DEVELOPMENTAL
141,426-430
BIOLOGY
(1990)
Differential Expression of Fetomodulin and Tissue Plasminogen to Characterize Parietal Endoderm Differentiation of F9 Embryonal Carcinoma Cells SUMI IMADA, HARUMI Division
of
Cell Biology,
Meiji Institute of Health Science,
YAMAGUCHI, Meiji
Accepted
Milk June
Activator
AND MASARU IMADA
Products
Company
Inc., 540
Naruda, Odawara, Japan 250
19, 1990
Fetomodulin is a surface marker protein of differentiated F9 embryonal carcinoma cells. Gene cloning has recently identified it as thrombomodulin which binds thrombin and proteolytically activates protein C. Activity assays and RNA blotting were adopted to analyze F9 cell differentiation with specific reference to another well-characterized marker, tissue plasminogen activator. Retinoic acid induced primitive endoderm differentiation of F9 cells and simultaneously activated tissue plasminogen activator synthesis. This differentiation, however, did not result in fetomodulin expression. When primitive endoderm cells were exposed to 1 mMdibutyry1 cyclic AMP, the tissue plasminogen activator level rose further within 6 hr. In contrast, the cofactor activity of fetomodulin stayed below a detectable level for as long as 15 hr and then increased with time. Expression of the two marker proteins appeared to be regulated differently. o 1990 Academic
Press, Inc.
bin. Second, it results in some lOOO-fold potentiation of thrombin to activate protein C by proteolytic digestion. Activated protein C is responsible for degrading essential coagulating factors, Va and VIIIa. Thrombomodulin, thus, plays a key role in maintaining blood homeostasis (reviewed by Esmon, 1989). Although this surface antigen is expressed in a variety of cells besides endothelium at late stages of embryonic development, whether it functions as a thrombin receptor or is endowed with yet unknown functions during development remains to be clarified. This point was previously discussed (Imada et al., 1990). Purified fetomodulin binds thrombin and acts as a cofactor in protein C activation, which functionally identifies fetomodulin as thrombomodulin (hereafter referred to as FM/TM). In the present paper we used specific antibodies to demonstrate that increased cofactor activity of differentiated F9 cells was exclusively due to induction of FM/TM. We then compared the appearance of FM/TM and tissue plasminogen activator at protein and RNA levels in order to distinguish gene regulation of the two differentiation markers.
INTRODUCTION
Mechanisms of differentiation became amenable to direct biochemical analyses in part by development of methods to manipulate cell fate in vitro and to identify progeny cells with specific differentiation markers. For example, the F9 embryonal carcinoma cell system has offered an opportunity to study cell differentiation in early embryogenesis. Stable, undifferentiated F9 stem cell population can be converted into primitive endoderm by treatment with retinoic acid (FS/RA; Strickland and Mahdavi, 1978). Primitive endoderm can be further induced to differentiate into parietal endoderm by treatment with N6,02’-dibutyryl adenosine-3’:5’-cyclic monophosphoric acid (dbcAMP) (FS/RA:cAMP; Strickland et al., 1980). We identified a differentiation marker protein, fetomodulin, which appeared when F9/ RA cells were exposed to reagents such as dbcAMP, cholera toxin, or a phosphodiesterase inhibitor that could elevate cytoplasmic cyclic AMP concentrations. Localization of fetomodulin in parietal endoderm was also corroborated in vivo by immunohistochemieal staining of embryos (Imada et aZ., 198’7). Fetomodulin was recently identified as thrombomodulin by cloning and sequencing its complementary and genomic DNA (Imada et al, 1990). Thrombomodulin is a surface thrombin receptor commonly localized in vascular endothelium (Esmon and Owen, 1981). Thrombin-thrombomodulin interaction can markedly modulate substrate specificity of thrombin in a dual fashion. First, it inactivates fibrinogenolytic activity of throm0012-1606/90 $3.00 Copyright All rights
0 1990 by Academic Press. Inc. of reproduction in any form reserved.
MATERIALS
AND METHODS
Cell Culture and in Vitro ~krentiatim F9 embryonal carcinoma were maintained in Ham’s serum. For activity assays, well culture dishes (Coster) 426
cells (Bernstine et ak, 1973) F12 with 10% fetal bovine F9 cells were plated on 24at 500 cells/cm2 on Day 0.
