In VitroCall.Dev.Biol.28A:449-454,June 1992 © 1992TissueCu'ttureAssociation 0883-8364/92 $01.50+0.00
GROWTH OF A HUMAN YOLK SAC TUMOR CELL LINE WITH YOLK SAC-DERIVED FUNCTIONS IN SELENIUM-SUPPLEMENTED CHEMICALLY DEFINED SYNTHETIC MEDIUM KIYOSHI OHKAWA, TAKASHI HATANO, NAOKO TAKIZAWA,KAZUE SHINMOTO, KYOSUKE YAMADA, MAKOTO MATSUDA, KOJI TAKADA, ANDYUTAKA TSUKADA
Department of Biochemistry (K. 0., T. H., N. T., K. S., M. M.) and Anatomy (ll) (K. Y.), Jikei University School of Medicine, 3-25-8, Nishi-Shinbashi, Minato-ku, Tokyo, 105; and Department of Biomedical Research Division (K. T., Y. T.), SRL (Ir~c), Hachioji, Tokyo, 192, Japan (Received 20 September 1991; accepted 10 January 1992)
SUMMARY
A human yolk sac tumor cell line, TG 1, which was established from a testicular yolk sac tumor, was found to replicate continuously in a chemically defined medium supplemented with Na2SeO 3 (ISRPMI). TG1 produced several plasma proteins and growth factors: albumin, alpha-fetoprotein (AFP), ferritin, carcinoembryonic antigen, beta-2-microglobulin, polyamine, neuron specific enolase, tissue polypeptide antigen, transferrin (TO', epidermal growth factor, and platelet derived growth factor. By analysis of lentil lectin (LcHA)-affinity electrophoresis, to examine the microheterogeneity of carbohydrate chains of synthetic glycoproteins, TG1 cells cultured with ISRPMI produced only LcHA reactive Tf and AFP based on core fucose attached to asparagine-linked N-acetylglucosamine residues instead of LcHA-nonreactive Tf and AFP produced by TG1 cells cultured with fetal bovine serum (FBS)-containing medium, o~1-6 Fucosyhransferase activity was significantly greater in the TG1 cells cultured with ISRPMI (39.9 + 1.5 pmol • h -1 • mg -1 protein) than cultured with FBS-containing media (18.2 + 1.2 pmol - h -1- mg-1 protein). These results have indicated that the selective increase of 0~1-6 fucosyltransferase occurred when the cells were cultured with the FBS-free synthetic media.
Key words: yolk sac tumor; chemically defined medium; transferrin; alpha-fetoprotein; fucosyhransferase; lectin-affinity electrophoresis. ties to bovine ones on the growth of target cells (6,13,15,19). Moreover, there has been no comparative study concerning the changed molecular properties of the synthetic proteins by the cells between the FBS-containing and FBS-free culture conditions. In the present study, we report the production of some plasma proteins and growth factors by cells of a human yolk sac tumor line in culture in FBS-free chemically defined synthetic medium. Additionally, the altered expression of glycosyltransferase activity determined by the enzyme assay as well as by immunochemical assay for detection of microheterogeneity of carbohydrate moiety of the acceptor proteins was examined during the cell adaptation in FBS-free chemically defined synthetic medium.
INTRODUCTION It has been well-established that human yolk sac tumor cells are able to synthesize several human plasma proteins in vitro (10,12,20). Human yolk sac tumors xenotransplanted into athymic nude mice also synthesize and secrete the human plasma proteins into mice sera (9,14,,16). They are alpha-fetoprotein (AFP), albumin (Alb), transferrin (TO, alpha-l-antitrypsin, transthyretin, hemopexin, beta-2-microglobuhn, plasminogen, apolipoprotein A1 and B, complement C4 and C5, retinol binding protein, alpha-2-HS glycoprotein, haptoglobin, antihemophilic factor, carcinoembryonic antigen (CEA), and so on. To examine the functions and characteristics of yolk sac in human embryogenesis, human yolk sac tumor cell line culturing in vitro is a powerful and suitable tool because of limitations on the availability and use of human early embryonic specimens for research purposes. Some cell lines derived from human yolk sac tumors reported were usually maintained in culture media containing fetal bovine serum (FBS). It was difficult to maintain the cultivated cells in vitro only by chemically defined synthetic medium without addition of FBS; however, a variety of advantages for cell culture in chemically defined medium was also demonstrated to explore novel autocrine growth substances and to investigate the direct and distinct effects of some growth factors, which are often similar proper-
MATERIALS AND METHODS Cell line. Human testicularyolk sac tumor cellline,TGI, used in this study was establishedfrom AFP-producing human testicularyolk sac tumor, JTG-I, transplanted,and maintained in athymic nude mice (16,17). TGI cellswere maintained in RPMI 1640 supplemented with 1 0 % FBS (Xavier Invest.Wacol, Australia),2.2 mg/ml sodium bicarbonate,and 2 m M L-glutamine. TGI cellswere also able to replicatewith low plating efficiencybut continuouslyin a chemicallydefined,serum- or protein-free
medium (ISRPMI) (13), which was supplemented with 3 )< I0 -a M Na2SeOa (Sigma, St. Louis, MO). Conditioned media obtained from each culture were collected and stocked - 8 0 ° C until use. Detection of human plasma proteins and growth factors. Conditioned 449
450
OHKAWA ET AL.
media from culturing with ISRPMI were mainly used for analysis. Alb (RIA, Nihon DPC, Tokyo, Japan), AFP (RIA, Daiichi Isotope, Tokyo, Japan), ferritin (RIA), CEA (R1A), beta-2-microglobulin (EIA, Wako, Tokyo, Japan), polyamine (enzyme assay, Tokuyama Soda, Tokyo, Japan), neuron specific enolase (NSE) (RIA, Cambridge Research, Biochemicals, UK), tissue polypeptide antigen (TPA) (RIA), epidermal growth factor (EGF) (RIA, Amersham), and platetet-derived growth factor (PDGF) (RIA, Amersham) were assayed.
5
Metabolic labeling and immanoprecipitation of AFP and Tf synthesized by TG1 cells. Cultures were labeled for 24 h with [14C]-i-leucine (5 pCi/ml, 12.5 GBq/mmol, Amersham) in the L-leucine-free Eagle's minimum essential medium (GIBCO, Grand Island, NY) supplemented with NazSeO3. Cellfree labeled conditioned media were dialyzed extensively against lO mM sodium phosphate, pH 7.0, 0.15 M NaC1, 1 mM phenylmethylsulfonylfluoride, 0.2% NaN3 (PBS), concentrated by centrifugation (Amicon, Danvers, MA). Immunoprecipitations were performed as described previously (19) using either goat anti-human Tf IgG (Cappel, Organon Teknika, West Chester, PA) or horse anti-human AFP IgG (14) as the first antibodies, and rabbit antibody against goat or horse lgG as the second antibodies, respectively. Analysis of precipitates was performed by sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) according to Laemmli (11) followed by fluorography (2). Preimmune horse or goat IgG was used as control.
