[CANCER RESEARCH 59, 4715–4719, September 15, 1999]

The Human Embryonal Carcinoma Marker Antigen TRA-1-60 Is a SialylatedKeratan Sulfate Proteoglycan

Graeme Badcock, Christine Pigott, John Goepel, and Peter W. Andrews1

Departments of Biomedical Science [G. B., C. P., P. W. A.] and Pathology [J. G.], University of Sheffield, Western Bank, Sheffield S10 2TN, United Kingdom

ABSTRACT

Human embryonal carcinoma (EC) cells are the stem cells of terato-carcinomas, and they are key components of germ cell tumors (GCTs).They express several high molecular weight glycoprotein antigens that aredown-regulated upon differentiation. One of these antigens, defined bymonoclonal antibody TRA-1-60, can be detected in the serum of GCTpatients and provides a useful complement to the established serummarkers human chorionic gonadotropin and a-fetoprotein, especially inthose patients without elevated serum human chorionic gonadotropin ora-fetoprotein. To examine the relationship of the TRA-1-60-defined anti-gen to similar antigens defined by other monoclonal antibodies, we havecarried out comparative Western blot and immunoprecipitation analysesof human GCT-derived cell lines with monoclonal antibodies TRA-1-60,TRA-1-81, GCTM2, and K21. The TRA-1-60 antigen was detected byWestern blot analysis in extracts of all human EC cell lines and in clinicalspecimens of GCT tested as a diffuse band with a molecular weight of>200,000. A similar but noticeably fainter band was detected in GCTcomposed of seminoma only. The antigen was not expressed by GCT-derived lines without an EC phenotype. Affinity bead-purified TRA-1-60,TRA-1-81, GCTM2 and K21 antigens reacted in Western blot analysiswith each of the other antibodies tested, indicating that the epitopesrecognized by each antibody are carried by the same molecular species.This molecule could be metabolically labeled with inorganic [35S]sulfateand was degraded by keratanase. Glycopeptides produced from affinity-purified TRA-1-60 antigen by extensive digestion with Pronase exhibiteda molecular weight in excess of 10,000 and were degraded by keratanase.The TRA-1-60 epitope was destroyed by digestion with neuraminidase,but the epitopes defined by TRA-1-81, GCTM2, and K21 were not. Ourresults indicate that human EC cells generally express a cell surfacesialylated keratan sulfate proteoglycan that is subject to modification toyield a variety of epitopes, one of which is recognized by the monoclonalantibody TRA-1-60. Sensitivity to milk alkaline digestion suggests that theoligosaccharides of this proteoglycan areO-linked to a core polypeptide.

INTRODUCTION

Testicular germ cell tumors, which include EC,2 teratocarcinoma,and seminoma, are the most common solid tumors of young men (1,2). Their incidence has increased 2–3-fold in most Western countriesover the past 50 years (3). The reasons for this increasing incidenceare unclear, although a role for environmental estrogen analoguesaffecting germ cell developmentin utero has been proposed. Fortu-nately, these tumors are highly susceptible to chemotherapy, andcurrent treatment regimens, including monitoring of serum tumormarker levels, are very successful (4). Nevertheless, treatment failuresoccur, and even successful treatment is often accompanied by signif-icant morbidity.

The serum tumor markers HCG and AFP are especially valuabletools for monitoring the treatment of patients (5, 6). They are widely

used in clinical practice to detect tumor recurrence after primarytreatment and to minimize the need for chemotherapy. However, asignificant proportion of patients with GCT are negative for eitherHCG or AFP or for both markers. Furthermore, the expression pat-terns of these markers may change over time as the tumors evolve,with initially positive tumors sometimes becoming “marker negative”during treatment and follow-up. This probably reflects the fact that, inthese histologically complex tumors, HCG and AFP are most likelyproduced by trophoblastic and yolk sac cells, respectively, rather thanby the EC stem cells themselves. Whether HCG and/or AFP aredetectable in the serum will depend upon the proportion of thetrophoblastic and yolk sac elements present in the tumor.

EC cells are probably the single most important malignant compo-nent of nonseminomatous germ cell tumors and are thought to giverise to the other cell types of teratomas by differentiation. Therefore,a serum marker more directly linked to the presence of these stemcells would be a valuable complement to the existing markers, such asHCG and AFP. Several cell surface antigens characteristically ex-pressed by human EC cells have been defined by a variety of mono-clonal antibodies (e.g., Refs. 7–11). One of these antigens, TRA-1-60(8), is detectable in the serum of GCT patients whose tumors containan EC component (12, 13). Moreover, as anticipated, serum TRA-1-60 antigen is often detectable even when HCG or AFP are absent,and it has, therefore, been suggested as an additional marker tocomplement serum HCG and AFP levels.

