoclonal antibody PKK-1 (Labsystems Oy, Finland) as primary antibodies. FITC (Auorescein isothiocyanate)-labeled goat F(ab’), anti-mouseIgG was used as secondary antibody. For the staining of actin filaments, NBD-phallacidin was used. To stain for HAM-4 and actin filaments, hepatocytes were fixed with cold methanol (10 min). For staining of microtubules and cytokeratin filaments, hepatocytes were first immersed in extraction buffer (30-60 min) before fixation in methanol. Extraction buffer [19] contained 20% glycerol, l-2% Triton X-100,10 mM KCl, 5 mM ethylene glycol bis(&aminoethyl ether)N,N’-tetraacetic acid, 1 mM PMSF, and 25 mM imidazole, pH 7.0. Inhibitors Colchicine and lumicolchicine. Colchicine and P-lumicolchicine were dissolved in washing or culture medium to give a final concentration of 100 PM. Colcemid. A stock solution (1 mM in distilled water) was diluted to give a final concentration of 10 PM in culture or washing medium. Nocodazole. A stock solution (1 n&f nocodazole in dimethyl sulfoxide (DMSO)) was diluted in washing or culture medium to give a final concentration of 10 PM. One percent DMSO affected the distribution of neither HAM-4 antigen nor cytoskeletal elements. RESULTS
Reformation
of BC and Distribution
of HAM-4
Antigen
More than 90% of hepatocytes were present as single cells immediately after dissociation and washing. Initially, HAM-4 stained the dissociated hepatocyte outline uniformly. After l-l.5 h, most cells had adhered to the glass coverslips and lo-30% of them were touching each other and formed pairs of cells. However, these pairs had rarely formed closed-intercellular-spaces and hepatocytes remained rounded in appearance. Within
HEPATOCYTES
51
l-3 h of seeding, HAM-4 staining became localized to the area of attachment to neighboring cells, while staining diminished on other surfaces (Fig. 1A). After 3-6 h in culture, hepatocytes began to flatten onto the glass substratum and contacted one another more frequently. At this time, many pairs of cells already showed closed-intercellular-spaces (BC-structures) and various different stages of BC-reconstruction were revealed by TEM. These included pairs of hepatocytes just touching each other, pairs of hepatocytes which were in contact and showed slight thickening of adjacent surface membranes but which retained a defined space between them (LC; linear cell contact according to Wanson et al. [ ll]), and pairs of hepatocytes with closed-intercellular-spaces, where the typical microvilli of bile canaliculi in uiuo were observed. Such microvilli were also observed on free surface membranes adjacent to LC in some cell pairs and on either side of LC, before intercellular space closed completely (Fig. 2). During the increase in cell-to-cell contact between 3 and 6 h in culture, HAM-4 staining became more localized and intense in the adherent region between hepatocytes (Fig. 1B). At later times, strong fluorescence of HAM-4 staining was clearly restricted to the closed-intercellular-space between two hepatocytes (BC-structure), showing segregation of HAM-4 antigen onto the BC-surface. No staining was observed in the region where LC was thought to develop (Fig. lC, shown by an arrow). After 12-24 or 74 h in culture, the hepatocytes formed a flat monolayer in which each cell was in contact with several others forming a cord-line. BC-structures fused and developed into long tubular canals within the cell sheet. At these stages, HAM-4 stained the outline of the tubular BC-structures (Fig. 1C) and their branches (Fig. 1D).