Lectin-affinity electrophoresis and antibody-a~nity blotting analysis for fractionation of molecular species of Tf and AFP synthesized by TGt cells. Lectin-affinity electrophoresis of Tf or AFP was performed on 1.0% agarose gel plates (Litex type HSA, Glostrup, Denmark) in barbital/barbital Na buffer containing either O. 1 mg/cm 2 concanavalin A (ConA, Hohaen, Tokyo, Japan) or 20/.tg/cm ~ lentil lectin (LcHA, Hohnen) at 10 ° C, as reported previously (18,19). After electrophoresis of conditioned media, AFP or Tf bands separated in lectin-containing gels were specifically transferred to nitrocellulose papers, which were precoated with affinity-purified polyclonal horse anti-human AFP [gG or goat anti-human Tf lgG (19). The nitrocellulose papers were then washed twice in 0.05% Tween 20 in 20 mM Tris-HCl, pH 7.5, 0.5 M NaC1, 0.01% sodium merthiolate (TBS), treated with rabbit anti-human AFP or Tf IgG (Dakopatts, Glostrup, Denmark), followed by treatment with swine antibody to rabbit IgG labeled with horseradish peroxidase (Dakopatts). Color was developed by incubating the papers with 3,3'-diaminobenzidine tetrachloride (Polyscience, Warrington, PA) and HzO2 in TBS. The color-developed papers were washed, dried, and scanned with densitometer (Flying-spot scanner, CS-9000, Shimadzu, Tokyo, Japan) at 492 nm. Preparation of otl-6fucosyltransferase fiom TG1 cells. Cells either cultured with ]SRPMI or FBS-containing media were harvested using a rubber scraper and washed in PBS 3 times at 4 ° C. A pellet of approximately 1 × 107 cells was suspended in 0.5 ml of 0.1% Triton X-100 and sonicated at 4 ° C. After centrifugation at 1000 g for 10 min at 4 ° C, the pellet was resuspended in 1 m l of O.1 M cacodylate buffer, pH 6.0, 40 mM MgCb, 10 mM NaC1, 2% Triton CF54, and 10% glycerol. The suspension was sonicared for 8 s 12 times and then centrifuged at 11 000 g for 15 rain at 4 ° C. The supernatant solution was used as the enzyme source. Protein determination was performed by using Bio-Rad protein assay kit. Determination of o~l-6fucosyltransferase (EC2.4.1.68) activity. Enzyme activity was determined mainly by the method described in Voynow et al. (23) as the following incubation mixture (115 ttl): 25 ttl of 0.2 M Tris-HCl, pH 8.0, 5 gl of 25 mM guanosine triphosphate, 5 pl of 10% Triton X-100, 5 pl of 0.4 M MgC12, 5 #1 of 5% NAN3, 20 #1 of 1 #mol/ml glycopeptide in water. For preparation of the asialo-agalacto Tf glycopeptide from human sero Tf (Sigma), the procedure described in Schachter et al. (21) was mainly used, with slight modifications, and 1 #mol/ml in water was made. Then 50 ttl of enzyme solution was put into the Eppendorf tube in which guanosine diphosphate-i-fucose [fucose-14C(U)](GDP-L[l'~C]fucose, 0.05 /.tCi, 232 mCi/mmol) had been dried under N2 gas stream. Suitable incubation times were predetermined by examination of the reaction time course for periods up to 18 h. After incubation at 37" C, the mixture was chilled and then applied to a column (5 X 50 ram) of AG1-X8 (Bio-Rad, Richmond, CA). The column was washed in :2.0 ml water and the unbound fraction was added to 20 p.| of 0.1 M CaC|z and 0.1 M MnCI=. The mixture was then applied to a column (5 × 50 mm) of LcHA-Sepharose 4B (Pharmacia, Piscataway, NJ) by reapplying the fall through twice. The column was washed with 10 ml of 10 mMTris-HCl, pH
t'
i 3
0
4
8
Days after plating FIC. 1. Growth curves of TG1 cells in FBS-containing medium (solid squares) and in ISRPMI (solid circles) in which cells were directly seeded and cultured at the initial time.
8.0, O. 15 M NaC1, 1 mM CaClz, 1 mM MnCI2, and I mM MgC1z, and eluted with 10 ml of 0.5 M a-methyl-D-mannoside in the same buffer. The radioactivity of the eluate was determined by liquid scintillation counter (Beckman, CS6000[C, Fullerton, CA). The radioactivity measured was used to calculate the amount of fucosylated product formed. ~ESULTS
Growth characteristics of TG1 cells. TG1 cells grew in 1 0 % FBS-containing media with a population doubling time of 68.2 h, and plating efficiency was approximately 80%. In contrast, when the cells were plated in ISRPMI directly, most of cells failed to attach to the dish and plating efficiency was extremely low (6%), as shown in Fig. 1. After latent time of the first 2 days, the cells gradually grew with population doubling time of 70.7 h. After Passage 2, cells were routinely plated into 5 0 % conditioned ISRPMI of TG1 culture containing ISRPMI, and the medium was replaced with fresh medium after 2 days of culture and conditioned ISRPMI had begun to collect and store after 1 mo. of culture. Plasma proteins and growth factors production in ISRPMI. Plasma proteins and some growth factors produced by TG1 were analyzed on conditioned ISRPMI and were listed as follows: Alb 4 0 0 ng" m1-1 • 10 -6 cells. 3 -1 days, AFP 6 6 0 ng" m1-1" 10 -6 cells. 3 -1 days, Tf 2 0 0 ng. m1-1. 10 -6 cells- 3 -1 days, CEA 0.5 n g - m l - t . 10 -6 c e l l s . 3 - i days, ferritin 5 ng. m1-1- 10-6cells • 3 - t days, beta-2-microglobulin 2 0 0 ng-m1-1- 1 0 - 6 c e l l s - 3 -1
451
YOLK SAC TUMOR IN DEFINED MEDIUM
MW(kDa).
electrophoretic profile was detected between conditioned media obtained from FBS-containing culture and from ISRPMI culture. LcHA nonreactive, Lt3, band disappeared in the conditioned ISRPMI, in which Tf was composed of only strongly reactive (Ltl) and weakly reactive (Lt2) bands. Disappearance of LcHA-nonreactire band after TGI culturing with ISRPMI was also found in the subfraction of AFP (Figs. 4 and 5). In conditioned ISRPMI, AFP consisted of only two components that were strongly (Lal) and weakly reactive (La2) to LcHA, and LcHA-nonreactive AFP (La3), which was detectable in conditioned FBS-containing media, was not observed. By subfractionation of AFP using erythroagglutinating phytohemagglutinin (EPHA)-affinity electrophoresis, which detects specifically the bisecting GlcNAc at the core ~-mannose of biantennary structures of the molecules, no detectable difference of the electrophoretic profiles was observed between AFP derived from conditioned FBS-containing medium and conditioned ISRPMI (data not shown).
92.4 69" 46-+ 30--*
Fucosyltransferase activity of TG1 cell cultures with or without FBS. In preliminary assays, the separation of I14C]fucosedabeled
AFP
Tf
FIG. 2. Fluorograph of metabolicallylabeled proteins synthesized and secreted by TG1 cells. Immunoprecipitationby anti-human AFP antibody and anti-human Tf antibody shown in lane 1 (AFP)and lane 2 (TJ), respectively. MW = molecular size markers.
asialo-agalactoTf glycopeptide from GDP-L-[14C]fucose was rapidly effected using small columns of AGI-X8 and more than 95% of isotope-labeled glycopeptides were effectively bound and eluted by a-methyl-D-mannoside using LcHA-Sepharose 4B. In addition, the formation of labeled product was linear up to 4 h, and a standard incubation time of 2 h was used in the determination of activity (Fig. 6). The activity of al-6fucosyhransferase was significantly greater in the TGI cells cultured in ISRPMI (39.9 + 1.5 pmol. h -1" mg-1
days, polyamine 3.16 #mol.1-1. 10 -6 cells. 3 -1 days, NSE 4.2 ng-m1-1. 10 _6 cells.3 -1 days, TPA 30 000 U" 1- l - 10 -6 cells-3 -1 days, EGF, 1.5 pg. m1-1. 10 -6 cells-3 -1 days, and PDGF 2.0 pg" m1-1. 10 -6 cells. 3 -1 days. The immunoprecipitates using anti-AFP or anti-Tf antibody also demonstrated the band corresponding to their molecular weight of 70 000 for AFP and 80 000 for Tf, respectively (Fig. 2).