The TRA-1-60 antigen, like several other similar antigens report-edly expressed by human EC cells, is associated with a polydisperseglycoprotein with a molecular weight in excess of 200,000 (8, 14).Although the EC cell marker antigens SSEA3 and SSEA4 are glyco-lipids (9), TRA-1-81, GCTM2, K4, and K21, for example, are mono-clonal antibodies that identify proteins that are of similar size to thatrecognized by TRA-1-60, with similar, though not identical, patternsof expression on panels of tumor-derived cell lines (8, 15). However,the relationship of the TRA-1-60 antigen to these other similar anti-gens remains obscure. Sequential immunoprecipitation experimentshave suggested that the TRA-1-60 and TRA-1-81 epitopes are asso-ciated with distinct molecular species (8). Similarly, the K4- andK21-defined epitopes, which are reported to be carried by polylac-tosamine oligosaccharides, have also been detected associated withdistinct polypeptides (11). The antibody GCTM2 has been reported torecognize an epitope associated with a cell surface keratan sulfate, butwhether the other antigens are also related to keratan sulfate has notbeen ascertained (16).

We have now examined the relationship of the TRA-1-60-definedantigen from human EC cells to the antigens defined by monoclonalantibodies TRA-1-81, GCTM2, and K21. Our results suggest thatthese antibodies all recognize distinct epitopes formed by modifica-tions of a common keratan sulfate core molecule that is characteris-tically expressed by human EC cells.

MATERIALS AND METHODS

Cell Culture. Cells were cultured in DMEM supplemented with 10% fetalbovine serum at 37°C under a humidified atmosphere of 10% CO2 in air. Theywere subcultured and reseeded at 53 106 cells per 75-cm2 culture flask after

Received 3/11/99; accepted 7/21/99.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby markedadvertisementin accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1 To whom requests for reprints should be addressed. Phone: 44-(0)114-222-4173;Fax: 44-(0)114-276-5413; E-mail: P.W.Andrews@sheffield.ac.uk.

2 The abbreviations used are: EC, embryonal carcinoma; HCG, human chorionicgonadotropin; AFP,a-fetoprotein; HRP, horseradish peroxidase; PMSF, phenylmethyl-sulfonyl fluoride; PGC, primordial germ cell; ES, embryonic stem; SSEA1, stage-specificembryonic antigen-1.

4715

Research. on August 20, 2020. © 1999 American Association for Cancercancerres.aacrjournals.org Downloaded from

harvesting with 0.25% trypsin in Dulbecco’s PBS containing 1 mM EDTA. Anexception was the NTERA2 pluripotent EC cell line, which was harvested forpassage by scraping as described previously (17). NTERA2 EC cells wereinduced to differentiate by culture in medium containing 1025 M all-trans-retinoic acid (Eastman Kodak, Rochester, NY; Ref. 18). The germ cell tumorcell lines with an EC phenotype used were: 2102Ep cl.4D3 (7), 833KE (19,20), 1218E (20), 1156QE (20), and NTERA2 cl.D1 (17). In addition, thefollowing cell lines were used: 1411H, a yolk sac carcinoma line (21); 577 MF,a malignant teratoma cell line without an EC phenotype (20); BeWo, agestational choriocarcinoma line (22); and HT29, a colorectal carcinoma line(23).

Antibodies. The antibodies TRA-1-60 and TRA-1-81 were obtained asculture supernatants from the respective hybridomas, as described previously(8). GCTM2 (10) was provided by Dr. Martin Pera (Monash University,Clayton, Victoria, Australia), and K21 (11) was provided by Dr. WolfgangRettig (Memorial Sloan Kettering Institute, New York, NY). HRP-conjugatedgoat antimouse IgM (Sigma Chemical Co., Poole, Dorset, United Kingdom)was used as a detecting antibody in Western blots.

Affinity Purification Using Dynabeads. M-450 Dynabeads, precoatedwith rat antimouse IgM (m chain; 4.5-mm-diameter magnetic polystyrenebeads; Dynal, Wirral, United Kingdom), were incubated for 1 h with primaryantibody. After washing with PBS containing 0.1% BSA, these antibody-coated beads were incubated for 1 h with a lysate of 2102Ep EC cells preparedin 0.1% octylglucoside (Calbiochem), containing 0.1% BSA to prevent non-specific binding. After extensive washing, the antigen-antibody complex wasdissociated from the beads by boiling in electrophoresis loading buffer; afterremoval of the beads using a magnet, the resulting extract was subjected toSDS-PAGE.

Electrophoresis and Western Blot Analysis.Cells were lysed in 0.1%octylglucoside (Calbiochem) in 0.05M Tris-HCl (pH 6.8) with 0.1M PMSFand subjected to SDS-PAGE according to the procedure of Laemmli (24). ForWestern blot analysis, proteins were transferred electrophoretically from anSDS-polyacrylamide gel to a nitrocellulose membrane (0.45-mm pore size),which was then blocked with a 5% (w/v) solution of domestic fat-free milkpowder containing 0.05% polysorbitan (Tween 20) in PBS for 2 h andsuccessively incubated with primary monoclonal antibody in PBS (2 h at roomtemperature) and goat antimouse immunoglobulin conjugated with HRP. Theblots were then developed using the Bio-Rad HRP Immuno-detection kitcontaining 4-chloro-1-napthol as the substrate for the HRP enzyme. For35S-labeled samples, the gel was fixed and treated with the fluorographicreagent Amplify (Amersham, Little Chalfont, United Kingdom) for 20 minprior to exposure to Fuji RX X-ray film.