IMADA,
YAMAGUCHI,
AND II&ADA
FS/RA cultures received 1 PM all-trans-retinoic acid (RA) from Day 1 to Day 4 and were rinsed and incubated for 2 hr in RA-free medium. Cells were grown for 4 more days after medium refreshment. FS/RA:cAMP cultures received 1 mMdbcAMP at various times after removing RA. Assays were performed on Day 8. RNA was prepared from similarly treated cultures grown on two 150mm culture dishes. FM/TM was purified from cultures seeded at 2000 cells/cm’ and was grown for 5 days with 1 pMRA and 1 mMdbcAMP. In the present study, we only tested dbcAMP because addition of cholera toxin, a phosphodiesterase inhibitor, or dbcAMP similarly induced FM/TM expression (Imada et ab, 1987). Antibodies
Monoclonal antibodies, Rl-242 and Rl-1622, and a polyclonal rabbit antibody were described (Imada et aL, 1987,199O). A hybridoma-secreting IgG2a subclass antibody, R2-20111, was prepared by fusing NS-1 myeloma cells with spleen cells of a rat immunized with affinitypurified FM/TM (S. Imada et al., unpublished results). Immunoglobulins were purified by ammonium sulfate precipitation followed by DEAE Affigel-Blue chromatography (Bio-Rad). Radiolabeling and Immunoprecipitation
Cells were lysed in 50 mM Tris-HCl (pH 7.4) with 0.5% Triton X-100, 150 mM NaCl, and 1 mM phenylmethylsulfonyl fluoride. Lysates were then cleared by centrifugation at 5000 rpm for 5 min. r5S]Methioninelabeled proteins were immunoprecipitated by incubating 50 ~1 of the lysate for 1 hr at 37°C with 1 ~1 of 20 mg/ml of immunoglobulin and 50 ~1 of 50 mM Tris-HCl (pH 7.4) containing 10 mg/ml bovine serum albumin, 0.5% Triton X-100, 1% sodium deoxycholate, 0.5% sodium dodecyl sulfate (SDS), and 1 mM phenylmethylsulfonyl fluoride. Goat anti-rat (Fab’), (10 ~1; Cappel) was used as a second antibody. IgG Sorb (The Enzyme Center) was washed three times with 0.5 ml of 10 mM Tris-HCl (pH 7.2) containing 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, and 0.25% SDS and proteins were extracted in SDS sample buffer. Other methods were previously described (Imada and Imada, 1982; Imada et aL, 1987, 1990). Cofactor Activity
Parietal Endodm
L@?k-rentiation
(Hoechst, Japan) to the cell-free supernatant, it was incubated with 50 nmol of substrate, Boc-Leu-Ser-ThrArg-AMC (Peptide Institute), for 10 min at 37°C. Released 7-amino-4-methyl-coumarine (AMC) was fluorometrically determined. Unless otherwise specified, activity was represented by nanomoles of AMC released per 10 min per culture. tPA and Protein Assay
Cell monolayers after the cofactor assay were lysed in 100 ~1 of 50 mM Tris-HCl (pH 7.4) with 0.5% Triton X-100 and 150 mM NaCl, and the cleared supernatant was subjected to protein and tPA assays. Plasminogendependent fibrinolysis on fibrin-agar plates was used according to Astrup and Mullertz (1952). Activity was calibrated against standard human tPA samples obtained from the Division of Pharmaceutical R&D of this institute. The activity was represented by WHO international units per culture or per milligrams of protein. Protein contents were determined with BCA protein assay reagent (Pierce Chemicals) using bovine serum albumin (Sigma) as a reference. RNA Preparation and Slot Hybridization
Total RNA was isolated according to Maniatis et al. (1982). Samples containing 90 pg of RNA in water were incubated with 3 vol of 6.15 Mformaldehyde in 10X SSC for 15 min at 65°C. They were applied to a slot blotter (Minifold II, Schleicher & Schuell) and hybridized with nick-translated probes for 15 hr at 42°C in a solution containing 50% (v/v) formamide, 5~ SSC, 25 mM sodium phosphate (pH 6.5), 2~ Denhardt’s solution, and 500 pg/ml of carrier calf thymus DNA. After washing once with 2~ SSC containing 0.1% SDS for 30 min at 65°C and twice with O.lX‘SSC containing 0.1% SDS for 30 min at 42”C, filters were contacted to Kodak X-ray films (XAR-5) with intensifying screens (DuPont; Cronex Lightening Plus) at -80°C. Mouse actin probe, p91 (Minty et aL, 1981), and mouse tPA probe, pTAM (Rickles et ah, 1988), were gifts from Drs. Buckingham and Strickland, respectively. A 0.4-kb fragment of FM/ TM cDNA generated by double restriction digestion of FM21 with EcoRI and PstI (Imada et al., 1990) was recloned in pUC19 and used as an FM/TM probe. RESULTS
Assay of Protein C Activation
An assay using intact cell cultures was performed according to Owen and Esmon (1981). Briefly, cell monolayers were incubated for 1 hr at 37°C with 0.5 units of bovine thrombin (Mochida Seiyaku) and 9 pg of human protein C (American Diagnostica Inc.) in the presence of 3.5 mM CaCl,. After adding 90 pg of anti-thrombin III
427
Expression of FM/TM
Messenger RNA
Northern blot analysis of FM/TM messenger RNA gave rise to a single band of 3.7 kb in differentiated F9 cells (Imada et aZ.,1990). Steady-state levels of FM/TM and tPA messages were then examined by hybridizing total RNA with specific probes for FM/TM, tPA, or ac-
428
DEVELOPMENTAL BIOLOGY
VOLUME 141.1990
C
b
TABLE 2 THROMBIN-DEPENDENT ACTIVATION OF PROTEIN C IN MONOLAYER CULTURES Cells None F9 FS/RA FS/RA:cAMP
FIG. 1. Slot blot hybridization. Lane a, hybridized with FM/TM probe. Lane b, hybridized with tPA probe. Lane c, hybridized with actin probe. 1, F9 cells; 2, FS/RA cells; 3, FS/RA:2d-CAMP cells; 4, FS/RA:4d-CAMP cells; 5, FS/RA:5d-CAMP cells. Cells in samples three to five were treated with dbcAMP for 2,4, and 5 days until Day 9.
tin (Fig. 1; Table 1). The FM/TM message stayed at the same basal level in F9 and FS/RA cells and rose after dbcAMP treatment. On the other hand, the tPA message was detectably increased in FS/RA cells and it was further elevated in FS/RA:cAMP cells as reported by Rickles et ah (1988). It appeared that the increment of the tPA message in FS/RA:2d-CAMP cells was somewhat greater than that of the FM/TM message. The results raised a possibility that tPA and FM/TM genes were independently regulated. However, since quantitation by hybridization is laborious and insensitive to small differences, we explored activity assays to follow the time course of their expression. Cofactor Activity
of F9 Cell Cultures
Protein C activation was measured with F9 monolayer cultures before and after induction of differentiation. It is known that thrombin alone can activate protein C, although much less efficiently, in the absence of a TABLE 1 DENSITOMETRY OF RNA SLOT BLOT ANALYSIS Cells F9 FS/RA FS/RA:2d-CAMP* FS/RA:4d-CAMP* FS/RA:5d-CAMP”
FM/TM 0.08 0.08 0.20 0.72 0.78
tPA
Actin
FM/TM a actin
tPA” actin
0.09 0.27 0.93 1.93 1.98
1.12 1.19 0.93 1.23 1.11
0.07 0.07 0.22 0.59 0.70
0.08 0.23 1.00 1.61 1.78
Note. The autoradiogram shown in Fig. 1 was traced by a densitometer. a Relative enrichment of FM/TM and tPA messages was expressed by taking ratios to actin message. * FS/RA cells were treated with dbcAMP for 2,4, and 5 days until Day 9.
