Biol Cell (1991) 73, 49-56 © Elsevier, Paris
49 Original
article
Modulations of the epithelial phenotype and functional activity of cultured bovine tracheal gland cells: dependence on the culture medium and passage number R a c h i d B e n a l i ~, F l o r e n c e D u p u i t ', M a r t i n e C h e v i l l a r d ~, J a c k y J a c q u o t ~, B e r n a r d H a y e 2, l~dith P u c h e l l e ~ 11NSERM U314; ~Laboratoire de Biochimie, Universit6 de Reims, 51092 Reims Cedex, France (Received 19 April 1991; accepted 17 July 1991)
Summary - Bovine tracheal gland (BTG) cells in culture show an ¢pithelial-fibroblastoid transition after several passages. To investigate these BTG cell phenotype changes, we studied the effects of both the culture medium and passage number on the expression of epithelial cytoskeletal proteins and glandular serous cell markers. We also analyzed the intr~xellular cAMP level in the basal state and after adrenergic stimulation. Three culture media were used: 1) serum-free defined medium (SFDM); 2) medium supp|cmented with 2% Ultroser G; and 3) medium suPI~lemented with 10% fetal calf serum (FCS). Using immunofluorescence microscopy, we showed that, in the first 4 passages whatever the culture conditions, BTG cells expressed immunoreactivities to cytokeratin filaments and desmoplak~ns I and If, whereas vimentin filaments were not detected. After four passages, BTG cells cultured in 10% FCS or 2% Ultroser G became progressively fibroblastoid and showed immunoreactivities to both vimentin and cytokeratin intermediate filaments. No immunoreactivity to vimentin filaments was observed on BTG cells cultured in a SFDM. Using biochemical analysis, we showed that basal levels of cAMP in cultured BTG cells and lysozyme secretion by these cells vat3' according to the culture medium and passage number. It was hit!her in BTG cells cultured in a SFDM compared to that recovered front cells cultured in medium supplemented with Ultroser G or FCS. Whatever the culture medium, BTG cells responded to stimulation by isoproterenol. However, the results of stimulation in a SFDM were higher than in Uitroser G or FCS supplemented medium. We conclude that the BTG epithelial cell organization and the regulation of biosynthesis of secretory proteins by these cells in culture depend on both the culture medium and passage number. The maintenance of the epithelial serous phenotype of BTG cells during the first passages (1 to 4) suggests that the study of the regulation of these gland cells should be limited to these early passages. bovine tracheal gland cells I difft~ntiation I culture medium I passage number
Introduction
Airway secretions play an important role in the defense of the respiratory tract. In airway epithelium, submucosal gland secretory cells are considered as the main source of secreted mucus [24]. In order to analyze the regulatory mechanisms of airway mucous secretion, different research groups have developed systems to culture bovine [3, 16] and human [33, 37] tracheal gland ceils. Recently we have described a method for the isolation and culture of bovine tracheal gland (BTG) cells [3] and reported that they modified their phenotype according to the culture medium and passage number [4]. After four passages we noticed that the BTG cells exhibited an epithelial-fibroblastoid transition. Morphological changes of a given cell type, from an epithelial to a fibroblastoid appearance, have been reported following addition of antibodies against cell adhesion molecules [2], transfecticm of cells with oncogenes [20], addition of EGF, transforming growth factor-~ or "epithelial scatter factor" to cell cultures [1, 9, 34], lowering of the extracellular Ca" + concentration [22], or plating the epithelial cells on certain substrata [35]. Recently, Boyer el al [7] reported that the culture medium may be responsible for the epithelialfibroblastoid transition of the rat bladder carcinoma-
derived cell line NBT II. Using human umbilical vein endothelial cells, Dichek et al [13] have shown variability in the level of expression of messenger RNA according to the passage number and time in culture. To d e f n e the factors that may modulate BTG cell morphological changes in culture, we have examined the effect of the culture medium and passage number on the expression of cytoskeletal proteins (cytokeratin, vimentin and desmoplakin) during the time of epithelial-fibroblastoid transition of these cells. Intracellular cyclic A M P (cAMP) regulates many critical differentiated functions of tracheal epithelial cells [15, 26]. To determine the modification of the relationship between the structure and functional activity of BTG cells in culture, we have investigated herein the effects of the culture medium and passage number on intracellular levels of cAMP in cultured BTG cells during steady state and after ~-adrenergic stimulation. The changes in functional activity of the BTG cells were also analyzed by reference to the secretion of lysozyme, which represents a good protein marker for secretory activity of tracheal gland serous cells in culture [36]. Our results demonstrate that the BTG epithelial cell organization and functional activity depend on both the culture medium chosen and their passage number.
50
R Benali et al
Materials and methods
Total protein were determined by the Bio-Rad protein microassay procedure [6] using bovine serum albumin as a standard. All other results are expressed as fmol per/~g total protein
Cell culture conditions Isolation and culture of BTG cells were performed according to the technique previously described [3]. Briefly, BTG cells were isolated by enzymatic digestion and cultured on human placental collagen [10] in the presence of: 1) a 1 : 1 medium of Dulbecco's Modified Eagle's/Ham's Fl2 (D/FI2) supplemented with 10070 fetal calf serum (FCS); 2) D/FI2 supplemented with 2070 Ultroser G (a serum substitute); 3) a serum-free defined medium (SFDM) containing D/FI2, 10 ng/ml epidermal growth factor (EGF), 1 pg/ml transferrin, 1 pg/ml insulin, 0.5 pg/ml hydrocortisone and 10 ng/ml retinoic acid. Cultures were serially passaged by trypsinization and cultured cells from passages 4 to 8 were examined. Cells were seeded at a density of 3 x 104 cells/cm 2.
lmmunofluorescence For indirect immunofluorescence staining of cytokeratin, vimentin and desmoplakin, B:FG cells, grown on glass coverslips coated with human placental collagen, were fixed for 30 min at - 2 0 ° C in methanol and washed in phosphate-buffered saline (PBS) pH 7.4. Then the biotin streptavidin fluorescein technique was applied. Mouse monoclonal antibodies to cytokeratins 19, 18 and 8 (PKK 1) were purchased from Labsystems and used at a 1 : 10 dilution. Mouse monoclonal antibody to vimentin was purchased from Amersham and used at a 1 : 50 dilution. Mouse monoclonal antibodies to desmoplakins I and II were purchased from Boehringer Mannheim and used at 1 : 10 dilution. Streptavidin-fluorescein and biotinylated goat anti-mouse IgG were obtained from Amersham and used at 1:25 dilution. All the dilutions were performed in a 1°70 bovine serum albumin (BSA)-PBS solution pH 7.4. Before immunostaining, cells were incubated in a 1070 BSAPBS solution for 10 min. Primary antibodies were applied for 1 h, followed by two 5-rain washes in PBS, a l-h incubation with biotinylated goat anti-mouse IgG, two 5-min washes in PBS, a 40-min incubation with streptavidin-fluorescein and two final washes of 5 min each in PBS. All the incubations were carried out at room temperature. The glass coverslips were mounted on the slide with antifading glycerol and the stained cells were examined under a Zeiss Axiophot fluorescence microscope.
Treatment of cells and cyclic AMP assay Cells from passage 4 and passage 8 were cultured in 12-well cluster plastic dishes and used when confluent. The monolayers were rinsed twice with D/FI2 containing HEPES (20 mM) and incubated for 10 min at 37°C in 900 ~1 of the D/FI2 containing HEPES (N-2-hydroxyethyl-piperazine-N'-2-ethanesulfonic acid, 20 mM) and IBMX (isobutylmethylxanthine, 10-SM), in order to study the basal levels of cAMP, and for 0, 5, 10, 15, 20 and 25 min in 900/~1 of the D/FI2 containing HEPES (20 mM), IBMX (10-5M) and isoproterenol (10-5M), in order to study the effects of fl-adrenergic stimulation on the cAMP accumulation. After the various incubation periods, cAMP production was stopped by adding 100 tzi of 10 N periodic acid. Cells were then scraped from the culture dishes and intracellular cAMP levels were determined by radioimmunoassay using the method of Cailla et ai [8]. All determinations were done in duplicate and the mean was calculated.