Metabolic Labeling with [ 35S]Sodium Sulfate. 2102Ep cells were cul-tured until just confluent in a 75-cm2 flask, washed once with sulfate-depletedmedium (sulfate-depleted DMEM/Ham’s F-12; Life Technologies, Inc.) andcultured for a further 24 h in 10 ml of sulfate depleted medium including 10%dialyzed FCS and 125mCi of [35S]sodium sulfate (DuPont-NEN, Stevenage,United Kingdom). Cells were then washed with PBS containing 0.1 mM PMSFand 10 mM magnesium sulfate and lysed at a proportion of 23 107 cells/ml in0.05M Tris-HCl containing 0.1% octylglucoside, 0.1 mM PMSF, and 1% BSAas a carrier protein.

Enzymatic Treatment. Sialidase (Vibrio cholerae),a-fucosidase (chickenliver), endo-b-galactosidase (Bacteroides fragilis), peptide-N-glycosidase F(Flavobacterium meningosepticum), endo-a-N-acetylgalactosaminidase(Streptococcus pneumoniae), chondroitinase ABC (Proteus vulgaris), andheparinase I (Flavobacterium heparinum) were all obtained from OxfordGlycosystems Ltd. (Abingdon, United Kingdom). Keratanase (Pseudomonassp.) was obtained from ICN (Thame, United Kingdom). Cell lysates wereprepared by osmotic shock in water for 5 min on ice; after centrifugation at10,0003 g for 15 min to remove cell debris, the lysate was adjusted to theionic strength and pH appropriate for each enzyme [heparinase I, 50 mM

sodium acetate, 2.5 mM calcium acetate, and 50 mM NaCl (pH 7.0); chon-droitinase ABC, 50 mM Tris-HCl (pH 7.3); keratanase, 50M Tris-HCl (pH7.4); and endo-b-galactosidase, 50 mM sodium citrate/phosphate (pH 5.7)].

Glycopeptide Analysis. Antigen labeled with35S was purified using Dyna-beads and digested with Pronase (Calbiochem; 1 mg/ml) in 0.05M Tris-HCl(pH 8.0) at 37°C for 24 h. A further equal volume of enzyme solution was thenadded, and the incubation was repeated for another 24 h (25). After heating to100°C for 5 min to destroy the Pronase activity, the digest was either chro-

matographed directly on G-50 Sephadex in 0.1M NaCl and 0.05M Tris-HCl(pH 7.2 or adjusted to pH 7.4, the optimum for keratanase) and incubated fora further 24 h with 2 units of keratanase at 37°C before chromatography.

RESULTS

To confirm that the TRA-1-60 antigen is generally expressed in asimilar form by human EC cells, we carried out Western blot analysisof lysates from a series of GCT-derived cell lines (Fig. 1A) andclinical biopsies of testicular GCT (Fig. 1B). As anticipated, all of theEC cell lines tested (2102Ep, NTERA2, TERA1, 833KE, 1218E, and1156QE) expressed detectable TRA-1-60 antigen and, in each case,the antigen appeared as a similar diffuse band with a molecular weightof .200,000. A similar band was also detected in lysates of three offive teratocarcinomas, three of four seminomas, and one of one mixedtumor with EC and seminoma components. There were some minordifferences in the pattern of the antigen within the diffuse bandbetween samples, which might indicate some heterogeneity betweendifferent cells. In contrast to the EC cells, no antigen was detected inextracts of the cell lines that do not exhibit an EC phenotype, includ-ing: differentiated derivatives of NTERA2 EC cells; a yolk sac car-cinoma line, 1411H; a malignant teratoma line, 577 MF; a choriocar-cinoma line, BeWo; and a colorectal carcinoma cell line, HT29. Theseresults correlate well with previous studies by immunobinding assayof cell lines and immunohistochemistry of tumor samples (8, 15); itseems likely that the TRA-1-60 antigen detectable in the teratocarci-noma samples was due to the presence of EC elements in thesetumors. It is notable that the antigen band was relatively weak in eachof the seminoma samples, which may indicate a lower level ofexpression in these cells and explain why TRA-1-60 antigen is readilydetectable in the serum of patients with histologically recognizableEC elements but not in the serum of seminoma patients (12).

Other monoclonal antibodies that recognize antigens expressed byhuman EC cells and are associated with a similar diffuse high mo-lecular weight protein upon SDS-PAGE, including TRA-1-81 (8),GCTM2 (10), and K21 (11). The similar size and expression pattern