Thrombin + + + + -
AMC released” 3.4c nd” 1.6 nd 1.0 nd 21.3 nd
tPA activity* -d nd nd 1.1 2.1 15 21
Note. Protein C activation was examined in the presence or absence of thrombin with or without cultured cells. ’ The amount of AMC released in 10 min incubation is indicated in nanomoles. This should correlate to levels of activated protein C. * Presented in WHO international units. ’ Protein C can be activated by thrombin alone. The amount of AMC released in the absence of cells was considered the background level. d Not done. ’ Not detectable.
cofactor. Therefore, the amount of AMC released by thrombin in the absence of cells is the background level. As shown in Table 2, AMC release was not observed in the absence of thrombin, indicating that cells are devoid of nonspecific proteases that can activate protein C. Note that a significant increase in activity was associated with FS/RA:cAMP cells and that F9 and FS/RA cells had little or no activity. As for tPA, however, F9/ RA cells were positive and this was further augmented in FS/RA:cAMP cells, consistent with the results of hybridization and the report of Strickland et al. (1980).
Eflect of Antibodies Added to Monolage~ Cultures
Antibodies that could inhibit thrombin-FM/TM interaction should quantitatively block AMC release if it was solely due to FM/TM. Before such blocking assays, antibodies were tested for their specificity by immunoprecipitating radiolabeled cell surface proteins and’total cellular proteins. The four antibodies, polyclonal rabbit antibody, R2-20111, Rl-242, and Rl-1622, reacted with a single protein that comigrated with affinitypurified FM/TM. Results are presented in Fig. 2 and in a previous report (Imada et aL, 198’?). A polyclonal rabbit antibody and a monoclonal antibody, R2-20111, inhibited FM/TM binding with thrombin while Rl-1622 and Rl-242 did not (Imada et al, 1990; unpublished results). When polyclonal antibody or R2-20111 was added to cultures, activity was almost quantitatively eliminated while normal rabbit immunoglobulin and Rl-1622 were virtually ineffective (Table 3). We concluded that FM/
&ADA,
A
1
2
3
YAMAGUCHI,
AND IMADA
Parietal Endodem
429
B@rentiation
4
205, 60 116, 97* 40 66r 66~
20 45t
45,
24
FIG. 2. Immunonrecipitation with anti-FM/TM antibodies. Rabbit sera or ascites fluid were used to immunoprecipitate radioiodinated surface proteins (A). Metabolically labeled proteins (B) were immunoprecipitated with purified immunoglobulins as described under Materials and Methods. Positions of size marker proteins are shown by arrowheads in the left of figures and molecular sizes are indicated in kilodaltons. Immunoaffinity-purified FM/TM gave rise to a single band with an apparent molecular size of 116 kDa. In each figure, samples in lane 1 to lane 4 were immunoprecipitated with polyclonal rabbit anti FM/TM serum, normal rabbit serum, R2-20111, and normal rat immunoglobulin, respectively.
TM was exclusively responsible for the increase of AMC release in differentiated F9 cells. Time Course of Appearance and tPA Activities
of Cofactor
One millimolar dbcAMP was added at various times after removing RA from FS/RA cultures and cofactor and tPA activities were simultaneously assayed on Day 8 (or 96 hr in Fig. 3). A time course of marker protein induction was thus reconstructed. FS/RA cells had little or no cofactor activity and they remained inactive for up
TABLE INHIBITION
OF ACTIVITY
3
IN CULTURE
BY ANTIBODIES
Immunoglobulin added
Cofactor activity (nmol AMC/mg protein)
None Rabbit anti-FM/TM Normal rabbit R2-20111 Rl-1622
64.6 2.7 53.9 6.3 65.3
Inhibition (%I 0 98 17 92 0
Note. FS/RA:cAMP cells were preincubated for 30 min at 37°C with or without 100 pg of immunoglobulin in 200 ~1 of Hank’s balanced salt solution. The cofactor assay was then followed in the presence of immunoglobulins.
40
72
96
time
FIG. 3. Time course of FM/TM and tPA expression in parietal endoderm differentiation. Dibutyryl cyclic AMP was added at various times after RA treatment and cofactor and tPA activities were determined 96 hr afterward with duplicate or triplicate samples. Time of dbcAMP treatment is shown in abscissa in hours. 0, cofactor activity shown in left ordinate in nanomoles AMC released per milligram protein. l , tPA activity shown in right ordinate in international units per milligram protein.