Lysozyme assay Culture media from confluent BTG cells were concentrated by pressure filtration over Amicon YM-2 ultrafilters (Amicon Corp, Lexington, MA) and dialyzed against deionized water. Lysozyme enzymatic activity was evaluated spectrophotometrically by measuring the initial rate of lysis of Micrococcus luteus suspension (Worthington Biochemical Corp) according to Shugar's method [32] modified by Jacquot et al [23]. The bovine lysozyme-type enzymatic activity was defined as units of enzyme activity per microgram of protein and expressed as equivalent of hen eggwhite lysozyme (Worthington).
Statistical analysis For cAMP assay, variance analysis was used to compare differences in the values according to the culture medium, passage number and incubation time. For lysozyme and total protein secreted by cultured BTG cells, the results were compared using the Student t-test. Values are expressed as means + SD of 3 determinations. P < 0.05 was considered as the limit of signification.
Results
Cell morphology under different culture conditions Bovine tracheal gland cells were cultured in a S F D M or in a medium supplemented either with 2% Ultroser G or with 10% FCS. Before passage 5, all the BTG cells cultured under these three culture conditions assumed an epithelial morphology, forming dense colonies o f tightly packed polyglonal cells (figs la, 2a) and showed immunoreactivities to cytokeratin filaments (figs lb, 2b) and desmoplakins I and II (figs lc, 2c). Vimentin filaments were not detected in BTG cells whatever the culture conditions (figs ld, 2d). However, at passage 5, about 50% of BTG cells cultured in 10% FCS or 2 0 Ultroser G became fibroblastoid: they flattened, exhibited a spindle shape, and crawled one over another (fig 3a) and showed immunoreactivities to both vimentin and cytokeratin f.]aments (fig 3b, c). At passage 8, 100% of BTG cells cultured under these two culture conditions were fibroblastoid (fig 2e) and expressed no immunoreactivities to cytokeratin filaments (fig 2f) and to desmoplakins I and II (fig 2g). They expressed immunoreactivity to vimentin filaments (fig 2h). In contrast, all the BTG cells cultured in a SFDM from passage 5 to passage 8 retained an epithelial morphology (fig le) and expressed immunoreactivities to cymkeratin filaments (fig 1f) and to desmoplakins I and Ii (fig lg) but not to vimentin (fig lh).
c A M P levels under different culture conditions The intracellular c A M P levels of the BTG cells in the basal state and after isoproterenol stimulation in the three cul-
Fig 1. Light micrographs of BTG cells cultured in a serum-free defined medium at passages 4 (a-d) and 8 (e-h). a, e. Phase contrast micrographs of confluent monolayers of cultured BTG cells; cells demonstrate an epithelial morphology of tightly packed polygonal cells. Bar -- 100 ~m. b, f. Immunofluorescence staining of cytokeratin filaments in cells as detected by monoclonal antibodies to cytokeratins 19, 18, and 8 (PKKI). All cells showed immunoreactivity to PKKI antibodies. Bar -- 25 pm. c, g. Immunofluorescence staining of desmoplakins along cells boundaries of cells as detected by monoclonal antibodies to desmoplakins I and Il. Bar = 20/~m. d, h. Immunofluorescence staining of vimentin filaments in cells. No immunoreactivity to vimentin was detected. Bar = 25/zm.
Phenotype of epithelial cells in culture
51
52
R Benali et al
.
•
''~----
--'--
-
~-7-
Phenotype of epithelial cells in culture
53
Fig 2, Light micrographs of BTG cells cultured in medium supplemented with either 2% Ultroser G or 10070FCS at passages 4 (a-d) and 8 (e-h). a, e. Phase contrast micrographs of cells. In a, cells demonstrate an epithelial morphology of tighly packed polygonal cells, in e, they exhibit a fibroblastoid morphology. Bars = 100/zm. b, f. lmmunofluorescence staining of cytokeratin filaments in cells. In b, all cells were positive with anti-cytokeratin PKKI antibody. In f, no immunoreactivity to cytokeratin was detected. Bar = 25/~m. e, g. Immunofluorescence staining of desmoplakins along cells boundaries in cells. Note the disappearance of immunoreactivity from cellular junctions and cell periphery at passage 8 kg). Bar = 25 pm. d, h. lmmunofluorescence staining of vimentin filaments in cells. In d, no immunoreactivity to vimentin was detected. In h, all cells were positive with anti-vimentin antibody. Bar =
25 ~ m .
ture conditions and at different passages were examined. Figure 4 shows the comparison of basal levels o f c A M P o f BTG cells grown in a SFDM or in D / F I 2 medium supplemented either with 2070 Ultroser G or with 10o70 FCS at passage 4 and passage 8. Variance analysis showed a significant variation of c A M P levels according to the culture medium (P < 0.01) and to the passage number (P < 0.001). At passage 4, intracellular level of cAMP grown in a SFDM was 130 fmol per pg protein, whereas it was only 90 fmol per pg protein when BTG cells were grown in D / F I 2 supplemented with 2o70 Ultroser G and 70 fmol per/~g protein when they were grown in D / F I 2 medium supplemented with 10o70 FCS. At passage 8, intracellular c A M P levels of BTG cells were reduced by 50°70 compared to that recovered at passage 4, whatever the culture medium. During the isoproterenol stimulation period (fig 5), the intracellular cAMP levels increased with incubation time with a maximum effect after a 15-rain period and varied significantly according to the culture medium (P < 0.001) and to the passage number (P < 0.001) as demonstrated by variance analysis. At passage 4 (fig 5A) the intracellular c A M P level in cells cultured in SFDM was 2.5 and 3.0-fold higher compared to that recovered from cells cultured in 2o7o Ultroser G and 10°70 FCS, respectively. At passage 8 (fig 5B), intracellular cAMP responses o f BTG cells were reduced by 50o70 (fig 5B) compared to that recovered at passage 4, whatever the culture medium, but the patterns of response were similar. L ysozyme secretion The amounts of iysozyme secreted by BTG cells main° tained in the three culture conditions and at different passages were determined. Figure 6 shows the comparison of lysozyme amounts secreted by BTG cells grown in SFDM or in D / F I 2 medium supplemented either with 2°/o Ultroser G or with 10o70 FCS at passage 4. BTG cells cultured in SFDM secreted 350 ng of lysozyme/106 cells/ 72 h incubation (expressed as equivalent of hen egg-white
Fig 3. Light micrographs of BTG cells cultured in medium supplemented either with 2% Ultroser G or with 10% FCS at passage 5. a. Phase constrast m~crograph of confluent monolayers of BTG cells. Some cells still demonstrate an epithelial morphology of tightly packed p¢!ygonal cells (sofid arrow] whereas other cells are fibroblastoid (open arrow). Bar = 100/~m. b. Double immunofluorescence staining of vimentin filaments (FITC) in fibroblastoid cells and cytokeratin filaments (rhodamin) in tightly polygonal celIs. Bar = 50/~m. e. Double immunofluorescence staining showing the presence of both vimentin (FITC) and cytokeratin (rhodamin) filaments in fibroblastoid cells. Bar = 10 pm.
54
R Benali et a!
cAMP
cAMP
10 "15 M/pg proteins 5000
10 "15 M/pg proteins 150
~000 3000
100
2000 1000
50
0 SFMO
US G
FOS
Fig 4. Basal levelsof cAMP of BTG cells in the presence of IBMX (10 -5 M) for 10 min at passages 4 (P4) and 8 (P8) on day 8 of culture. Varianceanalysisshowed a significantvariation of cAMP production according to the culture medium (P < 0.01) and the passage number (P < 0.001).
