Influences of extracellular matrix components on the growth and differentiation of ruminal epithelial cells in primary culture

Research in Veterinary Science 1993, 54, 102-109

Influences of extracellular matrix components on the growth and differentiation of ruminal epithelial cells in primary culture P. GALFI, Department of Physiology and Biochemistry, University of Veterinary Sciences, H1400 Budapest, PO Box 2, Hungary, G. G~BEL, H. MARTENS, Department of Veterinary

Physiology, Free University of Berlin, Koserstrasse 20, D-IO00 Berlin 33, Germany

Primary cultures of ovine ruminal epithelial cells were made to study the influence of collagen types I and IV and of medium supplementation with various hormones and Na-n-butyrate on cell morphology and growth characteristics. Both collagen type I and type IV led to increased cell proliferation with the stimulatory effect being more pronounced in collagen IV. In cultures grown on collagen I, both non-stratified and stratified colonies were found, whereas cultures grown on collagen IV showed predominantly stratified growth. Cells in both stratified and non-stratified colonies were positive for cytokeratin antibody. In non-stratified colonies, positive staining with fibronectin antibodies (FN-15) was found in a network over and around the cells. It is suggested that the non-stratified ruminal epithelial cells are in some respects similar to a 'non-differentiating keratinocyte' strain, derived from newborn foreskin epidermis. Cells in stratified colonies bound Ulex europaeus (UEA0 lectin which has been shown to be specific for differentiated epithelial cells in ruminal mucosa. Supplementation of culture medium with glucagon and insulin increased the total cell-overgrown area of collagen I cultures, whereas this effect was absent in cultures grown on collagen IV. In both cultures grown on collagen I or IV, hydroeortisone led to an increase in total cell-overgrown area, whereas Na-n-butyrate inhibited proliferation.

THE ruminant forestomach is lined with a keratinised, stratified, non-glandular, squamous epithelium (Lavker et al 1969) which is histologically similar to epidermis (Henrikson 1970). Despite the histological similarities, there are

striking functional differences between epidermis and rumen epithelium. The epithelium is involved in absorption, metabolism of intermediary compounds, bacterial adherence and protection of underlying tissues from the abrasive digesta mass (Keynes and Harrison 1970, Whanger and Church 1970, Cheng et al 1979, Bergman 1990, Gabel and Martens 1991). Furthermore, a change of diet may induce either proliferative or regressive changes in the rumen epithelium. These changes are, in turn, associated with altered functions of the epithelium (Dirksen et al 1984, G~ibel et al 1987). One method of studying the underlying mechanisms of the aforementioned processes is to isolate cells and keep them in culture. Isolation and cultivation of rumen epithelial cells has been repeatedly described (G~ilfi et al 1981a, Inooka et al 1984, Baldwin and Jesse 1991). Under these conditions, the cells form stratifed colonies and cornified envelopes can be detected during their terminal differentiation (G~ilfi et al 1981b, 1983). However, it is not known how proliferation, stratification and differentiation are influenced by extracellular matrix components or by substances present in the medium. Therefore, it was the aim of this study to evaluate how these processes are influenced by extracellular matrix proteins, hormones and Na-n-butyrate, substances known to modulate growth and differentiation in other cell types. In addition, the cultured cells were characterised by the use of monoclonal antibodies against cytoskeletal proteins and by lectins which have been shown to be an indicator of differentiation (Watt 1988).

102

Culture of sheep ruminal epithelium Materials and methods

Cell dissociation Seven adult blackhead sheep fed hay ad libitum, were used. After killing and exsanguination, 500 to 1000 papillae were excised from the atrium ruminis. The papillae were stored in phosphate buffered saline (PBS) without Ca++ and Mg++ and supplemented with penicillin 400 iu ml-1 and streptomycin 400 jag ml-1 for one hour at 4°C for sterilisation. Thereafter, the tissue was dissociated by fractional trypsinisation (0.25 per cent trypsin and 0.02 per cent Na2EDTA; G/tiff et al 1981a). The isolated cells were examined and classified by morphological characteristics as described earlier (G/tiff et al 1981a, 1982). The cell suspension of the third and, or, fourth fraction was filtered through a gauze mesh and cells were collected by centrifugation at 800 g for 10 minutes. The supernatant was discarded and the cell pellet resuspended in PBS with penicillin/streptomycin. Centrifugation and resuspension were repeated three times.