Determination of two major glycoproteins (AFP and Tf) with microheterogeneity of lectin-affmity before and after culturing TG1 cells with ISRPMI. By this method, AFP and Tf were detectable as low as 0.6 and 1 ng, respectively, as reported previously (18,19) and the results were compatible with the profiles obtained from lectin-affinity crossed immunoelectrophoresis. Using this sensitive method after ConA-affinity electrophoresis, Tf from normal human serum was divided into three bands (Fig. 3), i.e. strongly reactive major (Ctl)," moderately reactive minor (Ct2), and nonreactive, slight but detectable band (Ct3). These results were also confirmed by densitometric analysis (Fig. 4). Tf from TG1 cultured with FBS formed relatively major Ctl and minor Ct2 bands. Tf from TG1 cultured without FBS (ISRPMI) was also split into strongly reactive minor (Ctl) and weakly reactive major (Ct2) bands. Nonreactive Ct3 band was not detectable in both cases (Fig. 3 and 4). In the case of AFP separated by ConA-affinity electrophoresis, ConA reactive major (Cal) and nonreactive minor (Ca2) bands were detectable both in conditioned FBS-containing media and ISRPMI (Figs. 4 and 5). By LcHA-affinity electrophoresis, Tf in normal human serum was divided into three bands (Figs. 3 and 4), nonreactive major (Lt3), weakly reactive minor (Lt2), and strongly reactive but trace (Ltl). Tf from TG1 cultured with FBS-containing media was also composed of three bands; however, Ltl increased as compared with that of normal serum. In contrast, an obvious difference of the
Q1 Ct2
Ct3 NHS
ConA
i
TGIFBS-(ISRPMI) rD~':i
TGIFBS+ + +
NHS LcHA
TG 1FBS-(ISRPMI) TGIFBS+ Ltl Lt2
Lt3
Fro. 3. AnalyticalTf bands separated by ConAor LcHA-atfinityelectrophoresis of Tf from TG1 cultured with or without FBS-containingmedium and from normal human serum (NHS). Each 1-#1 sample was applied and eleorophoresis was run until bromophenol blue migrated 9.0 cm from the origin (-). Ctl-Ct3 and Ltl-Lt3 are described in the text. FBS+:10% FBS containing RPMI, FBS-(ISRPMI):RPMIsupplemented with Na2SeOs without FBS.
452
OIqKAWA ET AL.
"If ConA ++ ...... r"-++--r" ..........
~+ ~
~-'-+ . . . . .
"~"-+-
i
i--;-,.'...~
r-
+l.... ~ .......
/ j
i--
!: +!! ,i i::
.....
";
LcHA
+'
...... • .+; • + • .~
+
lt~tO..-r----i
~+ • i-.-
. t .~. + i-+.
+ ...... t"" ~
" +"
4.+
_.+. ~ L.
!.~!:+-....+
+
i
...,:..,.. .+ .i .
~..~ ,
......
..
~-. ~..... ~ " ~ ! ~ . ~ . . , ~
.++..+...+..
+. . •. . . +
~..
+e+ + +_
~- ..::+ : :+. '+ '~ +- l:+ ++....... + +..+ [+..: + + + , -;__+,
.... £
"
.'"- t-
r---+
~
v+ ~
....
-t
t+++:+j + ..+ + . . . . . .
+1........t '++![-", +,:
t
+
r
;
+
i,cHA
-.+~. ~ .~ ...+..~+ ~. -+..! .....
-+-i---~'-~ -~-+!'r" ~ i + ..... ~..~.....-....-....-..+....~._,~
,." ",~
....-~-+-~+----.:-.i.-4-.-~
~:_+...+-.r.~__ +_...~_.+
TG1FBS+
ConA
+++
+..-
-'-+.:--!-----i-#'+-!.---.-.: •.-+..~.-..~....~.-.:....'-...~..+..u. --.--?--T-:-..-7++-7-.:--7/....~
AFP
+
-i
....+-+..-+...+-4-..~...~--~
•-4..-.~.+.-~..+.~ -.~__~._.L_.::
........
r" ", "-'~+ --: ....
..... t'" ".."' "t" i- .t.. + . + ..!....
::::::::::::::::::::::::::::: • L...r....4...++.^+ + . . ~
..~.....
::.!::i,. i .i i--!--i -'~--i-
i--4-~.--~
-~-~--~
- -+ -.,÷-. --~ .-.4.,+-i.-...L-.-4..
TGIFBSi
i'+
- 1.- "i
+.""~+++"÷+++'~+ ~.......'._.i_..-....-....-....-.__: .~
:!.-::!:::~.::+ I ~,--~7~?'ml' +!-; i....+..4, i rl
" "+
~-..-:-...---~--+-..-..-.~-4-.+
.i..+--÷-I--~-.-+-..+ ....
.:::~..~:::.~:i.i:::.:,t:...~iiii~::::i ::
~
:
:
:
:
.
.
~
..~...r....~....~
......
.+..+...~
.
- . "!"'"~""~
++÷+ .: ~ : ~
~
~
........
:::!+"i:-:!':~:~:~ +"'t "'+" +'"::i"~::.i: "! ....
i:!~:!:~ I-'-'.÷ ...+...i
~:i"il
1
.... ~-.....~.....+.....~....;....+....i....i... .... ~--~...4 .... ~-- ~ .4- --!..
..... ~+-
i i.:
. ..~.~..-L..i~i..i~.~T~Zi~i.iI
'+~'+-++"+~--~+~L-+4 "
~- 4 - -+-.~-,++,-.4..-.L -.+-.~
,
"
, +
+-- ~~ '.4--~-.~- -~-.~ -.".--.4 --
t"
"4-
+
FIG. 4. Densitometric analysis was carried out according to Materials and Methods. See Figs, 3 and 5 for other keys.
protein) than those cultured in FBS-containing media (18.2 + 1.2 pmol- h - 1 . mg -1 protein).
DISCUSSION
In the present study, it was possible to eliminate serum and other proteins from culture medium for the growth of yolk sac tumor cells by adding a trace amounts of selenium to a chemically defined medium, ISRPMI. It is important to note that no adaptation procedures were required for the growth of the celt line after replacing serum with Na2SeO s. Many human germ cell tumors have been established as long-term monolayer cultures (1,3,8,1 O, 12,20,24); however, this is the first report of culture of the yolk sac tumor cells with serum- or protein-free chemically defined medium. The TG1 ceil line synthesized and secreted many plasma proteins and growth factors in the conditioned media as previously reported (1,3,8,10,12,20,24). Synthesized Tf might be one of the
essential substances for proliferating TG1 cells in vitro. A similar result was reported previously (19). Ferritin, polyamine, NSE, TPA, EGF, and PDGF were first detected and reported in a yolk sac tumor cell line (TG1, present case) cultured in vitro. To utilize the conditioned media with serum- or protein-free cultures is of great advantage for immunoassay to detect the very small amounts of substances and to eliminate the confusion of crossreactivity of the target substances to the similar substances often present in bovine serum, Wheia the TG1 cells were xenotransplanted into athymic nude mice, 13 human plasma proteins were detectable in plasma or cystic fluid in the tumor mass of mice bearing TG1 cells as reported elsewhere (14,16). These results suggest that TG1 and other yolk sac tumor cell lines and some embryonal carcinoma cell lines are able to maintain the same characteristics of human yolk sac (7) in early embryogenesis. By LcHA-affinity electrophoresis, elevation of fucosylation was demonstrated in two secretory glycoproteins, AFP as marker pro-
YOLK SAC TUMOR IN DEFINED MEDIUM tein for yolk sac tumor and Tf as one of the growth factors for prohferating cells, after the TG1 cells cultured with ISRPMI. LcHA-nonreactive fractions of AFP and Tf, which were detected in conditioned FBS-containing media, disappeared in conditioned ISRPMI, i.e. both AFP and Tf in the conditioned ISRPMI were composed of LcHA-strongly and weakly reactive two fractions. Conversely, no obvious difference in reactivity of AFP and Tf to CuriA as well as AFP to EPHA was noted between two conditioned media, and the electrophoretic profiles of AFP were almost similar to those of previous reports (18,22 ), The oligosaccharide structure required for LcHA activity lies in the core fucose attached to asparaginelinked N-acetylglucosamine residues (Fuc al-6GlcNAc-Asn). This additional attachment of fucose to GlcNAc does not influence the reactivity to ConA, however, the increase of tri- and tetra-antennary glycosylation of carbohydrate moieties of Tf molecules diminishes the affinity to ConA (5). This fact concluded that the fucosylated Tf in bi- and tri-antennary glycans increased in Tf molecule synthesized by TG1 cells. These