Fig. 1. Western blot analysis of TRA-1-60 reactivity of detergent lysates of GCT-derived cell lines (A) and clinical biopsies of testicular GCT (B). The lysates prepared inoctylglucoside were electrophoresed SDS on a 7% polyacrylamide gel.A, Lane 1, 2102Ep(EC); Lane 2, NTERA-2 (EC);Lane 3, TERA-1 (EC);Lane 4, 833KE (EC);Lane 5,1218E (EC);Lanes 6–8, NTERA-2 cells induced to differentiate by retinoic acid for 7,14, and 21 days, respectively;Lane 9, 1156QE (EC);Lane 10, 1411H (yolk sac carci-noma);Lane 11, 577 MF (malignant teratoma);Lane 12, BeWo (choriocarcinoma); andLane 13, HT29 (colon carcinoma). Note that all of the EC cell lines tested (2102Ep,NTERA2, TERA1, 833KE, 1218E, and 1156QE) expressed a diffuse TRA-1-60 reactiveband with molecular weight in excess ofMr 200,000.B, Lane 1, lymphoma;Lanes 2–6,teratocarcinomas;Lanes 7–10, seminomas; andLane 11, mixed pathology tumor includ-ing seminoma and teratocarcinoma. Note that a similar pattern of reactivity is seen in threeof five teratocarcinomas, three of four seminomas, and one of one mixed pathologysamples of testicular GCT.

4716

KERATAN SULFATE ANTIGENS IN HUMAN EMBRYONAL CARCINOMAS

Research. on August 20, 2020. © 1999 American Association for Cancercancerres.aacrjournals.org Downloaded from

of these antigens suggest that they may be related. To investigate this,we isolated the antigens detected by TRA-1-60, TRA-1-81, GCTM2,and K21 by affinity bead purification and carried out Western blotswith each antibody in turn (Fig. 2). In each case, it was evident thatthe antigen recognized by one of the antibodies also carries epitopesthat are detected by each of the others.

These results and earlier results suggesting association with differ-ent molecular entities (8) would be compatible if a single core proteinis subject to posttranslational modification. For example, glycosyla-tion could yield the different epitopes, several of which may never-theless be carried by the same core polypeptide; however, somemolecular species might carry only a subset of possible epitopes. TheK21 epitope has previously been reported to be associated witholigosaccharide structures, whereas the GCTM2 epitope has beenreported to recognize a polypeptide epitope of a keratan sulfateproteoglycan. To test whether the TRA-1-60 epitope is also associatedwith an oligosaccharide, we investigated whether it is sensitive to avariety of enzymatic or chemical degradations.

Digestion of 2102Ep cell lysates witha-fucosidase, endo-b-galac-tosidase, or endo-a-acetyl galactosaminidase had no detectable effectupon the TRA-1-60 antigen (data not shown). We have previouslyshown that TRA-1-60 reactivity of cells is destroyed by neuramini-dase digestion (14). When Western blots of 2102Ep EC cell lysateswere incubated with neuraminidase, prior to probing with antibody,the TRA-1-60 epitope was also destroyed (Fig. 3). On the other hand,

the reactivity of the TRA-1-81, GCTM2, and K21 epitopes with theirrespective antibodies was enhanced, suggesting that sialylation blocksor sterically hinders these epitopes. Lower molecular weight forms ofGCTM2 and K21, in particular, were also revealed by neuraminidasetreatment.

Because the GCTM2 antigen was reported to be a keratan sulfateproteoglycan (16), we determined that the TRA-1-60 and TRA-1-81antigens can also be similarly metabolically labeled with inorganic[35S]sulfate. When these three antigens were isolated by affinityimmunobead purification from 2102Ep EC cells that had been grownin the presence of inorganic [35S]sulfate and digested with neuramin-idase prior to electrophoresis, all three showed similar upward shiftsin apparent molecular weight (Fig. 4). Likewise, the K21 antigen hasalso been reported to be metabolically labeled by inorganic sulfate andto show an increase in apparent molecular weight when digested withneuraminidase (11).

Taken together, these results suggest that the TRA-1-60 antigen isclosely related to the other similar high molecular weight glycoproteinantigens expressed by human EC cells and that all represent modifi-cations to a core keratan sulfate proteoglycan. To examine this hy-pothesis, we incubated affinity bead-purified TRA-1-60 antigen withseveral enzymes that are specific for degrading various proteoglycans,prior to electrophoresis and Western blot analysis. No effect wasobserved with hyaluronidase, heparinase, or chondroitinase (data notshown), but the antigenic activity was abolished by keratanase (Fig.5A). When35S-labeled TRA-1-60 antigen was similarly incubatedwith keratanase, it was also degraded (Fig. 5B). For comparison, in thelatter experiment, a parallel aliquot of35S-labeled antigen was incu-bated with neuraminidase because trace neuraminidase contaminationof keratanase was reported by the manufacturer. As shown previously,neuraminidase increased the apparent molecular weight of the labeledTRA-1-60 antigen (Fig. 5B, Lane 4) and did not degrade it in the sameway as did keratanase (Fig. 5B, Lane 5).

To investigate further the nature of the oligosaccharide carbohy-drate chains associated with the TRA-1-60 antigen, we digestedaffinity bead-purified35S-labeled TRA-1-60 antigen with Pronase.This digestion did not yield any discrete peptides, and all of the labelran with the dye front upon SDS-PAGE (data not shown). When thePronase-digested antigen was subjected to gel filtration on SephadexG50, all of the label eluted in the void volume, indicating association

Fig. 2. Coexpression of TRA-1-60, TRA-1-81, GCTM2, and K21 reactive epitopes onthe same molecular species. Antigen was purified from a lysate of 2102Ep human EC cellsusing beads armed with the respective antibodies, as indicated. The resulting boundantigen was then solublized, electrophoresed, subjected to Western blot, and probed witheach antibody.Lanes 1, TRA-1-60;Lanes 2, TRA-1-81;Lanes 3, GCTM2; andLanes 4,K21. As can be seen, the antigen isolated with any one of the antibodies was reactive witheach of the others on Western blot analysis.