103

Engelbrecht-Holm-Swarm ~ , o u s e sarcoma; Sigma). Collagen was prepared as described by Friedman et al (1981, type I) or Sawada et al (1986, type IV). Briefly, 0.001 to 40 lag collagen was dissolved in either 1 ml of 0-1 M acetic acid (type I) or in 1 ml of 0-044 M acetic acid (type IV). One ml of these solutions was pipetted into the culture dishes and allowed to dry at room temperature. Before use, the collagen coated dishes were washed with PBs to remove acetic acid.

Supplementation of media with hormones and Na-n-butyrate Various substances were added on the third day of culture. Hydrocortisone (Sigma) was dissolved in ethanol (Merck), and insulin (from sheep) (Sigma), glucagon (NoVo Industrie) and Na-n-butyrate (aDH Chemicals) were dissolved in MEM-Hanks' solution modified as described above to the following final concentrations: 10-7 M hydrocortisone (0.004 per cent ethanol); 1.6 x 10-9 M insulin; 3 x 10-12 M glucagon; 5 × 10- 3 M Na-n-butyrate.

Cell culture Dissociated cells were suspended in M-199 Hanks' medium supplemented with 15 per cent fetal bovine serum (FBS), 20mM HEPES,0.66 mM L-glutamine, 50 ~tg m1-1 gentamicin, 100 ktg ml-1 penicillin, streptomycin 100 iu ml-1 and seeded into 35 mm plastic tissue culture dishes (Coming Glass Works), on 15 x 15 mm coverslips, or into cell culture inserts (22 mm in diameter, Becton Dickinson) placed in 35 mm culture dishes. Media were discarded and replaced with M-199 Hanks' solution supplemented with 15 per cent FBS on the second day of culturing. On the third day of culturing, M-199 Hanks' solution was exchanged for MEM-Hanks' nutrient medium supplemented with 5 per cent FBS,20 mM HEPES, 0.66 mM L-glutamine, 50 lag ml-1 gentamicin, 100 iu ml-1 pencillin, 100 lag ml-1 streptomycin. This medium was used for further culturing and was changed twice weekly. A total of 0.4 ml cm -2 nutrient medium was added to the dishes.

Collagen coating The culture dishes were coated with collagen type I or collagen type IV (type I from calf skin, type IV from basement membrane of

Immunohistochemical procedure The following antibodies (mouse monoclonals) were used for immunofluorescence. LU5, reactive with an epitope which is common to all cytokeratins with a cytokeratin-associated protein (Overbeck et al 1985) (Boehringer Mannheim); anti-basic cytokeratin, reactive with an epitope common to cytokeratins of the basic (type II) subfamily; LE61, reactive with cytokeratin 18 of mesenchymal origin (Lane 1982); anti-t~-sm-1, reactive with the a-actin isoform of smooth muscle (Skalli et al 1986); anti-desmoplakin I and II, reactive with desmoplakins; Vim 3B4, reactive with vimentin (Boehringer Mannheim); fibronectin antibodies (FN-15) (Sigma) reactive with fibronectin and with a peptic fragment (11.5 kDa) from human fibronectin; Fn'C-conjugated second antibodies were obtained from Sigma and Jackson Immunoresearch Laboratories (GB). Furthermore, Ulex europaeus I lectin (tJEAI, Sigma) conjugated with fluorescein isothyocyanate (FITC)was used. This lectin was demonstrated to be reactive with keratinising ceils of epidermis (Watt 1988) and with cells in the upper cell layers of rumen epithelium (S. Neogr/tdy, A. Pusztai and P. G/tiff, unpublished observations).