Cytoskeleton In addition to HAM-4 antigen, cytoskeletal structures (microtubules, cytokeratin filaments, and actin filwere studied. Microtubule distribution aments) changed in the late stages of culture after BC-reconstruction while cytokeratin and actin filament distribution did not change. Immediately after dissociation, microtubules were abundant near the periphery of the dissociated hepatocytes (Fig. 3A). Cytokeratin and actin filaments were also distributed uniformly at the periphery of the cells (data not shown). Microtubules in solitary adherent hepatocytes were evenly distributed, exhibiting radial rays extending in all directions (data not shown). In contrast, in paired hepatocytes microtubules showed uneven distribution,
52
SAWADA
ET AL
FIG. 1. HAM-4 indirect immunofluorescence on cultured hepatocytes. X1200. (A) Two hepatocytes forming a couplet after 1.5 h of culture, show strong Buorescence at the region of cell-to-cell contact and the open surface of the contact region. (B) Strong fluorescence is restricted to the region of cell-to-cell contact in a couplet after 1.5 h of culture. (C) Fluorescence is restricted to the surface of a BC-structure in a couplet after 12 h of culture. Linear contact is thought to occur between the BC-surface and the outer surface of the couplet where no fluorescence is detected (arrow). (D) Fluorescence reveals the contour of a BC-structure formed in a hepatocyte sheet after 74 h ofculture. The BC-structure is expanded and branched.
being particularly abundant and sometimes forming networks in the cytoplasm near cell-to-cell contact regions (Fig. 3B). This dense distribution remained even
in cell pairs which had already reconstructed small BCstructures (Fig. 3C). Cytokeratin filaments remained underlying the whole cell surface both during and
FIG. 2. Transmission electron microscopy of cultured hepatocytes during the formation of BC (6 h). Bar, 1 Wm. (A) A couplet in which surface differentiation has occurred in the area adjacent to a linear contact (arrow). Microvilli are more concentrated in this area, and the area is still open to the outside at this stage. (B) Surface differentiation in the area adjacent to a linear contact. Microvilli show the typical shape of a bile canaliculus.
SURFACE
ANTIGEN
OF CULTURED
53
HEPATOCYTES
FIG. 3. Anti-tubulin indirect immunofluorescence during BC formation. X1200 (A) Hepatocytes 1.5 b after dissociation. Local concentration is not seen. (B) A couplet of hepatocytes after 12 h of culture. The fluorescence is concentrated under a BC-structure. (C) A couplet observed after 24 h of culture, extended widely over the coverslip. Microtubules are concentrated under the BC-structure. (D) A sheet of hepatocytes after 68 h of culture. Microtubules have extended, forming a network in the cells. Microtubules are not concentrated near B&structures (*). n, nucleus.
after BC-reconstruction although they were slightly more abundant at the edges of BC-structures in some cases (Fig. 4A). Some cytoplasmic filaments were also observed. In contrast, actin filaments were specifically localized around BC-structures (data not shown). In the late stages of culture, when BC-structures had expanded and developed into tubular canals, the dense arrays of microtubules beneath the BC-structures disappeared and a relatively uniform microtubular network was observed throughout the cytoplasm (Fig. 3D). The distribution of actin and cytokeratin filaments did not change after BC-structures had reformed (data not shown).
The Effect of Inhibitors
on HAM-4
and Cytoskeletons
In the presence of colchicine, ,&lumicolchicine, colcemid, and nocodazole, hepatocytes were slow to attach and spread on the glass surface and did not form a flat monolayer. However, BC-structures formed even in the presence of these inhibitors (Fig. 5A) but with l-3 h delay compared with control cultures. No microtubules were detected by indirect immunofluorescence after cells were washed and cultured in the presence of inhibitors. However, actin filaments (Fig. 5C) and cytokeratin filaments (Fig. 4B) showed similar distribution in cultures both with and without inhibitors; both types of
SAWADA
ET AL
tribute to the unequal distribution of surface molecules in epithelial cells. In previous studies, the nonuniform distribution of some surface molecules was associated with cell adhesion molecules [20,21] or cell-to-cell junctional structures [22,23], and an example in which tight junctions were not involved has also been reported [24]. With regard to the cytoskeleton, epithelial cell surface molecules have previously been divided into two groups: the group of molecules whose redistribution was affected by microtubule inhibitors [25] and the other which was not affected by such inhibitors [14, 25, 261. In this study, HAM-4 antigen redistribution did not require tight junctions and junctional complexes, since redistribution of this antigen was observed before for-
FIG. 4. Indirect immunofluorescence of cytokeratin filaments. ~1200. (A) Four hepatocytes with BC-structures (arrows) cultured for 12 h. The fluorescence is especially concentrated at the BC-structures and the region of cell-to-cell contact. The fluorescence is also evident along the whole periphery of the cells. (B) A couplet with a BC-structure (arrow) after 12 h of culture in the presence of 100 pi%4 colchicine. The cells are outlined by fluorescence, which is especially strong around the BC-structure.