to 15 hr of dbcAMP treatment. However, after this apparent lag period, the cofactor activity sharply increased with time until it leveled off in 3 days. In contrast, tPA was already positive in FS/RA cells and it was also increased, although much earlier than the cofactor, after a reproducible depression in 3-hr treated cultures. We interpreted the results to indicate that a lag of about 15 hr was required before FS/RA:cAMP cells became positive for FM./TM and that temporal differences existed in regulation of FM/TM and tPA expression. DISCUSSION
Since FM/TM is identical with thrombomodulin (Imada et aL, 1990), we exploited its functional assay as a measurement of FM/TM in culture. We concluded from blocking experiments with antibodies that differentiated F9 cells expressed cell surface FM/TM in an active form and that protein C activation assay was specific for FM/TM. FM/TM in FS/RA cells was below the level of detection by cofactor activity assay and RNA blotting, consistent with results of immunohistochemical staining and protein analysis (Imada et aZ., 1987). This may distinguish FM/TM from other parietal endoderm markers, including tPA, that are positive in FS/RA cells. FM/TM also differed from tPA in kinetics of expression during parietal endoderm differentiation. FM/ TM developed after a lag of 15 hr while tPA increased
430
DEVELOPMENTAL BIOLOGY
with a much shorter delay following exposure to dbcAMP. Therefore, we hypothesize that FM/TM and tPA are regulated differently, although more sensitive assays may be required to unequivocally prove this point. It must also be pointed out that we measured FM/ TM and tPA localized on cell surface and in cytoplasm, respectively. However, a temporal difference of their appearance seems to be too large to be accounted for only by a translocation phenomenon. In a similar experiment, Rickles et al. (1988) have reported that the tPA message stayed at a basal level for at least 12 hr while we found tPA activity to rise within 6 hr. This difference may be partly due to the fact that we induced parietal endoderm differentiation of FS/RA cells in the absence of RA while they treated FS/RA cells with RA and dbcAMP. RA apparently attenuated functional differentiation induced by dbcAMP. A negative cooperative effect of RA and cyclic AMP was suggested by down-regulation of the retinoic acid receptor messenger by cyclic AMP in the presence of RA (Hu and Gudas, 1990). Parietal endoderm cell lineage is uniquely suited to analyze mechanisms of differentiation. F9 cell differentiation in substratum-attached cultures mimics the in vivo situation where primitive endoderm converts into parietal endoderm. Visceral endoderm obtained in aggregate F9 cultures (Hogan et al, 1981) can transdifferentiate into parietal endoderm (Hogan and Tilly, 1981; Grover and Adamson, 1986; Casanova and Grabel, 1988) and similar transdifferentiation was observed in vivo when visceral endoderm was injected into blastocyst (Gardner, 1982). Differentiation from visceral to parieta1 endoderm may be induced by attachment to adherent substratum such as trophoblastic giant cells or matrix components. FM/TM may serve as a new tool to characterize natural inducers of differentiation and to dissect molecular events of parietal endoderm differentiation. The authors thank Drs. Buckingham and Strickland for their generous gifts of plasmids containing actin and tPA genes, respectively. The technical assistance of Ms. M. Uyeno, Y. Mizutani, and T. Akizawa is deeply appreciated. REFERENCES ASTRUP, T., and MULLERTZ, S. (1952). The fibrin plate method for estimating fibrinolytic activity. Arch. Biochem Biophys 40,346-351. BERNSTINE, E. G., HOOPER, M. L., GRANDCHAMP, S., and EPHRUSSI, B.
VOLUME 141,199O
(1973). Alkaline phosphatase activity in mouse teratoma. Proc. Natl. Ad. Sci. USA 70.3899-3903. CASANOVA, J. E., and GRABEL, L. B. (1988). The role of cell interaction in the differentiation of teratocarcinoma derived parietal and visceral endoderm. Dev. Biol 129,124-139. ESMON, C. T., and OWEN, W. G. (1981). Identification of an endothelial cell cofactor for thrombin-catalyzed activation of protein C. Proc. Natl Acad Ski. USA 78.2249-2252. ESMON, C. T. (1989). The role of protein C and thrombomodulin in the regulation of blood coagulation. J. Biol Chem. 264,4743-4746. GARDNER, R. L. (1982). Investigation of cell lineage and differentiation in the extraembryonic endoderm of the mouse embryo. J. Emlnyol. Exp. Mwphol. 68,175-198. GROVER, A., and ADAMSON, E. D. (1986). Evidence for the existence of an early common biochemical pathway in the differentiation of F9 cells into visceral or perietal endoderm: Modulation by cyclic AMP. Deu. Biol. 114,492-503. HOGAN, B. L. M., and TILLY, R. (1981). Cell interactions and endoderm differentiation in cultured mouse embryos. J. Embryol Exp. Morphol. 62,379-394. HOGAN, B. L. M., TAYLOR, A., and ADAMSON, E. (1981). Cell interactions modulate embryonal carcinoma cell differentiation into parieta1 or visceral endoderm. Nature (London) 291,235-237. Hu, L., and GUDAS, L. J. (1990). Cyclic AMP analogs and retinoic acid influence the expression of retinoic acid receptor 01,p and y mRNAs in F9 teratocarcinoma cells. Mol. CeU. Biol. 10,391-396. &ADA, S., and IMADA, M. (1982). Increase of a surface glycoprotein by cyclic AMP in Chinese hamster ovary cells: Dependence on cell-cell interaction. J. Biol. Chem 257,9108-9113. &ADA, M., IYADA, S., IWASAKI, H., KUME, A., YAMAGUCHI, H., and MOORE, E. E. (1987) Fetomodulin: Marker surface protein of fetal development which is modulatable by cyclic AMP. Dev. Biol. 122, 483-491. IMADA, S., YAMAGUCHI, H., NAGUMO, M., KATAYANAGI, S., IWASAKI, H., and IMADA M. (1990). Identification of fetomodulin, a surface marker protein of fetal development, as thrombomodulin by gene cloning and functional assays. Den Biol. 140,113-122. MANIATIS, T., FRITSCH, E. F., and SAMBROOK, J. (1982). “Molecular Cloning, A Laboratory Manual.” Cold Spring Harbor Laboratory, New York. MINTY, A. J., CARAVA’ITI, M., ROBERT, B., COHEN, A., DAUBAS, P., WEYDERT, A., GROS, F., and BUCKINGHAM, M. E. (1981). Mouse actin messenger RNAs: Construction and characterization of a recombinant plasmid molecule containing a complementary DNA transcript of mouse ol-actin mRNA. J. Biol Chem. 256,1008-1014. OWEN, W. G., and ESMON, C. T. (1981). Functional properties of an endothelial cell cofactor for thrombin-catalyzed activation of protein C. J. Biol. Ckem 256,5532-5535. RICKLES, R. J., DARROW, A. L., and STRICKLAND, S. (1988). Molecular cloning of complementary DNA to mouse tissue plasminogen activator mRNA and its expression during F9 teratocarcinoma cell differentiation. J. Biol. Chem. 263,1563-1569. STRICKLAND, S., and MAHDAVI, V. (1978). The induction of differentiation in teratocarcinoma stem cells by retinoic acid. CeU 15,393-403. STRICKLAND, S., SMITH, K. K., and MAROTM, K. R. (1980). Hormonal induction of differentiation in teratocarcinoma stem cells: Generation of parietal endoderm by retinoic acid and dibutyryl CAMP. CeU 21,347-355.