results were also confirmed by the fact that the enzyme activity of a 1-6fucosyltransferase, which transports fucose to position 6 of asparagine-linked GlcNAc of N-glycans, was significantly greater in the TG1 cells cuhured with ISRPMI than in the cells with FBS-containing media. The enzyme activity correlated to the results obtained from LcHA-affinity electrophoresis. These results suggested the enhanced cellular expression of al-6fucosyltransferase after the change of culture condition from FBS-containing to serum-free medium. Abnormal fucosylation of the carbohydrate moiety of the proteins has been demonstrated in some proteins, such as Tf (4,19,26), AFP (27), and 'y-glutamyl transpeptidase (25) during malignant transformation of the cells. The increased level of fucosyltransferase activity and successive accumulation of the glycoproteins with abnormal fucosylation of the carbohydrate moieties is probably one of the oneofetal characteristics in the expression of cancer cells and the acquirement of growth characteristics without necessity of FBS may be the ultimate malignant transformation of the cancer cells. These facts clearly indicates
C,a2
Ca1
TGIFBS+ ConA "I'G 1FBS-(ISRPMI)
-
-ITGIFBS+
LcHA
~
~ii~
;:~ ~
~
TGIFBS-(ISRPMI)
t t 1 Lal La2 La3
FIG. 5. Analytical AFP bands separated by ConA or LcHA-affinityelectrophoresis from TG1. Each 3-gl sample was eleetrophoresed until bromophenol blue migrated 4.5 cm from the origin (-), Cal-Ca3 and Lal-La3 are described in the text. See Fig. 3. for other keys,
453
0
! 0
2
4
. 18
hour FIG. 6. Time-based graph of formation of [14C]-L-fucosein al-6fucosyltransferase assay. Incubation were performed using GDP-L-[14C]fucose as donor and asialo-agalacto Tf glycopeptide as acceptor for the various times shown. Each point represents the mean + SD of quadruplicate examinations,
that human cancer cell lines stably cultured with protein-free synthetic medium would be a useful material for the biomedical research. REFERENCES 1. Andrews, P. W.; Goodfellow, P. N.; Shevinsky, L. H., et al. Cell-surface antigens of a clonal human embryonal carcinoma cell line: morphological and antigenic differentiation in euhure. Int. J. Cancer 29:523-531: 1982. 2. Bonner, W. M.; Laskey, R. A. A film detection method for tritium-labeled proteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem. 46:83-88; 1974. 3. Bronson, D. L.; Andrews, P. W.; Soher, D., et al. Cell line derived frora a metastasis of a human testieular germ cell tumor. Cancer Res. 40:2500-2506; 1980. 4, Campion, B,; Leger, D.; Wieruszeski, J-M., et al. Presence of fucosylated triantennary, tetraantennary and pentaantennary glycans in transferrin synthesized by the human hepatocarcinuma cell line Hep G2, Eur. J. Biochem. 184:405-413; 1989. 5. Cummings, R. D.; Kornfeld, S. Fracti°nati°n °f asparagine-linked °ligosaccharides by serial lectin-agarose affinity chromatography. A rapid, sensitive, and specific technique, J. Biol. Chem. 257:1123511240; 1982. 6. Darlington, G. J.; Kelly, J. H.; Buffone, G. J, Growth and hepatospecific gene expression of human hepatoma cells in a defined medium. In Vitro Cell. Dev. Biol, 23:349-354; 1987, 7. Gitlin, D." Perricelli, A. Synthesis of human serum albumin, prealbumin, o~-fetoprotein, cqantitrypsin and transferrin by the human yolk sac. Nature 228:995-997; 1970. 8. Grossman, H. B.; Wedemeyer, G. UM-TC-1, a new human testicular carcinoma cell line. Cancer J. 1:22-25; 1986. 9. Hata, J.; Ueyama, Y.; Tamaoki, N., et aL Human yolk sac tumor serially transplanted in nude mice: its morphologic and functional properties. Cancer 46:2446-2455; 1980.
454
OHKAWA ET AL.
10. lshiwata, I.; Ishiguro, T.; Ishiwata, C., et al. Establishment and characterization of a human ovarian endodermal sinus tumor cell line producing specific type of a-fetoprotein subfraction. Gynecol. Oneol. 25:281-293; 1986. 11. Laemmli, U. K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685; 1970. 12. Lanford, R. E.; Bronson, D. L.; Esflack, L. E., et al. Plasma protein and apolipoprotein synthesis by human yolk sac carcinoma cells in vitro. In Vitro Cell. Dev. Biol. 27A:205-210; 1990. 13. Nakabayashi, H.; Taketa, K.; Miyano, K., et al. Growth of human hepatoma cell lines with differentiated functions in chemically defined medium. Cancer Res. 42:3858-3863; 1982. 14. Nishi, S.; Nishita, T.; Ohkawa, K. Production of 13 plasma proteins by human testieular yolk sac tumor transplanted into nude mice. Tumor Biol. 12:184-188; 1991. 15. Ogasawara, M.; Sirbasku, D. A. A new serum-free method of measuring growth factor activities for human cancer cells in culture. In Vitro Cell. Dev. Biol. 24:911-920; 1988. 16. Ohkawa, K.; Koike, K.; Amasaki, H., et al. Heteru-transplantation of human embryonal carcinoma into the nude mouse: functional and morphologic examination. Asia Oceania J. Obstet. Gyaecol. 6:7681; 1980. 17. Ohkawa, K.; Hibi, N.; Tsukada, Y. Evaluation of a conjugate of purified antibodies against human AFP-dextran-daunorubiein to human AFP-producing yolk sac tumor cell lines. Cancer lmmunol. Immunother. 22:81-86; 1986. 18. Ohkawa, K.; Abe, T.; Tsukada, Y., e! al. An improved sensitive assay method for the heterogeneity of c~-fetoprotein: possible application for early differential diagnosis. J. Clin. Chem. Clin. Biochem. 27:337-341; 1989.
19. Ohkawa, K.; Takada, K.; Takizawa, N., et al. Clear cell carcinoma of the human ovary synthesizes and secretes a transferrin with microheterogeneity of lectin affinity. FEBS Lett. 270:19-23; 1990. 20. Pera, M. F.; Blasco Lafita, M. J.; Mills, J. Cultured stem-cells from human testicular teratomas: the nature of human embryonal carcinoma, and its comparison with two types of yolk-sac carcinoma. Int. J. Cancer 40.'334-343; 1987. 21. Schachter, H.; Narasimhan, S.; Gleeson, P., et al. Glycosyltransferases involved in elongation of N-glycosidically linked oligosaccharides of the complex or N-acetyllactosamine type. Methods Enzymol. 98:98-134; 1983. 22. Taketa, K.; Ichikawa, E.; Nakabayashi, H., et al. Further resolution of human ot-fetoprotein by affinity electrophoresis with erythroagglufining phytobemagglutinin of Phaseolus vulgaris lectin. Tumor Biol. 6:519-531; 1985. 23. Voynow, J. A.; Scanlin, T. F.; Glick, M. C. A quantitative method for GDP-L-Fuc:N-Acetyl-~-D-glucosaminide a 1-6fucosyhransferase activity with lectin affinity chromatography. Anal. Biochem. 168:367373; 1988. 24. Yamamoto, T.; Komatsubara, S.; Suzuki, T., et al. In vitro cultivation of human testicular embryonal carcinoma and establishment of a new cell line. Gann 70:677-680; 1979. 25. Yamashita, K.; Hitoi, A.; Taniguchi, N., et al. Comparative study oftbe sugar chains of 3,-glutamyhranspeptidase purified from rat liver and rat AH-66 hepatoma. Cancer Res. 43:5059-5063; 1983. 26. Yamashita, K.; Koide, N.; Endo, T., et al. Altered glycosylation of serum transferrin of patients with hepatocellular carcinoma. J. Biol. Chem. 264:2415-2423; 1989. 27. Yoshima, H.; Mizuochi, T.; Ishii, M., et al. Structure of the asparaginelinked sugar chains of ot-fetoprotein purified from human ascites fluid. Cancer Res. 40:4276-4281; 1980.