filament were observed beneath the surface of reconstructed BC-structures after 12 h of culture. HAM-4 staining decreased significantly during washing, disappeared during culture, and was not detected even on the surface of BC-structures (Fig. 5B) in the presence of inhibitors. However, P-lumicolchicine (100 &Z) affected neither BC-reconstruction nor distribution of HAM-4-staining, which was concentrated at the BC-surface (Figs. 5D and 5E). DISCUSSION The redistribution of HAM-4 antigen started at a very early stage in BC-reformation and finally localized at the BC-surface. This process appeared to be associated with the presence during BC-reconstruction of densely packed microtubules at points of cell-to-cell contact since microtubule inhibitors distrupted HAM-4 antigen distribution. Intercellular junctional structures and cytoskeleton are two possible major factors which may directly con-
FIG. 5. Hepatocytes cultured in 100 FM colchicine or fl-lumicolchicine. x1200. (A, B, C) Hepatocytes cultured for 12 h in colchicine. BC-structure (*) is evident by phase-contrast microscopy (A), and actin filaments (C, staining with NBD-phallacidin) are concentrated just beneath the surface of the BC-structure. However, no specific accumulation of HAM-4 antigen (B, HAM-4 staining) is detected at the B&structure. (D, E) A couplet observed after 12 h of culture in @lumicolchicine. The BC-structure (s) has expanded considerably as shown by phase-contrast microscopy (D), and a specifically rich distribution of HAM-4 antigen is detected (E, HAM-4 staining). n, nucleus.
SURFACE
ANTIGEN
OF CULTURED
mation of tight junctions, which in turn has been reported to occur several hours after BC-reformation [ll, 121. The appearance of typical bile-canalicular microvilli on open surface areas suggested that not only HAM-4 redistribution but also the reconstruction of some other characteristic BC features might start before the BC-surface became enclosed and bordered by intercellular junctional structures. The role of LC, an intercellular contact established at an early stage, remains obscure. In this study, the redistribution of HAM-4 antigen exhibited a close relationship to microtubule distribution. Actin and cytokeratin filaments were not sufficient to support HAM-4 redistribution when microtubule function had been disrupted, There was no evidence for a direct association between HAM-4 antigen and microtubules. The shedding of HAM-4 antigen did not increase significantly in the presence of colchicine (data not shown). However, a possible hypothesis is that the lack of HAM-4 antigen redistribution and the decrease in HAM-4 staining may relate to the disruption of intracellular-molecular-transport, which may be associated with microtubules in hepatocytes as in some other cells [27-291. The distribution of microtubules during BC-reconstruction appears to support this hypothesis. Consequently, we speculate that the redistribution of HAM-4 antigen may occur by the shedding of the antigen from the whole of the cell surface and the specific targetting of new molecules to BC-surface under the control of the microtubule system. However, the possibility that superficial movement of HAM-4 antigen molecules occurs during this redistribution has not been eliminated. Recently, Kobayashi proved that HAM-4 recognizes dipeptidyl peptidase IV (personal communication). This enzyme is found abundantly on the bile-canalicular surface [30]. Thus, the data which we present here may identify the distribution of this enzyme. REFERENCES 1. 2. 3.
Toda, G., Oka, H., Oda, T., and Ikeda, Y. (1975) Bioc~m. Biophys. Actu 413,52-64. Wisher, M. H., and Evans, W. H. (1975) Biochem. J. 146,375388. Blitzer, B. L., and Boyer, J. L. (1978) J. Clin. Znuest. 62, 11041108.
Received March 15, 1991 Revised version received November
8, 1991
4. 5. 6. 7. 8. 9. 10. 11. 12.