Fig. 3. Neuraminidase sensitivity of the TRA-1-60 antigen in comparison to theTRA-1-81, K21, and GCTM2 antigens. Following electrophoresis of 2102Ep human ECcell lysates, Western blots were incubated with neuraminidase prior to probing withantibody. As shown, the reactivity with TRA-1-60 was eliminated, whereas reactivity withTRA-1-81, GCTM2, and K21 was enhanced.

Fig. 4. Neuraminidase sensitivity of35S-labeled TRA-1-60, TRA-1-81, and GCTM2antigens. Lysates were prepared from 2102Ep human EC cells metabolically labeled withinorganic 35S, and the antigens that were reactive with TRA-1-60, TRA-1-81, andGCTM2 were affinity-purified and analyzed by PAGE. An aliquot of purified antigen wasdigested with neuraminidase prior to SDS-PAGE. In each case, the antigen was stronglylabeled with35S. Upon digestion with neuraminidase before electrophoresis, the apparentsize of each antigen was similarly shifted upward.

4717

KERATAN SULFATE ANTIGENS IN HUMAN EMBRYONAL CARCINOMAS

Research. on August 20, 2020. © 1999 American Association for Cancercancerres.aacrjournals.org Downloaded from

with glycopeptides with molecular weights in excess of 10,000 (Fig.6A). However, when the Pronase digest was further digested withkeratanase, the35S-labeled glycopeptides were substantially de-graded, confirming the keratan sulfate character of the TRA-1-60antigen (Fig. 6B).

To determine whether the oligosaccharides with which the TRA-1-60 epitope appears to be associated areO- or N-linked, we firsttested whether the affinity-purified antigen was susceptible to diges-tion with either peptide-N-glycanase F or endo-a-N-acetylgalac-tosaminidase (O-glycanase). Neither enzyme affected its antigenicityor electrophoretic mobility (data not shown). However, the antigenwas completely destroyed by mild alkaline digestion, which is con-sistent with the presence ofO-glycolytic linkages (Fig. 7).

DISCUSSION

The commonly prevailing view of the origin of GCT is that theyarise from transformed PGCs during embryonic development. The

evidence for this in humans is largely circumstantial (26). However, inthe laboratory mice of the 129/Sv strain, which develop spontaneoustesticular teratomas, strong experimental evidence supports the hy-pothesis that these tumors originate from germ cells about the timethat they enter the genital ridge (27). Significant differences in thepathology of human and mouse GCT, notably, the absence of semi-nomas in the mouse, might suggest that the 129/Sv mouse does notprovide a good model of the human disease. On the other hand, wehave suggested that, in both cases, generation of a tumor couldinvolve two steps (28): first, a block to PGC differentiation and theirtransformation into permanently mitotic cells; and second, a switch inphenotype from that of PGC to that of ES cells, “short-circuiting” thechanges that are normally the consequence of gametogenesis andfertilization. We would suggest that the second step is stochastic, witha high probability of occurring in the mouse but low probability inhumans: in this model, the transformed PGCs would correspond tocarcinomasin situ and seminomas, whereas the second step generatesEC cells as caricatures (29) of normal ES cells.

Certainly, there is a close similarity in the gene expression patternof PGCs and ES/EC cells. For example, both express high levels ofalkaline phosphatase (30–32) and, in the mouse, the lactoseries car-bohydrate antigen, SSEA1 (33). They also characteristically expressthe stem cell transcription factor Oct-4 (34). Unlike the mouse, humanEC cells do not express SSEA1 (20). By contrast, the TRA-1-60antigen is strongly expressed by human but not by mouse EC cells (8).Recently, TRA-1-60 has also been shown to be expressed by monkey(35) and by human ES cell lines (36, 37), supporting the view thathuman EC cells do resemble human ES cells. Furthermore, the TRA-1-60 antigen is expressed by human PGC as well as by carcinomainsitucells (38) and seminoma cells (15), albeit at a lower level than theexpression by EC cells. Thus, in this regard, the pattern of TRA-1-60expression in humans is reminiscent of SSEA1 expression in themouse, although, whereas there is no evidence that the TRA-1-60antigen is associated with glycolipids (39), SSEA1 has been associ-ated with both glycolipids and high molecular weight glycoproteins(40).

Our results confirm previous data on the expression of TRA-1-60by human EC cells and germ cell tumors and demonstrate that theepitope is associated with a polydisperse macromolecule of similarsize range in all cases, including seminoma. Previous studies failed todetect TRA-1-60 antigen in the serum of seminoma patients but foundhigh levels in patients with histologically identifiable EC (12). Theresults presented here suggest that this difference is more likely areflection of different levels of expression rather than the epitopebeing associated with distinct molecular forms of the antigen.