GROWTH OF A HUMAN YOLK SAC TUMOR CELL LINE WITH YOLK SAC-DERIVED FUNCTIONS IN SELENIUM-SUPPLEMENTED CHEMICALLY DEFINED SYNTHETIC MEDIUM KIYOSHI OHKAWA, TAKASHI HATANO, NAOKO TAKIZAWA,KAZUE SHINMOTO, KYOSUKE YAMADA, MAKOTO MATSUDA, KOJI TAKADA, ANDYUTAKA TSUKADA
Department of Biochemistry (K. 0., T. H., N. T., K. S., M. M.) and Anatomy (ll) (K. Y.), Jikei University School of Medicine, 3-25-8, Nishi-Shinbashi, Minato-ku, Tokyo, 105; and Department of Biomedical Research Division (K. T., Y. T.), SRL (Ir~c), Hachioji, Tokyo, 192, Japan (Received 20 September 1991; accepted 10 January 1992)
SUMMARY
A human yolk sac tumor cell line, TG 1, which was established from a testicular yolk sac tumor, was found to replicate continuously in a chemically defined medium supplemented with Na2SeO 3 (ISRPMI). TG1 produced several plasma proteins and growth factors: albumin, alpha-fetoprotein (AFP), ferritin, carcinoembryonic antigen, beta-2-microglobulin, polyamine, neuron specific enolase, tissue polypeptide antigen, transferrin (TO', epidermal growth factor, and platelet derived growth factor. By analysis of lentil lectin (LcHA)-affinity electrophoresis, to examine the microheterogeneity of carbohydrate chains of synthetic glycoproteins, TG1 cells cultured with ISRPMI produced only LcHA reactive Tf and AFP based on core fucose attached to asparagine-linked N-acetylglucosamine residues instead of LcHA-nonreactive Tf and AFP produced by TG1 cells cultured with fetal bovine serum (FBS)-containing medium, o~1-6 Fucosyhransferase activity was significantly greater in the TG1 cells cultured with ISRPMI (39.9 + 1.5 pmol • h -1 • mg -1 protein) than cultured with FBS-containing media (18.2 + 1.2 pmol - h -1- mg-1 protein). These results have indicated that the selective increase of 0~1-6 fucosyltransferase occurred when the cells were cultured with the FBS-free synthetic media.
Key words: yolk sac tumor; chemically defined medium; transferrin; alpha-fetoprotein; fucosyhransferase; lectin-affinity electrophoresis. ties to bovine ones on the growth of target cells (6,13,15,19). Moreover, there has been no comparative study concerning the changed molecular properties of the synthetic proteins by the cells between the FBS-containing and FBS-free culture conditions. In the present study, we report the production of some plasma proteins and growth factors by cells of a human yolk sac tumor line in culture in FBS-free chemically defined synthetic medium. Additionally, the altered expression of glycosyltransferase activity determined by the enzyme assay as well as by immunochemical assay for detection of microheterogeneity of carbohydrate moiety of the acceptor proteins was examined during the cell adaptation in FBS-free chemically defined synthetic medium.
INTRODUCTION It has been well-established that human yolk sac tumor cells are able to synthesize several human plasma proteins in vitro (10,12,20). Human yolk sac tumors xenotransplanted into athymic nude mice also synthesize and secrete the human plasma proteins into mice sera (9,14,,16). They are alpha-fetoprotein (AFP), albumin (Alb), transferrin (TO, alpha-l-antitrypsin, transthyretin, hemopexin, beta-2-microglobuhn, plasminogen, apolipoprotein A1 and B, complement C4 and C5, retinol binding protein, alpha-2-HS glycoprotein, haptoglobin, antihemophilic factor, carcinoembryonic antigen (CEA), and so on. To examine the functions and characteristics of yolk sac in human embryogenesis, human yolk sac tumor cell line culturing in vitro is a powerful and suitable tool because of limitations on the availability and use of human early embryonic specimens for research purposes. Some cell lines derived from human yolk sac tumors reported were usually maintained in culture media containing fetal bovine serum (FBS). It was difficult to maintain the cultivated cells in vitro only by chemically defined synthetic medium without addition of FBS; however, a variety of advantages for cell culture in chemically defined medium was also demonstrated to explore novel autocrine growth substances and to investigate the direct and distinct effects of some growth factors, which are often similar proper-
MATERIALS AND METHODS Cell line. Human testicularyolk sac tumor cellline,TGI, used in this study was establishedfrom AFP-producing human testicularyolk sac tumor, JTG-I, transplanted,and maintained in athymic nude mice (16,17). TGI cellswere maintained in RPMI 1640 supplemented with 1 0 % FBS (Xavier Invest.Wacol, Australia),2.2 mg/ml sodium bicarbonate,and 2 m M L-glutamine. TGI cellswere also able to replicatewith low plating efficiencybut continuouslyin a chemicallydefined,serum- or protein-free
medium (ISRPMI) (13), which was supplemented with 3 )< I0 -a M Na2SeOa (Sigma, St. Louis, MO). Conditioned media obtained from each culture were collected and stocked - 8 0 ° C until use. Detection of human plasma proteins and growth factors. Conditioned 449
450
OHKAWA ET AL.
media from culturing with ISRPMI were mainly used for analysis. Alb (RIA, Nihon DPC, Tokyo, Japan), AFP (RIA, Daiichi Isotope, Tokyo, Japan), ferritin (RIA), CEA (R1A), beta-2-microglobulin (EIA, Wako, Tokyo, Japan), polyamine (enzyme assay, Tokuyama Soda, Tokyo, Japan), neuron specific enolase (NSE) (RIA, Cambridge Research, Biochemicals, UK), tissue polypeptide antigen (TPA) (RIA), epidermal growth factor (EGF) (RIA, Amersham), and platetet-derived growth factor (PDGF) (RIA, Amersham) were assayed.
5
Metabolic labeling and immanoprecipitation of AFP and Tf synthesized by TG1 cells. Cultures were labeled for 24 h with [14C]-i-leucine (5 pCi/ml, 12.5 GBq/mmol, Amersham) in the L-leucine-free Eagle's minimum essential medium (GIBCO, Grand Island, NY) supplemented with NazSeO3. Cellfree labeled conditioned media were dialyzed extensively against lO mM sodium phosphate, pH 7.0, 0.15 M NaC1, 1 mM phenylmethylsulfonylfluoride, 0.2% NaN3 (PBS), concentrated by centrifugation (Amicon, Danvers, MA). Immunoprecipitations were performed as described previously (19) using either goat anti-human Tf IgG (Cappel, Organon Teknika, West Chester, PA) or horse anti-human AFP IgG (14) as the first antibodies, and rabbit antibody against goat or horse lgG as the second antibodies, respectively. Analysis of precipitates was performed by sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) according to Laemmli (11) followed by fluorography (2). Preimmune horse or goat IgG was used as control.