104

P. G61fi, G. Gabel, 11. Martens

For indirect immunofluorescence of ruminal mucosa, tissue specimens were washed in PBSand fLxed in 70 per cent ethanol. The fixed material was embedded in paraffin, 5 lain thick sections were prepared, finally deparaffinised and washed in PBS before investigation. For staining with FITC-conjugated UEAI, 5 ~ thick cryostat sections from unfixed papillae were prepared. Cultured cells were washed in PBS and fixed in acetone (-20°C, 10 minutes). Cells were not fixed for staining with FITC-conjugated UEAI. Sections of ruminal mucosa and cultured cells were incubated with the first antibody or UEA~FITCfor 45 minutes at room temperature, washed thoroughly in PBS,and incubated with the appropriate FITC-conjugated second antibody, washed again in PBs, mounted in glycerol:PBS (1:2), and examined using a fluorescent microscope (Zeiss Axiovert 35M).

Determination of cells forming cornified envelopes To measure the number of cells with cornified envelopes desquamated into the medium, the supernatant medium was collected by pipette on day 17 of culturing. The cells were treated with sodium dodecyl sulphate (SDS) and ~-mercaptoethanol as described by Sun and Green (1976) and King et al (1986), then counted in a haemocytometer. Cornified envelopes are regarded as insoluble in solutions containing SDS and 13-mercaptoethanol (Sun and Green 1976, King et al 1986).

500 400 V

g

t/

300

/ /

r

o

200 o o

100

I

I

~

I

-4

-3

-2

-1

I

0

Collagen [log (gg cm-Z)] FIG t : Effect of coating the culture dishes either with collagen type I or type IV on cell-overgrown area. Horizontal axis represents the amount of collagen cm-2 of culture dish on a logarithmic scale. Each point represents the mean (with SEM)of four separate experiments made in triplicate

condition. For comparison of multiple means, an analysis of variance was first carried out on the data. If this indicated there was a significant difference between means, the Newman-Keuls multiple test (Haiger 1982) was employed to determine which of the means differed from the others. Results

Collagen coating Determination of cell-overgrown area Cells attaching to the surface of the culture dishes were fixed by adding 10 per cent PBSbuffered formalin solution to the dishes for 10 minutes and thereafter stained with a 10 per cent Giemsa solution for 30 minutes. The size of the cell-overgrown area was determined by an electronic planimeter (MOP-AM 02, Kontron).

Statistics Each treatment was tested in four different primary cultures in triplicate. The values obtained from the three observations made in each primary culture were averaged. From these data, the mean value + SEMwas calculated for each experimental

Treatment of culture dishes with collagen was necessary for cell spread (Fig 1). Independent of the collagen type, increases in the cell-overgrown area were seen when each cm2 of the culture dish was coated with more than 0.2 gg collagen. As shown in Fig 1, collagen IV seemed to be more effective in promoting cell growth at 2 gg cm-2 or more. In the further studies presented below, ruminal epithelial cells were grown on dishes coated with collagen at a concentration of 4 gg cm -2 (type I) or 2 gg cm-2 (type IV). In cultures grown on collagen I, both slightly and strongly stained colonies could be seen. In cultures grown on collagen IV, mainly strong staining was observed (Figs 2a, b). The subconfluent cells slightly stained by Giemsa had ruffled

Culture of sheep ruminal epithelium

FIG 2: Nine-day-old cultures of rumen epithelial cells grown either on collagen I or collagen IV. Giemsa staining. (a) Culture dish (diameter 35ram) with cells grown on collagen I. Both strongly and slightly stained areas can be seen. (b) Cells grown on collagen IV. The culture dish is covered mainly by colonies strongly stained by giemsa. (c) Light microscopy of slightly stained areas. The cells have ruffled membranes and grow only one cell layer thick. (d) Light microscopy of strongly stained areas. The cells form compact stratified colonies and desquamated cells (arrow) can be seen. Each bar = 50 I.tm

membranes and grew only one cell layer thick. The confluent cells strongly stained by Giemsa formed compact, stratified colonies and desquamated cells could be seen (Figs 2c, d). In the studies described below, colonies are referred to as non-stratified if they were slightly stained by Giemsa and if the cells showed morphological characteristics as shown in Fig 2b. Strongly stained areas with cells showing characteristics as shown in Fig 2d are referred to as stratified.