5
10
15
20
25
10
15
20
25
5000 4000
(~)
3000 2000
lysozyme). This lysozyme content was 7- to 10-fold higher and significantly different (P < 0.001) compared with that recovered from BTG cells cultured in 10% FCS or 207o Ultroser (3 and represents more than 5.0°7o of total proteins secreted in the culture medium (fig 7). After passage 4, the amount of lysozyme secreted by BTG cells cultured in a SFDM decreased significantly (P < 0.001) and represented only 0,7% of total proteins secreted in the culture medium (fig 7). However, no significant change in the amount of total proteins secreted by BTG cells in a SFDM from passage 4 to 8 was observed (fig 7). Discussion Bovine tracheal gland cells cultured in the presence of FCS or Ultroser G showed an epithelial-fibroblastoid transition after 4 passages. In this work we have examined the expression of cytoskeletal and desmosomai proteins in cultured BTG cells using selected monoclonal antibodies. It appears clear from our study that the phenotype of cultured BTG cells varies dramatically according to the culture medium used and passage number. We have observed after 4 passages that culture medium supplemented with 2% Uitroser G or with 10% FCS induced morphological changes resembling those observed in embryonic cells undergoing epitheliai-fibroblastoid transition [14, l?], During the first four passages, BTG cells form a cobblestone-like monolayer of polygonal cells and express immunoreactivities to cytokeratin intermediate filaments and desmoplakins, whereas after 4 passages, the cells dissociated, assumed a spindle-like shape and expressed immunoreactivity to both cytokeratin and vimentin. Around passage 8 after the epithelial-fibroblastoid transition, the desmoplakin distribution along cell boundaries was dramatically reduced and the expression of the cytokeratin intermediate filaments disappeared completely. No modu-
1000 0 0
5
Incubation time (rain) Fig 5. Effects of isoproterenol (10 -s M) in the presence of IBMX (10 -s M) on cAMP formation in BTG cells at passages 4 (A) and 8 (13)on day 8 of culture. Variance analysis showed a significant variation of cAMP production according to the incubation time (P < 0.001), the culture medium (P < 0.001) and the passage number (P < 0,001),
Lysozyme (ng/lO" cells172h) 400
300
200
100
O
SFOM
FCS
US G
Fig 6. Lysoz.vmesecretion by BTG cells cultured in a serum-free defined medium (SFDM), medium supplemented with 2% Ultroser G (US G) or 10% fetal calf serum (FCS). Values represent mean + SD of 3 determinations. (* P < 0.001).
Phenotype of epithelial cells in culture
Proteins (pg/lO s cells/72h)
Lysozyme/total proteins (%) 6
5
4 3 2 1
0
0 4
5
6 7 Passage number
8
Fig 7. Lysozyme secretion by BTG cells cultured in a serum-free defined medium expressed as percentage of total protein secreted in the medium ( • - - • ) and total protein secreted by BTG cells cultured in a serum-free defined medium ( D - - ~ ) at different passages. Values represent mean + SD of 3 determinations. (* P < 0,001).
lation of the epithelial phenotype of BTG cells was observed when these cells were cultured in a serum-free defined medium. Vimentin intermediate filaments were formed in the fibroblastoid BTG cells. It is possible, however, that the epitheliai-fibroblastoid transition and formation of vimentin filaments in these cells was induced by soluble factors contained in fetal calf serum or in the Ultroser G. A rapid induction of vimentin intermediate filaments formation has been observed often in primary cultures of epithelial cells [31] including rat hepatocytes [18] as well as human urothelial cells. The advent of vimentin is known to be influenced by various components such as externally added retinoid compounds as well as certain growth factors and hormones [25, 29, 30]. It has been therefore the general opinion that induction of vimentin is a result of culture conditions [II, 21, 27, 38]. Our results indicate that the bovine tracheal gland epithelial cell organization depends on both the culture medium and passage number, and the use of a serum-free defined medium seems to be essential to maintain the differentiated epithelial phenotype of these cells. Similar results indicating an influence of the culture medium on epithelial cell phenotype have been reported recently by Boyer et ai [7]. According to the last mentioned authors, the exposure of the rat bladder carcinoma cell line NBT-II to medium containing Ultroser G, shows an epithelial-fibroblastoid transition. Extensive alterations of desmosomal structures and proteins occurred during the epithelial-fibroblastoid transition of NBT-II cells. The pathway leading to desmosomal alterations include phosphorylation and protein synthesis. The mechanisms by
55
which Ultroser G acts to promote desmosomal changes is most probably a multistep cascade. Boyer et al [7] have suggested that solubilization of desmoplakins is a relatively early event in cell dissociation and in epithelialfibroblastoid transition. Benitez-King et al [5] have shown for MDCK cells that during the in vitro formation of an epithelial monolayer cytoskeleton configuration, cell shape and cell contacts modulate cytoskeletal protein posttranslational modifications and there is a correlation between the phosphorylation patterns of intermediate filaments proteins and cytoskeletal protein synthesis. Intracellular cyclic AMP (cAMP) is known to regulate many critical differentiated functions of tracheal epithelial cells [15, 26]. In the present study, we have shown that basal levels of cAMP in cultured BTG cells vary according to the culture medium and passage number. It is higher in BTG cells cultured in a SFDM compared to that recovered from cells cultured in medium supplemented with Ultroser G or FCS. Whatever the culture medium, BTG cells respond to stimulation by isoproterenol. However, the response to stimulation in a SFDM is higher than in Ultroser G or FCS supplemented medium. The difference in response to stimulation by isoproterenol may be due to a decrease of ~-adrenergic receptor concentration under these culture conditions or to the inhibitory factors conrained in Ultroser G or in FCS. Recently, it has been shown [12] that dexamethasone, which is present at a high concentration in Uitroser G, decreases ~-adrenergic receptor concentration and inhibits ~-adrenergic responses during explant culture of human fetal lung. Analysis of lysozyme biosynthesis and secretion by BTG cells also showed differences according to the culture medium used. Secretion of lysozyme by BTG cells was 7- to 10-fold higher in a SFDI~! .han in Ultroser G or FCS supplemented medium. This suggests that lysozyme synthesis and secretion depend on specific factors present in culture medium. The mechanism~ by which Ultroser G and FCS modify the protein synthesis remain to be elucidated. In the present study, we have also been able to demonstrate the effect of passage number on lysozyme secretion by BTG cells, since the amount of this protein decreased markedly after 4 passages. The modifications in the amount of lysozyme secreted by BTG cells after 4 passages may, therefore, result from regulation at the level of both RNA transcription and the intracellular degradation of protein. These data are consistent with recent studies on human umbilical vein endothelial cells [13], and on pig skin and corneal fibroblasts [19, 28], which showed modulations of gene expression during in vitro cells 'ageing'. In conclusion, we have shown that the bovine tracheal gland epithelial cell phenotype in culture depends on both the culture medium used and passage number. Early passages of cells cultured in a serum-free defined medium appear to be the optimal conditions to investigate the normal regulation of biosynthesis of secretory proteins produced by bovine tracheal glands cells.
Acknowledgments
The authors are grateful to Mrs J Hinnrasky for her technical support and to Mrs A Chaveriat for typing this manuscript. This work was supported by the contract INSERM/SYNTHELABO n ° 87118 and by Association Francaise de Lutte contre la Mucoviscidose (AFLM).