Previous studies of human EC cells have indicated that they char-acteristically express high molecular weight oligosaccharides associ-

Fig. 5. Sensitivity to keratanase.A, affinity-purified TRA-1-60 antigen was incubatedin the absence (Lane 1) or presence (Lane 2) of keratanase, after which it was analyzedby electrophoresis and Western blot. Note that the antigen was almost completelydestroyed by keratanase.B, 35S-labeled affinity purified TRA-1-60 antigen was incubatedwith no enzyme (Lane 3), neuraminidase (Lane 4), or keratanase (Lane 5). Note themarked upward shift in apparent molecular weight after neuraminidase treatment andsubstantial diminution of the labeled TRA-1-60 antigen after keratanase treatment.

Fig. 6. 35S-labeled TRA-1-60 antigen was affinity-purified and subject to extensivedigestion with Pronase (two periods of 24 h with 1 mg/ml Pronase at 37°C). The resultingglycopeptides were chromatographed on G50 Sephadex (A).An aliquot of the Pronase-digested TRA-1-60 antigen was further incubated with 2 units of keratanase at 37°C for24 h and chromatographed on the same column (B). Note that the Pronase-digested[35S]glycopeptides were almost exclusively found in the void column (Mr .10,000),whereas these were substantially degraded by digestion with keratanase.

Fig. 7. TRA-1-60 antigenicity was destroyed by alkaline hydrolysis, suggestingOlinkage of carbohydrate chains. Alkaline hydrolysis of the TRA-1-60 antigen 2102Ep celllysates were incubated at 45°C for 24 h in PBS (pH 7.0;Lane 1) or under variousconditions of alkaline hydrolysis: 0.05M NaOH, 25°C, 24 h (Lane 2); 0.1M NaOH, 25°C,16 h (Lane 3); and 0.3M NaOH, 45°C, 5 h (Lane 4).

4718

KERATAN SULFATE ANTIGENS IN HUMAN EMBRYONAL CARCINOMAS

Research. on August 20, 2020. © 1999 American Association for Cancercancerres.aacrjournals.org Downloaded from

ated with cell surface glycoproteins (41). Our results suggest thatmany of the various related high molecular weight glycoproteinantigens that are typically expressed by human EC cells, including theTRA-1-60 antigen, are variants of a related macromolecule, namely,a keratan sulfate. Whereas the precise epitope structure recognized byTRA-1-60 remains to be established, it seems most likely that itrepresents a terminal modification to the core-sulfated,O-linked lac-tosaminoglycan that constitutes the carbohydrate of keratan sulfate.How this antigen is shed by EC cells and whether all forms are shedare unclear. Previously, we failed to detect by Western blot the shedform of the TRA-1-60 antigen in culture supernatants of EC cell lines,although it is associated with equally high molecular weight aggre-gates by gel filtration chromatography (14). This suggests that thenature of the molecule is altered on shedding.

Previous data have clearly indicated that serum TRA-1-60 antigenis a useful marker to complement others, notably AFP and HCG, usedto monitor treatment of GCT patients (12, 13). The GCTM2 antigenhas also been reported in serum of GCT patients (42). A knowledgeof the structure of the core keratan sulfate and its associated epitopesmay help in designing simpler assay systems. Furthermore, the com-mon expression of this family of keratan sulfate proteoglycans on ECcells and probably on early ES cells suggests that it may play animportant role in embryonic development and in development andprogression of GCT. The parallel finding with SSEA1 expression inthe mouse and the observation that SSEA-1 and TRA-1-60 antigensboth involve high molecular weight lactosaminoglycan core carbohy-drates suggest the hypothesis that TRA-1-60 and its related moleculesmay perform in humans whatever functions are performed by SSEA1-reactive glycoproteins in the mouse.

REFERENCES

1. Damjanov, I. Pathogenesis of testicular germ cell tumours. Eur. Urol.,23: 2–7, 1993.2. Dixon, F. J., and Moore, R. A. Atlas of Tumor Pathology: Tumors of the male sex

organs (monograph). Washington, DC: Armed Forces Institute of Pathology, 1952.3. Skakkebæk, E. E., Rajpert-de Meyts, E., Jørgensen, N., Carlsen, E., Petersen, P. M.,

Giwercman, A., Andersen, A-G., Jensen, T. K., Andersson, A-M., and Muller, J.Germ cell cancer and disorders of spermatogenesis: an environmental connection?Acta Pathol. Microbiol. Immunol. Scand.,106: 3–12, 1998.

4. Horwich, A. Testicular germ cell tumours: an introductory overview.In: A. Horwich(ed.), Testicular Cancer, pp. 1–17. London: Chapman and Hall, 1996.

5. Bosl, G. J., Geller, N. L., and Cirrincione, C. Serum tumor markers in patients withmetastatic germ cell tumors of the testis. A 10 year experience. Am. J. Med.,75:29–35, 1983.

6. Mason, M. D. Tumour markers.In: A. Horwich (ed.), Testicular Cancer, pp. 35–51.London: Chapman and Hall, 1996.

7. Andrews, P. W., Goodfellow, P. N., Shevinsky, L. H., Bronson, D., and Knowles,B. B. Cell-surface antigens of a clonal human embryonal carcinoma cell line:morphological and antigenic differentiation in culture. Int. J. Cancer,29: 523–531,1982.