55
HEPATOCYTES
Wall, D. A., and Hubbard, A. L. (1981) J. Cell Biol. SO, 687-696. Hubbard, A. L., Wall, D. A., and Ma, A. (1983) J. Cell Biol. 96, 217-229. Hubbard, A. L., Bartles, J. R., and Braiterman, L. T. (1985) J. Cell Biol. 100, 1115-1125. Fukumoto, T., Kimura, H., Naito, M., Miyamoto, M., Yamashita, A., and Sugiyama, H. (1984) Mol. Immunol. 21,285-291. Tamakoshi, K., Fukumoto, T., Kanai, K., and Yamashita, A. (1985) Clin. Exp. Zmmunol. 60, 373-380. Yamaguchi, K., Fujikura, Y., Sawada, T., and Fukumoto, (1987) J. Electron. Microsc. 36, 300. [abstract]
T.
Yamaguchi, K., Fujikura, Y., Kuniki, H., Itoh, K., Tamakoshi, K., and Fukumoto, T. (1991) Cell Struct. Funct. 16, 303-313. Wanson, J. C., Drochmans, P., Mosselmans, R., and Ronveaux, M. F. (1977) J. Cell Biol. 77,858-877. Sakisaka, S., Yoshida, H., Matsumoto, H., Ikejigi, N., Eguchi, T., Kawaguchi, M., and Tanikawa, K. (1981) J. Clin. Electron Mi-
crosc. 14, 404-405. 13.
Tanaka, K., Sato, M., Tomita, Biochem. 84,937-946.
14.
Gautam, A., Oi-Cheng, Ng., and Boyer, J. L. (1987) Hepatology 7,216-223. Brabander, M. D., Wanson, J. C., Mosselmans, R., Geuens, G., and Drochmans, P. (1978) Biol. Cell. 31, 127-140. Berry, M. N., and Friend, D. S. (1969) J. Cell Biol. 43, 506-520. Seglen, P. 0. (1976) Methods Cell Biol. 13, 29-83. Karnovsky, M. J. (1967) J. Cell Biol. 35, 213-236. Sawada, T., and Schatten, G. (1988) Cell Motil. 9,219-230.
15. 16. 17. 18. 19.
Y., and Ichihara,
A. (1978) J.
20.
Imhof, B. A., Vollmers, H. P., Goodman, W. (1983) Cell 35,667-675.
21.
Behrens, J., Birchmeier, W., Goodman, S. L., and Imhof, B. A. (1987) J. Cell Biol. 101, 1307-1313. Herzlinger, D. A., and Ojakian, G. K. (1984) J. Cell Biol. 98, 1777-1787. Hoi-Sang, U., Saier, M. N., Jr., and Ellisman, M. H. (1979) Exp. Cell Biol. 122, 384-391. Vega-Salas, D. E., Salas, P. J. I., Gundersen, D., and RodriguesBoulan, E. (1987) J. Cell Biol. 104,905-916. Eilers, U. K., Klumperman, J., and Hauri, H-P. (1989) J. Cell Biol. 108, 13-22. Salas, P. J. I., Misel, D. E., VegaSalas, D. E., Gundersen, D., Cereijido, M., and Rodrigues-Boulan, E. (1986) J. Cell Biol. 102, 1853-1867. Allen, R. D., Weiss, D. G., Hayden, J. H., Brown, D. T., Fujiwake, H., and Simpson, M. (1985) J. Cell Biol. 100,1736-1752. Vale, R. D., Schnapp, B. J., Reese, T. S., and Sheetz, M. P. (1985) Cell 40, 559-569. Burgess, T. L., and Kelly, R. B. (1987) Annu. Reu. Cell Biol. 3, 243-293. Fukui, Y., Yamamoto, A., Kyoden, T., Kato, K., and Tashiro, Y. (1990) Cell Struct. Funct. 15, 117-125.
22. 23. 24. 25. 26.
27. 28. 29. 30.
S. L., and Birchmeier,