Lectin-affinity electrophoresis and antibody-a~nity blotting analysis for fractionation of molecular species of Tf and AFP synthesized by TGt cells. Lectin-affinity electrophoresis of Tf or AFP was performed on 1.0% agarose gel plates (Litex type HSA, Glostrup, Denmark) in barbital/barbital Na buffer containing either O. 1 mg/cm 2 concanavalin A (ConA, Hohaen, Tokyo, Japan) or 20/.tg/cm ~ lentil lectin (LcHA, Hohnen) at 10 ° C, as reported previously (18,19). After electrophoresis of conditioned media, AFP or Tf bands separated in lectin-containing gels were specifically transferred to nitrocellulose papers, which were precoated with affinity-purified polyclonal horse anti-human AFP [gG or goat anti-human Tf lgG (19). The nitrocellulose papers were then washed twice in 0.05% Tween 20 in 20 mM Tris-HCl, pH 7.5, 0.5 M NaC1, 0.01% sodium merthiolate (TBS), treated with rabbit anti-human AFP or Tf IgG (Dakopatts, Glostrup, Denmark), followed by treatment with swine antibody to rabbit IgG labeled with horseradish peroxidase (Dakopatts). Color was developed by incubating the papers with 3,3'-diaminobenzidine tetrachloride (Polyscience, Warrington, PA) and HzO2 in TBS. The color-developed papers were washed, dried, and scanned with densitometer (Flying-spot scanner, CS-9000, Shimadzu, Tokyo, Japan) at 492 nm. Preparation of otl-6fucosyltransferase fiom TG1 cells. Cells either cultured with ]SRPMI or FBS-containing media were harvested using a rubber scraper and washed in PBS 3 times at 4 ° C. A pellet of approximately 1 × 107 cells was suspended in 0.5 ml of 0.1% Triton X-100 and sonicated at 4 ° C. After centrifugation at 1000 g for 10 min at 4 ° C, the pellet was resuspended in 1 m l of O.1 M cacodylate buffer, pH 6.0, 40 mM MgCb, 10 mM NaC1, 2% Triton CF54, and 10% glycerol. The suspension was sonicared for 8 s 12 times and then centrifuged at 11 000 g for 15 rain at 4 ° C. The supernatant solution was used as the enzyme source. Protein determination was performed by using Bio-Rad protein assay kit. Determination of o~l-6fucosyltransferase (EC2.4.1.68) activity. Enzyme activity was determined mainly by the method described in Voynow et al. (23) as the following incubation mixture (115 ttl): 25 ttl of 0.2 M Tris-HCl, pH 8.0, 5 gl of 25 mM guanosine triphosphate, 5 pl of 10% Triton X-100, 5 pl of 0.4 M MgC12, 5 #1 of 5% NAN3, 20 #1 of 1 #mol/ml glycopeptide in water. For preparation of the asialo-agalacto Tf glycopeptide from human sero Tf (Sigma), the procedure described in Schachter et al. (21) was mainly used, with slight modifications, and 1 #mol/ml in water was made. Then 50 ttl of enzyme solution was put into the Eppendorf tube in which guanosine diphosphate-i-fucose [fucose-14C(U)](GDP-L[l'~C]fucose, 0.05 /.tCi, 232 mCi/mmol) had been dried under N2 gas stream. Suitable incubation times were predetermined by examination of the reaction time course for periods up to 18 h. After incubation at 37" C, the mixture was chilled and then applied to a column (5 X 50 ram) of AG1-X8 (Bio-Rad, Richmond, CA). The column was washed in :2.0 ml water and the unbound fraction was added to 20 p.| of 0.1 M CaC|z and 0.1 M MnCI=. The mixture was then applied to a column (5 × 50 mm) of LcHA-Sepharose 4B (Pharmacia, Piscataway, NJ) by reapplying the fall through twice. The column was washed with 10 ml of 10 mMTris-HCl, pH
t'
i 3
0
4
8
Days after plating FIC. 1. Growth curves of TG1 cells in FBS-containing medium (solid squares) and in ISRPMI (solid circles) in which cells were directly seeded and cultured at the initial time.
8.0, O. 15 M NaC1, 1 mM CaClz, 1 mM MnCI2, and I mM MgC1z, and eluted with 10 ml of 0.5 M a-methyl-D-mannoside in the same buffer. The radioactivity of the eluate was determined by liquid scintillation counter (Beckman, CS6000[C, Fullerton, CA). The radioactivity measured was used to calculate the amount of fucosylated product formed. ~ESULTS
Growth characteristics of TG1 cells. TG1 cells grew in 1 0 % FBS-containing media with a population doubling time of 68.2 h, and plating efficiency was approximately 80%. In contrast, when the cells were plated in ISRPMI directly, most of cells failed to attach to the dish and plating efficiency was extremely low (6%), as shown in Fig. 1. After latent time of the first 2 days, the cells gradually grew with population doubling time of 70.7 h. After Passage 2, cells were routinely plated into 5 0 % conditioned ISRPMI of TG1 culture containing ISRPMI, and the medium was replaced with fresh medium after 2 days of culture and conditioned ISRPMI had begun to collect and store after 1 mo. of culture. Plasma proteins and growth factors production in ISRPMI. Plasma proteins and some growth factors produced by TG1 were analyzed on conditioned ISRPMI and were listed as follows: Alb 4 0 0 ng" m1-1 • 10 -6 cells. 3 -1 days, AFP 6 6 0 ng" m1-1" 10 -6 cells. 3 -1 days, Tf 2 0 0 ng. m1-1. 10 -6 cells- 3 -1 days, CEA 0.5 n g - m l - t . 10 -6 c e l l s . 3 - i days, ferritin 5 ng. m1-1- 10-6cells • 3 - t days, beta-2-microglobulin 2 0 0 ng-m1-1- 1 0 - 6 c e l l s - 3 -1
451
YOLK SAC TUMOR IN DEFINED MEDIUM
MW(kDa).
electrophoretic profile was detected between conditioned media obtained from FBS-containing culture and from ISRPMI culture. LcHA nonreactive, Lt3, band disappeared in the conditioned ISRPMI, in which Tf was composed of only strongly reactive (Ltl) and weakly reactive (Lt2) bands. Disappearance of LcHA-nonreactire band after TGI culturing with ISRPMI was also found in the subfraction of AFP (Figs. 4 and 5). In conditioned ISRPMI, AFP consisted of only two components that were strongly (Lal) and weakly reactive (La2) to LcHA, and LcHA-nonreactive AFP (La3), which was detectable in conditioned FBS-containing media, was not observed. By subfractionation of AFP using erythroagglutinating phytohemagglutinin (EPHA)-affinity electrophoresis, which detects specifically the bisecting GlcNAc at the core ~-mannose of biantennary structures of the molecules, no detectable difference of the electrophoretic profiles was observed between AFP derived from conditioned FBS-containing medium and conditioned ISRPMI (data not shown).
92.4 69" 46-+ 30--*
Fucosyltransferase activity of TG1 cell cultures with or without FBS. In preliminary assays, the separation of I14C]fucosedabeled
AFP
Tf
FIG. 2. Fluorograph of metabolicallylabeled proteins synthesized and secreted by TG1 cells. Immunoprecipitationby anti-human AFP antibody and anti-human Tf antibody shown in lane 1 (AFP)and lane 2 (TJ), respectively. MW = molecular size markers.
asialo-agalactoTf glycopeptide from GDP-L-[14C]fucose was rapidly effected using small columns of AGI-X8 and more than 95% of isotope-labeled glycopeptides were effectively bound and eluted by a-methyl-D-mannoside using LcHA-Sepharose 4B. In addition, the formation of labeled product was linear up to 4 h, and a standard incubation time of 2 h was used in the determination of activity (Fig. 6). The activity of al-6fucosyhransferase was significantly greater in the TGI cells cultured in ISRPMI (39.9 + 1.5 pmol. h -1" mg-1
days, polyamine 3.16 #mol.1-1. 10 -6 cells. 3 -1 days, NSE 4.2 ng-m1-1. 10 _6 cells.3 -1 days, TPA 30 000 U" 1- l - 10 -6 cells-3 -1 days, EGF, 1.5 pg. m1-1. 10 -6 cells-3 -1 days, and PDGF 2.0 pg" m1-1. 10 -6 cells. 3 -1 days. The immunoprecipitates using anti-AFP or anti-Tf antibody also demonstrated the band corresponding to their molecular weight of 70 000 for AFP and 80 000 for Tf, respectively (Fig. 2).