141,426-430
BIOLOGY
(1990)
Differential Expression of Fetomodulin and Tissue Plasminogen to Characterize Parietal Endoderm Differentiation of F9 Embryonal Carcinoma Cells SUMI IMADA, HARUMI Division
of
Cell Biology,
Meiji Institute of Health Science,
YAMAGUCHI, Meiji
Accepted
Milk June
Activator
AND MASARU IMADA
Products
Company
Inc., 540
Naruda, Odawara, Japan 250
19, 1990
Fetomodulin is a surface marker protein of differentiated F9 embryonal carcinoma cells. Gene cloning has recently identified it as thrombomodulin which binds thrombin and proteolytically activates protein C. Activity assays and RNA blotting were adopted to analyze F9 cell differentiation with specific reference to another well-characterized marker, tissue plasminogen activator. Retinoic acid induced primitive endoderm differentiation of F9 cells and simultaneously activated tissue plasminogen activator synthesis. This differentiation, however, did not result in fetomodulin expression. When primitive endoderm cells were exposed to 1 mMdibutyry1 cyclic AMP, the tissue plasminogen activator level rose further within 6 hr. In contrast, the cofactor activity of fetomodulin stayed below a detectable level for as long as 15 hr and then increased with time. Expression of the two marker proteins appeared to be regulated differently. o 1990 Academic
Press, Inc.
bin. Second, it results in some lOOO-fold potentiation of thrombin to activate protein C by proteolytic digestion. Activated protein C is responsible for degrading essential coagulating factors, Va and VIIIa. Thrombomodulin, thus, plays a key role in maintaining blood homeostasis (reviewed by Esmon, 1989). Although this surface antigen is expressed in a variety of cells besides endothelium at late stages of embryonic development, whether it functions as a thrombin receptor or is endowed with yet unknown functions during development remains to be clarified. This point was previously discussed (Imada et al., 1990). Purified fetomodulin binds thrombin and acts as a cofactor in protein C activation, which functionally identifies fetomodulin as thrombomodulin (hereafter referred to as FM/TM). In the present paper we used specific antibodies to demonstrate that increased cofactor activity of differentiated F9 cells was exclusively due to induction of FM/TM. We then compared the appearance of FM/TM and tissue plasminogen activator at protein and RNA levels in order to distinguish gene regulation of the two differentiation markers.
INTRODUCTION
Mechanisms of differentiation became amenable to direct biochemical analyses in part by development of methods to manipulate cell fate in vitro and to identify progeny cells with specific differentiation markers. For example, the F9 embryonal carcinoma cell system has offered an opportunity to study cell differentiation in early embryogenesis. Stable, undifferentiated F9 stem cell population can be converted into primitive endoderm by treatment with retinoic acid (FS/RA; Strickland and Mahdavi, 1978). Primitive endoderm can be further induced to differentiate into parietal endoderm by treatment with N6,02’-dibutyryl adenosine-3’:5’-cyclic monophosphoric acid (dbcAMP) (FS/RA:cAMP; Strickland et al., 1980). We identified a differentiation marker protein, fetomodulin, which appeared when F9/ RA cells were exposed to reagents such as dbcAMP, cholera toxin, or a phosphodiesterase inhibitor that could elevate cytoplasmic cyclic AMP concentrations. Localization of fetomodulin in parietal endoderm was also corroborated in vivo by immunohistochemieal staining of embryos (Imada et aZ., 198’7). Fetomodulin was recently identified as thrombomodulin by cloning and sequencing its complementary and genomic DNA (Imada et al, 1990). Thrombomodulin is a surface thrombin receptor commonly localized in vascular endothelium (Esmon and Owen, 1981). Thrombin-thrombomodulin interaction can markedly modulate substrate specificity of thrombin in a dual fashion. First, it inactivates fibrinogenolytic activity of throm0012-1606/90 $3.00 Copyright All rights
0 1990 by Academic Press. Inc. of reproduction in any form reserved.
MATERIALS
AND METHODS
Cell Culture and in Vitro ~krentiatim F9 embryonal carcinoma were maintained in Ham’s serum. For activity assays, well culture dishes (Coster) 426
cells (Bernstine et ak, 1973) F12 with 10% fetal bovine F9 cells were plated on 24at 500 cells/cm2 on Day 0.