105

FIG 3: Immunohistochemical identification of keratins and m-smooth muscle actin in sections of ruminal papillae (a-c), in non-stratified (df), and in stratified colonies (g-i). (a), (d) and (g) show keratin filaments stained with anti-basic cytokeratin (a) or with anti-cytokeratin pan, LU5, (d, g). Note alignment of filaments in areas of cell contact (d, arrow). No staining was observed with (b, e, h) LE61. Actin filaments were stained in a rich vascularised connective tissue core of papilla with (c) anti-e~sm-'l ; cell cultureswere not stained (f, i). E Epithelium, LP Lamina propria. Bars = 200 I.tm (a-c) or 50 I.tm (d-g)

Positive and negative reactions found in the cultured cells were qualitatively identical to those found in the epithelial layers of ruminal papillae (Fig 3a-c), indicating that the cultured cells were of epithelial origin. The absence of simple epithelial keratins (LE61) and a-actin indicate that the cultured cells were neither fibroblasts nor of Immunohistochemical findings myoepithelial origin. The positive staining with To investigate the epithelial origin of non-strat- ~-sm-1 in the subepithelial layers of ruminal ified and stratified colonies, indirect immunoflu- papillae points to a rich vascularisation in the orescence staining was carried out by using anti- connective tissue core of the papilla (Fig 3c). The coalignment of keratin filaments found bodies against cytokeratins. Positive staining was observed in both types of colonies with LU5 (Fig in adjacent cells (Fig 3d) suggested the presence 3d, g), which is reactive with all cytokeratins of desmosomes, intercellular junctions charac(Overbeck et al 1985), and also with anti-basic teristic of epithelial cells. Using antibodies against cytokeratin, reactive with cytokeratins of the desmoplakins I and II, positive immunofluoresbasic subfamily (not shown). The cells were nei- cent staining of the cells was observed at the sites ther stained by LE61 (Fig 3e, h), which reacts of cell contact (Fig 4a). Furthermore, those cells with mesenchymal origin cytokeratin (Lane 1982) in the non-stratified areas having contact with nor by anti-a-sm-1 (Fig 3f, i), reactive with c~- each other showed positive staining in a network over and around the cells with FN-15 (Fig 4b). actin of smooth muscle (Skalli et al 1986).

106

P. Gdtlfi, G. Ggibel, H. Martens

FIG 4: Non-stratified colonies. (a) Desmoplakin staining with antidesmoplakin. Bar = 501~m. (b) Fibronectin staining with FN-15

FIG 6: Direct immunofluorescent staining of glycoproteins with FITC conjugated lectins (UEAI-FITC). (a) Section of ruminal mucosa. The arrowheads indicate lamina basalis of epithelial region. E Epitheliu m; Le Lamina propria. Bar = 50 ~m. (b) Non-stratified colonies. (c) Stratified colonies. Bar = 50 ~m

Effect of hormones and Na-n-butyrate In cultures grown on collagen I, all tested hormones caused a significant increase in the total area covered by cells (Table 1). After supplementation with hydrocortisone, the area with non-stratified growth decreased significantly with a concomitant increase in stratified growth (Table l). Addition of Na-n-butyrate resulted in a reduction of cell-overgrown area accompanied by an almost complete absence of stratified growth (Table 1). In cultures grown on collagen IV, all hormones tested except hydrocortisone showed no effect on the total cell-overgrown area (Table 2). The ratio of stratified to non-stratified areas was uniformly greater than 20:1. As in the collagen I cultures, addition of Na-n-butyrate led to an inhibition of growth with an almost complete disappearance of stratified colonies (Table 2). The number of cells with cornified envelopes desquamated into the medium was determined by treating the medium with solutions containing SDS and [~-mercaptoethanol. Addition of Na-n-

FIG 5: Identification of vimentin in ruminal papillae and in cultures or rumen epithelial cells by staining with Vim 3B4 antibody. (a) Cross section of ruminal papilla. Positive reaction was seen in the connective tissue. E Epithelium; LP Lamina propria. Bar = 100 ~m. (b) Cross section of a stratified culture grown on permeable supports (Falcon). Arrowheads indicate positive reaction in the basal cell layer. Bar = 50 p,m. (c) Non-stratified colonies grown on a coverslip. Bar = 50 t~m

Positive staining against vimentin (Vim 3B4) was observed in both stratified (not shown) and non-stratified colonies (Fig 5c). Immunostaining of cross sections of stratified areas suggested that positive reaction with Vim 3B4 antibody is limited to the cell layer which is directly attached to the membrane (Fig 5b). In contrast to the cultured cells, the epithelial layers of ruminal papillae showed negative reactions with Vim 3B4 antibody (Fig 5a). UEAI-FITCshowed positive reaction in the stratified colonies only (Fig 6b-c), suggesting a higher degree of differentiation and keratinisation of these cells.