Determination of two major glycoproteins (AFP and Tf) with microheterogeneity of lectin-affmity before and after culturing TG1 cells with ISRPMI. By this method, AFP and Tf were detectable as low as 0.6 and 1 ng, respectively, as reported previously (18,19) and the results were compatible with the profiles obtained from lectin-affinity crossed immunoelectrophoresis. Using this sensitive method after ConA-affinity electrophoresis, Tf from normal human serum was divided into three bands (Fig. 3), i.e. strongly reactive major (Ctl)," moderately reactive minor (Ct2), and nonreactive, slight but detectable band (Ct3). These results were also confirmed by densitometric analysis (Fig. 4). Tf from TG1 cultured with FBS formed relatively major Ctl and minor Ct2 bands. Tf from TG1 cultured without FBS (ISRPMI) was also split into strongly reactive minor (Ctl) and weakly reactive major (Ct2) bands. Nonreactive Ct3 band was not detectable in both cases (Fig. 3 and 4). In the case of AFP separated by ConA-affinity electrophoresis, ConA reactive major (Cal) and nonreactive minor (Ca2) bands were detectable both in conditioned FBS-containing media and ISRPMI (Figs. 4 and 5). By LcHA-affinity electrophoresis, Tf in normal human serum was divided into three bands (Figs. 3 and 4), nonreactive major (Lt3), weakly reactive minor (Lt2), and strongly reactive but trace (Ltl). Tf from TG1 cultured with FBS-containing media was also composed of three bands; however, Ltl increased as compared with that of normal serum. In contrast, an obvious difference of the
Q1 Ct2
Ct3 NHS
ConA
i
TGIFBS-(ISRPMI) rD~':i
TGIFBS+ + +
NHS LcHA
TG 1FBS-(ISRPMI) TGIFBS+ Ltl Lt2
Lt3
Fro. 3. AnalyticalTf bands separated by ConAor LcHA-atfinityelectrophoresis of Tf from TG1 cultured with or without FBS-containingmedium and from normal human serum (NHS). Each 1-#1 sample was applied and eleorophoresis was run until bromophenol blue migrated 9.0 cm from the origin (-). Ctl-Ct3 and Ltl-Lt3 are described in the text. FBS+:10% FBS containing RPMI, FBS-(ISRPMI):RPMIsupplemented with Na2SeOs without FBS.
452
OIqKAWA ET AL.
"If ConA ++ ...... r"-++--r" ..........
~+ ~
~-'-+ . . . . .
"~"-+-
i
i--;-,.'...~
r-
+l.... ~ .......
/ j
i--
!: +!! ,i i::
.....
";
LcHA
+'
...... • .+; • + • .~
+
lt~tO..-r----i
~+ • i-.-
. t .~. + i-+.
+ ...... t"" ~
" +"
4.+
_.+. ~ L.
!.~!:+-....+
+
i
...,:..,.. .+ .i .
~..~ ,
......
..
~-. ~..... ~ " ~ ! ~ . ~ . . , ~
.++..+...+..
+. . •. . . +
~..
+e+ + +_
~- ..::+ : :+. '+ '~ +- l:+ ++....... + +..+ [+..: + + + , -;__+,
.... £
"
.'"- t-
r---+
~
v+ ~
....
-t
t+++:+j + ..+ + . . . . . .
+1........t '++![-", +,:
t
+
r
;
+
i,cHA
-.+~. ~ .~ ...+..~+ ~. -+..! .....
-+-i---~'-~ -~-+!'r" ~ i + ..... ~..~.....-....-....-..+....~._,~
,." ",~
....-~-+-~+----.:-.i.-4-.-~
~:_+...+-.r.~__ +_...~_.+
TG1FBS+
ConA
+++
+..-
-'-+.:--!-----i-#'+-!.---.-.: •.-+..~.-..~....~.-.:....'-...~..+..u. --.--?--T-:-..-7++-7-.:--7/....~
AFP
+
-i
....+-+..-+...+-4-..~...~--~
•-4..-.~.+.-~..+.~ -.~__~._.L_.::
........
r" ", "-'~+ --: ....
..... t'" ".."' "t" i- .t.. + . + ..!....
::::::::::::::::::::::::::::: • L...r....4...++.^+ + . . ~
..~.....
::.!::i,. i .i i--!--i -'~--i-
i--4-~.--~
-~-~--~
- -+ -.,÷-. --~ .-.4.,+-i.-...L-.-4..
TGIFBSi
i'+
- 1.- "i
+.""~+++"÷+++'~+ ~.......'._.i_..-....-....-....-.__: .~
:!.-::!:::~.::+ I ~,--~7~?'ml' +!-; i....+..4, i rl
" "+
~-..-:-...---~--+-..-..-.~-4-.+
.i..+--÷-I--~-.-+-..+ ....
.:::~..~:::.~:i.i:::.:,t:...~iiii~::::i ::
~
:
:
:
:
.
.
~
..~...r....~....~
......
.+..+...~
.
- . "!"'"~""~
++÷+ .: ~ : ~
~
~
........
:::!+"i:-:!':~:~:~ +"'t "'+" +'"::i"~::.i: "! ....
i:!~:!:~ I-'-'.÷ ...+...i
~:i"il
1
.... ~-.....~.....+.....~....;....+....i....i... .... ~--~...4 .... ~-- ~ .4- --!..
..... ~+-
i i.:
. ..~.~..-L..i~i..i~.~T~Zi~i.iI
'+~'+-++"+~--~+~L-+4 "
~- 4 - -+-.~-,++,-.4..-.L -.+-.~
,
"
, +
+-- ~~ '.4--~-.~- -~-.~ -.".--.4 --
t"
"4-
+
FIG. 4. Densitometric analysis was carried out according to Materials and Methods. See Figs, 3 and 5 for other keys.
protein) than those cultured in FBS-containing media (18.2 + 1.2 pmol- h - 1 . mg -1 protein).
DISCUSSION
In the present study, it was possible to eliminate serum and other proteins from culture medium for the growth of yolk sac tumor cells by adding a trace amounts of selenium to a chemically defined medium, ISRPMI. It is important to note that no adaptation procedures were required for the growth of the celt line after replacing serum with Na2SeO s. Many human germ cell tumors have been established as long-term monolayer cultures (1,3,8,1 O, 12,20,24); however, this is the first report of culture of the yolk sac tumor cells with serum- or protein-free chemically defined medium. The TG1 ceil line synthesized and secreted many plasma proteins and growth factors in the conditioned media as previously reported (1,3,8,10,12,20,24). Synthesized Tf might be one of the
essential substances for proliferating TG1 cells in vitro. A similar result was reported previously (19). Ferritin, polyamine, NSE, TPA, EGF, and PDGF were first detected and reported in a yolk sac tumor cell line (TG1, present case) cultured in vitro. To utilize the conditioned media with serum- or protein-free cultures is of great advantage for immunoassay to detect the very small amounts of substances and to eliminate the confusion of crossreactivity of the target substances to the similar substances often present in bovine serum, Wheia the TG1 cells were xenotransplanted into athymic nude mice, 13 human plasma proteins were detectable in plasma or cystic fluid in the tumor mass of mice bearing TG1 cells as reported elsewhere (14,16). These results suggest that TG1 and other yolk sac tumor cell lines and some embryonal carcinoma cell lines are able to maintain the same characteristics of human yolk sac (7) in early embryogenesis. By LcHA-affinity electrophoresis, elevation of fucosylation was demonstrated in two secretory glycoproteins, AFP as marker pro-
YOLK SAC TUMOR IN DEFINED MEDIUM tein for yolk sac tumor and Tf as one of the growth factors for prohferating cells, after the TG1 cells cultured with ISRPMI. LcHA-nonreactive fractions of AFP and Tf, which were detected in conditioned FBS-containing media, disappeared in conditioned ISRPMI, i.e. both AFP and Tf in the conditioned ISRPMI were composed of LcHA-strongly and weakly reactive two fractions. Conversely, no obvious difference in reactivity of AFP and Tf to CuriA as well as AFP to EPHA was noted between two conditioned media, and the electrophoretic profiles of AFP were almost similar to those of previous reports (18,22 ), The oligosaccharide structure required for LcHA activity lies in the core fucose attached to asparaginelinked N-acetylglucosamine residues (Fuc al-6GlcNAc-Asn). This additional attachment of fucose to GlcNAc does not influence the reactivity to ConA, however, the increase of tri- and tetra-antennary glycosylation of carbohydrate moieties of Tf molecules diminishes the affinity to ConA (5). This fact concluded that the fucosylated Tf in bi- and tri-antennary glycans increased in Tf molecule synthesized by TG1 cells. These