IMADA,
YAMAGUCHI,
AND II&ADA
FS/RA cultures received 1 PM all-trans-retinoic acid (RA) from Day 1 to Day 4 and were rinsed and incubated for 2 hr in RA-free medium. Cells were grown for 4 more days after medium refreshment. FS/RA:cAMP cultures received 1 mMdbcAMP at various times after removing RA. Assays were performed on Day 8. RNA was prepared from similarly treated cultures grown on two 150mm culture dishes. FM/TM was purified from cultures seeded at 2000 cells/cm’ and was grown for 5 days with 1 pMRA and 1 mMdbcAMP. In the present study, we only tested dbcAMP because addition of cholera toxin, a phosphodiesterase inhibitor, or dbcAMP similarly induced FM/TM expression (Imada et ab, 1987). Antibodies
Monoclonal antibodies, Rl-242 and Rl-1622, and a polyclonal rabbit antibody were described (Imada et aL, 1987,199O). A hybridoma-secreting IgG2a subclass antibody, R2-20111, was prepared by fusing NS-1 myeloma cells with spleen cells of a rat immunized with affinitypurified FM/TM (S. Imada et al., unpublished results). Immunoglobulins were purified by ammonium sulfate precipitation followed by DEAE Affigel-Blue chromatography (Bio-Rad). Radiolabeling and Immunoprecipitation
Cells were lysed in 50 mM Tris-HCl (pH 7.4) with 0.5% Triton X-100, 150 mM NaCl, and 1 mM phenylmethylsulfonyl fluoride. Lysates were then cleared by centrifugation at 5000 rpm for 5 min. r5S]Methioninelabeled proteins were immunoprecipitated by incubating 50 ~1 of the lysate for 1 hr at 37°C with 1 ~1 of 20 mg/ml of immunoglobulin and 50 ~1 of 50 mM Tris-HCl (pH 7.4) containing 10 mg/ml bovine serum albumin, 0.5% Triton X-100, 1% sodium deoxycholate, 0.5% sodium dodecyl sulfate (SDS), and 1 mM phenylmethylsulfonyl fluoride. Goat anti-rat (Fab’), (10 ~1; Cappel) was used as a second antibody. IgG Sorb (The Enzyme Center) was washed three times with 0.5 ml of 10 mM Tris-HCl (pH 7.2) containing 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, and 0.25% SDS and proteins were extracted in SDS sample buffer. Other methods were previously described (Imada and Imada, 1982; Imada et aL, 1987, 1990). Cofactor Activity
Parietal Endodm
L@?k-rentiation
(Hoechst, Japan) to the cell-free supernatant, it was incubated with 50 nmol of substrate, Boc-Leu-Ser-ThrArg-AMC (Peptide Institute), for 10 min at 37°C. Released 7-amino-4-methyl-coumarine (AMC) was fluorometrically determined. Unless otherwise specified, activity was represented by nanomoles of AMC released per 10 min per culture. tPA and Protein Assay
Cell monolayers after the cofactor assay were lysed in 100 ~1 of 50 mM Tris-HCl (pH 7.4) with 0.5% Triton X-100 and 150 mM NaCl, and the cleared supernatant was subjected to protein and tPA assays. Plasminogendependent fibrinolysis on fibrin-agar plates was used according to Astrup and Mullertz (1952). Activity was calibrated against standard human tPA samples obtained from the Division of Pharmaceutical R&D of this institute. The activity was represented by WHO international units per culture or per milligrams of protein. Protein contents were determined with BCA protein assay reagent (Pierce Chemicals) using bovine serum albumin (Sigma) as a reference. RNA Preparation and Slot Hybridization
Total RNA was isolated according to Maniatis et al. (1982). Samples containing 90 pg of RNA in water were incubated with 3 vol of 6.15 Mformaldehyde in 10X SSC for 15 min at 65°C. They were applied to a slot blotter (Minifold II, Schleicher & Schuell) and hybridized with nick-translated probes for 15 hr at 42°C in a solution containing 50% (v/v) formamide, 5~ SSC, 25 mM sodium phosphate (pH 6.5), 2~ Denhardt’s solution, and 500 pg/ml of carrier calf thymus DNA. After washing once with 2~ SSC containing 0.1% SDS for 30 min at 65°C and twice with O.lX‘SSC containing 0.1% SDS for 30 min at 42”C, filters were contacted to Kodak X-ray films (XAR-5) with intensifying screens (DuPont; Cronex Lightening Plus) at -80°C. Mouse actin probe, p91 (Minty et aL, 1981), and mouse tPA probe, pTAM (Rickles et ah, 1988), were gifts from Drs. Buckingham and Strickland, respectively. A 0.4-kb fragment of FM/ TM cDNA generated by double restriction digestion of FM21 with EcoRI and PstI (Imada et al., 1990) was recloned in pUC19 and used as an FM/TM probe. RESULTS
Assay of Protein C Activation
An assay using intact cell cultures was performed according to Owen and Esmon (1981). Briefly, cell monolayers were incubated for 1 hr at 37°C with 0.5 units of bovine thrombin (Mochida Seiyaku) and 9 pg of human protein C (American Diagnostica Inc.) in the presence of 3.5 mM CaCl,. After adding 90 pg of anti-thrombin III
427
Expression of FM/TM
Messenger RNA
Northern blot analysis of FM/TM messenger RNA gave rise to a single band of 3.7 kb in differentiated F9 cells (Imada et aZ.,1990). Steady-state levels of FM/TM and tPA messages were then examined by hybridizing total RNA with specific probes for FM/TM, tPA, or ac-
428
DEVELOPMENTAL BIOLOGY
VOLUME 141.1990
C
b
TABLE 2 THROMBIN-DEPENDENT ACTIVATION OF PROTEIN C IN MONOLAYER CULTURES Cells None F9 FS/RA FS/RA:cAMP
FIG. 1. Slot blot hybridization. Lane a, hybridized with FM/TM probe. Lane b, hybridized with tPA probe. Lane c, hybridized with actin probe. 1, F9 cells; 2, FS/RA cells; 3, FS/RA:2d-CAMP cells; 4, FS/RA:4d-CAMP cells; 5, FS/RA:5d-CAMP cells. Cells in samples three to five were treated with dbcAMP for 2,4, and 5 days until Day 9.
tin (Fig. 1; Table 1). The FM/TM message stayed at the same basal level in F9 and FS/RA cells and rose after dbcAMP treatment. On the other hand, the tPA message was detectably increased in FS/RA cells and it was further elevated in FS/RA:cAMP cells as reported by Rickles et ah (1988). It appeared that the increment of the tPA message in FS/RA:2d-CAMP cells was somewhat greater than that of the FM/TM message. The results raised a possibility that tPA and FM/TM genes were independently regulated. However, since quantitation by hybridization is laborious and insensitive to small differences, we explored activity assays to follow the time course of their expression. Cofactor Activity
of F9 Cell Cultures
Protein C activation was measured with F9 monolayer cultures before and after induction of differentiation. It is known that thrombin alone can activate protein C, although much less efficiently, in the absence of a TABLE 1 DENSITOMETRY OF RNA SLOT BLOT ANALYSIS Cells F9 FS/RA FS/RA:2d-CAMP* FS/RA:4d-CAMP* FS/RA:5d-CAMP”
FM/TM 0.08 0.08 0.20 0.72 0.78
tPA
Actin
FM/TM a actin
tPA” actin
0.09 0.27 0.93 1.93 1.98
1.12 1.19 0.93 1.23 1.11
0.07 0.07 0.22 0.59 0.70
0.08 0.23 1.00 1.61 1.78
Note. The autoradiogram shown in Fig. 1 was traced by a densitometer. a Relative enrichment of FM/TM and tPA messages was expressed by taking ratios to actin message. * FS/RA cells were treated with dbcAMP for 2,4, and 5 days until Day 9.