TABLE 1: Cells grown on collagen I for 17 days. Influence of hormones and Na-n-butyrate on cell overgrown area and on the number of cells with cornified envelope in the incubation medium Treatment Control Glucagon (3x10-12 M) Insulin (1-6x10-9 M) Hydrocortisone (10 -7 M) Na-n-butyrate (5×10 -3 M)

Celt-overgrown area (mm 2) Stratified Non-stratified Total 136 189 246 370

+ 18 a _+.+_11 ab + 18 b 4- 25c ND

i07 185 122 29 82

+ + + + +

11 a 10 b 10 a 9c 11 a

243 374 367 399 82

_+ 9 a _+ 14 b + 21 b +_24 b + 11 c

Cells with cornified envelope (cell number × 103)/dish 3.08_+ 0.49 ab 1.11 _+0.28 a 5.38+ 0.55b c 7.53+ 0.59 c 15-5+ 3"4 c

a,b,c,d,e Different superscripts within one column represent significant differences of at least P<0.05 Data are given as means with SEM of four separate experiments made in triplicate NO Stratified areas were not detected

Culture of sheep ruminal epithelium

107

TABLE 2: Cells grown on collagen IV. Influence of hormones and Na-n-butyrate on cell overgrown area and on the number of cells with cornifled envelope in the incubation medium

Treatment Control Glucagon (3x10-12M) Insulin (1.6x10-9 M) Hydrocortisone (10-7 M) Na-n-butyrate (5x10~ M)

Cell-overgrown area (mm 2) Stratified Non-stratified 370 382 356 702

+ 19a + 24a + 23 a -i- 54 b ND

ND ND ND ND 112 _+ 17

Cells with cornified envelope (cell number × 103)/dish 5.7-1- 0.7 a 4"9+0'8 a 8'2 -+ 1 '9 a 28"7_+ 3"3b 5 4 ' 3 - 7.1c

a,b.c Different superscripts within one column represent significant differences of at least P<0.05 Data are given as means with SEMof four primary cultures made in triplicate ND Not determined, area was smaller than 5 per cent of total area

butyrate to the medium increased the number of cells with cornified envelopes both in cultures grown on collagen I and on collagen IV (Tables 1 and 2). Hydrocortisone led to an increase in the number of cells with cornified envelopes, with the effect being more pronounced in cultures grown on collagen IV (Tables 1 and 2). Discussion

In accordance with earlier observations (G~tlfi et al 1988, G~ilfi and Neogr~idy 1988), it was shown that ruminal epithelial cells grown in culture form both stratifed and non-stratified colonies. By immunofluorescence staining, the cells of both types of colonies were identified as being of epithelial origin. However, morphology and differentiation processes seemed to differ between the two types of cells. FIa'C-conjugated UEAI lectin is shown to be reactive with differentiated and keratinised cells in the epidermis (Watt 1988) and in the epithelium of ruminal mucosa (S. Neogr~tdy, A. Pusztai and P. G~dfi, unpublished observations). In accordance with these findings, the present study showed that UEAI binding in the epithelial layers of ruminal mucosa increased from stratum spinosum to stratum corneum, that is, in the direction of differentiation (Fig 6a). Thus, the greater binding of lectin in stratified colonies of the cell cultures as shown in Fig 6c indicates a higher degree of differentiation in these cells. An insoluble protein envelope, the cornified envelope, is laid down underneath the plasma membrane late in terminal differentiation (Watt 1988, Schmidt et a11989). There are several layers of cornified cells in normal epidermis but, in culture, cells containing envelopes tend to slough off into the medium, and cornified layers do not accumulate (Sun and Green 1976). Therefore, in