56
R Benali et at
References
1 Barrandon Y, Green H (1987) Cell migration is essential for sustained growth of keratinocyte colonies; the roles of tra~sforming growth factor-a and epidermal growth factor. C,.li 50, 1131-1137 2 Behrnes J, Birchmeier W, Goodman SL, Imhof BA (1985) Dissociation of Madin-Dabry canine epithelial cells by the monoclonal antibody anti Arc-I : mechanistic aspects and identification of the antigen as a component related to uvomurulin. J Cell Biol 101, 1307-1315 3 Benali R, Dupuit F, Jacquot J, Fuche¥ C, I-linnrasky J, Ploton D, Puchelle E (1989) Growth and characterization of isolated bovine tracheal gland cells in culture. Influence of a reconstituted basement membrane raat~-ix. Bb)i Cell 66, 263-270 4 Benali R, Dupuit F, Jacquot J, Haye B, Pucheile E (1990) Influence of the culture medium and passage numbers on the intracellular formation of cyclic AMP by cultured bovine tracheal gland serous cells. Am Rev Respir Dis (abstract) 141, A88 5 Benitez-King G, Cazares F, Meza I (1989) Synthesis and phosphorylation of cytoskeletal proteins during the in vitro biogenesis of MDCK cell monolayers. J Cell Sci 93, 53-61 6 Bio-Rad Protein Assay (1984) Bio-Rad technical bulletin n ° 1069, Bio-Rad Laboratories, Richmond, Ca, USA 7 Boyer B, Tucker GC, Vall~s AM, Franke WW, Thiery JP (1989) Rearrangements of desmosomal and cytoskeletal proteins during the transition from epithelial to fibroblastoid organization in cultured rat bladder carcinoma cells. J Cell Biol 109, 1495-1509 8 Caila HL, Racine-Weisbuch MS, Delaage MA (1973) Adenosine Y, 5' cyclic monophosphate assay at 10-msmol level. Anal Biochem 56, 394-407 9 Chinkers M, McKauna JA, Cohen S (1979) Rapid induction of morphological changes in human carcinoma cells A431 by epidermal growth factor. J Cell Bio183,260-265 10 Coleman DL, Tuet IK, Widdicombe JH (1984) Electrical properties of dog tracheal epithelial cellsgrown in monolayer culture. Am J Physioi 246, C355-359 1! Dairkee SH, Blayney CM, Asarnov DM, Smith SH, Hackett AJ (1985) Early expression of vimentin in human mammary cultures. In Vitro 21,321-327 12 Davis DJ, Jacobs MM, Ballard PL, Gonzales LK, Roberts JM (1990)~-Adrenergic receptors and cAMP response increase during explant culture of human fetal lung: partial inhibition by dexamethasone. Pediatr Res 28, 190-195 13 Dichek D, Quertermous T (1989) Variability in messenger RNA levels in human umbilical vein endothelial cells of different lineage and time in culture. In Vitro 25, 2889-292 14 Edelman GM (1984) Cell adhesion and morphogenesis: the regulatory hypothesis. Proc Natl Acad Sci USA 81, 1460-1464 15 Eling TE, Danilowicz RM, Henke DC, Sivarajak K, Yankaskas JR, Boucher RC (1986) Arachidonic acid metabolism by canine tracheal epithelial cells: product formation and relationships to chloride secretion. J Biol Chem 261, 1 2 8 4 - i 289 16 Finkbeiner WE, Nadei JA, Basbaum CB (1986) Establishment and characterization of a cell line derived from bovine tracheal glands. In Vitro 22, 561-567 17 Franke WW, Grund C, Jackson BW, lllmensee K (1983) Formation of cytoskeletal elements during mouse embryogenesis. IV. Ultrastructure of primary mesenchymal cells and their cell-cell interactions. D(fferentiation 25, 121-141 18 Franke WW, Mayer D, Schmid E, Denk H, Borenfreund E (1981) Differences of expression of cytoskeletal proteins in cultured rat hepatocytes and hepatoma cells. Exp Cell Res 134, 345-365 19 Fini ME, Girard MT (1990) The pattern of metalloproteinase
expression by corneal fibroblasts is altered by passage in culture. J Cell Sci 97, 373-383 20 Garcia I, Sordat B, Rauccio-Farinon E, Dunand M, Kraehenbuhl JP, Diggelman H (1986) Estabfishment of. two rabbit mammary epithefial cell fines with distinct oncogenic potential and differentiated phenotype after microinjection of transforming genes. Mol Cell Bioi 6, 1974-1982 21 Hall HG, Bissel MJ (1986) Characterization of the intermediate flament proteins of murine mammary gland epithelial cells. Exp Cell Res 162, 379-389 22 Hennings H, Mickael D, Cheng C, Steinert P, Holbrook K, Yuspa SH (1980) Calcium regulation of growth and differentiation of mouse epidermal cells in culture. Cell 19, 245-254 23 Jacquot J, Tournier JM, Puchelle E (1985) In vitro evidence that human airway lysozyme is cleaved and inactivated by Pseudomonas aeruginosa elastase and not by human leucocyte elastase. Infect lmmun 47, 55-560 24 Kaliner M, Schelhamer JH, Borson B, Nadel J, Patow C, Marom Z (1986) Human respiratory mucus. A m Rev Respir Dis 134, 612-621 25 Kim KH, Stellmach V, Javors J, Fuchs E (1987) Regulation of human mesothelial cell differentiation: opposing roles of retinoids and epidermal growth factor in the expression of intermediate filaments proteins. J Cell Bio1105, 3039-3051 26 Lazarus SC, Basbaum CB, Gold WM (1984) Localization of cAMP in dog and cat trachea: effects of ~-adrenergic agonist. A m J Physiol 247, C327-C334 27 Mackay AM, Tracy RP, Craighead JE (1990) Cytokeratin expression in rat mesothelial cells in vitro is controlled by the extracellular matrix. J Cell Sci 95, 97-107 28 Martin M, El Nabout R, Lafuma C, Crechet F, Remy J (1990) Fibronectin and collagen gene expression during in vitro ageing of pig skin fibroblasts. Exp CelIRes 191, 8-13 29 Rheinwald JG, O'Connell TM (1985) Intermediate filament protein as distinguishing markers of cell type and differentiated state in cultured human urinary tract epithelia. Ann N Y Acad Sci 455-259-267 30 Rheinwald JG, O'Connell TM, Conneli ND, Rybak SM, Allen-Hoffmann BL, La Rocca P J, Wu YJ, Rehwoldt SM (1984) Expression of specific keratin subsets and vimentin in normal human epithelial cells. A function of cell type and conditions of growth during serial culture. Cancer Cells (Cold Spring Harbor) I, 217-228 31 Schmid E, Schiller DL, Grund C, Stadler J, Franke WW (1983) Tissue type-specific expression of intermediate iliament proteins in a cultured epithelial cell line from bovine mammary gland. J Cell Biol 96, 37-50 32 Shugar D (1952) Measurement of lysozyme activity and ultraviolet inactivation of lysozyme. Biochem Biophys Acta 8, 302-309 33 Sommerhoff CP, Finkbeiner WE (1990) Human tracheobronchial submucosal gland cells in culture. A m J Respir Cell Mol Biol 2, 41-50 34 Stoker M, Gherardi E, Perryman M, Gray J (1987) Scatter factor is a fibroblast-derived modulator of epithelial cell mobility. Nature 327, 239-242 35 Tchao R (1982) Novel forms of epithelial cell motility on collagen and on glass surfaces. Cell Motil 4, 333-341 36 Tom-Moy M, Basbaum CB, Nadel JA (1983) Localization and release of lysozyme from ferret trachea: Effects of adrenergic and cholinergic drugs. Cell Tissue Res 228, 549-562 37 Tournier JM, Merten M, Meckler Y, Hinnrasky J, Fuchey C, Pucheile E (1990) Culture and characterization of human tracheal gland cells. A m Rev Respir Dis 141, , 1280-1288 38 Yang J, Balakrishnan A, Hamamoto S, Beattie CW, Gupta TK, Welling SR, Nandi S (1986) Different mitogenic and phenotypic responses of human breast epithelial cells grown in two versus three dimensions. Exp Cell Res 167, 563-569
49 Original
article
Modulations of the epithelial phenotype and functional activity of cultured bovine tracheal gland cells: dependence on the culture medium and passage number R a c h i d B e n a l i ~, F l o r e n c e D u p u i t ', M a r t i n e C h e v i l l a r d ~, J a c k y J a c q u o t ~, B e r n a r d H a y e 2, l~dith P u c h e l l e ~ 11NSERM U314; ~Laboratoire de Biochimie, Universit6 de Reims, 51092 Reims Cedex, France (Received 19 April 1991; accepted 17 July 1991)