Thrombin + + + + -
AMC released” 3.4c nd” 1.6 nd 1.0 nd 21.3 nd
tPA activity* -d nd nd 1.1 2.1 15 21
Note. Protein C activation was examined in the presence or absence of thrombin with or without cultured cells. ’ The amount of AMC released in 10 min incubation is indicated in nanomoles. This should correlate to levels of activated protein C. * Presented in WHO international units. ’ Protein C can be activated by thrombin alone. The amount of AMC released in the absence of cells was considered the background level. d Not done. ’ Not detectable.
cofactor. Therefore, the amount of AMC released by thrombin in the absence of cells is the background level. As shown in Table 2, AMC release was not observed in the absence of thrombin, indicating that cells are devoid of nonspecific proteases that can activate protein C. Note that a significant increase in activity was associated with FS/RA:cAMP cells and that F9 and FS/RA cells had little or no activity. As for tPA, however, F9/ RA cells were positive and this was further augmented in FS/RA:cAMP cells, consistent with the results of hybridization and the report of Strickland et al. (1980).
Eflect of Antibodies Added to Monolage~ Cultures
Antibodies that could inhibit thrombin-FM/TM interaction should quantitatively block AMC release if it was solely due to FM/TM. Before such blocking assays, antibodies were tested for their specificity by immunoprecipitating radiolabeled cell surface proteins and’total cellular proteins. The four antibodies, polyclonal rabbit antibody, R2-20111, Rl-242, and Rl-1622, reacted with a single protein that comigrated with affinitypurified FM/TM. Results are presented in Fig. 2 and in a previous report (Imada et aL, 198’?). A polyclonal rabbit antibody and a monoclonal antibody, R2-20111, inhibited FM/TM binding with thrombin while Rl-1622 and Rl-242 did not (Imada et al, 1990; unpublished results). When polyclonal antibody or R2-20111 was added to cultures, activity was almost quantitatively eliminated while normal rabbit immunoglobulin and Rl-1622 were virtually ineffective (Table 3). We concluded that FM/
&ADA,
A
1
2
3
YAMAGUCHI,
AND IMADA
Parietal Endodem
429
B@rentiation
4
205, 60 116, 97* 40 66r 66~
20 45t
45,
24
FIG. 2. Immunonrecipitation with anti-FM/TM antibodies. Rabbit sera or ascites fluid were used to immunoprecipitate radioiodinated surface proteins (A). Metabolically labeled proteins (B) were immunoprecipitated with purified immunoglobulins as described under Materials and Methods. Positions of size marker proteins are shown by arrowheads in the left of figures and molecular sizes are indicated in kilodaltons. Immunoaffinity-purified FM/TM gave rise to a single band with an apparent molecular size of 116 kDa. In each figure, samples in lane 1 to lane 4 were immunoprecipitated with polyclonal rabbit anti FM/TM serum, normal rabbit serum, R2-20111, and normal rat immunoglobulin, respectively.
TM was exclusively responsible for the increase of AMC release in differentiated F9 cells. Time Course of Appearance and tPA Activities
of Cofactor
One millimolar dbcAMP was added at various times after removing RA from FS/RA cultures and cofactor and tPA activities were simultaneously assayed on Day 8 (or 96 hr in Fig. 3). A time course of marker protein induction was thus reconstructed. FS/RA cells had little or no cofactor activity and they remained inactive for up
TABLE INHIBITION
OF ACTIVITY
3
IN CULTURE
BY ANTIBODIES
Immunoglobulin added
Cofactor activity (nmol AMC/mg protein)
None Rabbit anti-FM/TM Normal rabbit R2-20111 Rl-1622
64.6 2.7 53.9 6.3 65.3
Inhibition (%I 0 98 17 92 0
Note. FS/RA:cAMP cells were preincubated for 30 min at 37°C with or without 100 pg of immunoglobulin in 200 ~1 of Hank’s balanced salt solution. The cofactor assay was then followed in the presence of immunoglobulins.
40
72
96
time
FIG. 3. Time course of FM/TM and tPA expression in parietal endoderm differentiation. Dibutyryl cyclic AMP was added at various times after RA treatment and cofactor and tPA activities were determined 96 hr afterward with duplicate or triplicate samples. Time of dbcAMP treatment is shown in abscissa in hours. 0, cofactor activity shown in left ordinate in nanomoles AMC released per milligram protein. l , tPA activity shown in right ordinate in international units per milligram protein.