results were also confirmed by the fact that the enzyme activity of a 1-6fucosyltransferase, which transports fucose to position 6 of asparagine-linked GlcNAc of N-glycans, was significantly greater in the TG1 cells cuhured with ISRPMI than in the cells with FBS-containing media. The enzyme activity correlated to the results obtained from LcHA-affinity electrophoresis. These results suggested the enhanced cellular expression of al-6fucosyltransferase after the change of culture condition from FBS-containing to serum-free medium. Abnormal fucosylation of the carbohydrate moiety of the proteins has been demonstrated in some proteins, such as Tf (4,19,26), AFP (27), and 'y-glutamyl transpeptidase (25) during malignant transformation of the cells. The increased level of fucosyltransferase activity and successive accumulation of the glycoproteins with abnormal fucosylation of the carbohydrate moieties is probably one of the oneofetal characteristics in the expression of cancer cells and the acquirement of growth characteristics without necessity of FBS may be the ultimate malignant transformation of the cancer cells. These facts clearly indicates
C,a2
Ca1
TGIFBS+ ConA "I'G 1FBS-(ISRPMI)
-
-ITGIFBS+
LcHA
~
~ii~
;:~ ~
~
TGIFBS-(ISRPMI)
t t 1 Lal La2 La3
FIG. 5. Analytical AFP bands separated by ConA or LcHA-affinityelectrophoresis from TG1. Each 3-gl sample was eleetrophoresed until bromophenol blue migrated 4.5 cm from the origin (-), Cal-Ca3 and Lal-La3 are described in the text. See Fig. 3. for other keys,
453
0
! 0
2
4
. 18
hour FIG. 6. Time-based graph of formation of [14C]-L-fucosein al-6fucosyltransferase assay. Incubation were performed using GDP-L-[14C]fucose as donor and asialo-agalacto Tf glycopeptide as acceptor for the various times shown. Each point represents the mean + SD of quadruplicate examinations,
that human cancer cell lines stably cultured with protein-free synthetic medium would be a useful material for the biomedical research. REFERENCES 1. Andrews, P. W.; Goodfellow, P. N.; Shevinsky, L. H., et al. Cell-surface antigens of a clonal human embryonal carcinoma cell line: morphological and antigenic differentiation in euhure. Int. J. Cancer 29:523-531: 1982. 2. Bonner, W. M.; Laskey, R. A. A film detection method for tritium-labeled proteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem. 46:83-88; 1974. 3. Bronson, D. L.; Andrews, P. W.; Soher, D., et al. Cell line derived frora a metastasis of a human testieular germ cell tumor. Cancer Res. 40:2500-2506; 1980. 4, Campion, B,; Leger, D.; Wieruszeski, J-M., et al. Presence of fucosylated triantennary, tetraantennary and pentaantennary glycans in transferrin synthesized by the human hepatocarcinuma cell line Hep G2, Eur. J. Biochem. 184:405-413; 1989. 5. Cummings, R. D.; Kornfeld, S. Fracti°nati°n °f asparagine-linked °ligosaccharides by serial lectin-agarose affinity chromatography. A rapid, sensitive, and specific technique, J. Biol. Chem. 257:1123511240; 1982. 6. Darlington, G. J.; Kelly, J. H.; Buffone, G. J, Growth and hepatospecific gene expression of human hepatoma cells in a defined medium. In Vitro Cell. Dev. Biol, 23:349-354; 1987, 7. Gitlin, D." Perricelli, A. Synthesis of human serum albumin, prealbumin, o~-fetoprotein, cqantitrypsin and transferrin by the human yolk sac. Nature 228:995-997; 1970. 8. Grossman, H. B.; Wedemeyer, G. UM-TC-1, a new human testicular carcinoma cell line. Cancer J. 1:22-25; 1986. 9. Hata, J.; Ueyama, Y.; Tamaoki, N., et aL Human yolk sac tumor serially transplanted in nude mice: its morphologic and functional properties. Cancer 46:2446-2455; 1980.
454
OHKAWA ET AL.
10. lshiwata, I.; Ishiguro, T.; Ishiwata, C., et al. Establishment and characterization of a human ovarian endodermal sinus tumor cell line producing specific type of a-fetoprotein subfraction. Gynecol. Oneol. 25:281-293; 1986. 11. Laemmli, U. K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685; 1970. 12. Lanford, R. E.; Bronson, D. L.; Esflack, L. E., et al. Plasma protein and apolipoprotein synthesis by human yolk sac carcinoma cells in vitro. In Vitro Cell. Dev. Biol. 27A:205-210; 1990. 13. Nakabayashi, H.; Taketa, K.; Miyano, K., et al. Growth of human hepatoma cell lines with differentiated functions in chemically defined medium. Cancer Res. 42:3858-3863; 1982. 14. Nishi, S.; Nishita, T.; Ohkawa, K. Production of 13 plasma proteins by human testieular yolk sac tumor transplanted into nude mice. Tumor Biol. 12:184-188; 1991. 15. Ogasawara, M.; Sirbasku, D. A. A new serum-free method of measuring growth factor activities for human cancer cells in culture. In Vitro Cell. Dev. Biol. 24:911-920; 1988. 16. Ohkawa, K.; Koike, K.; Amasaki, H., et al. Heteru-transplantation of human embryonal carcinoma into the nude mouse: functional and morphologic examination. Asia Oceania J. Obstet. Gyaecol. 6:7681; 1980. 17. Ohkawa, K.; Hibi, N.; Tsukada, Y. Evaluation of a conjugate of purified antibodies against human AFP-dextran-daunorubiein to human AFP-producing yolk sac tumor cell lines. Cancer lmmunol. Immunother. 22:81-86; 1986. 18. Ohkawa, K.; Abe, T.; Tsukada, Y., e! al. An improved sensitive assay method for the heterogeneity of c~-fetoprotein: possible application for early differential diagnosis. J. Clin. Chem. Clin. Biochem. 27:337-341; 1989.
19. Ohkawa, K.; Takada, K.; Takizawa, N., et al. Clear cell carcinoma of the human ovary synthesizes and secretes a transferrin with microheterogeneity of lectin affinity. FEBS Lett. 270:19-23; 1990. 20. Pera, M. F.; Blasco Lafita, M. J.; Mills, J. Cultured stem-cells from human testicular teratomas: the nature of human embryonal carcinoma, and its comparison with two types of yolk-sac carcinoma. Int. J. Cancer 40.'334-343; 1987. 21. Schachter, H.; Narasimhan, S.; Gleeson, P., et al. Glycosyltransferases involved in elongation of N-glycosidically linked oligosaccharides of the complex or N-acetyllactosamine type. Methods Enzymol. 98:98-134; 1983. 22. Taketa, K.; Ichikawa, E.; Nakabayashi, H., et al. Further resolution of human ot-fetoprotein by affinity electrophoresis with erythroagglufining phytobemagglutinin of Phaseolus vulgaris lectin. Tumor Biol. 6:519-531; 1985. 23. Voynow, J. A.; Scanlin, T. F.; Glick, M. C. A quantitative method for GDP-L-Fuc:N-Acetyl-~-D-glucosaminide a 1-6fucosyhransferase activity with lectin affinity chromatography. Anal. Biochem. 168:367373; 1988. 24. Yamamoto, T.; Komatsubara, S.; Suzuki, T., et al. In vitro cultivation of human testicular embryonal carcinoma and establishment of a new cell line. Gann 70:677-680; 1979. 25. Yamashita, K.; Hitoi, A.; Taniguchi, N., et al. Comparative study oftbe sugar chains of 3,-glutamyhranspeptidase purified from rat liver and rat AH-66 hepatoma. Cancer Res. 43:5059-5063; 1983. 26. Yamashita, K.; Koide, N.; Endo, T., et al. Altered glycosylation of serum transferrin of patients with hepatocellular carcinoma. J. Biol. Chem. 264:2415-2423; 1989. 27. Yoshima, H.; Mizuochi, T.; Ishii, M., et al. Structure of the asparaginelinked sugar chains of ot-fetoprotein purified from human ascites fluid. Cancer Res. 40:4276-4281; 1980.