Summary - Bovine tracheal gland (BTG) cells in culture show an ¢pithelial-fibroblastoid transition after several passages. To investigate these BTG cell phenotype changes, we studied the effects of both the culture medium and passage number on the expression of epithelial cytoskeletal proteins and glandular serous cell markers. We also analyzed the intr~xellular cAMP level in the basal state and after adrenergic stimulation. Three culture media were used: 1) serum-free defined medium (SFDM); 2) medium supp|cmented with 2% Ultroser G; and 3) medium suPI~lemented with 10% fetal calf serum (FCS). Using immunofluorescence microscopy, we showed that, in the first 4 passages whatever the culture conditions, BTG cells expressed immunoreactivities to cytokeratin filaments and desmoplak~ns I and If, whereas vimentin filaments were not detected. After four passages, BTG cells cultured in 10% FCS or 2% Ultroser G became progressively fibroblastoid and showed immunoreactivities to both vimentin and cytokeratin intermediate filaments. No immunoreactivity to vimentin filaments was observed on BTG cells cultured in a SFDM. Using biochemical analysis, we showed that basal levels of cAMP in cultured BTG cells and lysozyme secretion by these cells vat3' according to the culture medium and passage number. It was hit!her in BTG cells cultured in a SFDM compared to that recovered front cells cultured in medium supplemented with Ultroser G or FCS. Whatever the culture medium, BTG cells responded to stimulation by isoproterenol. However, the results of stimulation in a SFDM were higher than in Uitroser G or FCS supplemented medium. We conclude that the BTG epithelial cell organization and the regulation of biosynthesis of secretory proteins by these cells in culture depend on both the culture medium and passage number. The maintenance of the epithelial serous phenotype of BTG cells during the first passages (1 to 4) suggests that the study of the regulation of these gland cells should be limited to these early passages. bovine tracheal gland cells I difft~ntiation I culture medium I passage number
Introduction
Airway secretions play an important role in the defense of the respiratory tract. In airway epithelium, submucosal gland secretory cells are considered as the main source of secreted mucus [24]. In order to analyze the regulatory mechanisms of airway mucous secretion, different research groups have developed systems to culture bovine [3, 16] and human [33, 37] tracheal gland ceils. Recently we have described a method for the isolation and culture of bovine tracheal gland (BTG) cells [3] and reported that they modified their phenotype according to the culture medium and passage number [4]. After four passages we noticed that the BTG cells exhibited an epithelial-fibroblastoid transition. Morphological changes of a given cell type, from an epithelial to a fibroblastoid appearance, have been reported following addition of antibodies against cell adhesion molecules [2], transfecticm of cells with oncogenes [20], addition of EGF, transforming growth factor-~ or "epithelial scatter factor" to cell cultures [1, 9, 34], lowering of the extracellular Ca" + concentration [22], or plating the epithelial cells on certain substrata [35]. Recently, Boyer el al [7] reported that the culture medium may be responsible for the epithelialfibroblastoid transition of the rat bladder carcinoma-
derived cell line NBT II. Using human umbilical vein endothelial cells, Dichek et al [13] have shown variability in the level of expression of messenger RNA according to the passage number and time in culture. To d e f n e the factors that may modulate BTG cell morphological changes in culture, we have examined the effect of the culture medium and passage number on the expression of cytoskeletal proteins (cytokeratin, vimentin and desmoplakin) during the time of epithelial-fibroblastoid transition of these cells. Intracellular cyclic A M P (cAMP) regulates many critical differentiated functions of tracheal epithelial cells [15, 26]. To determine the modification of the relationship between the structure and functional activity of BTG cells in culture, we have investigated herein the effects of the culture medium and passage number on intracellular levels of cAMP in cultured BTG cells during steady state and after ~-adrenergic stimulation. The changes in functional activity of the BTG cells were also analyzed by reference to the secretion of lysozyme, which represents a good protein marker for secretory activity of tracheal gland serous cells in culture [36]. Our results demonstrate that the BTG epithelial cell organization and functional activity depend on both the culture medium chosen and their passage number.
50
R Benali et al
Materials and methods
Total protein were determined by the Bio-Rad protein microassay procedure [6] using bovine serum albumin as a standard. All other results are expressed as fmol per/~g total protein
Cell culture conditions Isolation and culture of BTG cells were performed according to the technique previously described [3]. Briefly, BTG cells were isolated by enzymatic digestion and cultured on human placental collagen [10] in the presence of: 1) a 1 : 1 medium of Dulbecco's Modified Eagle's/Ham's Fl2 (D/FI2) supplemented with 10070 fetal calf serum (FCS); 2) D/FI2 supplemented with 2070 Ultroser G (a serum substitute); 3) a serum-free defined medium (SFDM) containing D/FI2, 10 ng/ml epidermal growth factor (EGF), 1 pg/ml transferrin, 1 pg/ml insulin, 0.5 pg/ml hydrocortisone and 10 ng/ml retinoic acid. Cultures were serially passaged by trypsinization and cultured cells from passages 4 to 8 were examined. Cells were seeded at a density of 3 x 104 cells/cm 2.
lmmunofluorescence For indirect immunofluorescence staining of cytokeratin, vimentin and desmoplakin, B:FG cells, grown on glass coverslips coated with human placental collagen, were fixed for 30 min at - 2 0 ° C in methanol and washed in phosphate-buffered saline (PBS) pH 7.4. Then the biotin streptavidin fluorescein technique was applied. Mouse monoclonal antibodies to cytokeratins 19, 18 and 8 (PKK 1) were purchased from Labsystems and used at a 1 : 10 dilution. Mouse monoclonal antibody to vimentin was purchased from Amersham and used at a 1 : 50 dilution. Mouse monoclonal antibodies to desmoplakins I and II were purchased from Boehringer Mannheim and used at 1 : 10 dilution. Streptavidin-fluorescein and biotinylated goat anti-mouse IgG were obtained from Amersham and used at 1:25 dilution. All the dilutions were performed in a 1°70 bovine serum albumin (BSA)-PBS solution pH 7.4. Before immunostaining, cells were incubated in a 1070 BSAPBS solution for 10 min. Primary antibodies were applied for 1 h, followed by two 5-rain washes in PBS, a l-h incubation with biotinylated goat anti-mouse IgG, two 5-min washes in PBS, a 40-min incubation with streptavidin-fluorescein and two final washes of 5 min each in PBS. All the incubations were carried out at room temperature. The glass coverslips were mounted on the slide with antifading glycerol and the stained cells were examined under a Zeiss Axiophot fluorescence microscope.
Treatment of cells and cyclic AMP assay Cells from passage 4 and passage 8 were cultured in 12-well cluster plastic dishes and used when confluent. The monolayers were rinsed twice with D/FI2 containing HEPES (20 mM) and incubated for 10 min at 37°C in 900 ~1 of the D/FI2 containing HEPES (N-2-hydroxyethyl-piperazine-N'-2-ethanesulfonic acid, 20 mM) and IBMX (isobutylmethylxanthine, 10-SM), in order to study the basal levels of cAMP, and for 0, 5, 10, 15, 20 and 25 min in 900/~1 of the D/FI2 containing HEPES (20 mM), IBMX (10-5M) and isoproterenol (10-5M), in order to study the effects of fl-adrenergic stimulation on the cAMP accumulation. After the various incubation periods, cAMP production was stopped by adding 100 tzi of 10 N periodic acid. Cells were then scraped from the culture dishes and intracellular cAMP levels were determined by radioimmunoassay using the method of Cailla et ai [8]. All determinations were done in duplicate and the mean was calculated.