to 15 hr of dbcAMP treatment. However, after this apparent lag period, the cofactor activity sharply increased with time until it leveled off in 3 days. In contrast, tPA was already positive in FS/RA cells and it was also increased, although much earlier than the cofactor, after a reproducible depression in 3-hr treated cultures. We interpreted the results to indicate that a lag of about 15 hr was required before FS/RA:cAMP cells became positive for FM./TM and that temporal differences existed in regulation of FM/TM and tPA expression. DISCUSSION
Since FM/TM is identical with thrombomodulin (Imada et aL, 1990), we exploited its functional assay as a measurement of FM/TM in culture. We concluded from blocking experiments with antibodies that differentiated F9 cells expressed cell surface FM/TM in an active form and that protein C activation assay was specific for FM/TM. FM/TM in FS/RA cells was below the level of detection by cofactor activity assay and RNA blotting, consistent with results of immunohistochemical staining and protein analysis (Imada et aZ., 1987). This may distinguish FM/TM from other parietal endoderm markers, including tPA, that are positive in FS/RA cells. FM/TM also differed from tPA in kinetics of expression during parietal endoderm differentiation. FM/ TM developed after a lag of 15 hr while tPA increased
430
DEVELOPMENTAL BIOLOGY
with a much shorter delay following exposure to dbcAMP. Therefore, we hypothesize that FM/TM and tPA are regulated differently, although more sensitive assays may be required to unequivocally prove this point. It must also be pointed out that we measured FM/ TM and tPA localized on cell surface and in cytoplasm, respectively. However, a temporal difference of their appearance seems to be too large to be accounted for only by a translocation phenomenon. In a similar experiment, Rickles et al. (1988) have reported that the tPA message stayed at a basal level for at least 12 hr while we found tPA activity to rise within 6 hr. This difference may be partly due to the fact that we induced parietal endoderm differentiation of FS/RA cells in the absence of RA while they treated FS/RA cells with RA and dbcAMP. RA apparently attenuated functional differentiation induced by dbcAMP. A negative cooperative effect of RA and cyclic AMP was suggested by down-regulation of the retinoic acid receptor messenger by cyclic AMP in the presence of RA (Hu and Gudas, 1990). Parietal endoderm cell lineage is uniquely suited to analyze mechanisms of differentiation. F9 cell differentiation in substratum-attached cultures mimics the in vivo situation where primitive endoderm converts into parietal endoderm. Visceral endoderm obtained in aggregate F9 cultures (Hogan et al, 1981) can transdifferentiate into parietal endoderm (Hogan and Tilly, 1981; Grover and Adamson, 1986; Casanova and Grabel, 1988) and similar transdifferentiation was observed in vivo when visceral endoderm was injected into blastocyst (Gardner, 1982). Differentiation from visceral to parieta1 endoderm may be induced by attachment to adherent substratum such as trophoblastic giant cells or matrix components. FM/TM may serve as a new tool to characterize natural inducers of differentiation and to dissect molecular events of parietal endoderm differentiation. The authors thank Drs. Buckingham and Strickland for their generous gifts of plasmids containing actin and tPA genes, respectively. The technical assistance of Ms. M. Uyeno, Y. Mizutani, and T. Akizawa is deeply appreciated. REFERENCES ASTRUP, T., and MULLERTZ, S. (1952). The fibrin plate method for estimating fibrinolytic activity. Arch. Biochem Biophys 40,346-351. BERNSTINE, E. G., HOOPER, M. L., GRANDCHAMP, S., and EPHRUSSI, B.
VOLUME 141,199O
(1973). Alkaline phosphatase activity in mouse teratoma. Proc. Natl. Ad. Sci. USA 70.3899-3903. CASANOVA, J. E., and GRABEL, L. B. (1988). The role of cell interaction in the differentiation of teratocarcinoma derived parietal and visceral endoderm. Dev. Biol 129,124-139. ESMON, C. T., and OWEN, W. G. (1981). Identification of an endothelial cell cofactor for thrombin-catalyzed activation of protein C. Proc. Natl Acad Ski. USA 78.2249-2252. ESMON, C. T. (1989). The role of protein C and thrombomodulin in the regulation of blood coagulation. J. Biol Chem. 264,4743-4746. GARDNER, R. L. (1982). Investigation of cell lineage and differentiation in the extraembryonic endoderm of the mouse embryo. J. Emlnyol. Exp. Mwphol. 68,175-198. GROVER, A., and ADAMSON, E. D. (1986). Evidence for the existence of an early common biochemical pathway in the differentiation of F9 cells into visceral or perietal endoderm: Modulation by cyclic AMP. Deu. Biol. 114,492-503. HOGAN, B. L. M., and TILLY, R. (1981). Cell interactions and endoderm differentiation in cultured mouse embryos. J. Embryol Exp. Morphol. 62,379-394. HOGAN, B. L. M., TAYLOR, A., and ADAMSON, E. (1981). Cell interactions modulate embryonal carcinoma cell differentiation into parieta1 or visceral endoderm. Nature (London) 291,235-237. Hu, L., and GUDAS, L. J. (1990). Cyclic AMP analogs and retinoic acid influence the expression of retinoic acid receptor 01,p and y mRNAs in F9 teratocarcinoma cells. Mol. CeU. Biol. 10,391-396. &ADA, S., and IMADA, M. (1982). Increase of a surface glycoprotein by cyclic AMP in Chinese hamster ovary cells: Dependence on cell-cell interaction. J. Biol. Chem 257,9108-9113. &ADA, M., IYADA, S., IWASAKI, H., KUME, A., YAMAGUCHI, H., and MOORE, E. E. (1987) Fetomodulin: Marker surface protein of fetal development which is modulatable by cyclic AMP. Dev. Biol. 122, 483-491. IMADA, S., YAMAGUCHI, H., NAGUMO, M., KATAYANAGI, S., IWASAKI, H., and IMADA M. (1990). Identification of fetomodulin, a surface marker protein of fetal development, as thrombomodulin by gene cloning and functional assays. Den Biol. 140,113-122. MANIATIS, T., FRITSCH, E. F., and SAMBROOK, J. (1982). “Molecular Cloning, A Laboratory Manual.” Cold Spring Harbor Laboratory, New York. MINTY, A. J., CARAVA’ITI, M., ROBERT, B., COHEN, A., DAUBAS, P., WEYDERT, A., GROS, F., and BUCKINGHAM, M. E. (1981). Mouse actin messenger RNAs: Construction and characterization of a recombinant plasmid molecule containing a complementary DNA transcript of mouse ol-actin mRNA. J. Biol Chem. 256,1008-1014. OWEN, W. G., and ESMON, C. T. (1981). Functional properties of an endothelial cell cofactor for thrombin-catalyzed activation of protein C. J. Biol. Ckem 256,5532-5535. RICKLES, R. J., DARROW, A. L., and STRICKLAND, S. (1988). Molecular cloning of complementary DNA to mouse tissue plasminogen activator mRNA and its expression during F9 teratocarcinoma cell differentiation. J. Biol. Chem. 263,1563-1569. STRICKLAND, S., and MAHDAVI, V. (1978). The induction of differentiation in teratocarcinoma stem cells by retinoic acid. CeU 15,393-403. STRICKLAND, S., SMITH, K. K., and MAROTM, K. R. (1980). Hormonal induction of differentiation in teratocarcinoma stem cells: Generation of parietal endoderm by retinoic acid and dibutyryl CAMP. CeU 21,347-355.