addition to lectin-binding, the number of cells with cornified envelopes desquamated into the medium can be taken as an indication of terminal differentiation. With the exception of butyrate treated cultures, the number of cells forming cornified envelopes was greater in cultures grown on collagen IV. Thus, the greater proportion of stratified growth in collagen IV cultures seemed to be accompanied by a greater number of cells forming cornified envelopes, indicating a higher degree of differentiation in stratified colonies in accordance with the higher lectin binding. As shown in Fig 2, cells of the non-stratified colonies differed morphologically from those in the stratified colonies. In addition to the morphological differences, it could be shown that the non-stratified cells were able to form desmosomes and produce extracellular matrix proteins like fibronectin (Fig 4). In some respects, properties of cells in the non-stratified colonies seemed to be similar.to those of a non-differentiating strain of human keratinocytes (NDK) detected by Adams and Watt (1988). These cells differ in morphology from normal keratinocytes; they fail to stratify or to undergo terminal differentiation and are able to produce extracellular matrix proteins like fibronectin and collagen IV (Adams and Watt 1988, Watt 1988). As regards cultures of rumen epithelial cells, the exact origin of the non-stratified cells has not been determined. As for the epidermal NDI( ceils, it is suggested that they may be a minor epithelial population or, for some reason, their appearance has been induced in culture (Adams and Watt 1991). For promoting cell growth, coating the culture dishes either with collagen I or IV seemed to be an important factor and collagen IV seemed to be more effective. In addition to proliferation,

108

P. Gdlfi, G. G?ibel, H. Martens

stratification seemed also to be intensified in cultures grown on collagen IV. While both nonstratified and stratified growth could be observed in cultures grown on collagen I, non-stratified areas made up less than 5 per cent of the total area (with the exception of butyrate treatment) of the cultures grown on collagen IV. Apart from extracellular matrix proteins, various hormones (Tsao et al 1982, King et al 1986, Taub 1990) are also known to influence growth and differentiation. In the present study, insulin, glucagon and hydrocortisone stimulated cell growth. However, the stimulatory effect of insulin and glucagon was absent in cultures grown on collagen IV. These results are in agreement with findings of Mohanam et al (1988) in cultured mammary cells who found that the effect of growth factors depended on the presence of collagen I, whereas this effect was absent in cultures grown on collagen IV. To explain the interrelations between growth factors and the type of collagen, at least two possibilities must be taken into account. First, treatment with growth factors may lead to synthesis of collagen IV (Mohanam et al 1988), which has been shown to have a greater stimulatory effect on cell growth and differentiation (Fig 1). Secondly, cells growing on collagen IV in a stratified fashion with the development of cornified ceils may, therefore, be less responsive to hormonal treatment. In contrast to insulin and glucagon, hydrocortisone was effective in promoting growth and differentiation not only in collagen I cultures but also in cultures grown on collagen IV. This may be due to the fact that hydrocortisone is more lipophilic and acts via cytosolic receptors, thus, the diffusional barrier of stratification induced by collagen IV may be less important. In various cell types, butyrate terminates proliferation and induces differentiation (Kruh 1982, Schmidt et al 1989, Rephaeli et al 1991, Howard et al 1991). The present study confirmed the antiproliferative effect of butyrate, since areas of cell overgrowth were diminished in the presence of butyrate (Table 1 and 2). Signs of differentiation were not visible in the cells attached to the collagen but were in those cells desquamated into the medium. Among those cells, the number of cells with cornified envelopes was greatly increased. Thus, in butyrate treated cultures, almost all cells seemed to have undergone

terminal differentiation resulting in the formation ofcornified envelopes and ultimate desquamation into the supernatant medium. Therefore, only those cells which were not able to form cornified envelopes remained attached to the collagen. In conclusion, it is known that normal ruminal epithelium undergoes intensive proliferative and regressive changes throughout the entire cell life. These processes are mainly caused by the type of feeding and the amount of concentrates given to the animal. However, it is still controversial which of the dietary factors induces proliferation and differentiation of the epithelium. This study shows that, in cultured ruminal cells, these processes are strongly influenced by extracellular components. Thus, rumen epithelial cells kept in culture may serve as a model for studying the influence of dietary factors on proliferation, stratification and differentiation.