Lysozyme assay Culture media from confluent BTG cells were concentrated by pressure filtration over Amicon YM-2 ultrafilters (Amicon Corp, Lexington, MA) and dialyzed against deionized water. Lysozyme enzymatic activity was evaluated spectrophotometrically by measuring the initial rate of lysis of Micrococcus luteus suspension (Worthington Biochemical Corp) according to Shugar's method [32] modified by Jacquot et al [23]. The bovine lysozyme-type enzymatic activity was defined as units of enzyme activity per microgram of protein and expressed as equivalent of hen eggwhite lysozyme (Worthington).
Statistical analysis For cAMP assay, variance analysis was used to compare differences in the values according to the culture medium, passage number and incubation time. For lysozyme and total protein secreted by cultured BTG cells, the results were compared using the Student t-test. Values are expressed as means + SD of 3 determinations. P < 0.05 was considered as the limit of signification.
Results
Cell morphology under different culture conditions Bovine tracheal gland cells were cultured in a S F D M or in a medium supplemented either with 2% Ultroser G or with 10% FCS. Before passage 5, all the BTG cells cultured under these three culture conditions assumed an epithelial morphology, forming dense colonies o f tightly packed polyglonal cells (figs la, 2a) and showed immunoreactivities to cytokeratin filaments (figs lb, 2b) and desmoplakins I and II (figs lc, 2c). Vimentin filaments were not detected in BTG cells whatever the culture conditions (figs ld, 2d). However, at passage 5, about 50% of BTG cells cultured in 10% FCS or 2 0 Ultroser G became fibroblastoid: they flattened, exhibited a spindle shape, and crawled one over another (fig 3a) and showed immunoreactivities to both vimentin and cytokeratin f.]aments (fig 3b, c). At passage 8, 100% of BTG cells cultured under these two culture conditions were fibroblastoid (fig 2e) and expressed no immunoreactivities to cytokeratin filaments (fig 2f) and to desmoplakins I and II (fig 2g). They expressed immunoreactivity to vimentin filaments (fig 2h). In contrast, all the BTG cells cultured in a SFDM from passage 5 to passage 8 retained an epithelial morphology (fig le) and expressed immunoreactivities to cymkeratin filaments (fig 1f) and to desmoplakins I and Ii (fig lg) but not to vimentin (fig lh).
c A M P levels under different culture conditions The intracellular c A M P levels of the BTG cells in the basal state and after isoproterenol stimulation in the three cul-
Fig 1. Light micrographs of BTG cells cultured in a serum-free defined medium at passages 4 (a-d) and 8 (e-h). a, e. Phase contrast micrographs of confluent monolayers of cultured BTG cells; cells demonstrate an epithelial morphology of tightly packed polygonal cells. Bar -- 100 ~m. b, f. Immunofluorescence staining of cytokeratin filaments in cells as detected by monoclonal antibodies to cytokeratins 19, 18, and 8 (PKKI). All cells showed immunoreactivity to PKKI antibodies. Bar -- 25 pm. c, g. Immunofluorescence staining of desmoplakins along cells boundaries of cells as detected by monoclonal antibodies to desmoplakins I and Il. Bar = 20/~m. d, h. Immunofluorescence staining of vimentin filaments in cells. No immunoreactivity to vimentin was detected. Bar = 25/zm.
Phenotype of epithelial cells in culture
51
52
R Benali et al
.
•
''~----
--'--
-
~-7-
Phenotype of epithelial cells in culture
53
Fig 2, Light micrographs of BTG cells cultured in medium supplemented with either 2% Ultroser G or 10070FCS at passages 4 (a-d) and 8 (e-h). a, e. Phase contrast micrographs of cells. In a, cells demonstrate an epithelial morphology of tighly packed polygonal cells, in e, they exhibit a fibroblastoid morphology. Bars = 100/zm. b, f. lmmunofluorescence staining of cytokeratin filaments in cells. In b, all cells were positive with anti-cytokeratin PKKI antibody. In f, no immunoreactivity to cytokeratin was detected. Bar = 25/~m. e, g. Immunofluorescence staining of desmoplakins along cells boundaries in cells. Note the disappearance of immunoreactivity from cellular junctions and cell periphery at passage 8 kg). Bar = 25 pm. d, h. lmmunofluorescence staining of vimentin filaments in cells. In d, no immunoreactivity to vimentin was detected. In h, all cells were positive with anti-vimentin antibody. Bar =
25 ~ m .
ture conditions and at different passages were examined. Figure 4 shows the comparison of basal levels o f c A M P o f BTG cells grown in a SFDM or in D / F I 2 medium supplemented either with 2070 Ultroser G or with 10o70 FCS at passage 4 and passage 8. Variance analysis showed a significant variation of c A M P levels according to the culture medium (P < 0.01) and to the passage number (P < 0.001). At passage 4, intracellular level of cAMP grown in a SFDM was 130 fmol per pg protein, whereas it was only 90 fmol per pg protein when BTG cells were grown in D / F I 2 supplemented with 2o70 Ultroser G and 70 fmol per/~g protein when they were grown in D / F I 2 medium supplemented with 10o70 FCS. At passage 8, intracellular c A M P levels of BTG cells were reduced by 50°70 compared to that recovered at passage 4, whatever the culture medium. During the isoproterenol stimulation period (fig 5), the intracellular cAMP levels increased with incubation time with a maximum effect after a 15-rain period and varied significantly according to the culture medium (P < 0.001) and to the passage number (P < 0.001) as demonstrated by variance analysis. At passage 4 (fig 5A) the intracellular c A M P level in cells cultured in SFDM was 2.5 and 3.0-fold higher compared to that recovered from cells cultured in 2o7o Ultroser G and 10°70 FCS, respectively. At passage 8 (fig 5B), intracellular cAMP responses o f BTG cells were reduced by 50o70 (fig 5B) compared to that recovered at passage 4, whatever the culture medium, but the patterns of response were similar. L ysozyme secretion The amounts of iysozyme secreted by BTG cells main° tained in the three culture conditions and at different passages were determined. Figure 6 shows the comparison of lysozyme amounts secreted by BTG cells grown in SFDM or in D / F I 2 medium supplemented either with 2°/o Ultroser G or with 10o70 FCS at passage 4. BTG cells cultured in SFDM secreted 350 ng of lysozyme/106 cells/ 72 h incubation (expressed as equivalent of hen egg-white
Fig 3. Light micrographs of BTG cells cultured in medium supplemented either with 2% Ultroser G or with 10% FCS at passage 5. a. Phase constrast m~crograph of confluent monolayers of BTG cells. Some cells still demonstrate an epithelial morphology of tightly packed p¢!ygonal cells (sofid arrow] whereas other cells are fibroblastoid (open arrow). Bar = 100/~m. b. Double immunofluorescence staining of vimentin filaments (FITC) in fibroblastoid cells and cytokeratin filaments (rhodamin) in tightly polygonal celIs. Bar = 50/~m. e. Double immunofluorescence staining showing the presence of both vimentin (FITC) and cytokeratin (rhodamin) filaments in fibroblastoid cells. Bar = 10 pm.
54
R Benali et a!
cAMP
cAMP
10 "15 M/pg proteins 5000
10 "15 M/pg proteins 150
~000 3000
100
2000 1000
50
0 SFMO
US G
FOS
Fig 4. Basal levelsof cAMP of BTG cells in the presence of IBMX (10 -5 M) for 10 min at passages 4 (P4) and 8 (P8) on day 8 of culture. Varianceanalysisshowed a significantvariation of cAMP production according to the culture medium (P < 0.01) and the passage number (P < 0.001).