8. Andrews, P. W., Banting, G., Damjanov, I., Arnaud, D., and Avner, P. Threemonoclonal antibodies defining distinct differentiation antigens associated with dif-ferent high molecular weight polypeptides on the surface of human embryonalcarcinoma cells. Hybridoma,3: 347–361, 1984.

9. Kannagi, R., Cochran, N. A., Ishigami, F., Hakomori, S-i., Andrews, P. W., Knowles,B. B., and Solter, D. Stage-specific embryonic antigens SSEA-3 and -4 are epitopesof a unique globo-series ganglioside isolated from human teratocarcinoma cells.EMBO J.,2: 2355–2361, 1983.

10. Pera, M. F., Blasco-Lafita, M. J., Cooper, S., Mason, M. D., Mills, J., and Monaghan,P. Analysis of cell-differentiation lineage in human teratomas using new monoclonalantibodies to cytostructural antigens of embryonal carcinoma cells. Differentiation,39: 139–149, 1988.

11. Rettig, W. J., Cordon-Cardo, C., Ng, J. S. C., Oettgen, H. F., Old, L. J, and Lloyd,K. O. High-molecular-weight glycoproteins of human teratocarcinoma defined bymonoclonal antibodies to carbohydrate determinants. Cancer Res.,45: 815–821,1985.

12. Marrink, J., Andrews, P. W., Van Brummen, P. J., De Jong, H. J., Sleijfer, D. T.,Schraffordt Koops, H., and Oosterhuis, J. W. TRA-1-60: a new serum marker inpatients with germ-cell tumours. Int. J. Cancer,49: 368–372, 1991.

13. Gels, M. E., Marrink, J., Visser, P., Sleijfer, D. T., Droste, J. H. J., Hoekstra, H. J,Andrews, P. W., and Koops, H. S. Importance of a new tumour marker TRA-1-60 inthe follow-up of patients with clinical state I nonseminomatous testicular germ celltumours. Ann. Surg. Oncol.,4: 321–327, 1997.

14. Andrews, P. W., Marrink, J., Hirka, G., Von Keitz, A., Sleijfer, D. T., and Gonczol,E. The surface antigen phenotype of human embryonal carcinoma cells: modulationupon differentiation. Rec. Res. Cancer Res.,123: 63–83, 1991.

15. Andrews, P. W., Casper, J., Damjanov, I., Duggan-Keene, M., Giwercman, A., Hata,J., Von Keitz, A., Looijenga, L. H. J., Millan, J. L., Oosterhuis, J. W., Pera, M. F.,Sawada, M., Schmoll, H. J., Skakkebaek, N. E., and Van Putten, W. Comparativeanalysis of cell surface antigens expressed by cell lines derived from human germ celltumours. Int. J. Cancer,66: 806–816, 1996.

16. Cooper, S., Pera, M. F., Bennet, W., and Finch, J. T. A novel keratan sulphateproteoglycan from a human embryonal carcinoma cell line. Biochem. J.,286: 959–966, 1992.

17. Andrews, P. W., Damjanov, I,. Simon, D, Banting, G., Carlin, C., Dracopoli, N. C.,and Fogh, J. Pluripotent embryonal carcinoma clones derived from the humanteratocarcinoma cell line Tera-2: differentiationin vivo andin vitro. Lab. Invest.,50:147–162, 1984.

18. Andrews, P. W. Retinoic acid induces neuronal differentiation of a cloned humanembryonal carcinoma cell linein vitro. Dev. Biol., 103: 285–293, 1984.

19. Bronson, D., Andrews, P. W., Solter, D., Cervenka, D., Lange, J., and Fraley, E. E.A cell line derived from a metastasis of a human testicular germ-cell tumor. CancerRes.,40: 2500–2506, 1980.

20. Andrews, P. W., Bronson, D., Benham, F., Strickland, S., and Knowles, B. B. A.Comparative study of eight cell lines derived from human testicular teratocarcinoma.Int. J. Cancer,26: 269–280, 1980.

21. Vogelzang, N. J., Bronson, D. L., Savino, D., Vessella, R. L., and Frayley, E. A.Human embryonal-yolk sac carcinoma model system in athymic mice. Cancer(Phila.),55: 2584–2593, 1985.

22. Pattillo, R. A., and Gey, G. O. The establishment of a cell line of human hormonesynthesizing trophoblastic cellsin vitro. Cancer Res.,28: 1231–1236, 1968.

23. Fogh, J., and Trempe, G. New human tumor cell lines.In: J. Fogh (ed.), HumanTumor Cellsin Vitro, pp. 115–159. New York: Plenum Press, 1975.

24. Laemmli, U. K. Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature (Lond.),227: 680–685, 1970.

25. Cossu, G., Andrews, P. W., and Warren, L. Covalent binding of lactosaminoglycansand heparan sulphate to fibronectin synthesised by a human teratocarcinoma cell line.Biochem. Biophys. Res. Commun.,111: 952–957, 1983.