Acknowledgements P. G. was supported by grants from the Alexander von Humboldt Foundation and from the National Research fund (OTKA,topic number 1254). We thank F. M. Watt for her gift of the LE61 and the o~-sm-1 monoclonal antibodies, and E. Heinrich for the antidesmoplakin sera and anti-basic cytokeratin antibody.

References ADAMS, J. C. ~¢ WATT, F. M. (1988) An unusual strain of human keratinocytes which do not stratify or undergo terminal differentiation in culture. Journal of Cell Biology 107, 1927-1938 ADAMS, J. C. & WATT, F. M. (1991) Expression of 131,133, [~4, and 155 integrins by human epidermal keratinocytes and non-differentiating keratinocytes. Journal of Cell Biology 115, 829-841 BALDWIN, R. L. & JESSE, B. W. (1991) Technical note: isolation and charactefisation of sheep ruminal epithelial cells. Journal of Animal Science 69, 3603-3609 BERGMAN, E. N. (1990) Energy contributions of volatile fatty acids from the gastrointestinal tract in various species. Physiological Reviews 70, 567-590 CHENG, K. J., McCOWAN, R. P. & COSTERTON, J. W. (1979) Adherent epithelial bacteria in ruminants and their roles in digestive tract function. American Journal of Clinical Nutrition 32, 139-148 DIRKSEN, G., LIEBICH, H. G., BROSI, G., HAGEMEISTER, H. & MAYER, E. (1984) Morphologie der Pansenschleimhant und Fetts~iureresorption beim Rind - bedeutende Faktoren for Gesundheit und Leistung. Zentralblattflir Veterinarmedizin A 31, 414-430 FRIEDMAN, D. L., CLAUS, T. H., PILKIS, S. J. & PINE, G. E. (1981) Hormonal regulation of DNA synthesis in primary cultures of adult rat hepatocytes. Experimental Cell Research 135, 283-290 GABEL, G., MARTENS, H., SUENDERMANN, M. & GALFI, P. (1987) The effect of diet, intraruminal pH and osmolarity on sodium, chloride and magnesium absorption from the temporarily

Culture o f sheep r u m i n a l epithelium isolated and washed reticulo-rumen of sheep. Quarterly Journal of Experimental Physiology 72, 501-511 GABEL, G. & MARTENS, H. (1991) Transport of Na and CI across forestomach epithelium: mechanisms and interactions with shortchain fatty acids. Physiological Aspects of Digestion and Metabolism in Ruminants. Eds T. Tsuda, Y. Sasaki and R. Kawashima. Orlando, Florida, Academic Press. pp129-151 GALFI, P., NEOGRADY, S. & KUTAS, F. (1981a) Culture of epithelial cells from bovine ruminal muensa. Veterinary Research Communications 4, 295-300 GALFI, P., NEOGRADY, S. & KUTAS, F. (1981b) Keratinization of bovine ruminal epithelial cells in primary culture. Biology of the Cells 42, 103-108 GALFI, P., NEOGRADY, S. & KUTAS, F. (1982) A method for identifying the degree of differentiation of isolated run'final epithelial cells. Zentralblatt far Veterinarmedizin A 29, 477-480 GALFI, P., NEOGRADY, S., KUTAS, F. & VERESEGYHAZY, T. (1983) Keratinization of cultured ruminal epithelial cells treated with butyrate and lactate. Zentralblatt far Veterinarmedizin A 30, 775-781 GALFI, P., NEOGRADY, S. & KUTAS, F. (1988) A bend6nyfilkahfirtya adapfftci6j~t szabfilyoz6 61ettani t6nyezbk vizsg~llata sejtteny6szetben (Study of the physiological factors responsible for the adaptation of ruminal mucosa in cell culture). Magyar A'llatorvosok Lapja 43, 309-311 GALFI, P. & NEOGRADY, S. (1988) Epithelial and non-epithelial cell- and tissue culture from the ruminal mucosa. Asian-Australasian Journal of Animal Sciences 2, 143-144 HAIGER, A. (1982) Biometrische Methoden in tier Tierproduktion. Wien, Ostereichischer Agrarverlag. HENRIKSON, R. P. (1970) Ultrastructure of ovine ruminal epithelium and localization of sodium in the tissue. Journal of Ultrastructure Research 30, 385-401 HOWARD, M. A., WARDWELL, S. & ALBERT, D. M. (1991) Effect of butyrate and corticosteroids on retinoblastoma in vitro and in vivo. Investigative Ophthalmology and Visual Science 32, 1711-1713 INOOKA, S., OHWADA, S. & TAMATE, H. (1984) A simple primary culture of epithelial cells from rumen mucosa of cattle. Japanese Journal of Zootechnical Sciences 55, 517-520 KEYNES, R. D. & HARRISON, F. A. (1970) Transport of inorganic ions across the rumen epithelium. Physiology of Digestion and Metabolism in the Ruminant. Ed A. T. Phillipson. Newcastleupon-Tyne, Oriel Press. pp113-118 KING, I., MELLA, S. L. & SARTORELLI, A. C. (1986)A sensitive method of quantifying the terminal differentiation of cultured epidermal cells. Experimental Cell Research 167, 252-256 KRUH, J. (1982) Effects of sodium butyrate, a new pharmacological agent, on cells in culture. Molecular and Cellular Biochemistry 42, 65-82