5
10
15
20
25
10
15
20
25
5000 4000
(~)
3000 2000
lysozyme). This lysozyme content was 7- to 10-fold higher and significantly different (P < 0.001) compared with that recovered from BTG cells cultured in 10% FCS or 207o Ultroser (3 and represents more than 5.0°7o of total proteins secreted in the culture medium (fig 7). After passage 4, the amount of lysozyme secreted by BTG cells cultured in a SFDM decreased significantly (P < 0.001) and represented only 0,7% of total proteins secreted in the culture medium (fig 7). However, no significant change in the amount of total proteins secreted by BTG cells in a SFDM from passage 4 to 8 was observed (fig 7). Discussion Bovine tracheal gland cells cultured in the presence of FCS or Ultroser G showed an epithelial-fibroblastoid transition after 4 passages. In this work we have examined the expression of cytoskeletal and desmosomai proteins in cultured BTG cells using selected monoclonal antibodies. It appears clear from our study that the phenotype of cultured BTG cells varies dramatically according to the culture medium used and passage number. We have observed after 4 passages that culture medium supplemented with 2% Uitroser G or with 10% FCS induced morphological changes resembling those observed in embryonic cells undergoing epitheliai-fibroblastoid transition [14, l?], During the first four passages, BTG cells form a cobblestone-like monolayer of polygonal cells and express immunoreactivities to cytokeratin intermediate filaments and desmoplakins, whereas after 4 passages, the cells dissociated, assumed a spindle-like shape and expressed immunoreactivity to both cytokeratin and vimentin. Around passage 8 after the epithelial-fibroblastoid transition, the desmoplakin distribution along cell boundaries was dramatically reduced and the expression of the cytokeratin intermediate filaments disappeared completely. No modu-
1000 0 0
5
Incubation time (rain) Fig 5. Effects of isoproterenol (10 -s M) in the presence of IBMX (10 -s M) on cAMP formation in BTG cells at passages 4 (A) and 8 (13)on day 8 of culture. Variance analysis showed a significant variation of cAMP production according to the incubation time (P < 0.001), the culture medium (P < 0.001) and the passage number (P < 0,001),
Lysozyme (ng/lO" cells172h) 400
300
200
100
O
SFOM
FCS
US G
Fig 6. Lysoz.vmesecretion by BTG cells cultured in a serum-free defined medium (SFDM), medium supplemented with 2% Ultroser G (US G) or 10% fetal calf serum (FCS). Values represent mean + SD of 3 determinations. (* P < 0.001).
Phenotype of epithelial cells in culture
Proteins (pg/lO s cells/72h)
Lysozyme/total proteins (%) 6
5
4 3 2 1
0
0 4
5
6 7 Passage number
8
Fig 7. Lysozyme secretion by BTG cells cultured in a serum-free defined medium expressed as percentage of total protein secreted in the medium ( • - - • ) and total protein secreted by BTG cells cultured in a serum-free defined medium ( D - - ~ ) at different passages. Values represent mean + SD of 3 determinations. (* P < 0,001).
lation of the epithelial phenotype of BTG cells was observed when these cells were cultured in a serum-free defined medium. Vimentin intermediate filaments were formed in the fibroblastoid BTG cells. It is possible, however, that the epitheliai-fibroblastoid transition and formation of vimentin filaments in these cells was induced by soluble factors contained in fetal calf serum or in the Ultroser G. A rapid induction of vimentin intermediate filaments formation has been observed often in primary cultures of epithelial cells [31] including rat hepatocytes [18] as well as human urothelial cells. The advent of vimentin is known to be influenced by various components such as externally added retinoid compounds as well as certain growth factors and hormones [25, 29, 30]. It has been therefore the general opinion that induction of vimentin is a result of culture conditions [II, 21, 27, 38]. Our results indicate that the bovine tracheal gland epithelial cell organization depends on both the culture medium and passage number, and the use of a serum-free defined medium seems to be essential to maintain the differentiated epithelial phenotype of these cells. Similar results indicating an influence of the culture medium on epithelial cell phenotype have been reported recently by Boyer et ai [7]. According to the last mentioned authors, the exposure of the rat bladder carcinoma cell line NBT-II to medium containing Ultroser G, shows an epithelial-fibroblastoid transition. Extensive alterations of desmosomal structures and proteins occurred during the epithelial-fibroblastoid transition of NBT-II cells. The pathway leading to desmosomal alterations include phosphorylation and protein synthesis. The mechanisms by
55
which Ultroser G acts to promote desmosomal changes is most probably a multistep cascade. Boyer et al [7] have suggested that solubilization of desmoplakins is a relatively early event in cell dissociation and in epithelialfibroblastoid transition. Benitez-King et al [5] have shown for MDCK cells that during the in vitro formation of an epithelial monolayer cytoskeleton configuration, cell shape and cell contacts modulate cytoskeletal protein posttranslational modifications and there is a correlation between the phosphorylation patterns of intermediate filaments proteins and cytoskeletal protein synthesis. Intracellular cyclic AMP (cAMP) is known to regulate many critical differentiated functions of tracheal epithelial cells [15, 26]. In the present study, we have shown that basal levels of cAMP in cultured BTG cells vary according to the culture medium and passage number. It is higher in BTG cells cultured in a SFDM compared to that recovered from cells cultured in medium supplemented with Ultroser G or FCS. Whatever the culture medium, BTG cells respond to stimulation by isoproterenol. However, the response to stimulation in a SFDM is higher than in Ultroser G or FCS supplemented medium. The difference in response to stimulation by isoproterenol may be due to a decrease of ~-adrenergic receptor concentration under these culture conditions or to the inhibitory factors conrained in Ultroser G or in FCS. Recently, it has been shown [12] that dexamethasone, which is present at a high concentration in Uitroser G, decreases ~-adrenergic receptor concentration and inhibits ~-adrenergic responses during explant culture of human fetal lung. Analysis of lysozyme biosynthesis and secretion by BTG cells also showed differences according to the culture medium used. Secretion of lysozyme by BTG cells was 7- to 10-fold higher in a SFDI~! .han in Ultroser G or FCS supplemented medium. This suggests that lysozyme synthesis and secretion depend on specific factors present in culture medium. The mechanism~ by which Ultroser G and FCS modify the protein synthesis remain to be elucidated. In the present study, we have also been able to demonstrate the effect of passage number on lysozyme secretion by BTG cells, since the amount of this protein decreased markedly after 4 passages. The modifications in the amount of lysozyme secreted by BTG cells after 4 passages may, therefore, result from regulation at the level of both RNA transcription and the intracellular degradation of protein. These data are consistent with recent studies on human umbilical vein endothelial cells [13], and on pig skin and corneal fibroblasts [19, 28], which showed modulations of gene expression during in vitro cells 'ageing'. In conclusion, we have shown that the bovine tracheal gland epithelial cell phenotype in culture depends on both the culture medium used and passage number. Early passages of cells cultured in a serum-free defined medium appear to be the optimal conditions to investigate the normal regulation of biosynthesis of secretory proteins produced by bovine tracheal glands cells.
Acknowledgments
The authors are grateful to Mrs J Hinnrasky for her technical support and to Mrs A Chaveriat for typing this manuscript. This work was supported by the contract INSERM/SYNTHELABO n ° 87118 and by Association Francaise de Lutte contre la Mucoviscidose (AFLM).