26. Rajpert-de Meyts, E., Jørgensen, N., Brøndum-Nielsen, K., Muller, J., andSkakkebaek, N. E. Developmental arrest of germ cells in the pathogenesis of germcell neoplasia. Acta Pathol. Microbiol. Immunol. Scand.,106: 198–206, 1998.

27. Stevens, L. C. The biology of teratomas. Adv. Morphol.,6: 1–31, 1967.28. Andrews, P. W. Teratocarcinomas and human embryology: pluripotent human EC

cell lines. Acta Pathol. Microbiol. Immunol. Scand.,106: 158–168, 1998.29. Pierce, G. B. Teratocarcinoma. Introduction and perspectives.In: M. I. Sherman and

D. Solter (eds.), Teratomas and Differentiation, pp. 3–12. New York: AcademicPress, 1975.

30. Bernstine, E. G., Hooper, M. L., Grandchamp, S., and Ephrussi, B. Alkaline phos-phatase activity in mouse teratoma. Proc. Natl. Acad. Sci. USA,70: 3899–3902,1973.

31. Chiquoine, A. D. The identification, origin and migration of primordial germ cells inthe mouse embryo. Anat. Res.118: 135–149, 1956.

32. Benham, F. J., Andrews, P. W., Bronson, D. L., Knowles, B. B., and Harris, H.Alkaline phosphatase isozymes as possible markers of differentiation in humanteratocarcinoma cell lines. Dev. Biol.,88: 279–287, 1981.

33. Solter, D., and Knowles, B. B. Monoclonal antibody defining a stage specific mouseembryonic antigen SSEA-1. Proc. Natl. Acad. Sci. USA,75: 5565–5569, 1978.

34. Brehm, A., Ovitt, C. E., and Schler, H. R. Oct 4: more than just a POUerful markerof the mammalian germ line. Acta Pathol. Microbiol. Immunol. Scand.,106: 114–126, 1998.

35. Thomson, J. A., Kalishman, J., Golos, T. G., Durning, M., Harris, C. P., Becker,R. A., and Hearn, J. P. Isolation of a primate embryonic stem cell line. Proc. Natl.Acad. Sci. USA,92: 7844–7848, 1995.

36. Thomson, J. A., Itskovitz-Eldor, J., Shapiro, S. S., Waknitz, M. A., Swiergiel, J. J.,Marshall, V. S., and Jones, J. M. Embryonic stem cell lines derived from humanblastocysts. Science (Washington DC),282: 1145–1147, 1998.

37. Shamblott, M. J., Axelman, J., Wang, S., Bugg, E., Littlefield, J. W., Donovan, P. J.,Blumenthal, P. D., Huggins, G. R., and Gearhart, J. D. Derivation of pluripotent stemcells from cultured human primordial germ cells. Proc. Natl. Acad. Sci.,95: 13726–13731, 1998.

38. Giwercman, A., Andrews, P. W., Jørgensen, N., Mller, J., Graem, N., andSkakkebaek, N. E. Immunohistochemical expression of embryonal marker TRA-1-60in carcinomain situ and germ cell tumors of the testis. Cancer (Phila.),72: 1308–1314, 1993.

39. Fenderson, B. A., Radin, N., and Andrews, P. W. Differentiation antigens of humangerm cell tumours: distribution of carbohydrate epitopes on glycolipids and glyco-proteins analyzed using PDMP, an inhibitor of glycolipid synthesis. Eur. Urol.,23:30–37, 1993.

40. Childs, R. A., Pennington, J., Uemura, K., Scudder, P., Goodfellow, P. N., Evans,M. J., and Feizi, T. High molecular weight glycoproteins are the major carriers of thecarbohydrate differentiation antigens I, i, and SSEA-1 of mouse teratocarcinomacells. Biochem J.,215: 491–503, 1983.

41. Muramatsu, H., Muramatsu, T., and Avner, P. Biochemical properties of the high-molecular-weight glycopeptides released from the cell surface of human teratocarci-noma cells. Cancer Res.,42: 1749–1752, 1982.

42. Mason, M. D., Pera, M. F., and Cooper, S. Possible presence of an embryonalcarcinoma-associated proteoglycan in the serum of patients with testicular germ celltumours. Eur. J. Cancer,27: 300, 1991.

4719

KERATAN SULFATE ANTIGENS IN HUMAN EMBRYONAL CARCINOMAS

Research. on August 20, 2020. © 1999 American Association for Cancercancerres.aacrjournals.org Downloaded from

1999;59:4715-4719. Cancer Res   Graeme Badcock, Christine Pigott, John Goepel, et al.   a Sialylated Keratan Sulfate ProteoglycanThe Human Embryonal Carcinoma Marker Antigen TRA-1-60 Is

  Updated version

  http://cancerres.aacrjournals.org/content/59/18/4715

Access the most recent version of this article at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/59/18/4715.full#ref-list-1

This article cites 34 articles, 10 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/59/18/4715.full#related-urls

This article has been cited by 7 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/59/18/4715To request permission to re-use all or part of this article, use this link

Research. on August 20, 2020. © 1999 American Association for Cancercancerres.aacrjournals.org Downloaded from