109

LANE, E. B. (1982) Monoclonal antibodies provide specific intramolecular markers for the study of epithelial tonofilament organization. Journal of Cell Biology 92, 665-673 LAVKER, R., CHALUPA, W. & DICKEY, J. F. (1969) An electron microscopic investigation of rumen mucosa. Journal of Ultrastructure Research 28, 1-15 MOHANAM, S., SALOMON, D. S. & KIDWEL, W. R. (1988) Substratum modulation of epidermal growth factor receptor expression by normal mouse mammary cells. Journal of Dairy Science 71, 1507-1514 OVERBECK, J., STAHLI, C., GUDAT, F., CARMAN, H., LAUTENSCHLAGER, C., D(IRM(ILLER, U., TAKACS, B., MIGGIANO, V., STAEHELIN, T. & HEITZ, Ph. U. (1985) Immunohistochemical characterization of an antiepithelial monoclonal antibody (mAB lu-5). Virchows Archiv (Pathologie and Anatomie) 407, 1-12 REPHAELI, A., RABIDAZEH, E., AVIRAM, A., SHAKALI, M., RUSE, M. & NUDELMAN, A. (1991) Derivatives of butyric acid as potential anti-neoplastic agents. International Journal of Cancer 49, 66-72 SAWADA, N., TOMOMURA, A., SAT'FLER, C. A., SATFLER, G. L., KLEINMAN, H. K. & PITOT, H. C. (1986) Extracellular matrix components influence DNA synthesis of rat hepatocytes in primary culture. Experimental Cell Research 167, 458-470 SCHMIDT, R., CATHELINEAU, C., CAVEY, M. T., DIONISIUS, V., MICHEL, S. SHROOT, B. & REICHERT, U. (1989) Sodium butyrate selectively antagonizes the inhibitory effect of retinoids on coruified envelope formation in cultured human keratinocytes. Journal of Cellular Physiology 140, 281-287 SKALLI, O., ROPRAZ, P., TRZECIAK, A., BENZONANA, G., GILLESSEN, D. & GABBIANI, G. (1986) Amonoclonal antibody against a-smooth muscle actin: a new probe for smooth muscle differentiation. Journal of Cell Biology 103, 2787-2796 SUN, T. T. & GREEN, H. (1976) Differentiation of the epidermal keratinocyte in cell culture; Formation of the cornified envelope. Cell 9, 511-521 TAUB, M. (1990) The use of defined media in cell and tissue culture. Toxicology in Vitro 4, 213-225 TSAO, M. C., WALTHAL, B. J. & HAM, R. G. (1982) Clonal growth of normal human epidermal keratinocytes in a defined medium. Journal of Cell Physiology 110, 219-229 WATT, F. M. (1988) The epidermal keratinocyte. BioEssays 8, 163167 WHANGER, P. D. & CHURCH, D. C. (1970) Enzyme activities of ruminant digestive tract epithelial tissue. Journal of Animal Science 30, 944-948 Received October 3, 1991 Accepted July 27, 1992