56
R Benali et at
References
1 Barrandon Y, Green H (1987) Cell migration is essential for sustained growth of keratinocyte colonies; the roles of tra~sforming growth factor-a and epidermal growth factor. C,.li 50, 1131-1137 2 Behrnes J, Birchmeier W, Goodman SL, Imhof BA (1985) Dissociation of Madin-Dabry canine epithelial cells by the monoclonal antibody anti Arc-I : mechanistic aspects and identification of the antigen as a component related to uvomurulin. J Cell Biol 101, 1307-1315 3 Benali R, Dupuit F, Jacquot J, Fuche¥ C, I-linnrasky J, Ploton D, Puchelle E (1989) Growth and characterization of isolated bovine tracheal gland cells in culture. Influence of a reconstituted basement membrane raat~-ix. Bb)i Cell 66, 263-270 4 Benali R, Dupuit F, Jacquot J, Haye B, Pucheile E (1990) Influence of the culture medium and passage numbers on the intracellular formation of cyclic AMP by cultured bovine tracheal gland serous cells. Am Rev Respir Dis (abstract) 141, A88 5 Benitez-King G, Cazares F, Meza I (1989) Synthesis and phosphorylation of cytoskeletal proteins during the in vitro biogenesis of MDCK cell monolayers. J Cell Sci 93, 53-61 6 Bio-Rad Protein Assay (1984) Bio-Rad technical bulletin n ° 1069, Bio-Rad Laboratories, Richmond, Ca, USA 7 Boyer B, Tucker GC, Vall~s AM, Franke WW, Thiery JP (1989) Rearrangements of desmosomal and cytoskeletal proteins during the transition from epithelial to fibroblastoid organization in cultured rat bladder carcinoma cells. J Cell Biol 109, 1495-1509 8 Caila HL, Racine-Weisbuch MS, Delaage MA (1973) Adenosine Y, 5' cyclic monophosphate assay at 10-msmol level. Anal Biochem 56, 394-407 9 Chinkers M, McKauna JA, Cohen S (1979) Rapid induction of morphological changes in human carcinoma cells A431 by epidermal growth factor. J Cell Bio183,260-265 10 Coleman DL, Tuet IK, Widdicombe JH (1984) Electrical properties of dog tracheal epithelial cellsgrown in monolayer culture. Am J Physioi 246, C355-359 1! Dairkee SH, Blayney CM, Asarnov DM, Smith SH, Hackett AJ (1985) Early expression of vimentin in human mammary cultures. In Vitro 21,321-327 12 Davis DJ, Jacobs MM, Ballard PL, Gonzales LK, Roberts JM (1990)~-Adrenergic receptors and cAMP response increase during explant culture of human fetal lung: partial inhibition by dexamethasone. Pediatr Res 28, 190-195 13 Dichek D, Quertermous T (1989) Variability in messenger RNA levels in human umbilical vein endothelial cells of different lineage and time in culture. In Vitro 25, 2889-292 14 Edelman GM (1984) Cell adhesion and morphogenesis: the regulatory hypothesis. Proc Natl Acad Sci USA 81, 1460-1464 15 Eling TE, Danilowicz RM, Henke DC, Sivarajak K, Yankaskas JR, Boucher RC (1986) Arachidonic acid metabolism by canine tracheal epithelial cells: product formation and relationships to chloride secretion. J Biol Chem 261, 1 2 8 4 - i 289 16 Finkbeiner WE, Nadei JA, Basbaum CB (1986) Establishment and characterization of a cell line derived from bovine tracheal glands. In Vitro 22, 561-567 17 Franke WW, Grund C, Jackson BW, lllmensee K (1983) Formation of cytoskeletal elements during mouse embryogenesis. IV. Ultrastructure of primary mesenchymal cells and their cell-cell interactions. D(fferentiation 25, 121-141 18 Franke WW, Mayer D, Schmid E, Denk H, Borenfreund E (1981) Differences of expression of cytoskeletal proteins in cultured rat hepatocytes and hepatoma cells. Exp Cell Res 134, 345-365 19 Fini ME, Girard MT (1990) The pattern of metalloproteinase
expression by corneal fibroblasts is altered by passage in culture. J Cell Sci 97, 373-383 20 Garcia I, Sordat B, Rauccio-Farinon E, Dunand M, Kraehenbuhl JP, Diggelman H (1986) Estabfishment of. two rabbit mammary epithefial cell fines with distinct oncogenic potential and differentiated phenotype after microinjection of transforming genes. Mol Cell Bioi 6, 1974-1982 21 Hall HG, Bissel MJ (1986) Characterization of the intermediate flament proteins of murine mammary gland epithelial cells. Exp Cell Res 162, 379-389 22 Hennings H, Mickael D, Cheng C, Steinert P, Holbrook K, Yuspa SH (1980) Calcium regulation of growth and differentiation of mouse epidermal cells in culture. Cell 19, 245-254 23 Jacquot J, Tournier JM, Puchelle E (1985) In vitro evidence that human airway lysozyme is cleaved and inactivated by Pseudomonas aeruginosa elastase and not by human leucocyte elastase. Infect lmmun 47, 55-560 24 Kaliner M, Schelhamer JH, Borson B, Nadel J, Patow C, Marom Z (1986) Human respiratory mucus. A m Rev Respir Dis 134, 612-621 25 Kim KH, Stellmach V, Javors J, Fuchs E (1987) Regulation of human mesothelial cell differentiation: opposing roles of retinoids and epidermal growth factor in the expression of intermediate filaments proteins. J Cell Bio1105, 3039-3051 26 Lazarus SC, Basbaum CB, Gold WM (1984) Localization of cAMP in dog and cat trachea: effects of ~-adrenergic agonist. A m J Physiol 247, C327-C334 27 Mackay AM, Tracy RP, Craighead JE (1990) Cytokeratin expression in rat mesothelial cells in vitro is controlled by the extracellular matrix. J Cell Sci 95, 97-107 28 Martin M, El Nabout R, Lafuma C, Crechet F, Remy J (1990) Fibronectin and collagen gene expression during in vitro ageing of pig skin fibroblasts. Exp CelIRes 191, 8-13 29 Rheinwald JG, O'Connell TM (1985) Intermediate filament protein as distinguishing markers of cell type and differentiated state in cultured human urinary tract epithelia. Ann N Y Acad Sci 455-259-267 30 Rheinwald JG, O'Connell TM, Conneli ND, Rybak SM, Allen-Hoffmann BL, La Rocca P J, Wu YJ, Rehwoldt SM (1984) Expression of specific keratin subsets and vimentin in normal human epithelial cells. A function of cell type and conditions of growth during serial culture. Cancer Cells (Cold Spring Harbor) I, 217-228 31 Schmid E, Schiller DL, Grund C, Stadler J, Franke WW (1983) Tissue type-specific expression of intermediate iliament proteins in a cultured epithelial cell line from bovine mammary gland. J Cell Biol 96, 37-50 32 Shugar D (1952) Measurement of lysozyme activity and ultraviolet inactivation of lysozyme. Biochem Biophys Acta 8, 302-309 33 Sommerhoff CP, Finkbeiner WE (1990) Human tracheobronchial submucosal gland cells in culture. A m J Respir Cell Mol Biol 2, 41-50 34 Stoker M, Gherardi E, Perryman M, Gray J (1987) Scatter factor is a fibroblast-derived modulator of epithelial cell mobility. Nature 327, 239-242 35 Tchao R (1982) Novel forms of epithelial cell motility on collagen and on glass surfaces. Cell Motil 4, 333-341 36 Tom-Moy M, Basbaum CB, Nadel JA (1983) Localization and release of lysozyme from ferret trachea: Effects of adrenergic and cholinergic drugs. Cell Tissue Res 228, 549-562 37 Tournier JM, Merten M, Meckler Y, Hinnrasky J, Fuchey C, Pucheile E (1990) Culture and characterization of human tracheal gland cells. A m Rev Respir Dis 141, , 1280-1288 38 Yang J, Balakrishnan A, Hamamoto S, Beattie CW, Gupta TK, Welling SR, Nandi S (1986) Different mitogenic and phenotypic responses of human breast epithelial cells grown in two versus three dimensions. Exp Cell Res 167, 563-569
